MarkV Timers and Counters Inaccurate

S

Thread Starter

SMS

One of our GE frame 6B gas turbine MarkV control started inaccurate counting (maybe from 2015) just after reboot R, S and T controller at the same time. Inaccurate parameters are:

1. Total Active Energy (for example: it counts 0.1MWH instead of 35MWH within 1hr)

2. Total Reactive Energy LEAD

3. Total Reactive Energy LAG

4. Total Gas Fuel Flow

During last outage on January 2018, I downloaded "ALL" to r, s and t controllers with corrected (by GE) TOTD_Q, but there is no change. Is this a hardware issue, RAM or EEPROM? Other than that the turbine is running fine. All controllers having A7 status right now.
 
It's been a VERY long time, but I recall there was a not-so-well-known feature of the Mark V operator interface that had a special card in the CPU box that pulses could be connected to and used to increment counters. If I recall correctly, there was a configuration file that was used to say how many pulses were equal to how many [engineering units] for the various inputs.

Does one of the operator interfaces (do you have <I> or GE Mark V HMI (running some version of MS-Windows and CIMPLICITY)?) have a card installed in the CPU box that has a multi-headed cable connected to it? (You may have two cards with multi-headed cables if the operator interface has a time synch card, also.) Are the Megawatt hour meter and VArhour meter and fuel gas flow transducer connected to the card (not the time synch card)?

Next, you will have to look in the F:\UNITn directory to find the configuration file and examine it to see if it's working correctly.

Also, check the F:\OPTIONS.DAT file to see if the option has been properly enabled.

That's about all I can remember; sorry. It certainly seem like--from the information provided--that there is something amiss with the either the input card, or the configuration file for the input card, or the way the TOTD_Q.DAT file is configured. This option wasn't used very often, and when it was used it was generally done so by one of the GE MAs (Manufacturing Associates--John Brown; Thomassen; BHEL; etc.) and they generally had lots of problems making them work properly. Sounds like you were lucky for some time, and then something took at turn for the worse.

Hope this helps! I'll check some of my old documentation when I get a chance and if I find anything I'll pass it along. So, check back once in a while for the next few weeks; I'm traveling a lot and won't be home much to check my archives (paper archives!!!).
 
Thanks CSA for your help.

1. There is no special card in MarkV <I> (Primary Operator Interface) computer for timers and counters. It is DOS based. It is connected to MarkV control panel through ARCNET. We have local and remote <I> computers, both <I> shows inaccurate accumulated readings for analog signals which actually come from r or s or t or designated controller RAM, isn't it?

2. We have following Totalizer files in the unit directory:TOTD_B.DAT, TOTD_Q.DAT, TOTDQ.DAT, TOTT_B.BAK, TOTT_B.DAT, TOTT_B.SRC, TOTT_Q.BAK, TOTT_Q.DAT, TOTT_Q.SRC.I don't know to which file you want me to look at?

Signal list for the totalizer given in the file TOTT_Q.SRC:<pre>
; Specify 1,2,3, or 4.
; types 1 (=counter)
; 2 (=timer)
; 3 (=analog accumulator)
; 4 (=accumulated logic) ;adds a constant value to an accumulator when
; pickup of a trigger logic is detected.
; Must keep signals in order.
;
; Signal Trigger level for
;Type Name analog type (opt)
;---- ------------ --------------------------------------------------------
1 L30CMIS ; ACCUM_01 Manually Initiated Starts Counter
1 L30CTS ; ACCUM_02 Total Starts Counter
1 L30CFLS ; ACCUM_03 Fast Load Starts Counter (Not Used)
1 L30CFS ; ACCUM_04 Fired Starts Counter
1 L30CES ; ACCUM_05 Emergency Trips Counter
1 L30CCT ; ACCUM_06 Customer Trips Counter
1 L30CEO ; ACCUM_07 Exhaust Overtemperature Trip Counter
1 L30CRL1X ; ACCUM_08 Load Rate > 50% Base Load Counter
1 L30CRL2X ; ACCUM_09 Load Rate > 90% Base Load Counter
2 L30FT_I ; ACCUM_10 Total Isochrone Fired Hours (Not Used)
2 L4DE ; ACCUM_11 Diesel engine running time
2 L30FT_T ; ACCUM_12 Total Fired Timer Counter
2 L30FT_P ; ACCUM_13 Peak Fired Timer Counter
1 L30DW60 ; ACCUM_14 Normal Start Counter with Load>60%
1 L30DW80 ; ACCUM_15 Trips Counter with Load>80%
2 L30FT_PR ; ACCUM_16 DLNOx Primary Mode Fired Time
2 L30FT_LL ; ACCUM_17 DLNOx Lean-Lean Mode Fired Time
2 L30FT_S ; ACCUM_18 DLNOx Secondary Mode Fired Time
2 L30FT_PM ; ACCUM_19 DLNOx Premix Mode Fired Time
2 L30FT_EL ; ACCUM_20 DLNOx Extended Lean-Lean Mode Fired Time
3 FQG 128 ; ACCUM_21 Totalized Gas Fuel Flow
3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
3 DVARP_TOT 384 ; ACCUM_23 Total Lead MVARS hours
3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS hours</pre>
Note: Only analog accumulators are not working properly that I described in the earlier thread.

3. Yes, rewriting (with manually calculated value from 2015 to 2018) of TOTD_Q.DAT file done by BHEL-GE. After reboot of all processors I have seen the above four analog values changed to new written value. But they don't work accurately while machine runs.

4. Steps what I did: I copied new TOTD_Q.DAT file and pasted to unit data directory (f:\Unitn) then I downloaded ALL (which formats EEPROM first) to r, s and t. I stopped all three processors and then 1st restarted <R> and waited until A7, then I restarted <S> and waited until A7 and so on. Do I need to restart three Processors at a time?

