Getting a better understanding of gas turbine control


Thread Starter


I'm a rookie in gas turbine using GE MARK VI control system. I see there were a lot of threads replied by CSA, and i found him very knowledgeable.

I was wondering how can a rookie like me to have a better understanding of gas turbine control.

In one thread, CSA advised to read Control Sequence Programs (CSPs) and Control Specification. what is Control Specification? how can i find it?

Thank you.

5) Lp speed control system
6) exhaust temperature control system
7) 2 nd stage Nozzle control system
8) Lube oil and cooling system

Best regards

Welcome to the business. Mark V turbine control panels employ what's formally known as the CSP--Control Sequence Program. Mark VIs (and Mark VIes) employ what's formally known as "application code." Both are terms for the application-specific sequencing, the "logic", that's gets downloaded to the turbine control to make it particular for a specific application and site.

"Reading" the CSP or application code involves learning a new language--relay ladder diagram language. In the case of the Mark VI and the Mark VIe they use function blocks to perform the same operations that rungs used to do in relay ladder logic, but they still use a similar graphical representation to relay ladder diagram logic.

The Mark V CSP can be printed, and so can the Mark VI/VIe application code--but printed application code looks almost nothing like what one sees in Toolbox/ToolboxST and it consumes reams of paper and is usually considered a waste of time and natural resources to print it. Myself, I use MS-Windows screen printing capability to capture bits of code I need to make notes on and study to a MS-Word or MS-Wordpad document, and then print them.

I would be very happy to work through the single most important rung in the GE-design heavy duty gas turbine world: L4. If you have some patience, we can cover everything from what trips your turbine, to all of the start-check permissives required to get L4 to pick up. It's a great way to learn how GE SHOULD name signals and how to look at signals and quickly determine when the logic signal is a logic "1" or "0", and even if it's a timer or an inverse timer or associated with a physical input or output.

The Control Specification is a document that is provided with the Speedtronic turbine control panel. A Control Specification is specific to every turbine (or group of turbines installed and commissioned at the same time on a site). In the old days, GE used to provide the Control Specification, and the Piping Schematics (everyone else in the world calls them P&IDs) in Vol. III of the Service Manuals provided with the turbine and auxiliaries. But, they couldn't leave well enough alone, and it's changes several times over the last decade-and-a-half (which volume of the manual they provide it in). It's a very good document for most things, but I've seen some glaring errors in the Control Specification, such as Control Spec.'s which say the turbine has a diesel starting motor when it has an electric starting motor, and vice versa. So, read it for "intent", not "content"--meaning the description of the starting sequence, in this case, should talk about purging, firing, warm-up, and acceleration. That has to be done regardless of the type of starting means.

Also, many people find the Control Specification years after the unit was commissioned and think the Control Constant values listed in the Control Specification are GOSPEL and start changing the as-running values to match the Control Specification, and that' when the real "fun" begins. If the turbine's been running fine for months or even years, the Control Constant values in the Mark VI are probably just fine, though some could probably use a little "tuning." Just as with hardware ("Berg") jumpers, the positions you find them in on a card which has been working but is suspect are most likely the correct ones--not what some piece of paper says.

So, if you want to work on L4 in this forum, just let me know. It won't be the quickest thing we've every done, but it will be a good and valuable learning experience--I promise.

And, don't neglect the P&IDs--the Piping Schematic drawings. Those are about the single most important group of drawing a technician or operator can have. Every technician, and every operator, should have their own copy, with their own notes and markings.

Finally, one thing I'm coming to understand about newbies is that they don't have a good grasp of how field devices and instruments work. RTDs, in particular, but even thermocouples and pressure switches and temperature switches and limit switches and pressure transmitters--not to mention electro-hydraulic servo-valves and LVDTs. RTDs and T/Cs (thermocouples) have been covered in some detail on And none of them are particularly difficult to understand--but wiring them to any control system can require some basic knowledge which most people don't have and seem to believe isn't really important. So, if you have any questions about field devices and instruments, open a new thread and we can talk about them, if necessary--after you've researched the archives (past threads, using the 'Search' feature) and if you have any questions.

Looking forward to working with you, Neo.

Another favorite document of mine to understand the Mark V controls is the A010 drawing- Constants and Diagnostics 377a3378.

I am not familiar with the Mark V1 but is should have something similiar.


thanks for your detailed reply.

It would be great if i have an opportunity to work with you.

As i am confused about how to start, i think working on L4 is a good choice. And i am interested in of gas turbine control.

But what should i do exactly?

Can you give me guidance and told me what specific i should do.

I am willing to spend several hours a day working on it.

Okay! Let's get started. First, some "back-room" stuff. In order to post rungs so that they're legible on we have to use HTML tags, specifically, the 'pre' and '/pre' tags (enclosed by the < and > characters). The 'pre' HTML tag needs to be added to the reply Text immediately before the rung, and then the '/pre' HTML tag needs to be added immediately after the rung--this is necessary to put the forum software into fixed-pitch font mode. (You can usually see the HTML tags in the body of replies.)

So, by putting a 'pre' (with the < and > characters--which won't show in this reply properly on the main pages of the L4 rung and then following it with the '/pre' (with the < and > characters) we can display the L4 rung:<pre>
L4S L94T L4T L4
-----| |-----------|/|----------|/|------------( )
L4 |
-----| |------</pre>
Also, at least in the web browser that I use one can grab the lower right corner of the Text box and drag it to the right (and down) to get more horizontal (and vertical) space when replying. This will be useful when logic signal names are long, and when there are as many as eight horizontal elements in a rung. You can always use the 'Preview' and 'Edit' buttons to see how your post appears before submitting it.

Also, let's talk about how GE is <b><i>supposed</i></b> to write and choose logic signal names (per their own standard--realize that every company has its own standard, and we're just talking about GE here). In the early days of digital Speedtronic turbine control systems it was decided to make the signal name describe when the logic was a "1", or when it was "True." In general, this has been followed over the years, but it's not always followed and that's a real shame, because it's <b>very</b> useful when it is.

Next, we need to be sure everyone (you, Neo, and me) is clear about what normally open and normally closed means. A normally open contact associated with a logic signal will be OPEN when the associated logic is "0" and will be CLOSED when the associated logic is "1". A normally closed contact associated with a logic signal will be CLOSED when the associated logic is "0" and will be OPEN when the associated logic is "1".

To get a logic signal to be a logic "1" it's necessary for power to flow through a series-parallel string to the "coil" of the logic signal, so that means that the logic signals of normally open contacts must be a logic "1" and the logic signals of normally closed contacts must be a logic "0". As an example, let's "read" the L4 rung (because it is, in fact, a sentence).

<b>"When L4S <i>IS</i> a logic "1" OR when L4 <i>IS</i> a logic "1", AND when L94T <i>IS NOT</i> a logic "1" AND when L4T <i>IS NOT</i> a logic "1", then L4 <i>will be</i> a logic "1"."</b>

So, when L4S is a logic "1" its associated normally open contacts will be CLOSED, and when L4 is a logic "1" its associated normally open contacts will be CLOSED, and when L94T is NOT a logic "1" (that is, when it IS a logic "0") its associated normally closed contacts will be CLOSED, and when L4T IS NOT a logic "1" (that is, when it IS a logic "0") its associated normally closed contacts will be CLOSED. And when either L4 or L4S is a logic "1", and when both L94T and L4T are NOT logic "1"s, then L4 will be a logic "1".

I find it easiest to "read" rungs in this fashion--"imagining" what state the associated logic signals of normally open and normally closed contacts have to be in in order for power to flow through the normally open or normally closed contacts to "energize" the logic signal coil(s) at the right end of the rung. Others like to read them in different fashions--possibly like this: When L4S is TRUE OR when L4 is TRUE, AND when L94T is FALSE AND L4T is FALSE then L4 will be TRUE. But, it's the same thing--TRUE and FALSE are another way of describing when logic signals are "1" or "0" and which state the logic signals need to be in in order for power to flow through them to energize the rung's logic signal coil. There are lots of ways to read and interpret rungs--lets' just stick with this one for this discussion, and hopefully it will become clearer why this way works best for GE logic (application code in the Mark VI).

So, from the longname descriptions of signal names in Toolbox, we can see that L4S stands for "L4 Set" (or sometimes called "L4 Start"). It will be a logic "1" (True) when all of the start-check permissives are met and when all of the other final start-checks are completed when a START is initiated. (Don't worry; we'll go over all of them in due time--presuming you have the patience.)

