Mark 5 diagnostic alarms

Our D11 Steam Turbine is experiencing diagnostic alarms on 2 valve regulator/LVDT 'position diffhigh.' I have calibrated the LVDT 0-100% and the voltages have a max spread between the highest and lowes of .035v. I can't find any info in the GEK.
 
For some reason, the Mark V is seeing a difference that exceeds a preset (and undocumented) limit between the two inputs.

If you are measuring the inputs with a True AC RMS voltmeter and are seeing 0.35 VAC difference between the two LVDT inputs to Reg (SVO-Servo-valve Output) 5, then it's likely there is some problem with the output of one of the LVDTs. They are usually more closely matched, even if the zero-stroke settings aren't very close.

By the way, what are the zero-stroke voltages of both the LVDTs?

As for any GE manual explaining Mark V Diagnostic alarms, you must have been in the proximity of some wacky-tabacky smoke.
 
I used the voltages as listed in the AUTOCAL tool. This is what the Mark 5 thinks it is reading so I find no need to use a meter. The Diagnostic Alarm is: 01337 TCQA LVDT position diffhgh REG#2. Max dif is .035V, not .35V. Regulator 2 at 0 position read:
LUDT #1 Uoltage:(R)-2.087, (S)-2-074, (T)-2.063 V rms.
LUDT #2 Uoltage:(R)-2.084, (S)-2.085, (T)-2.082 V rms.
LUDT #3 Uoltage:(R)-2.068, (S)-2.071, (T)-2.055 V rms.

I have searched the GEKs to find a clue as to what would cause this alarm with no luck. Thanks for your response.
 
Mike Borrelli,

It would seem you have a GE Mark V HMI (running MS-Windows and CIMPLICITY). The Diagnostic Alarms are "coded" with respect to which processor they are associated with, and I don't recall the "key" to the code. I think it's something like dividing the Drop Number by 2048 or 4096 or something like that, with the whole number result being associated with <R>, <S>, <T> or <C>, and the remainder being the actual Diag. Alarm Drop number.

But in any case, you need to know which processor the Diag. Alarm is associated with, and then I would suggest replacing the TCQA card or the TCQC card for that processor. If you replace the TCQA card, you will have to recalibrate the LVDTs.... And I would also suggest doing the same for the TCQC, since if I recall correctly the analog-to-digital LVDT feedback conversion is partially done on that card; I'm pretty sure the LVDT excitation comes from that card at least.

I would suggest doing the card changes one at a time instead of the "shotgun" approach.

Please write back to let us know how you fare!
 
Mike,

Since all LVDT's are reading nearly same voltage it won't be a field problem.

You can try to swap the card where LVDT # 2 is connected with the other one (LVDT # 1 or 3).

Take care
G.Rajesh
 
I assume you have looked at your HELP_QD.DAT File in F:.

Our file states the following about that alarm.<pre>
1336-1343 ;TCQA LVDT position diff high REG#n (n = 1-8)

CAUSE The difference between the "calculated positions" of the LVDTs is greater than 5%. This
alarm applies only to regulators using more than one LVDT. The "calculated positions" is affected
by the gain and offset. Both the gain and offset is calculated by the IO configurator (or autocalibrate),
then downloaded to EEPROM.

EFFECT Alarm only

ACTION Perform autocalibrate, and update via IOCFG.</pre>
I would probably retry the autocalibration and possibly try a manual calibration, ensuring I had the correct LVDT's in the IO config file using IO.ASG file.
 
What are your 'Zero stroke' and '100% stroke' settings for the LVDT's for Regulator #2 in your IO configurator? (I.E., if type 49 Regulator, LVDT #3, 4 and 15)
 
I was not aware of this file: HELP_QD.DAT. Thanks for this simple yet huge tip (I've needed for years). My file matches yours. I don't trust autocals because they go so fast that they are not very accurate. I manually stroke the valve from -25 to +125 and wait a moment and grab the LVDT voltages and manually enter them. It is more time consuming but more accurate. the null bias and gain have not needed any adjustment but the help file points to them. Will the autocal adjust the bias and gains?
 
Top