Waiting for your support.
 
Hi,
Sometimes during a "Hardboot" - simultaneous power off of all controllers - counters or/and timers lose there MSW or LSW (Most/Least Significant Word) set. And you get different reading. Therefore its always advisable to do an upload (EEPROM UP) to TOTD first.

If you have SET_CNTR.exe file in your <I> machine you can try following procedure:

1) Printout/printscreen of all accumulators (MSWs&LSWs) through LFORCE

2) Write in command line: SET_CNTR TOTD_Q.DAT and insert desirable numbers. Maybe you will have to convert decimals to binaries.

3) EEPROM DOWN <Unit> <R,S,T> TOTD

4) Hardboot the system.

Jolek.
 
SMS,

Thanks for the clarifications.

Yes; the Designated Voter is where the Timer & Counter Values you see on the <I> display come from.

(The ARCNet-based network is called the StageLink.)

You have supplied the correct file. (The .BAK files are the back-up Intel hex-format files created when MK5MAKE.BAT is executed.) The .SRC files are what MK5MAKE.BAT compiles into the Intel hex-format .DAT files that get downloaded to the EEPROM of each processor. The _B files are just kind of "place-holders;" they don't don anything but they have to exist, basically just for symmetry purposes.<pre>
>3 FQG 128 ; ACCUM_21 Totalized Gas Fuel Flow
>3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
>3 DVARP_TOT 384 ; ACCUM_23 Total Lead MVARS hours
>3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS hours</pre>
What were the Analog Trigger Level values BEFORE BHEL recalculated them?

When the analog counters stopped working several years ago, was a PROM change done at that time or since that time? My thinking here is that a programming error in a new set of PROMs might have caused a problem. But getting GE to fix that is going to be difficult, if not impossible; they don't want to support the Mark V any more.

Have there been any changes to the CSP (Control Sequence Program) where the triggers are written to? If so, what was done.

I NEVER recommend powering-down the Mark V panel by shutting off the DC power supply breaker to the panel. It's just NOT good for the power supplies in the Mark V or the electronics. And, by the same token, it's not a good idea to re-apply power to the panel by leaving all the individual <P> core switches/cables plugged in and ON and the closing the DC power supply breaker to the Mark V panel. That's even worse, in my experience. (I know; a LOT of people who attended GE Mark V training saw the instructors do that in the lab, and did it themselves in the lab--but there was NO I/O connected to the panels in the lab, and there was no potential for 125 VDC battery grounds in the lab.)

Your re-booting sequence was fine. One of the problems with the Timers & Counters scheme in the Mark V is that the values in EEPROM only get written to once per hour (approximately). So, sometimes getting the EEPROM values to update properly can be a little problematic--not often, but in some cases.

I was always taught to start powering up the Mark V by powering-up <T> first, waiting until it reached A7 and then waiting for about 2 minutes, and then powering up <S> and waiting for about 2 minutes AFTER it reached A7, then powering-up <R> and waiting about 2 minutes after it reached A7 and then powering-up <C>. That always worked very well for me, and there's no rhyme or reason why it should--but it did. I think it's just patience--the waiting after the most recently booted processor gets to A7 before applying power to the next processor is key. What's six or seven more minutes, to allow all communications between processors to get sorted?

That's about all I can think of. If there was PROM change at the time the analog counters stopped working it's likely that there is a bug in the firmware on one of the PROMsets (probably on the DCC/SDCC card) that has caused the problem.

Hope this helps--at least a little bit!
 
It would be VERY unusual to find SET_CNTR.EXE in a G:\EXEC file.

I found this in some old documentation:<pre>Type 3 - Analog -
Time Averaged Analog Accumulator.
Configured with Variable Control Signal.
Totalizes the result of dividing the accumulated minute averages for the
designated Variable by the trigger level.
Trigger level is expressed in counts.
Writes RAM every minute.
Writes EEPROM at top of the hour.

Trigger Level Equation

trigger_level units
counts = ------------------- * 32768 counts/min *(carry_units)
full_scale_value units

Examples:
mwatthours
counts = ----------- * 32768 counts/min *(60min/hour)
mwatt

lbs
counts = -------- * 32768 counts/min *(min/60sec)
lbs/sec

If Engineering units are in 'unit' per second:

trigger_level units
counts = ---------------------------------- * 32768 counts/min
full_scale_value units * 60 sec/min

If Engineering units are 'unit' per minute:

trigger_level units
counts = ----------------------------------- * 32768 counts/min
full_scale_value units * 1 min/min

If Engineering units are 'unit' per hour:

trigger_level units
counts = ------------------------------------ * 32768 counts/min
full_scale_value units * 1/60 hour/min

Small trigger levels give large resolution.
Do not use Zero (0) for a trigger level.
Do not use a negative number for the trigger level,
remember, to totalize, ADD, not SUBTRACT.
Do not use a trigger level greater than 32767 counts,
it is considered as the two's complement which is a negative number.