Now, let's continue with the descriptions of the other names of elements of the L4 rung. L94T will be a logic "1" when a normal shutdown is in progress and the fuel is to be shut off <i>in two of the three control processors, <R>, <S> and <T> for a TMR control panel--or just in <R> for a SIMPLEX control panel</i>. L4T will be a logic "1" when any condition which trips the turbine <i>in two of the threee control processors, <R>, <S> and <T> for a TMR control panel--or just in <R> for a SIMPLEX control panel</i>. (There are MANY sub-rungs for L4T, and we'll go over all of them.)

So, what does this tell us about L4? In order to get L4 to go to a logic "1" (to "pick up"), L4S needs to be a logic "1" and L94T and L4T both need to be logic "0"s, and then L4 will also go to logic "1" which will close the normally open contact in parallel with L4S and "latch" or "seal-in" the rung--so that if EITHER L94T goes to a logic "1" OR L4T goes to a logic "1" L4 will "drop out", or go to a logic "0". It should also be clear that it's necessary for L4S to go to a logic "1" to get L4 to also go to a logic "1", but that once L4 is a logic "1" that L4S does NOT need to remain a logic "1" for L4 to remain a logic "1"--as long as both L94T and L4T both remain logic "0".

That's it. Once L4 is a logic "1" there are only two conditions which can "drop out" L4 (make it go to a logic "0")--L94T going to a logic "1" or L4T going to a logic "1". And in order to start, run and shutdown (normal shutdown) the turbine, L4 MUST be a logic "1", and if L4 transitions from logic "1" to logic "0" while the turbine is starting or running or shutting down then the fuel stop valve(s) will be commanded to close and the turbine will be tripped. (Fuel stop valves are not open during purging and prior to firing, but L4 is required to be a logic "1" in order to open fuel stop valves when necessary, and for them to remain open.)

And--presuming that both L94T and L4T are BOTH logic "0"s, which they must be in order to initiate a START--there is only one condition that will allow L4 to pick up (go to a logic "1")and that is when L4S goes to a logic "1". (That's because prior to a START, when a READY to START is indicated to the operator, L4 is a logic "0". This can be seen in the 'Start Check Permissive' display.)

Hopefully you can begin to see how the signal name being chosen to describe when the signal is a logic "1" can be helpful. And hopefully you can also begin to see how knowing how to read signal names (most of them--unfortunately not all, because standards aren't always adhered to) can be very beneficial to quickly reading and interpreting rungs and logic (application code in the Mark VI). There are some little "tricks"--which aren't documented anywhere, not even in GE (sad, but true)--which we will discover on this journey.

So, your assignment, Neo, is to post the L4S rung to this thread (using the HTML tags). And, then, please "read" (write the sentence) that describes when L4S will be a logic "1".

Also, please post the longnames for each of the logic signals in the rung. Here's an example:

L4 = Master Protective ("1" to Run)
L4S = L4 Set
L4T = Master Protective Trip ("1" to Trip)
L94T = Normal Shutdown Trip

(I believe that if you hover the cursor over a signal name in Toolbox the longname description will appear at the lower left corner of the Toolbox window.The longnames I wrote were exact quotes from a Mark VI site; they are not the same for every job. Just post what it says in your application code from Toolbox. Sometimes, there is no longname; and worse, criminally, the longname is wrong. That's life, though. We learn to deal with it.)

We will go through each element of the L4S rung (which has several "sub-rungs", and the L4T rung, which has MANY sub-rungs) so you will get very proficient at using the HTML tags--and "reading" signal names. We'll go over the L94T rung, too. And, we will discuss the various other types of proper (and possibly improper) GE signal names, as well as understand the conditions and permissives.

You're going to have to look up Control Constant values, and various device settings (from the Device Summary document provided with every GE-design heavy duty gas turbine). So, you will need to find the Device Summary document (it also used be in Vol. III, but who knows where "they" put it now...).

If you have any questions or comments on the above, let's deal with those first. If I introduce words or terms you're unfamiliar with, ask for clarification. If you're having some trouble with normally open vs. normally closed contacts, just be patient with yourself. It will begin to become clearer soon. Best not to get caught up in that discussion because it does get very clear after a while--it just takes longer for some people than others, but be patient.

This is going to be a joint effort, but to be of the most benefit to you we are going to use the exact rungs and signals in your Mark VI application code so that it's specific to your turbine and auxiliaries. Your job is to provide requested information--rungs, Control Constant values, pressure switch or temperature switch settings, etc.--and the rest we will sort out as we continue on this journey of discovery.

Remember, relay ladder logic is just another "language"--and it has "sentences" with verbs and predicates and subjects just like any other language. And it's nothing more than logic, sequencing. By learning how to "anticipate" when a logic signal will be a logic "1" (and by inference, when it won't be a logic "1"--that is, when it will be a logic "0") one can more easily read the rungs and begin to understand what is to be happening so that one can understand GE-design heavy duty gas turbine control and GE-design heavy duty gas turbine control philosophy.

Finally, L4 is just one way the turbine can be tripped--from the application code running in the control processors. There is also tripping which can be done through the <P> core and associated inputs to the <P> core--and we'll get to that to as we finish this discussion. Because it's important to understand this isn't the only way a turbine can be tripped--GE has several. Be patient, and we'll cover them all, and in the process learn more about the Mark VI.

I'm looking forward to this journey! Let's keep moving and we're going to "see" some amazing things. "Learning is finding out what you already knew," is one of my favourite quotes. When presented correctly, it's exactly what happens--one says to himself (or herself), "I knew that!" It's just that you hadn't thought it about it that way before. And with a few little hints, tips and tricks it can be very easy--and intuitive, which means, well, you already knew that!

By the way, I'm going to apologize in advance for making a few mistakes along the way. I'm not a good person to proofread my own writing (most people aren't good at proofreading their own writing), and I'm doing this in my spare time, and so I do--and will--make mistakes. Just hopefully, not too many of them. Please be patient with me.

Thank you for your reply.I am excited too, feeling i'm on a mysterious trip with you.

This time my assignment is L4S right?

If there are some mistakes and drawbacks or you have advice just let me know,and i'll try to do better next time. I hope to have a long and exciting journey with you.

Here comes the L4S:<pre>
L4SX L14HR L14Y L4s
L4SX = Master protective set auxiliary logic(confusing)
L14HR = HP zero speed signal
L14Y = time delay loss of master protective(not sure if it is needed)

So, if L4SX, L14HR and L14Y are 1, then L4s is 1, otherwise L14 remains 0.

Should i make further classification? But it is hard for me.

I think L14SX is relatively important, so i dig deeper.
Here comes the L14SX;<pre>
L3RS L1X L33cse L63QT L4SX
L3RS = Start up check present(not sure)
L1X = startup check stop master control-startup permiss(confusing)
L33cse = starting means cluth engaged
L63QT = lube oil gen low pressure voted

So, if L3RS, L1X, and L33cse are 1, and L63QT is 0, then the L4SX is 1.

Is it enough for this time? It seems that the logic is endless if i go deeper.

I have questions, and some of them are foolish to you.

1) I think L means the type of variables is logic, and do the number,like 4,14 means something?

2) I am a little confused L4, can you give me some specigication.

3) How is my assignment? I am not sure whether i do it in the way you ask.

Best regards,Neo.

This is a good start. Please do be careful when "copying" signal names; if the signal names is in all capital letters in Toolbox, you should use all capitals in your post. If the signal name is in lower-case letters, you should use lower-case letters in your post. And, please double-check to make sure you are typing the proper signal names (see below).

Yes. "L" means the signal is a logic--usually; GE violates that standard occasionally, too, but not very often, fortunately.

And, the numbers--they are VERY important. They correspond to GE's interpretation of the ANSI Device Numbering standard, and GE's interpretation is very close to the ANSI standard. For example, a "4" device is a 'master protective' or 'master control' element/device. A "14" is a speed level element/device. A "33" is a limit switch element/device. A "63" is a pressure sensing element/device (a pressure switch). A "94" is a shutdown (normal shutdown--not emergency shutdown, trip) element/device. A "5" is an emergency shutdown element/device (like an emergency stop pushbutton). It's VERY helpful to commit the ANSI device number standard to memory--or at least have a copy available when you are reading application code. It's critical, actually, and as you get more familiar with the device numbers you will begin to recognize certain functions/signals. You can find lots of ANSI lists on the world wide web with your preferred Internet search engine.

