Markvie - servo output current fault

We have two Markvie retrofit (TMR). one of the unit, we changed out gcv, srv and lfb servo valves during maintenance and carried out lift calibration. after this we started getting, L3GRVFLT Gas ratio valve position servo trouble and L3GRVSC_ALM Stop Ratio Valve servo current fault alarms appear. PCAA_1G2A card parameters show unequal. Build and download done.(We are using PCAA core analog card for servos)
These alarms disappear.

L3GCVFLT GCV servo trouble and L3GCVSC_ALM GCV servo current fault alarms appear. Again PCAA_1G2A parameters build and download. But fault did not reset.

Controller RST power turn off and on. Now GCV and SRV faults appear together.

No permissive to select Gas fuel.
GCV calibration tried after forcing L3ADJ1. Selected Regulator1  calibrate valve Calibrate.

<b><i>moderator's note:</i> &#61664;</b> is some sort of special character that does not show up correctly.

Window appears: Auto calibrate: Could not achieve calibration mode. Check if the calibration mode permissive is set.

Same case for SRV

LVDT position and Excitation values are normal in ToolBox
1Servo Output tab shows current of 0.15 % for 65GC (GCV) and 0.11% for 90SR (SRV). Whereas 65FP fuel bypass valve current shows 36%. (65GC and 90SR servo currents were around 30% before this trouble alarm)

GCV, SRV valve movement checked by removing servo coil wires from PCAA cards and injecting mA externally.

By injecting 1mA current, GCV valve opened fully from each coil.
Same way SRV also opened with 1 mA current in each coil.
LVDT feedback in TCAT card in full close position is around 0.70 volt for GCV and SRV. Also LVDT excitation value in ToolBox shows 7.2 V.

At this time, 65FP signal also lost and L3LFBSC_ALM Liquid fuel bypass valve servo current trouble alarm appear. Also Auxiliary Check (Servo valves) L3ACS.

Several times "Master Reset" done during the above period.

Startup is blocked by L3ACS Auxiliary Check (servos) alarm.
Controller RST power turn off and On and master reset. But situation remains same.

No abnormality in PCAA cards or UCSA controllers. No other diagnostic alarms also.

We downloaded back up software to controller, but result is same.
Is there anyone to help.....Your urgent response would be greatly appreciated....
 
This is another perfect example of the myth that changing servo-valves requires calibration or re-calibration--of anything. A servo-valve is NOT calibrated; only the LVDT feedback from a device which is positioned by an electro-hydraulic servo-valve. And, if there are no LVDTs then the "valve" can't be calibrated (not that the "valve" is calibrated to begin with).

And, before any "calibration" or "re-calibration" one should always check to see if the LVDTs require calibration/re-calibration by comparing the LVDT feedback to the actual position.

There is a logic permissive which must be forced to "1" for each of the servo-valve outputs in order to be able to use AutoCalibrate. I don't have a dongle and working copy of ToolboxST to be able to tell you precisely where to find the logic signal in ToolboxST, but you should be able to use the 'Help' feature to find the signal somewhere in the myriad of PCAA tabs. That should at least allow you to be able to complete a calibration without the alarm "Could not achieve calibration mode...."

But, if you only replaced the servo-valves it is NOT necessary to "calibrate" LVDT feedback.
 
Whilst there is no need to recallibrate the LVDTs if replacing a servo valve an important check that should be performed is to verify the null current of the new servo. The total summation should be about 0.8mA. The easiest way to do this is by selecting the manual position feature in the calibrate screen and open the valve to mid range. The setpoint and actual values should be within 0.5per cent. If the error is greater the null current setting should be trimmed or in worst case the servo replaced again.

However after replacing a servo valve i will normally check each coil independently of the speedtronics by using a Fluke milliamp source to drive each coil one at a time. At the same time the polarity and deadband around the null point can be checked. It's also a good idea to check the valve will close, or go to full bypass, if all coils are open. The worst case I have come across is a valve which went full open ie a positive bias current.
 