For example, to find trigger level to totalize gas fuel flow every 60
pounds:

From UNITDATA.DAT get scale type for FQG:

FQG 4998 002 0076 0000 0000 0001 0060 170C

From ENGLISH.SCA get maximum scale value:

#scale_data 76 64.0000 0.00000 2 #/sec LB_S
FQG - max scale 64 #/sec

60 #
counts = --------------------- * 32768 counts/min
64 #/sec * 60 sec/min

= 512 counts

In F:\UNITn\TOTT_Q.SRC:
3 FQG 512 ;Fuel Flow Gas</pre>This as also in the file:<pre>
Type 4 - Meter -
Digital Driven Analog Accumulator.
Configured with Logic Control Signal.
Totalizes the result of the trigger level multiplied by the number of
0 to 1 transitions of the designated Logic.
The trigger level is expressed in counts.

transition incremental value
counts = ----------------------------------
desired display incremental value

Keep trigger levels small to avoid filling up accumulator.
Do not use Zero (0) for a trigger level.
Do not use a negative number for the trigger level
remember to totalize, or ADD, not SUBTRACT.
Do not use a trigger level greater than 32767 counts
It is considered as the two's complement which is a negative number.
Writes EEPROM at top of the hour.
Maximum transition rate of 1 hertz.
Available with:
A systems - 5.12.
B systems - 6.02.

Example:
4 L33WATTHR 10 ;Megawatt-hours

Configuring Accumulators -
F:\UNITn\TOTT_Q.SRC
1 L30CMIS ;Manual starts
2 L30FT_G ;Gas fired hours
3 FQG 512 ;Fuel Flow Gas
4 L33WATTHR 10 ;Megawatt-hours</pre>
Then, after reading this, could it be that the animated display file is incorrect?<pre>Displaying Accumulators -
Use Animated Display Item fx1 or fx2:
fx1 xxx yyy color f2_name x2_name gain dec.
32 bit counter from two 16 bit Control Signals:
ACCUM_nn_MSW, ACCUM_nn_LSW
Gain - when counters are not per unit:
1/10th of hours counters - L30FT_G example
60 # s per count - FQG example
Decimal - number of displayed places.
Offset - not provided for obvious reasons.

Displaying Accumulators - <I> Animated Display example:
tx2 050 800 df " MANUALLY INITIATED STARTS "
fx2 +550 +000 df accum_01_msw accum_01_lsw 1 0
tx2 +150 +000 df " COUNTS "

tx2 -700 -075 df " TOTAL FIRED HOURS "
fx2 +550 +000 df accum_02_msw accum_02_lsw 0.1 1
tx2 +150 +000 df " HOURS "

tx2 -700 -075 df " TOTAL GAS FUEL "
fx2 +550 +000 df accum_03_msw accum_03_lsw 60 0
tx2 +150 +000 df " POUNDS "

tx2 -700 -075 df " TOTAL MEGAWATT HOURS "
fx2 +550 +000 df accum_04_msw accum_04_lsw 1 0
tx2 +150 +000 df " MWATTHRS "</pre>So, these are some ideas to look into.

Please write back to let us know what you find!
 
Thanks CSA,

1. Analog Trigger Value: (I don't understand what you suppose to mean)
Before editing values were:
Total Active Energy: May start from 2706939.6 MWH.
Total Reactive Energy LEAD: May start from 1599648.3 MVARH.
Total Reactive Energy LAG: May start from 1911.9 MVARH.
Total Gas Fuel Flow: May start from 305562 Nm3.

Calculated values by BHEL:
Total Active Energy: May start from 2942790.12 MWH.
Total Reactive Energy LEAD: May start from 1599648.3 MVARH. (Kept same as before)
Total Reactive Energy LAG: May start from 1911.9 MVARH. (Kept same as before)
Total Gas Fuel Flow: May start from 1029976542 Nm3.

2. On 13th March, 2016 (earlier I said 2015 in my 1st thread) it was noticed that analog counters stopped working. During CI on 11th March, 2016 AFPCS (96CS-1) was configured from CTBA analog input channel no. 4 to CTBA analog input channel 12 due to fault in channel 4.
From:<pre>
C_C_MAI04 AFPCS DPH2O ;C -CTBA-046 Inlet air total press transmitter [96CS-1]
To:
C_C_MAI12 AFPCS DPH2O ;C -CTBA-070 Inlet Air total pressure transmitter (96CS-1)</pre>
MK5MAKE and download (only CFG) was done without any error. Rebooting done by switching OFF all three processors from <PD> switch and then switched ON, one by one.

3. No PROM change done that time.

4. No CSP changed that time. Last time editing of CSP was done on 2010 during Alarm configuration. What I can remember, I got plenty of support from you to configure Alarms in MarkV that time on 2010 (after the change there was no error in analog counters). Here is the link below:
http://control.com/thread/1285664755

5. I never use DC breaker for powering down MarkV, instead I use <PD> switches.

6. On next shutdown I am going to follow your sequence as you mentioned in your earlier thread <T>, <S> and <R>. If I am lucky enough then problem will be fixed!!!! One thing I want to let you know that we have enough spare DCC cards with new EEPROM (U9) with same version.
 
>1. Analog Trigger Value: (I don't understand what you suppose to mean)

The value in the third column is the Analog Trigger Level (see the comments at the top of the file).

>2. On 13th March, 2016 (earlier I said 2015 in my 1st
>thread) it was noticed that analog counters stopped working.
>During CI on 11th March, 2016 AFPCS (96CS-1) was configured
>from CTBA analog input channel no. 4 to CTBA analog input
>channel 12 due to fault in channel 4.
>From:<pre>
>C_C_MAI04 AFPCS DPH2O ;C -CTBA-046 Inlet air
>total press transmitter [96CS-1]
>To:
>C_C_MAI12 AFPCS DPH2O ;C -CTBA-070 Inlet Air
>total pressure transmitter (96CS-1)</pre>
>MK5MAKE and download (only CFG) was done without any error.
>Rebooting done by switching OFF all three processors from
><PD> switch and then switched ON, one by one.