Please be very careful when copying information from Toolbox:

>Here comes the L4S:<pre>
> L4SX L14HR L14Y L4s
>-----||-----------||----------||------------( )</pre>
>L4SX = Master protective set auxiliary
>L14HR = HP zero speed signal
>L14Y = time delay loss of master
>protective(not sure if it is needed)

It should be:<pre>
L4SX L14HR <b>L4Y</b> L4S
-----||-----------||----------||------------( )</pre>
L4SX = Master protective set auxiliary logic
L14HR = HP zero speed signal
<b>L4Y</b> = time delay loss of master protective

Let's deal with the L14HR signal. "14" means it's a speed level signal, "H" means it's related to the HP shaft (and since your turbine is a single-shaft, I presume) the shaft is the HP shaft, and "R" means the shaft is 'at rest' or, 'at zero speed.' So, L14HR is a logic "1" when the turbine-generator shaft is at rest, or at zero speed.

L4SX, as you rightly noted, is important. In this case the "X" means it's a 'permissive' or 'auxiliary' to L4S. In earlier versions of GE digital Speedtronic turbine control systems, there was a limit of eight "elements" horizontal string in a rung. That means there could be no more than seven contacts (normally open and/or normally closed) and one coil--for a total of eight. So, if there were more than seven contacts necessary it was required to split the rung into multiple rungs, and name the "permissive" rungs by applying an "X" suffix to the main logic signal name. Since most of the application code was copied from earlier control systems, this was continued. And on some turbines, there may be additional contacts in either or both of the L4SX and L4S rungs.

So, in your mind you could envision the following:<pre>
---| |---| |----| |------|/|-----| |----| |-----( )</pre>

because it's exactly the same as the combination of L4SX and L4S.

Let's tackle L4Y. When you see a "Y" suffix on a signal name it almost always means an "inverse time delay" to its namesake, in this case L4. If you look up L4Y in Toolbox, it will look like this:<pre>
L4 L4Y
1.0 sec</pre>

By "inverse time delay" you can see that 1.0 second after L4 transitions from a logic "1" to a logic "0" (which is the inverse of a logic "1") that L4Y will be a logic "1". Or:

"1.0 second after L4 is <b>NOT</b> a logic "1", L4Y will be a logic "1"."

[If you look up L4Z, it will look like this:<pre>
L4 L4Z
-----| |-------------------------(T)
1.0 sec</pre>

A "Z" suffix almost always means a "time delay" to its namesake, in this case L4.

L4Z is read:

"1.0 second AFTER L4 <b>IS</b> a logic "1", L4Z will be a logic "1"."]

Let's read the combined L4SX/L4 rung:

"When L3RS is a logic "1" AND L1X is a logic "1" AND L33CSE is a logic "1" AND L63QT is NOT a logic "1" AND L14HR is a logic "1" AND L4Y is a logic "1", then L4S will be a logic "1"."

L3RS = Ready to Start
L1X = Auxiliary to START command
l33cse = Clutch-Starting Engaged (driven by a discrete (contact) input
L63QT = Low L.O. Pressure TRIP
L14HR = HP Shaft at Rest
L4Y = L4 Inverse Time Delay

I think (I hope!) that the signal you wrote as L33CSE was actually l33cse (all lower-case alpha characters), because that's GE's "standard" for signals that associated with inputs and outputs--to express the signal names in lower-case characters. It's not required--but it's intended to be helpful to people reading the application code to let them know those particular signal names with lower-case alpha characters are associated with physical inputs or outputs. (Again, GE doesn't always follow their own standards. But we learn to live with it.)

So, now, let's re-read the combined L4SX/L4S rung:

"1.0 seconds after L4 goes to a logic "0" AND the 'Ready to Start' signal is a logic "1" AND the START command auxiliary is a logic "1" and the Starting Clutch is engaged AND there is NOT a Low L.O. Pressure Trip and the shaft is at zero speed, then L4S will be a logic "1"."

In order to be able to start a GE-design heavy duty gas turbine, the operator interface must display a "READY TO START" indication. Most HMIs have a "Start-Check Permissive Display which can be used to determine why a Ready to Start indication is not being displayed. (We are going to go through all of that--just be patient.) Part of the 'Ready to Start' indication is that there are no turbine trips (L4T is a logic "0")--remember L4T from the L4 rung. (A "3" is a complete sequence element/device. In this case, the ready to start sequence is complete.) L3RS will be a logic "1" when the start-check ("<b>R</b>eady to <b>S</b>tart) sequence is complete.

When the "READY TO START" indication is seen AND when the operator selects START and executes the command then L1X will pick up (a "1" is a starting element/device per ANSI & GE). When L1X picks up, that starts several auxiliaries, including the Aux. L.O. Pump, which pressurizes the L.O. system after a few seconds (if the Aux. L.O. Pump is not already running). L1X is an auxiliary signal that is a logic "1" when a turbine START is active.

Also, L1X usually starts the Hydraulic Ratchet pump to engage the starting clutch, which should cause actuate limit switch 33CS-1 after the clutch halves are engaged to make l33cse go to a logic "1". The signal name tells you when it's going to be a logic "1"--when the starting clutch (the jaw clutch on most machines; a SSS clutch on some units) is <b>E</b>ngaged.

When the low L.O. pressure switches at the collector end of the generator (the furthest point away from the L.O. pumps) are actuated--meaning they have sensed minimal L.O. pressure--then logic signal L63QT will go to a logic "0". A 63 is a pressure sensing element/device, mostly a pressure switch per GE's interpretation. When you see a "Q" in a logic signal name it means oil, or flow. In this case it means L.O., and the "T" means 'Trip.' L63QT will be a logic "1" when the L.O. bearing header pressure--as sensed by the switches at the collector end of the generator--is below the minimum pressure required to protect the bearings. Again, the signal name tells you when it's going to be a logic "1"--when the L.O. Bearing Header pressure is low enough to initiate a turbine trip. In this case, we want it to be a logic "0" (because we need the normally closed contact to be closed to get power flow through the contact) so we want L63QT to be NOT a logic "1"--which is the same as saying when the L.O. pressure is NOT low enough to initiate a turbine trip. And, again, that will happen several seconds after the Aux. L.O. pump starts and pressurizes the bearing header and the pressure switches at the collector end of the generator are all actuated.

L14HR is a logic "1" when the HP shaft is at <b>R</b>est.

L4Y is a logic "1" 1.0 seconds after L4 is NOT a logic "1".

And that's L4S (and L4SX).

Regarding your confusion about L4, it's the signal that says "It's okay to run the turbine" when it's a logic "1". And when it's NOT a logic "1" (when it's a logic "0"), then the turbine is to be tripped. If you look at how many places in the application code L4 is used, you will--it's a LOT! Probably the most used signal, because it's the permissive to start run a lot of auxiliaries, to energize a lot of solenoids, to permit fuel to flow, and on and on. L4 must be a logic "1" to burn fuel in the turbine, and if it changes from logic "1" to logic "0" the fuel flow-rate will be stopped--tripping the turbine.

This logic signal, L4, should <b>>>>NEVER<<<</b> be forced to a logic "1". Never. Ever. NEVER. Not even when the turbine is at zero speed. If this signal is forced to a logic "1" when the turbine is not running, a lot of auxiliaries may start and run, and a LOT of alarms will be annunciated, and in the worst case, fuel might be accidentally admitted to the combustors. It's just not ever permissible to force L4. Even if some GE document tells you to do so--it's wrong. And, it should absolutely NEVER be forced to a logic "1" when the turbine is running--because if L4T is picked up (by, say, low L.O. pressure, or high vibration, or loss of flame, or exhaust overtemperature, etc.) the turbine will NOT trip. So, to anyone reading this: Never force L4 to a logic "1"--whether the turbine is running or not.

I hope it's becoming clearer how to "read" logic signal names. In general, the signal name tells you when the logic is going to be "1". And, armed with this information, it should be easier to read application code.

Don't try to think that you have to learn everything right now. You have this information, you can print it, and you can review it any time you wish. And, it's always available on

I don't believe you've told us what turbine you are working on at your site.... And what fuel(s) it burns.... And what kind of starting means it has....

Your next assignment is to examine L3RS. Post the rung, and the logic signal names from Toolbox, and write your sentence of how the rung works.

One of the contacts should be L3STCK, and the L3STCK rung should have several contacts like L3STCK0, L3STCK1, L3STCK2, etc. Please post the L3STCK rung from your Toolbox file.

Thank you sincerely for your detailed reply. Your comment makes it easier for me to read application code.

Firstly, i should make some clarifications:
1) The gas turbine is MS6001B,burning syngas.
2) The signal name of l33cse is in lower case letter.