309Eguy,

Let's be clear about the order of events to be taken when replacing a servo-valve on a GE-design heavy duty gas turbine:

1) Make sure the Speedtronic null bias current value is set to 2.67% (for TMR Mark V, Mark VI, or Mark VIe Speedtronic turbine control panels). This may mean changing the value and downloading it to the Speedtronic and re-booting the processors, as necessary. This is necessary to set the null bias current value to the default, starting position for a new (or refurbished) servo-valve.

2) Replace the servo-valve.

3) Verify the polarity of the servo current being applied to each individual coil of the servo-valve. (This has been documented in several threads on control.com.)

4) Verify the LVDT feedback is nearly equal to the actual device position. This means, measurements have to be taken on the device and compared to the LVDT feedback.

5) Presuming the LVDT feedback is nearly equal to the actual device position, if the LVDT feedback (actual device position) is not nearly equal to the reference position then make an adjustment to the null bias current value (which will require downloading, re-booting, and then confirming the change produced the desired result). Any adjustment should not exceed the range of servo current adjustment.

There are several things to note in this procedure. First, negative servo current increases the flow of fuel or air to the turbine. Positive servo current decreases the flow of fuel or air to the turbine. So, null bias current<b>when viewed on a Mark V, Mark VI, or Mark VIe display--</b>will be <b><i>negative</b></i>. Negative null bias current is required to overcome the null bias spring tension to keep the device at a steady state position.

Second, it's unfortunate that GE chose to use a positive value for null bias current in the configuration fields of the Speedtronic programming software, and let the Speedtronic invert the value when applying it to the servo-valve output. So, when you see a value in Toolbox, for example, of 2.67%, that means the Speedtronic will apply -2.67% to each servo current output.

Third, for Mark V, Mark VI, and Mark VIe GE chose to display servo current in percent of rated current, where rated current is -10 mA to +10 mA. So, 1% equals 0.1 mA; 10% equals 1 mA; and 100% equals 10 mA.

Fourth, Null bias current has an allowable <b>range</b>: -0.8 mA, +/-0.4 mA, or from -0.4 mA to -1.2 mA. Or, expressed in percent, -8%, +/-4%, or from -4% to -12%. Null bias current isn't a fixed number; we're dealing with the force required to overcome the tension of a mechanical spring and that's not always precise, though it's usually pretty constant assuming hydraulic oil temperature and the ambient temperature around the servo valve are relatively constant. [NOTE: For a TMR control panel, the null bias current is divided by three for each control processor. So, -0.8 mA becomes -0.267 mA, or -2.67% per processor.]

Finally, I have a couple of questions. Why use an external milliamp source to verify servo current polarity? Why not use the Speedtronic servo current? If you don't use the Speedtronic servo current, then shouldn't you check the polarity of the Speedtronic current at some point to make sure that it's the same as the polarity of the external source you used to perform the check?

And, last, why are you introducing a new "variable" (deadband) in this discussion? What do you mean by "...deadband around the null point..." in this context? Why does it need to be observed or documented? What is an acceptable "deadband" and what is an unacceptable "deadband"?
 
Point 1 thru 5 all sound perfect to me. The original poster appears not to know this. Other than telling him no need to calibrate the LVDT's, I thought I would tell him something useful.

> Finally, I have a couple of questions. Why use an external milliamp source to
> verify servo current polarity? Why not use the Speedtronic servo current?

Several reasons, 1st our test equipment is designed for task, routinely calibrated and certified, the speedtronics is not. 2nd it's an independent device to check against, if find the update rate in toolbox to slow. 3rd If we are replacing a servo i like to connect locally at the junction box and observe the valve/igv with my own eyes, rather than rely on toolbox trends. Once this is all done and reconnected i check with the speedtronics with the valve manually positioned to a mid range typically 50%. In turn i disconnect other two cores to check polarity of each core. If the valve closes you have a polarity issue.