I&#8217;m very surprised that worked, downloading only IOCFG. Because the hexadecimal address of AFPCS changed when it was relocated the CSP really shouldn’t have recognized the change. And, just because you edited IO.ASG that really doesn’t enable or scale C_C_MAI12 ; that has to be done in the I/O Configurator (and C_C_MAI04 needs to be disabled or there will be a Diagnostic Alarm). And, you should have had to download SEQ to <Q> to make the CSP change active, as well as IOCFG to <C>.

>5. I never use DC breaker for powering down MarkV, instead I use <PD> switches.

I don&#8217;t write just specifically to the original poster, but try to put in information for others reading the post so if they’re using bad practices they will think twice before doing so again.

6. I doubt changing the order of the download is going to make a difference.

Can you trend the trigger logic signals AND the accumulator values to see if they are toggling and incrementing properly?

TOTD_Q.DAT is the Timers & Counters data file. TOTT_Q.SRC is the Timers & Counters definition file. I suspect something is amiss with the EEPROM TOTT partitions in <R>, <S> and <T>. Have you downloaded TOTT recently? If not, I would definitely suggest trying that.

Keep us posted!!!
 
Thank you very much for your support.

>>1. Analog Trigger Value: (I don't understand what you
>suppose to mean)
>
>The value in the third column is the Analog Trigger Level
>(see the comments at the top of the file).

I found a discrepancy in TOTT_Q.SRC file of that faulty unit:<pre>
3 FQG +126 ; 128
3 DWATT_TOT +384 ; 384
3 DVARP_TOT +384 ; 384
3 DVARN_TOT +384 ; 384</pre>
+126 is definitely wrong I think. Again + symbol before each analog trigger level is also confusing!!!! May be the correction will be like this:<pre>
3 FQG 128 ; ACCUM_21 Totalized Gas Fuel Flow
3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
3 DVARP_TOT 384 ; ACCUM_23 Total Lead MVARS hours
3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS hours</pre>
>I'm very surprised that worked, downloading only
>IOCFG. Because the hexadecimal address of AFPCS changed when
>it was relocated the CSP really shouldn't have recognized
>the change. And, just because you edited IO.ASG that really
>doesn't enable or scale C_C_MAI12; that has to be done
>in the I/O Configurator (and C_C_MAI04 needs to be disabled
>or there will be a Diagnostic Alarm). And, you should have
>had to download SEQ to <Q> to make the CSP change active, as
>well as IOCFG to <C>.

Correct. All the changes done by LTSA expert that time. <C> IOCFG was modified then as per channel configuration.

>Can you trend the trigger logic signals AND the accumulator
>values to see if they are toggling and incrementing
>properly?

Trend is given below<pre>
1 T4 DWATT MW
2 T4 DWATT_TOT MW
3 T4 LTRUE LOGIC
4 T4 LZZ LOGIC
10-MAR 12:30:59.343 30.7 30.7 1 0
10-MAR 12:31:00.343 30.8 30.8 1 0
10-MAR 12:31:01.343 30.8 30.8 1 0
10-MAR 12:31:02.343 30.8 30.8 1 0
10-MAR 12:31:03.343 30.8 30.8 1 0
10-MAR 12:31:04.343 30.8 30.8 1 0
10-MAR 12:31:05.343 30.8 30.8 1 0
10-MAR 12:31:06.343 30.9 30.9 1 0
10-MAR 12:31:07.343 31.0 31.0 1 0
10-MAR 12:31:08.343 30.8 30.8 1 0
10-MAR 12:31:09.343 30.8 30.8 1 0
10-MAR 12:31:10.343 30.8 30.8 1 0
10-MAR 12:31:11.343 30.9 30.9 1 0
10-MAR 12:31:12.343 30.8 30.8 1 0
10-MAR 12:31:13.343 30.7 30.7 1 0
10-MAR 12:31:14.343 30.8 30.8 1 0
10-MAR 12:31:15.343 30.8 30.8 1 0</pre>
>TOTD_Q.DAT is the Timers & Counters data file. TOTT_Q.SRC is
>the Timers & Counters definition file. I suspect something
>is amiss with the EEPROM TOTT partitions in <R>, <S> and
><T>. Have you downloaded TOTT recently? If not, I would
>definitely suggest trying that.

No. I did not download TOTT last time. TOTT_Q.SRC of that unit is may be wrong and that's why that partition is not working properly. In next available shutdown I'll correct the TOTT_Q.SRC file and then download (TOTT)to the controllers to correct the totalizer partition as well.
 
>I found a discrepancy in TOTT_Q.SRC file of that faulty
>unit:<pre>
>3 FQG +126 ; 128
>3 DWATT_TOT +384 ; 384
>3 DVARP_TOT +384 ; 384
>3 DVARN_TOT +384 ; 384</pre>
>+126 is definitely wrong I think. Again + symbol before each
>analog trigger level is also confusing!!!! May be the
>correction will be like this:<pre>
>3 FQG 128 ; ACCUM_21 Totalized Gas
>Fuel Flow
>3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
>3 DVARP_TOT 384 ; ACCUM_23 Total Lead
>MVARS hours
>3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS
>hours</pre>
Hmmm.... I'm not sure which file you are taking the information above from? There should be a TOTT_Q.SRC from BEFORE the change made by BHEL, and one AFTER the change was made by BHEL. Did BHEL add the plus sign?

You also said that ONE of the units at your site is having a problem. Are they other units with Mark V control systems and the same totalizers? If so, what is the difference in the TOTT_Q.SRC files (the ones BEFORE and AFTER BHEL made the changes on the unit that's not working, versus one of the units that is working)?

>>Can you trend the trigger logic signals AND the accumulator
>>values to see if they are toggling and incrementing properly?