This time my assignment is to examine L3RS, and i find it a little complicated.
Here comes the L3RS:<pre>
L3RS1 = Start check stop ready to start logic1;
L3RS2 = Start check stop ready to start logic2;
L5VPRO_LATCH = Protective VPRO card trip_latch(Never heard it before)

SO,when Ready To Start 1 signal(L3RS1)and Ready To Start 2 signal (L3RS2) are logic 1,and
L5VPRO_LATCH is 0, the L3RS will be logic 1.

It will be necessary to examine L3RS1 and L3RS2, and this time i'll examine L3RS1.<pre>
L3STCK = Startup check stop stop start check permissive.

And the signal L3STCK:<pre>

L3STCK0 = Startup check stop start check logic 0;
L3STCK1 = Startup check stop start check logic 1;
L3STCK2 = Startup check stop start check logic 2;
L3STCK3 = Startup check stop start check logic 3;

In conclusion, the L3RS is divided in L3RS1 and L3RS2, and L3RS1 is devided in L3STCK0, L3STCK1 ,L3STCK2 and L3STCK3.

I was wondering what kind of signal will be grouped in L3STCK0, L3STCK1, L3STCK2 and L3STCK3 respectively.

L3STCK0(start check logic 0 for units without third fuel and no HRSG by pass)(?)<pre>
-------||----------||----------|/|----------|/|----------|/|-------|/|-------|/|------------( )
| |
| |
L27BZ | TRUE |
L27BN = AC BUS normal;
L27BZ = AC BUS undervoltage(confusing);
l3cp = startup check stop customer permissive to start (difference between l3cp and L1X?);
L86TCI = compressor inlet thermocouple disagree (not clear);
L3IGVFLT = IGV valve position servo trouble;
L86MP = startup check stop master prot startup lockout (not sure);
L5ESTOP1 = state of E-STOP 1 (confusing);
L45FP_STCK = Fire protection start permissive (not clear);

SO, if any of L27BN and L27BZ is logic 1, AND l3cp is logic 1, whenL86TCI, L3IGVFLT, L5WSTOP, L5ESTOP1and L45FP_STCK are logic 0, then L3STCK is logic 1.

I'll start with the signal L3IGVFLT. The L3IGVFT :<pre>
--------||------------( )
The signal L3IGVF1 and L3IGVF are outputs from L86GVTV2 block,and my interpretation is when the csgv is <31 and >35, the L3IGVFLT is logic 1.

It seems that the L3IGVFT signal will function only before gas turbine startup. (am not sure for i cannot figure out the signal L3IGV).

I am poor in electric knowledge, so the other signals are not easy for me. But i did what i can, i looked up the ANSI device numbers;

27: undervoltage relay
86: lockout relay

1. Does that mean signal names with 27/86 are related to the status of certain relay?
And what's the difference between undervoltage relay and lockout relay?

2. Can you explain the word lockout for me ,because it's hard for me to understand.
If the above questions are too easy, just let me know, and i'll try to find answer on my own.

I would like to continue next time,is it ok? Because it is not easy for me :}

And i'll try to do better next time.

Should i wait for you comment and guidance or just continue to examine the signal L3RS?

Best regards.

I am working on my response; L3STCK0 is a fairly extensive rung.

Please check the "sense" of L45FP_STCK (normally closed or normally open). It is confusing me.

It might be a good idea to put the rung in your reply.

Please take your time and be certain of the case of signal names and the sense of contacts.

Thanks for the information about the turbine at your site. What is the source of the syngas? Does the unit start and stop on syngas, also, or does it use another fuel (natural gas or liquid fuel (distillate) for starting and stopping and you have to transfer to syngas at some speed/load?

As for questions being too easy, sometimes the easiest ones are the hardest ones to answer. I learn a lot from teaching, and I like to say, "There is no such thing as a dumb question--just dumb answers."

A 27 may be related to a protective relay, or it might be related to the presence/absence of voltage in one of the MCCs or the battery charger. Undervoltage is undervoltage; it indicates the loss of or the lack of a minimum level of voltage for some condition.

A lock-out condition is one that requires some manual intervention to reset. The intent is to make certain that someone is aware of the problem and has taken appropriate action to resolve the condition before the alarm condition is unlatched, and many times before the turbine can be re-started.

An 86 device is a lock-out--which may be a relay, or it may be a latched alarm/trip condition as described above. An 86 lock-out relay is a relay that must be manually latched (against spring pressure that would otherwise open the latch). When one of several other relays operates (when a serious condition is detected), it operates an electric coil that unlatches the relay.

There are many different generator/transformer protective relays or relay functions each with its own unique ANSI device number.

So, if 2 of the 3 the VPROs in the <P> core initiate a trip, from the signal name L5VPRO_LATCH I would presume they are latched in the trip condition and L5VPRO_LATCH would go to a logic "1". Once the trip condition is resolved and cleared to reset the alarm and unlatch the trip one probably has to issue a 'Master Reset' from the HMI to make L5VPRO_LATCH to go back to logic "0"--which is the good condition. When L5VPRO_LATCH is a logic "1" the "READY TO START" indication can't be achieved.

Yes; the next logical rung would be L3RS1, which is "driven by" L3STCK, which is driven by the four start-check rungs: L3STCK0, L3STCK1, L3STCK2 and L3STCK3 (for your machine). If there wasn't a limit to the number of horizontal rung elements there would only need to be a single rung--probably L3STCK, but since there are well more than seven start-check permissives there has to be more than one rung, with "auxiliary" rungs used to get all the start-checks to finally pick up L3STCK.

And the next logical step would be to analyze L3STCK0. Though, at one point you listed L3STCK0 as "Startup check stop start check logic 0" and in another place you listed it as "start check logic 0 for units without third fuel and no HRSG by pass". Where did you find the two descriptions? (I guess the second description was from a 'Comment' in the application code.?.?.?)

I think you might have been in a great hurry when typing your response.... I think the L27BN and L27BZ should have been l27bn and l27bz and the contact for L45FP_STCK should have been normally open, not normally closed. This is because both l27bn and l27bz are usually associated with individual discrete (contact) inputs from relays in the generator protection panel.

And, if the signal name L45FP_STCK was chosen correctly, then it should be a logic "1" when the Fire Protection System is capable of detecting a fire, tripping the turbine, and discharging fire extinguishing agent. And we want the Fire Protection system to be capable of all of the above. If L45FP_STCK is a logic "1" when it's "ready" then the normally closed contact you drew would be open, and L3STCK0 would not got to a logic "1"--which is what we need to get a READY TO START.

l27bn = is a logic "1" when the Bus Voltage is Normal
l27bz = is a logic "1" when the Bus Voltage is NOT Normal (the "Z" in this case means 'inverse')
l3cp = logic "1" when the 'Customer Permissive' to START is a logic "1"
TRUE = ??? This is one I'm not familiar with, but I <b>think</b> it's always a logic "1"--ALWAYS
L86TCI = is a logic "1" when the redundant axial compressor inlet thermocouples disagree (have a wide discrepancy) by more than a setpoint
L3IGVFLT = is a logic "1" when an IGV Fault has been detected
L5WSTOP = (you didn't provide a description, and I'm not familiar with this one--at all)
L5SETOP1 = is a logic "1" when one of the E-STOP Push-button inputs to the <P> core is actuated (open)
L45FP_STCK = is (should be) a logic "1" when the Fire Protection Start-check Permissive is satisfied (unless the signal name was chosen incorrectly)

So, let's read the rung:

"If the bus voltage <b>IS</b> normal OR the bus voltage is <b>NOT</b> normal, AND if the Customer Permissive to Start <b>is</b> a logic "1" OR TRUE <b>is</b> a logic "1", AND the compressor inlet thermocouples DO <b>NOT</b> disagree with each other AND there is <b>NOT</b> an IGV fault AND there is <b>NOT</b> an L5WSTOP AND the first E-Stop input to the <P> core is <b>NOT</b> actuated AND the Fire Protection Start-check Permissive is <b>NOT</b> [???] satisfied, then L3STCK0 <b>will be</b> a logic "1".

The l27bn/l27bz parallel combination is a way to be sure the bus PTs are both sensing voltage or both NOT sensing voltage. (The "bus" is the voltage on the other side of the generator breaker--the grid voltage, the "running" voltage as it's often called.) The two logic signals should each be driven by separate discrete inputs (which is why I put the signals names in lower-case letters) from a bus under-voltage relay that is picked up when both PTs sense bus voltage, or when both PTs DO NOT sense bus voltage. If there is a problem sensing bus voltage there should be a Process Alarm to alert the operator to the condition.