> And, last, why are you introducing a new "variable" (deadband) in this discussion?

Call it what you want, but once you placed the valve to remain stationary at say a current of 0.8mA there is a dead band either side of this where the does not move. In practice i find this to be just under 100mA. This a gut feel (if that's allowed here) based on experience and familiarity but is another good reason to inject current locally when next to the valve and observe the response. As you rightly point out the null current is also dependent on the valve it's connected to, not only the servo.

This is the extent of checks I believe available when changing a servo. Based on your experience you can decide to cut steps out.

So what checks do you do when replacing a servo? They are not plug and play, walk away, hope for the best.
 
Advice was offered on how to possibly correct whatever situation was caused by the unnecessary attempt to calibrate the "servos". I do understand why people think servos are calibrated, but it's just not true--it's a myth. And it's myths like these that cause problems like the original poster is experiencing. I'm just trying to use this post as yet another example for others of what kinds of problems can be self-inflicted when performing unnecessary "calibrations" when replacing servos.

Nothing more and nothing less.

The original poster said he also used a milliamp source to confirm the servos worked. He was asking why he couldn't get AutoCalibrate to work to try to correct his 'Not Ready to Start' problem.

I often wonder if backing-up Mark VIe files actually includes LVDT calibration values, or, if the Mark VIe determines that if newer LVDT calibration values are present in the SPVO/PCAA when downloading older LVDT calibration values if it ignores the older values.... I have very limited experience with Mark VIe and ToolboxST but have had this same issue posed to me several times and have developed this question I hope to answer in the not-to-distant future.

As for the work you are talking about performing, first, using an external source to drive the servos, and second, using the Speedtronic to do the same thing again, that's a lot of extra work. And I'm not about creating any extra work for myself, or anyone else. I have seen people do this in the past, and when they don't observe the polarity of the current they are applying with the external source and the valve doesn't behave the way they think it should, they just cavalierly change polarity to get the device to do what they think it should--and declare everything okay. Then when they reconnect the Speedtronic and they have mismatched servo currents (because of one incorrect servo coil polarity) they don't understand how that could have happened since they just "tested" it.

And this bit about "purpose-built" devices being used to drive servo-valves is very disturbing. No; the Speedtronic servo-valve output wasn't designed to have control over the current output value, and yes the external source is designed specifically for that purpose. But, the servo is designed to be driven by the Speedtronic servo-valve output and not by an external source. And since all one is trying to do is verify polarity, why would one disconnect two leads from the Speedtronic to connect an external source (which may or may not be corrected with the proper polarity) and then reconnect the Speedtronic and have to perform the same polarity test again?

Lastly, since the servo-valve output range is only from -10 mA to +10 mA, I don't see how a "deadband" of 100 mA could exist around a null point.

As for what checks I do when replacing a servo, well Steps 1 through 5, though I rarely find it necessary to modify null bias current values from 2.67%. RARELY. Just recently I found it necessary to modify the null bias of an existing servo to more than 4% (-0.4 mA per processor) and suggested to the Customer they replace the servo, and upon replacement the servo worked perfectly with the default 2.67%, and some nagging Diagnostic Alarms and some intermittent 'device not following reference' Process Alarms were resolved as a result.

You're right; replacing a servo-valve is not plug and play. But it does NOT require "calibration" using AutoCalibrate, either. And the sooner this myth gets busted the better--for everyone except Moog and the companies that refurbish perfectly good servos that get replaced for no reason other than, "Well; let's replace the servo to see if that will stop the turbine from tripping on high exhaust temperature spreads!". (And I'm continually told that very reason when asked why a servo was replaced. And replacing a servo valve <b>NEVER</b> resolves a high exhaust temperature spread issue.)
 
> don't see how a "deadband" of 100 mA

CSA, Clearly an error, I meant to say 0.1mA.