>Trend is given below<pre>
> 1 T4 DWATT MW
> 2 T4 DWATT_TOT MW
> 3 T4 LTRUE LOGIC
> 4 T4 LZZ LOGIC </pre>
Soooo, LTRUE is a signal that's ALWAYS true (a logic "1"; picked-up), and LZZ is a "bit bucket" (something to write to when the state doesn't matter). Why did you trend those signals? Why not the four signals that are giving you problems? You trended ONE of four signals that are giving you problems.... AND, you need to use the MSW and LSW words used for totalizing the data (ACCUM_21_MSW and ACCUM_21_LSW; ACCUM_22_MSW and ACCUM_22_LSW; ACCUM_23_MSW and ACCUM_23_LSW; ACCUM_24_MSW and ACCUM_24_LSW). The trended numbers not make much sense, but it's worth a shot at this point.

And, if there's a Mark V with these totalizers working, what do the trended values look like for that unit?

>No. I did not download TOTT last time. TOTT_Q.SRC of that
>unit is may be wrong and that's why that partition is not
>working properly. In next available shutdown I'll correct
>the TOTT_Q.SRC file and then download (TOTT)to the
>controllers to correct the totalizer partition as well.

TOTT_Q.SRC is the Timer & Counter definition file. If there's something amiss with the definition file, it's probably NOT going to work. If it even exists, TOTD_Q.SRC doesn't do ANYTHING, if I remember correctly (it's just there for "symmetry"; don't ask). The only TOTD_Q.xxx file that does anything is TOTD_Q.DAT (for <I>s) or TOTD_Q.AP1 (for GE Mark V HMIs). And, they will MESS UP EVERTHING IN THE TOTALIZERS IN THE MARK V IF DOWNLOADED WITHOUT HAVING BEEN UPLOADED IMMEDIATELY PRIOR TO THE DOWNLOAD. TOTD_Q.DAT or TOTD_Q.AP1 should almost NEVER be downloaded, unless it was uploaded immediately prior to the download--OR, a fresh, new TOTD_Q.DAT or TOTD_Q.AP1 was obtained from GE or one of its packagers (BHEL, for example).

GE is nothing if not inconsistent; and this is one of those times. That's why it's almost NEVER advisable to download ALL--because it wipes out the EEPROM, and unless the TOTD_Q file has the values from immediately before it is downloaded, it can really mess up the totalizers. And, even FORMAT, if downloaded to all <Q> processors can cause problems; it should only ever be necessary to download FORMAT to a new EEPROM or when something is wrong with a single processor.

I know--NONE of this is documented exactly like above, but the function of each file, and the EEPROM Downloader, is explained in the Mark V Maintenance Manual, GEH-5980 (I'm pretty sure). It just takes reading and re-reading and re-re-reading and re-re-re-reading to make it all make sense. And, in the context of what one is trying or wanting to do. It's one of the things that gets people in a LOT of trouble with the Mark V--doing MARK5MAKEs and downloading, and downloading ALL--when it's NOT necessary or advisable. Just to be seen "doing something" isn't the right thing to do--especially when it can be the WRONG thing to do.

Try to think this through and get all of the proper information from the correct files, and identify which version of which file from what unit (working versus working). Don't be in a hurry--it hasn't worked for almost two years, right? What's a few more days or a couple of weeks even?
 
Thank you very much.

>Hmmm.... I'm not sure which file you are taking the
>information above from? There should be a TOTT_Q.SRC from
>BEFORE the change made by BHEL, and one AFTER the change was
>made by BHEL. Did BHEL add the plus sign?
>You also said that ONE of the units at your site is having a
>problem. Are they other units with Mark V control systems
>and the same totalizers? If so, what is the difference in
>the TOTT_Q.SRC files (the ones BEFORE and AFTER BHEL made
>the changes on the unit that's not working, versus one of
>the units that is working)?

OK. We have 4 Frame 6B units. GT4 analog accumulator is not working. TOTT_Q.SRC is as follow:<pre>
3 FQG +126 ; 128
3 DWATT_TOT +384 ; 384
3 DVARP_TOT +384 ; 384
3 DVARN_TOT +384 ; 384</pre>
But other good unit TOTT_Q.SRC analog accumulators are defined as follow:<pre>
3 FQG 128 ; ACCUM_21 Totalized Gas Fuel Flow
3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
3 DVARP_TOT 384 ; ACCUM_23 Total Lead MVARS hours
3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS hours</pre>
I sent copy of complete <C> directory to BHEL and they replied with corrected TOTD_Q.DAT file. There was no change in TOTT_Q.DAT file of unit specific directory. I copied TOTD_Q.DAT (new one with calculated value from 2016 to 2018) to unit directory and downloaded ALL as per their recommendation. Download All Consists:

FORMAT, SEQ, CONST, IOCFG, UBBL, HIST, EPA, MAOUT, EVENT, CHNG, BOI, TOTT, TOTD, CBLR.