Note that a complete lack of bus voltage does not prevent a start; many times for testing there is no bus voltage when a turbine is started, or some turbines have the ability to close on to a dead bus. So the lack of bus voltage does not prevent a start; but when there is no bus voltage (a complete lack of bus voltage) there is a process alarm for that to alert the operator to the condition, but, again it does not prevent a start. A recent thread on indicated some turbines start with a dead bus, and then close on the bus with a "dead" generator, and then apply excitation and slowly increase generator terminal voltage to "charge the line" (the "bus" line).

And, this is a good time to mention that every start-check permissive should be accompanied by a Process Alarm to alert the operator why a 'READY TO START' is not being displayed--<b>EXCEPT</b> one signal. And we'll get to that one soon enough.

The l3cp/TRUE parallel combination is a way to bypass a 'Customer Permissive to Start' discrete input which usually exists in most all Speedtronic systems. It might be from the DCS saying the HRSG drums are full and heat can be applied to the HRSG from the gas turbine exhaust, or it just might an operator selection from the DCS saying they are prepared for a gas turbine start. (This signal does not trip the turbine when the turbine is running if it goes to a logic "0"; it just must be a logic "1" to start the turbine. If it's not, there should be a Process Alarm to alert the operator there is no Customer Permissive to Start.) The TRUE is the software "bypass" to l3cp; if there is no discrete input connected to the Speedtronic it will never change state, so the software TRUE (which is ALWAYS a logic "1") is used to bypass l3cp when there is no signal from the "Customer." In the "old" days, we used to just put a hardwire jumper on the l3cp discrete input terminals to permanently make l3cp a logic "1"; this is just a software equivalent of that. The signal could have been just deleted from the rung; TRUE completely baffles some people since it's not explained in any GE documentation--anywhere.)

L86TCI a logic signal that warns the operator of a problem with the axial compressor inlet thermocouples, usually signal CTIF-1, CTIF-2 and sometimes CTIF-3 (Compressor Temperature-Inlet, Flange 1, -2 or -3). There is some logic in the application code that checks to make sure the three signals are within an allowable range of each other; if not, then a gas turbine start is not permitted. This is because axial compressor inlet temperature is used for biasing HP shaft speed for some very important functions (the resultant signal is TNHCOR, Turbine Speed-HP Shaft, Corrected). So, if the temperature measurements are not reasonably close to each other then there is a good possibility that TNHCOR can be adversely impacted which could affect turbine operation.

The "86" in the signal name is an indication that the condition requires a Master Reset to unlatch the alarm once the condition has been corrected. Signals with "86" in the name are generally called "lock-out" signals. Most do require a Master Reset; some do not (unfortunately; another occasion when GE doesn't adhere to its own standard).

L3IGVFLT is a logic signal that we can spend more time on if you wish. As you noted, it's checking to make sure the IGVs are within reasonable limits of where they should be prior to starting. The minimum mechanical stop on the IGV ring is set for approximately 32 DGA, so the lower limit should be something slightly less than that--and 31 DGA is a reasonable number. The minimum operating angle during starting is 34 DGA and the IGVs shouldn't be much more open that prior to a start, so 35 DGA is also a reasonable number. This is just a check to make sure the IGVs are in a reasonable position prior to starting (presuming the LVDT feedback is properly calibrated). The IGVs are one method to limit axial compressor surge/stall during starting by keeping them "closed" (34 DGA or so is considered closed). (Axial compressors are very unusual and have some very unique requirements.)

The description you provided, probably from Toolbox, can be misleading. A servo problem might be the cause of this logic signal being "1", but, it might just be that the IGVs were left in some position other than what they should be after a maintenance outage, or the IGV LVDT calibration is inaccurate (this happens a LOT, unfortunately)--which is <b>NOT</b> a servo-valve problem, it's a human problem. Sometimes, the IGV minimum mechanical stop is also not set correctly after a maintenance outage when the IGVs have been worked on, resulting in the IGVs going to 30.9 or 30.8 DGA--which is not horrible, but is approaching the limits of acceptability. When a START signal is initiated, the IGV Trip Solenoid, 20TV-1 will be energize and hydraulic pressure will move the IGVs to the minimum operating condition (34 DGA). So, if personnel are CERTAIN the IGV LVDTs are properly calibrated AND that the mechanical stop has not be properly set it's sometimes permissible to change the limit to prevent the alarm. But, any value less than approximately 30.5 DGA means the minimum mechanical stop should be repositioned (it was moved) or the IGV LVDTs have not been properly calibrated--which, again, is <b>NOT</b> a servo-valve problem; it's a measurement and calibration procedure problem which should be corrected before starting.

Also, I would be interested to know if the application code for your turbine requires a Master Reset to unlatch L3IGVF1 and L3IGVF2 once the condition has been resolved. If so, this would be an example of a lockout that doesn't have "86" in the name. (Isn't this fun?)

I would like to see the rung for L5WSTOP, as it's not something I'm familiar with. So, if you would, please, post it to your next reply.

L5SETOP1 is a logic signal that comes from the <P> core when one of E-Stop P/B (Pushbutton) circuits is NOT closed. (E-Stop P/B circuits are normally closed, and when the circuit opens (by pressing the the E-Stop P/B) the turbine is tripped. There are a couple, if I recall correctly, E-Stop P/B inputs so it's easier to identify which E-Stop P/B has been actuated (or possibly a wire in the circuit has come loose--it does happen!). Note again--this comes from <P> core. It is a hardware trip, as opposed to a software trip. It's meant to be a "positive" means of tripping the turbine (by opening the circuit) without having to go through software to trip the turbine. (There are also software trips, but I believe many insurance companies and some technical regulations and standards require a hardware method of tripping the turbine.) Hopefully, we will remember to cover this when we get to the L4T rungs, and other ways to trip the turbine.

L45FP_STCK; this one I believe is either an incorrectly chosen signal name, a typo, or the contact sense (normally open vs. normally closed) is typed wrong. The signal name indicates this signal will be a logic "1" when the fire protection circuit is "ready" to detect a fire, trip the turbine and discharge fire extinguishing agent (CO2 or water sprays). If the signal name is correct--the logic is "1" when it's active and in a "ready" condition then the contact sense as shown is wrong. If the contact sense is correct (from the application code), then this is an example of someone not choosing a signal name correctly....

Let's work on L3STCK1 next, but, also please post the L5WSTOP rung, and if the sense of the L45FP_STCK contact is correct as you have typed it, please post the L45FP_STCK rung, also. And, please check on the case of signals l27bn and l27bz; if they are indeed all capitals, then please check to see if they are driven by discrete inputs; this could also be another example of GE not adhering to their own standards.... (What's the purpose of having standards, eh?)

Glad to continue this finding journey, and thank you angain for your detailed reply.

First i have to apologize to you for several unnecessary mistakes i made. So, i'll start with some clarification:

1)The sense of L45FP_STCK is normally open.

2)There is no signal named L5WSTOP,it should be L86MP.(weird mistake,sorry again)

But the signal L27BN and L27BZ do in capital letter.

SO the L3STCK0 rung should be:<pre>
| |
| |
L27BZ | TRUE |
L27BN = AC BUS normal;
L27BZ = AC BUS undervoltage;
l3cp = startup check stop customer permissive to start ;
L86TCI = compressor inlet thermocouple disagree ;
L3IGVFLT = IGV valve position servo trouble;
L86MP = startup check stop master prot startup lockout(confusing) ;
L5ESTOP1 = state of E-STOP 1;
L45FP_STCK = Fire protection start permissive ;

I Checked the source of signal <pre>
true true
l27bn1---------||---------L27BN; l27bz1---------||---------L27BZ;</pre>
In your reply, you mentioned if the signal is driven by a discrete(contact) input it should be in lower case letter.
I am confused about the meaning of the word 'discrete'.

Do you mean that if one signal is directed from sensors it's name should be in lower case letter?

I am still a little confused about relay, as my understanding, a relay functions like a switch,when DO is logic 1, then it closed and the circuit is powered with electricity.

The (undervoltage) relay and (lockout) relay confuse me. Take signal L27BN as an example, how does a undervoltage relay related to L27BN?

Yeah, the application code in my turbine requires a Master Reset to unlock L32IGVF1 and L3IGVF2.