I don't see where the extra work is in my method. I start at device and work back to the speedtronics. Check the valve and servo, check field wiring, check speedtronics connection, check speedtronics control, maybe perhaps 30minutes of extra work if needed. I dont do it every single time and with experience and knowledge you know which steps to cut out depending on problems encountered and symptoms observed.

I've shared what i wanted to say. Hopefully the original poster will learn something from the comprehensive responses. I'll leave the last word to be had by your good self.
 
Dear CSA and 309EGUY,

not to take this post too far off course but I do want to chime in on 309EGUY's comments regarding servo deadband. On our site I also manage a Alstom steam turbine that uses Moog servos and a Balluff LVDT assembly for feedback. The Balluff converts the LVDT into a 4-20ma feedback for the controller. The ABB controller does not have any autocalibrate feature or trending so everything has to be done manually. Like 309EGUY I use Fluke 789 or 744 to source current to the servo, and us another fluke to measure feedback.

I do this anytime we replace a servo to measure its null current requirement and also to see what kind of deadband there is around the null current setting. This gives me a chance to see how the new or rebuilt servo compares to the one being replaced. In this case most of the time we are swapping these to coincide with a maintenance cycle. I do not recall what type of deadband current reading I have seen (its in my notes), but there is definitely some there. I figure it gives me some feeling of if the servo might be sticking due to varnish issue or otherwise.

Once I can see how the servo operates then I hook back up with the controller and set null current and check offset and gain to get the valve where I need it.

Anyway just throwing my 2 cents in. I do like the speedtronic trends and autocal features, but I think its nice to play and fully understand what the MKVI is doing behind the scenes sometimes With more of this I think we could quickly dispel some of the myths out there. But that takes information and training that ain't there for everyone unfortunately.

You guys have a great holiday, hopefully your not spending it holed up in a PEECC somewhere!!!
 
Dear CSA, 309Eguy and MIKEVI

Thank you all for the comments on valve lift calibration. It is really useful and AS CSA has pointed out, the myth of valve calibration should be taken off from everyone's mind.

All of you were focused on calibration, I did not get any response to my problem.

Anyway I did some more trouble shooting and all the parameters are normal now.

I removed the PCAA card power and ethernet cables and fixed back again. software back up done again. after this, all the alarms reset and servo currents show again.

Thanks you.
 
iqbal,

Thanks very much for the feedback. I find it very difficult to believe that there were NO Diagnostic Alarms or Diagnostic LEDs lit on the PCAA cards when the servo output current was, apparently, suicided. (Suiciding means the output terminals are shorted together to prevent any output current from being applied to the servo coil.) I've been told GE has been putting a lot of time on a project to reduce nuisance Diagnostic Alarms, but this condition would not (should not, rather) be one of them. If there is some condition resulting in the suiciding of the servo output currents that must be alarmed (logically speaking, of course; sometimes GE doesn't use conventional logic, though).

It's also interesting to note that when you downloaded an existing configuration (ostensibly including I/O Configuration parameters), the problem did not clear. Again, this leads me to believe that either LVDT calibration data is not saved to the master .tcw file when calibrations are performed, or something like the PCAA "rejects" older calibration data if it has newer calibration data, or something like that. One would think that by downloading what is believed to be a known, good configuration and rebooting that the incorrect calibration values would be overwritten, and <b>from the description provided</b> that does not appear to have been the case.

Or, it might just be that some bit got stuck during downloading and re-booting and needed to get unstuck with a "hard re-boot". Many people are VERY impatient with downloading and don't pay attention to alarms and error and warning messages in the Status pane during downloading.

Anyway, thanks again for the feedback. Sorry we weren't more helpful with the problem you had--which was caused by "The Myth".
 
dear iqbal and all,

i have the same problems in my plants. the unit trip by this alarms

servo fault trip L3ACS-ALM
sec gas control valve trouble L3GCVSFLT_ALM

i am do the troubleshooting. every thing normal except when make the calibration (all forced signal true).

this alarm appear:
auto calibrate:could not achieve calibration mode
check if the calibration mode permissive is set.
 
Top