Recently I observed Unit4 (faulty one) TOTT_Q.SRC is not same as other units where totalizers are working fine. To be more sure i rechecked the copied <C> directory that I sent to BHEL, there I got same wrong analog trigger level. Again I checked very old backup (taken on 2010) of that unit directory (unit4) and there I found TOTT_Q.SRC is same as other running units.Using GE maintenance manual GEH-5980 i got below table from that faulty unit.<pre>
CHECK T1 S
CHECKSUM TEST OF UNIT T4 PROCESSOR R: 22-OCT-1991 10:00:51
Partition offset size --------date-------- cksum -id- -status-
SEQ 8000 2786 15-JAN-2018 11:07:20 B626 SEQ OK
CONST 2000 1400 23-FEB-2018 17:30:35 A2C4 CNST OK
IOCFG 4000 0399 08-JAN-2018 20:06:40 D682 IO OK
UBBL 0000 0000 01-JAN-1980 00:00:00 0000 UBL <NO DATA>
HIST 0000 0000 01-JAN-1980 00:00:00 0000 HIST <NO DATA>
EPA 0000 0000 01-JAN-1980 00:00:00 0000 EPA <NO DATA>
MAOUT 0DD0 0084 11-MAR-2016 10:11:00 2E73 4-20 OK
EVENT 0E60 008C 11-MAR-2016 10:11:00 1F46 EVNT OK
CHNG 0F60 0004 11-MAR-2016 10:11:00 CAFE CHNG OK
BOI 12E0 0586 11-MAR-2016 10:11:00 3AC7 BOI OK
TOTT 0C00 00A9 13-MAR-2016 11:13:24 1C04 TOTT OK
TOTD 0200 0902 19-DEC-2017 18:05:58 0000 TOTD NO CKSUM
CBLR 1160 0006 11-MAR-2016 10:11:00 CAFE CBLR OK</pre>
Sorry again. I mentioned I downloaded All on January, 2018 actually date will be 19th December, 2017 and that time GT4 MI was running.

It is seen from the above table, the last change of TOTT was done on 13th March, 2016 at 11:13:24 hrs just before of GT4 startup (after CI) and LTSA expert was working on it. From operation shift diary I found GT4 start command given at 11:50 hrs and synchronized at 12:06 hrs. And on 13th March, 2016 during night shift, analog counters inaccuracy was noticed.

If I change the TOTT_Q.SRC file (same as other units), Compile(MK5MAKE) and download (both TOTT and TOTD)and initialize all three controllers with new TOTD written with corrected value then is there any chance to get the totalizer back in normal condition?

Sir, Waiting for your reply..
 
>>Hmmm.... I'm not sure which file you are taking the
>>information above from? There should be a TOTT_Q.SRC from
>>BEFORE the change made by BHEL, and one AFTER the change
>was
>>made by BHEL. Did BHEL add the plus sign?
>>You also said that ONE of the units at your site is having
>a
>>problem. Are they other units with Mark V control systems
>>and the same totalizers? If so, what is the difference in
>>the TOTT_Q.SRC files (the ones BEFORE and AFTER BHEL made
>>the changes on the unit that's not working, versus one of
>>the units that is working)?
>
>OK. We have 4 Frame 6B units. GT4 analog accumulator is not
>working. TOTT_Q.SRC is as follow:<pre>
>3 FQG +126 ; 128
>3 DWATT_TOT +384 ; 384
>3 DVARP_TOT +384 ; 384
>3 DVARN_TOT +384 ; 384</pre>
>But other good unit TOTT_Q.SRC analog accumulators are
>defined as follow:<pre>
>3 FQG 128 ; ACCUM_21 Totalized Gas
>Fuel Flow
>3 DWATT_TOT 384 ; ACCUM_22 Total MW hours
>3 DVARP_TOT 384 ; ACCUM_23 Total Lead
>MVARS hours
>3 DVARN_TOT 384 ; ACCUM_24 Total Lag MVARS
>hours</pre>


>I sent copy of complete <C> directory to BHEL and they
>replied with corrected TOTD_Q.DAT file. There was no change
>in TOTT_Q.DAT file of unit specific directory. I copied
>TOTD_Q.DAT (new one with calculated value from 2016 to 2018)
>to unit directory and downloaded ALL as per their
>recommendation. Download All Consists:
>
>FORMAT, SEQ, CONST, IOCFG, UBBL, HIST, EPA, MAOUT, EVENT,
>CHNG, BOI, TOTT, TOTD, CBLR.

I'm really having a very difficult time understanding what happened. Every Mark V and Mark V operator interface comes with a TOTD_Q.DAT file; and I don't even believe there is a TOTD_Q.SRC file (because TOTD_Q.DAT is an Intel hex format file of the totalizer data--and the file size NEVER changes, no matter how many times it's uploaded or downloaded or even when one receives a modified TOTD_Q.DAT file from GE or one of its suppliers (such as BHEL). The data in the file may change; the file date/time stamp may change--but the file size never changes. It's always the same. So, when you received an "updated" file from BHEL, even thought it was the same size as the file on the <I> that it is replacing, it <i>should</i> (theoretically if the update was done properly) have different data (timer and counter and accumulator values).

When TOTD_Q.DAT gets downloaded to the control processors (<R>, <S> & <T>), and then they all get re-booted (they all have to be shut down (NOT at the same time--you shut them down ONE at a time, and leave each one off until all three are shut down), and then you power them up one at a time (again--my method is to power up beginning with <T>, then after gets to I/O Status A7 and remains at A7 for approximately 2 minutes then I power-up <S> and wait for it to achieve A7 and remain there for 2 minutes before powering-up <R>).

BHEL should have just copied a working TOTT_Q.SRC (and even the TOTT_Q.DAT) from a working <I> to the <I> of the Mark V with the non-working Totalizers and downloaded it to EEPROM.

I'm not sure the plus sign is the cause of the issues, but it certainly could be.

I don't understand the "<C>" directory reference. Do you mean C: drive? All you had to do was send the contents of the UNIT1 directory of the F: drive (which is technically C:\SITE\UNIT1 (if I recall correctly). TOTD_<b>Q</b>.DAT has nothing to do at all with <C>. The left- and right carat symbols in Speedtronic (Mark*) turbine control systems refer to control panel hardware, and in the case of the initial Mark V operator interface, the <I> (and the <G> if used).