Last thing to clarify the second description of LSTCK0 do comes from Toolbox,and i find it misleading.
Here comes the LSTCK1:<pre>
l26qn L28FDSCK L430 L14HR L86HD l12hblt L3STCK1
l26qn=Lube oil tank temperature low low;
L28FDSCK=Flame detector trouble;
L430=Off modo selected;
L14HR=HP zero speed signal;
L86HD=Hydraulic protective trouble;
l12hblt=overspeed bolt trip;

If the 'lube oil tank temperature is lower than setpoint' AND 'any one of the flame detector detect no signal of flame before start AND 'off MODE' is not selected'AND'the HP-shaft' is at rest AND 'Hydraulic protective is not logic 1 'ANd'overspeed bolt trip is not logic 1', then L3STCK will be logic 1.

As to signal l26qn, does it from a temperature switch which outputs 1 when the temperature is lower than setpoint? I think the temperature of lube oil affect its performance.

It seems the signal is related to the fuel servo system ,and i think it is important. Should i examine it next time?

I am confused about L430 signal, does it an input from HMI? And just to normally shutdown gas turbine?

Up to now, i was confronted a lot of signals and i find it hard to sort it out? What should i focus on through this journey?

I will give a brief introduction of gas turbine on my site some time later. I will prepare it.

Best regards.

I reread the your response below.

"The l27bn/l27bz parallel combination is a way to be sure the bus PTs are both sensing voltage or both NOT sensing voltage. (The "bus" is the voltage on the other side of the generator breaker--the grid voltage, the "running" voltage as it's often called.) The two logic signals should each be driven by separate discrete inputs (which is why I put the signals names in lower-case letters) from a bus under-voltage relay that is picked up when both PTs sense bus voltage, or when both PTs DO NOT sense bus voltage. If there is a problem sensing bus voltage there should be a Process Alarm to alert the operator to the condition."

I think there are two PTs measuring the bus voltage,both of them output its logic signal through the status of a relay. The relay outputs signal l27bn is closed when the bus voltage is normal and the other relay that outputs signal l27bz is closed when the bus voltage is below setpoint.

So it is the hardware way of judging the status of bus voltage.
It can also be done in software way by implementing the compare block.

Is it right?

Best regards

Whoa, there. Let's not get too far ahead of ourselves.

You had asked about relays and discrete inputs (and by extension, discrete outputs) and you're talking about relays here like you have a good understanding of them.

A discrete input (or output) to a programmable control system (and the Mark VI is a purpose-built programmable control system--built, primarily, for the purpose of controlling GE-design heavy duty gas turbines) is one that is either "on" or "off", "1" or "0", "true" or "false", "energized" or "de-energized", "picked up" or "dropped out", "close" or "open", and so on. It has only one of two possible states.

A relay is simply a device for conveying information via contacts, that are either open or closed, when the relay is energized or de-energized. Relays usually have more than one set of contacts. The contacts can be electrically separate (have no common electrical circuits), or they can have electrically common contacts (so-called "c-form" contacts that have a normally open and normally closed terminal and share a COMmon terminal). In the case of an undervoltage relay, it is sending a discrete signal to the Speedtronic about the presence, or absence, of voltage. Additional contacts of the undervoltage relay may be used in other circuits for the same purpose (to indicate the presence or absence of voltage). Relays are pretty simple devices, however, they require a power source (110/220 VAC; 24 VAC; 24 VDC; 125 VDC; and many others depending on the manufacturer and application). They are electromechanical devices, and as such, they can fail when the electric coil fails, or when the contacts weld closed or the spring fails.

Now to your present question about L27BN and L27BZ. GE and it's packagers typically use two PTs to sense bus voltage, and two PTs to sense generator terminal voltage--both arranged in what's called an open-delta configuration. If any phase voltage is lost, or if there is a failure of any single PT then, in this case (the bus voltage) l27bn1 (I believe you said it was at your site, that drives L27BN) will not pick up. Conversely, both PTs have to indicate there is NO bus voltage before l27bz (which I believe you said drives L27BZ at your site) will indicate a complete loss of bus voltage. There really is no point in starting a gas turbine that drives a generator if there is a problem with the bus voltage because it is very likely that the gas turbine will have to be stopped to sort the problem.

As for whether or not the checks could be done with a COMPare block, well, yes. But then one would need to have additional AC voltage input circuits (at additional cost and complexity) when simple relays will do the job more than adequately by providing discrete inputs.

Discrete inputs have historically been much cheaper to implement than analog inputs (voltages; currents). Unfortunately, sometimes discrete inputs don't have as much "information" available (such as the actual voltage levels, versus a minimum or zero voltage level, or even a maximum voltage level in some cases).

Also, historically it was much easier to predict the failure mode of most types of discrete input devices (pressure switches; temperature switches; etc.). And GE is all about reliability and in the past transmitters weren't so reliable and could fail full on as well as full off.

Also, historically providing power (24 VDC or 110/220 VAC) for transmitters was expensive, as were transmitters, themselves. As was the cost and complexity of analog input circuits. For reliability, it would be necessary to have redundant analog inputs which would also increase cost and complexity. So, mostly discrete inputs were used--and very successfully, too. Older Speedtronic panels (pre-digital versions) only had limited analog inputs and outputs: speed (frequency, and prior to that voltage); temperature (thermocouples--more reliable than RTDs and don't require additional power supplies, though they did require cold junction compensation circuits (be VERY thankful we don't have to do THAT any more!); one or two pressure transmitters (fuel gas intervalve (P2) pressure and axial compressor discharge pressure (CPD: Compressor-Pressure, Discharge); LVDTs; servo-valve outputs. Almost every other signal was a discrete input or -output.

And this brings me to say things need to slow down just a little, well, maybe a little more than that. I have just been given a rush assignment at work and I need to devote time to that (since I get paid at work, and I don't get paid for this endeavour). Also, you have asked several questions about terms and words and we haven't had a chance to cover them in any detail yet. We need to keep some focus here so that we don't get ahead of ourselves. We need to be sure that all questions get answered before we jump to another topic or rung.

Control systems have gotten more and more sophisticated over the last few decades. And they have gotten both cheaper--and more expensive. I can recall when USD140,000.00 for a Mark IV Speedtronic turbine control panel was a LOT of money--and it wasn't all that capable as a programmable control system. Now, Mark VIes are nearly two- to three times as much, but they do have a LOT more capability, and they are much more expandable (inputs or outputs can be "easily" added to the control system (easily is a relative term--much easier than the Mark IV!).

But, one has to remember that as the control systems changed and improved over the decades, GE-design heavy duty gas turbines basically remained the same. The turbine designers and packagers were loathe to change inputs and outputs (lots of changes to layout drawings, and schematics, and manufacturing practices and sourcing and warehousing and parts lists and manuals). It can become VERY expensive to change inputs and outputs just because the control system capability has improved.

And, it's one of the things that have made GE-design heavy duty gas turbines so reliable and so desirable: They don't change very much. They are pretty constant things in a very fast-changing world. They are RELIABLE--when they are operated and maintained properly by people who are properly trained and have access to quality parts and services. The control philosophies GE has employed over the years have proven to be very robust and very scalable (meaning that as GE-design heavy duty gas turbines have gotten larger and more powerful the control systems were able to be expanded to maintain similar operation and protection philosophies and methods). So, when you ask, "Couldn't this be done with a COMPare block?" one has to think back to how things have evolved to the present condition when someone can ask a question like that.

Imagine how many wires would be have to be used to interconnect all these devices to a single control system. Imagine how complicated and large such a control system would have to be. I don't know about you, but I <b>LOATHE</b> European-style, high-density terminal boards--even when appropriate sized wires are used on them. I prefer to have terminal boards that have to use a voltmeter easily and without fear of shorting or grounding, and where I can see terminations visually and where I can handle wires for troubleshooting and repair purposes. And, you're proposing even more wires and more terminal boards and more printed circuit cards and more complexity.

Anyway, I'm enjoying our time and this journey, but I have to limit my time for the next couple of weeks. As I said before--don't try to ingest all of this right now, immediately. That's the great thing about this forum: it will be here for a long time. And the things that are written here will be here for a long time. And you can refer to them as your understanding grows. And you should be reading and re-reading them over time as your knowledge and experience grows. It's a LOT of information to try to take in all at once--we haven't even touched on inversion masking! And we haven't finished the start-check permissives. And we haven't started the trips!

So, let's dial it back a little bit, and let me get caught back up at my day job, and finish one rung before we start another. There's always time for more questions.

As for that second description of L3STCK0, GE is trying to use the same sequencing with slight modifications for all heavy duty gas turbines. Makes sense, right? If it works, why change it? And it works. Has for decades. Again, it's one of the things that make GE units so desirable and so reliable--the operation and control philosophies have not changed much over the years and it's very reliable.

That Comment from the application code is supposed to be for the requisition engineers who put the application code together in the factory. Sometimes (often) those engineers don't really have any field experience at all, and they make silly Comments. That's a silly Comment.

As for L3STCK1. You made a pretty huge mistake on the first signal, l26qn. Why would you want to start a turbine if: "the lube oil tank temperature is lower than setpoint" (your words from the sentence for L3STCK1)?

l26qn is a logic "1" when the L.O. Tank temperature is normal. It's driven by a discrete input from a temperature switch, 26QN-1. The switch contacts are closed when the tank temperature is "normal" (whatever the switch setpoint is from the Device Summary drawing, usually above approximately 60 deg F). And when the switch contacts are closed, l26qn will be a logic "1". This discrete input is NOT inverted. (We'll get to inversion masks shortly, but I want to get you started thinking about and anticipating the concept.) There will be a Process Alarm to indicate the L.O. Tank Temperature is Low.

L28FDSCK is a signal that is a logic "1" when none of the flame detectors is indicating flame when the unit is not running. And, one doesn't want to start a turbine when one or more of the flame detectors indicates flame when fuel isn't flowing.... There will be a Process Alarm to say the flame detectors are not in a ready-to-start condition.

L43O. The contact is normally closed, and L43O is a logic "1" when OFF mode <i>is</i> selected, so to have power flow through the normally closed contact L43O would have to a logic "0", or when OFF mode <b>is NOT</b> selected. Here's the <b>ONLY</b> start-check permissive that doesn't have a Process Alarm to alert the operator. Why? Because one doesn't need a READY TO START indication if the unit is in OFF mode. In other words, to get a READY TO START indication the operator has to select CRANK, FIRE, AUTO or one of the other available modes, but when OFF is selected the unit can't be started anyway. It's not a alarm condition when OFF mode is selected. So, operators just have to know: When they want or need a READY TO START they have to be in some mode other than OFF.

L43O doesn't come from the HMI. The HMI sends a command to the Speedtronic to switch to OFF mode, and if all of the permissives are satisfied to allow OFF mode to be selected then logic signal L43O is set to a logic "1". The HMI doesn't do any control or protection--it just sends commands to the Speedtronic, and monitors the operation of the unit (via signals from the Speedtronic), and, of course--it displays alarms from the Speedtronic. LOTS of alarms. Too many alarms, most would say--but that's because of poor configuration and/or poor maintenance or lack of understanding or just lack of response to alarms. Most operators, and their supervisors/managers, just don't pay any attention to any alarms as long as the turbine runs and doesn't trip. They think the Speedtronic will protect the turbine and trip it if it's necessary--and for the most part, it will. But there are some conditions that require operator action, and most operators (and technicians--and operations supervisors/managers) don't know when an alarm is indicative of a possibly larger problem. And, that's just poor training--and lack of personal motivation--both of which are permitted in too many plants.

L14HR is a logic "1" when the unit is at rest, or when it's at zero speed. MANY people look at the logic that drives L14HR and they see that it doesn't actually change state at 0.000 percent speed, but something slightly higher, and they think that's wrong (??!!!??!) and so they change it to 0.000% and then the Speedtronic doesn't work. Passive speed pick-ups aren't very good at very low RPMs and that's why the speed level isn't set at exactly 0.000% (0 RPM). I think it's set for something like 3 RPM, which is just fine for this purpose--which is to protect the jaw clutch from being damaged if it's engaged when the turbine shaft is spinning. That's (probably) why your unit has to wait to get down to "zero" speed to start--because it has a jaw clutch which shouldn't be engaged when the turbine shaft is still spinning (and 3 RPM isn't very fast!).

L86HD. This is a condition and Process Alarm that confuses many people--because of the alarm message text: Hydraulic Protective Trouble. For present-day GE-design heavy duty gas turbines this alarm has NOTHING to do with the Hydraulic System. It is related to the Trip Oil system, and the presence of Trip Oil pressure in the system when there should NOT be pressure. It is a lock-out--signified by the "86" in the signal name, which means that when the Trip Oil pressure condition is resolved a Master Reset is required to unlatch the alarm, and permit a READY TO START.

l12hblt is driven by a discrete input (the signal name is all in lower-case letters), and it's a logic "1" when the mechanical overspeed bolt is actuated. This input <b>IS</b> inverted, which means it is a logic "1" when the contacts of the overspeed bolt are open. As explained, the contacts of the L.O. Tank temperature switch, 26QN-1, are closed when the tank temperature is normal, and l26qn is a logic "1". It is NOT an inverted input. In this case, when the contacts of 12HBLT are open l12hblt is a logic "1"--which is the opposite (inverted) of what one would expect would happen.

Why does GE use inverted logic? It's a poor-man's method of something called "contact supervision." What's contact supervision? It's a method whereby the control system monitors discrete input circuits to ensure a wire in the circuit has not come loose, or has not been cut/damaged. Most contact supervision circuits use "end-of-line" resistors placed across the physical switch contact terminals and the control system is always checking to see if there is current flowing in the circuit--whether or not the switch contacts are open or closed. More current flows when the switch contacts are closed, and only a small amount of current flows when the switch contacts are open, and when no current is flowing the control system alarms to alert the operator to a bad circuit/wiring.

GE decided decades ago to standardize on circuits that were closed under normal operating circumstances, and opened to alarm or trip the turbine. To detect a closed contact, the circuit has to be intact (the wiring has to be good; the terminations have to be good) in order for current to flow--that's how the control system detects closed contacts. Open contacts result in no current flow.

Let's say the overspeed bolt limit switch contacts were configured to close when the overspeed bolt actuated (because of a turbine overspeed). And, let's say that one of the wires of that circuit had come loose from its crimp and had fallen away from the terminal. If the turbine oversped and the bolt limit switch contacts closed the control system would never detect the contact had closed--because of the disconnected wire. No current would flow in the circuit when the overspeed bolt contacts closed, so the control system would never know the contacts had closed. NOT good.

Now, let's say the overspeed bolt limit switch contacts were configured to open when the overspeed bolt actuated. Under normal circumstances, when the overspeed bolt limit switch contacts were closed current is always flowing in the circuit and the turbine thinks everything is just fine. And, let's say the turbine is running just fine at 73.4% of rated load, and suddenly one of the wires in that circuit vibrated loose from its crimp and became disconnected (it doesn't have to be a sudden disconnection--it just sounds better to say "suddenly"...). Even though the overspeed bolt limit switch contacts are still closed (there was no overspeed), because the circuit is now open there is no current flowing--just like if the switch contacts had opened. If the control system interprets no current flow as bad then it can take appropriate action, alarm and/or trip. In this way, the operator, and the technician, are aware of a bad circuit--could be a failed switch, just as easily as a loose wire, but the circuit is not intact and that's not a good condition under any circumstances.

So, GE's answer to ensuring that circuits are in good condition (before there was even contact supervision available in any programmable control system) was to "invert" the input. In this way, the signal goes to a logic "1" when the external circuit is open (no current flowing). The signal names are generally chosen to indicate when there is a problem (alarm or trip), and this matches the inversion philosophy--a logic "1" when the external circuit is open.

This is big one: inversion masking. Lots of people--myself included for years--used to say it was done to be "fail-safe." That's one way to think of it (not a good way in my opinion, because it doesn't really describe how or why--it's just a quick answer that sounds good).

Now, some people will say, "Well--if the turbine alarms or trips every time a wire comes loose, that's not very reliable!" And, they would be right. Wires should not come loose, but they do. Switches should not fail, but they do. So, GE's answer to that was to use redundant switches. It's not likely that two wires will come loose at the same time, or two switches will fail at the same time. But, if either did happen, then that's certainly a condition that should be resolved. But, if a wire comes loose or a switch contact fails in one of two redundant circuits and the logic (application code) requires two open circuits to alarm or trip the turbine on a serious fault then the turbine won't trip. GE also wrote logic to alert the operator/technician to a "disagreement" between redundant discrete inputs so the problem could be investigated and resolved.

So, it's about reliability. Making sure that circuits and switch contacts are serviceable and can and will alarm or trip on serious trouble. Inversion masking is something that generally takes some time to "absorb" and comprehend, but once one internalizes the concept and the practice its something that helps understand and improve lots of other aspects of control and automation. And, can help explain a lot of nuisance alarms and trips on other systems where a standard wasn't adopted and followed (and it happens a LOT in many control systems!).