>Recently I observed Unit4 (faulty one) TOTT_Q.SRC is not
>same as other units where totalizers are working fine. To be
>more sure i rechecked the copied <C> directory that I sent
>to BHEL, there I got same wrong analog trigger level. Again
>I checked very old backup (taken on 2010) of that unit
>directory (unit4) and there I found TOTT_Q.SRC is same as
>other running units.Using GE maintenance manual GEH-5980 i
>got below table from that faulty unit.<pre>
>CHECK T1 S
>CHECKSUM TEST OF UNIT T4 PROCESSOR R: 22-OCT-1991 10:00:51
>Partition offset size --------date-------- cksum -id-
>-status-
>SEQ 8000 2786 15-JAN-2018 11:07:20 B626 SEQ OK
>CONST 2000 1400 23-FEB-2018 17:30:35 A2C4 CNST OK
>IOCFG 4000 0399 08-JAN-2018 20:06:40 D682 IO OK
>UBBL 0000 0000 01-JAN-1980 00:00:00 0000 UBL <NO
>DATA>
>HIST 0000 0000 01-JAN-1980 00:00:00 0000 HIST <NO
>DATA>
>EPA 0000 0000 01-JAN-1980 00:00:00 0000 EPA <NO
>DATA>
>MAOUT 0DD0 0084 11-MAR-2016 10:11:00 2E73 4-20 OK
>EVENT 0E60 008C 11-MAR-2016 10:11:00 1F46 EVNT OK
>CHNG 0F60 0004 11-MAR-2016 10:11:00 CAFE CHNG OK
>BOI 12E0 0586 11-MAR-2016 10:11:00 3AC7 BOI OK
>TOTT 0C00 00A9 13-MAR-2016 11:13:24 1C04 TOTT OK
>TOTD 0200 0902 19-DEC-2017 18:05:58 0000 TOTD NO
>CKSUM
>CBLR 1160 0006 11-MAR-2016 10:11:00 CAFE CBLR
> OK</pre>
>
>It is seen from the above table, the last change of TOTT was
>done on 13th March, 2016 at 11:13:24 hrs just before of GT4
>startup (after CI) and LTSA expert was working on it. From
>operation shift diary I found GT4 start command given at
>11:50 hrs and synchronized at 12:06 hrs. And on 13th March,
>2016 during night shift, analog counters inaccuracy was
>noticed.

It is seen from the above table that it seems there is something amiss with the EEPROM on the Mark V--and you can verify that by running the same command on one of the Mark Vs that has the working Totalizers. I'm very confident that you will see VERY different results from what you see above. That would indicate, as another respondent has said, that it's possible the Totalizer registers on the EEPROM got corrupted.

>If I change the TOTT_Q.SRC file (same as other units),
>Compile(MK5MAKE) and download (both TOTT and TOTD)and
>initialize all three controllers with new TOTD written with
>corrected value then is there any chance to get the
>totalizer back in normal condition?

To say it AGAIN: You DO NOT need to run MK5MAKE if all you're doing is compiling the TOTT_Q.SRC file. The Table Compiler is one of the executables that runs when the batch file MK5MAKE.BAT is started. So, if you haven't made any changes to any other file that TOTT_Q.SRC, then JUST RUN the Table Compiler with the following command:<pre> TABLE_C TOTT</pre>and press ENTER. That will simply create ONE NEW complied table file: TOTT_Q.DAT.

In reality, you don't even have to do that; you can just copy TOTT_Q.DAT from of the other three <I>s that have working Totalizers into the F:\UNIT1 directors of the <I> of the Mark V that has the non-working Totalizers.

Now, because of the issues with the CHEKKsum "report" above, I would suggest first using the EEPROM Downloader to download FORMAT (and ONLY FORMAT) to all three processors, one at a time. Then re-boot <T> by shutting off the power to it for approximately 30 seconds--it WILL NOT go back to A7 when it is powered back up, because you just "erased" the EEPROM with the FORMAT command. Then use the EEPROM Downloader to download ALL (YES, it includes FORMAT also/again! but that's okay!!!), and PAY ATTENTION TO THE MESSAGES DURING THE DOWNLOAD and report any errors back here!

Then, Download FORMAT to <S> and when that's done POWER DOWN <S> (by shutting off the power to it for approximately 30 seconds). It will not go back to A7, again, but you just erased the EEPROM with the FORMAT command. Download ALL to <S> PAYING ATTENTION TO THE ERROR MESSAGES and reporting any back here, and then cycle the power to <S> off for about 30 seconds, and it should go back to A7.

Repeat the process to <R> (Download FORMAT only; re-boot <R>; Download ALL; re-boot <R>.) It should also go back to A7.

At this point you have reformatted the EEPROMS on each of the control processors (TWICE), and have downloaded the "good" TOTT and all of the other EEPROM partitions to all of the control processors and re-booted them.

I would advise starting this just after the top of the hour, say, beginning at 10 minutes past the hour, because as the information posted previously said, Totalizer data is written once per hour to the EEPROM, and just after the top of the hour.

If the turbine isn't running, this should be enough. Run the CHECKsum report again, and all three processors should report very similar, if not identical results.

You could even copy TOTD_Q.DAT from one of the working Mark V <I>s to the non-working Mark V <I> and try downloading it.

You mentioned previously you had spare EEPROM chips (the U9 chip on the DCC or SDCC card). You could try replacing the three EEPROM chips on each of the DCC or SDCC cards of <R>, <S> & <T>. HOWEVER, if you do that, you are going to have to set the VOTER IDs of all three processors using the LCC/SLCC keypad.... There IS a procedure for doing this in GEH-5980, and it's been documented before on control.com (and recently, too, if I'm not mistaken). This might be a good thing to do before just about anything else--given all the problems you're experiencing.