Re-stating the sentence for L3STCK1:

"When the L.O. Tank Temperature <b>IS</b> normal AND when the flame detectors <b>ARE</b> in a ready-to-start condition <b>AND</b> when OFF mode <b>is NOT</b> selected <b>AND</b> when the turbine shaft <b>IS</b> at zero speed (at rest) <b>AND</b> when there <b>is NOT</b> a Trip Oil pressure problem <b>AND</b> when the mechanical overspeed bolt limit switch <b>is NOT</b> de-actuated <b>THEN</b> L3STCK1 <b>will be</b> a logic "1"."

That should take care of L3STCK1.

One more comment about l2hblt--it doesn't trip the turbine. This logic signal does <b>NOT</b> (and SHOULD NOT) trip the turbine. It's simply a limit switch that indicates when the mechanical overspeed bolt has been actuated (which de-actuates the limit switch). The mechanical overspeed bolt will "dump" Trip Oil pressure--which is what trips the turbine. The limit switch just says, "The mechanical overspeed bolt is NOT actuated," or, "The mechanical overspeed bolt is in the RESET position, ready to indicate when the bolt has been actuated." Which is the position the mechanical overspeed bolt needs to be in to start the turbine.

Why shouldn't this limit switch be used to trip the turbine in the Speedtronic turbine control panel? Because, it's a single limit switch--there's no redundancy. If it fails, vibrates loose, or a wire is disconnected, then if this input is used to trip the turbine--it will trip the turbine. It's extremely rare that a single discrete input will be used to trip the turbine--extremely rare. It reduces reliability (using a single switch). The only purpose of this switch is to indicate the overspeed bolt has been actuated--or, that is has not, as it's being used in the L3STCK1 rung.

Now, have we satisfied all your queries about relays and lock-outs and such?

Do you want to spend any more time on L3STCK1 before proceeding to L3STCK2? If not, let's continue.

Do you have any questions about inversion masking? (You can have questions about it later, too. But if you have them now, let's try to answer them now. It's a very important concept, and it's key to troubleshooting and understanding GE turbine control systems--at least the digital ones.)

Let me know if you're okay with what we're doing here. Is there a way to make it easier? Or better?

Very glad to receive your response. I appreciate your patient and detailed reply.

I have been enjoying this journey with you as your comment is always enlightening and helpful.
You are a very patient and excellent teacher indeed.

You asked me whether i am okay with what we are doing, and my answer is yes,certainly.

And you asked me if there is a way to make things to make things easier. I think i should do some adjustments to make this journey easier for you.

1) I think i should focus on the main line,like present i should focus on L4S. I should set some questions that are not important at present aside, as we will have opportunity to discuss those questions latter (I wish:))

2) If your comment wanna to cover every side of a signal ,it will take a lot of time and effort of you, and i think it will disturb your day work. In order to avoid this situation, i think it will be better if you just comment the important information i should know at this present.

3) We could keep in touch every day, but just examine no more than one rung. And i will split my response my reply in several section, and if you are busy, you can just reply one of them,and continue next time.

In general, I think this journey is vary exciting, but it is difficult to start, but i think it will be easier if we continue.

And thanks again for what you have done as i and this forum don't pay :)
1) As to L3STCK1,I made a pretty mistake about l26qn.
I thought that the toolbox description refers to the logic 1 situation.

l26qn=Lube oil tank temperature low.
l26qn L3STCK1

So i think when lube oil temperature is lower than set point ,it will be logic 1.

2) I do have questions about inversion masking, like when we should use inversion masking.Why l26qn is not inversed when l12hblt has to be inverted??? Does it related to the type of relay,NC or NO.How it helps us to read application code.

I think take examples would make things easier to understand.

I think we should keep on moving, and i'll just write down the rung and examine it next time.</pre>
L86CB L4Y L3BHSTCK L30CC_STCK l33cl1 l45fp1 l27f
| |
| |

L86CB = Surge protection trouble startup lockout;
L3BHSTCK = Comp operating limit BH start check perm logic;
L30CC_STCK = Fire protection trouble-GT start inhibited;
l33cl1 = Fire protection aux/turb/laod compt CO2 locked;
l45fp1 = Fire protection aux/turb/laod compt fire pre-detec;
l27f = Fire protection trip relay undervoltage;

Best regards

>1) I think i should focus on the main
>line,like present i should focus on L4S.
>I should set some questions that are not
>important at present aside, as we will
>have opportunity to discuss those
>questions latter (I wish:))

If you have questions, let's try to answer them. If we need to come back to them later, we can. I do believe, though, that the more we continue the more things will become clearer--but, if the questions are preventing you from understanding something let's try to answer them.

>2) If your comment wanna to cover every
>side of a signal ,it will take a lot of
>time and effort of you, and i think it
>will disturb your day work. In order to
>avoid this situation, i think it will be
>better if you just comment the important
>information i should know at this

I think the important thing to get out of this exercise is how to "read" signal names and logic (rungs). DO NOT trust the longname description from Toolbox--just as with l26qn. When you wrote the sentence for the rung, it doesn't make sense to start the turbine with low lube oil tank temperature. You're saying, "If I can't trust the longname description in Toolbox, what can I trust?" Well, you have to analyze the information in front of you, and use any other resources you have to make sense of the situation.

For example, if you went to the VCRC input where 26QN-1 is connected, you would see the input is NOT inverted (it's "Normal"). And, if you went to the Device Summary document, which lists the setpoints for most of the devices provided with the turbine and auxiliaries (but, unfortunately, not all of them....) you would see the switch is (likely) set for 60 deg F (approx 16 deg C) INCreasing, and the NO (Normally Open) contacts of the device are to be used for the indication. This means that as the L.O. Tank Temperature increases above approximately 60 deg F the NO contacts will close. (The Device Summary should also list the DECreasing setpoint, something around 50 deg F, for a deadband on the switch of approx. 10 deg F. This means that once the L.O. Tank temperature had risen above 60 deg F, closing the NO contacts, that the L.O. Tank temperature would have to drop below approx. 50 deg F to cause the NO contacts to open, or "reset" as many people like to say--meaning go back to the "un-actuated" state.)

So, knowing that the input is NOT inverted, and that the switch contacts are closed above approximately 60 deg F, one can deduce that the longname description of the signal is wrong--that is, it does NOT describe when the signal name is going to be a logic "1".

You asked why this input is not inverted, but l12hblt is inverted. The input could have been inverted, and it would work just fine. Of course, the signal name would be exactly opposite of what it should be because it would be a logic "1" when the L.O. Tank temperature was less than normal. Sometimes these things, even though they could go one way or the other (with a signal name change to make it easier) are just continued for decades--and this is one of them.

I believe if you will post the L26QN_ALM rung (I'm guessing at the alarm signal name; it will likely be different--unless I am very lucky today) you will see it looks something like this:<pre> l26qn L26QN_ALM
-----|/|-------------( )</pre>
This rung, if it matches your application code, is a violation of another one of GE's control philosophies: That the logic that drives an alarm should be a logic "1" to make the alarm active. Do two wrongs make a right? In this case, it seems to. ;-) We learn to live with it.

A former colleague of mine used to say, "This ain't rocket science." And, it's not. Does it have to be exact? No. Should it be exact? It would certainly help. Is it critical to proper operation of the unit? No. Would it take much to fix it? No--but, at GE there is no one who really understands all of stuff and is in charge or reviewing it to be sure it's accurate and consistent and so long-standing mistakes like this just keep getting repeated, decade after decade. Even if someone realizes it should be changed, that person likely is afraid to do so--because most of the people writing the application code have little or no hands-on, field experience.

And yet, someone has changed the longname signal description--because for decades it was always "Lube Tank Temperature Normal", and in the application code at your site it's "Lube oil tank temperature low" (which is the alarmed condition of the input--the one that will prevent a READY TO START....). Again, the people who really do understand this stuff at GE realize--it's not critical. It's always been that way. Should it be changed? Probably. But, "If it ain't broke, don't fix it." The circuit has worked for decades even if it's not exactly as per philosophy, so it just gets neglected.

Anyway, I can't read the next rung as you wrote it--you clicked on 'Submit Post' with a mistake in the html tags (you used /pre BEFORE and after the text; it needs to be pre before, and /pre after). Please correct the text and re-submit. You also forgot to put L4Y's longname description in the list of signals in the rung.

One of the reasons I'm spending a little more time on signal descriptions is that a lot of other people read these posts, and it might help them to understand GE-design heavy duty gas turbine operation and control philosophy if there is a little more detail. So, please be patient with me.