I don't know what else to say or do. It could be the plus signs; it could not. No one knows for sure. Again, you could take a TOTT_Q.SRC from one of the working Mark V <I>s and copy it to the non-working Mark V <I> and download it. You could also take TOTD_Q.DAT from a working Mark V <I>--then you wouldn't even have to run the Table Compiler.

But, at this point my money is on an issue with the EEPROM chips. Or the DCC/SDCC cards (probably the RAM, and maybe even only the RAM or EEPROM chip on <R>, the Designated Processor). If you replace the entire DCC/SDCC card, you will still have to set the Voter ID.

Please write back to let us know how you fare! This shouldn't be so difficult; my apologies if I typed a filename incorrectly--it's one of the hazards of working with the Mark V. (So many files, many with similar names; some which do nothing, and others which are critical but don't ever seem to change size when one would think they would (when uploaded to!).)

Hope this helps!
 
Thanks Sir.

Okay. I will do the changes in the following way. Correct me I am wrong.

1. Ensure S/D and cooldown OFF condition of that nonworking totalizer GT.

2. Copy TOTT_Q.DAT file from other unit where totalizer working fine.

3. Paste to F:\Unit4 directory.

4. No MK5MAKE

5. EEPROM DOWN T4 T FORMAT
EEPROM DOWN T4 S FORMAT
EEPROM DOWN T4 T FORMAT

6. Shut off power supply of T 1st from <PD> switch. Wait 30 seconds- A7 will not come in this case. As all DAT files are already erased.

7. Again EEPROM DOWN T4 T ALL (which will format again and download all necessary DAT files.)

8. Hard reboot of T from <PD> switch.

9. I will keep close look on LCC display for any message. If <T> is bakc in A7 state then I will proceed to the next step.

10. Same procedure for <S> and <R>.

11. If all processors come back in A7 state then I will run "CHECKsum" for all processor to see three are identical or not.

12. If I face any error I will post in this thread immediately.

If above procedure does not work then I will go for changing EEPROM of DCC cards of all three processors. Steps:

1. Ensure S/D and cooldown OFF condition of that nonworking totalizer GT.

2. Switch OFF three processors power from <PD> switches.

3. Install U9 to all three SDCC cards of <T>, <S> and <R>.

4. Switch ON <T> from <PD> and assign voterID "T" from LCC keypad following GEH-5980.

5. As TOTT_Q.DAT already copied from other good unit I will download "ALL" to <T> and reboot <T>.

6. After reaching A7 of <T> then I will switch ON <S>.

7. Assign VoterID "S" and download "ALL" and then reboot, wait untill A7.

8. Same procedure for <R>.

9. If all processors come back in A7 state then I will run "CHECKsum" for all processor to see three are identical or not.

10. If I face any error I will post in this thread immediately.

If changing U9 doesn't help much then I will go for replacing DCC of all three and procedure will be almost similar to above steps. Again if I am wrong please let me know.

May be I can do a simulation check by injecting 4-20mA to following channel at shutdown condition of T4 to see DWATT accumulator works or not.

Q_Q_MAI16 DWATT MWATT ;Q -QTBA-065 Generator watts [96GW-1T]

Thanks again.
 
It is DONE!!!!

Downloaded TOTT_Q.DAT copied from other working totalizer unit, switched OFF all three and then rebooted <T>, <S> and <R> sequence.

Now I need only to rewrite the TOTD_Q.DAT file as per requirement. In this issue I am gonna talk again with BHEL.

Thank you very much.
 
sms,

Sorry; I missed your last post.

Thanks for the feedback!

Congratulations for getting the download done.

However, can you tell if the accumulators are incrementing properly now--even if the data values might not be correct?

TOTD_Q.DAT is the Totalizer Data file. It is in Intel hexadecimal format; there is no TOTD_Q.SRC file. The Table Compiler does not compile a TOTD_Q.SRC into TOTD_Q.DAT to be downloaded to the Mark V control processors. TOTD_Q.DAT is being written to approximately once every hour by the control processors. It can be uploaded from the Mark V to the operator interface, but there is no way to view the contents in engineering units/values.

You can obtain a "new" or updated TOTD_Q.DAT from a GE-authorized supplier if you supply them with the values you want to start FROM in the file. You would then copy the file to the unit-specific directory (F:\UNIT1, for example) and use the EEPROM Downloader to download it to all three control processors, then power them all down (one at a time), and power them each back up one at a time (using the method I recommended and which apparently you used).

TOTT_Q.SRC is the definition for the Totalizer function; it's an editable ASCII text file. It gets compiled by the Table Compiler (which is run every time MK5MAKE.BAT is run, or can be run individually from the command prompt) into TOTT_Q.DAT. It does NOT have any data or values in it--just the definitions of the various accumulators and their scaling, as appropriate.

I hope you were successful in your efforts to restore the function of the four accumulators which were not previously working for a short period of time.

Again, thanks for the feedback!
 
> However, can you tell if the accumulators are incrementing
> properly now--even if the data values might not be correct?

Yap, values are incrementing properly. I applied 20mA to the DWATT terminals for both <R> and <S> (for TMR configuration) to get 60 MW for the DWATT tag(may be scaled from 0 to 60 MW for 4-20mA), and on timers counters page "TOTAL ACTIVE ENERGY" increment was 1 MW/min. I noticed HMI <I> fetches timers and counters data in every minute from RAM.

TOTD_Q.DAT needs to be written with required data. And I already requested to BHEL. GT is still kept in down condition to get downloaded with the corrected TOTD file.

Thanks again for your support.
 
Hi sms,

in this point, you forget the <R> in the first line:<pre>
5. EEPROM DOWN T4 T <b>>R<</b> FORMAT
EEPROM DOWN T4 S FORMAT
EEPROM DOWN T4 T FORMAT</pre>
 
Top