Troubleshooting a MKV 125 Vdc Ground


Thread Starter


MKV <I> "BATTERY 125 V DC GROUND" alarm.
(L64D_P = 1, causing L64D =1)
Battery Charger is also indicating a ground.

MKV DIAGC - TCQA Voltages:
P125v = -1.42 Vdc
N125v = 129.14Vdc

We pulled and replaced isolation jumpers (5 each, JB1 - JB5) from following cards:
Cycled power to <R>, <S>, <T>, <X>, <Y> & <Z>

...could not isolate ground.

Shut down MKV completely by opening the DC breaker to MKV panel. The ground alarm did clear on the battery charger but came back in as soon as the DC breaker was re-closed.

Any suggestion of where to look next would be greatly appreciated.
Follow up...

Isolated all SOL outputs. Ground still in.
Isolated GCP (also on MKV DC Circuit), powered down MKV panel, disconnected all cards and MKV power supply, read 770 Ohms across <PD> TCPD's DCHI and DCLO terminals. Replaced TCPD card. New card reading many K OHMs. Powered up... Ground still there.

Andrew Turner


Is this a new installation?

If it has been set up and has run sucessfully in the past I suggest you do the following.

The Ground is very solid from the readings So you need to completely disconnect all the outputs from the Power distribution board (inside PD) That is all the Twisted pairs which are generally plug in... Connect a Fluke at the bottom of this board so you can monitor the DC input. i.e. Put one leg on earth and the other on either of the inputs... you need to see 63Volts here. When you do one by one strat to reconnect until you see the ground come back... this will isolate which circuit it is on... Let me know how you get on...
If your battery charger is indicating a ground, then you potentially have two problems: The Mk V 125 VDC battery supply is supposed to be UNGROUNDED--and that includes any ground reference circuit in the battery charger.

This problem has occurred many times when someone other than the turbine-generator packager has provided the 125 VDC source--they leave the battery charger ground detection circuit active/enabled. (Sometimes it even happens when the turbine-generator packager, such as Stewart & Stevenson or Nuovo Pignone or Alsthom doesn't know about the ungrounded battery supply "requirement.")

This circuit is then in parallel with the one in the Mk V, which means that higher ground resistances will result in erroneous ground detection readings by the Mark V. The Mk Vs circuit is also in parallel with the battery charger's ground detection circuit--which causes the battery charger to also falsely indicate nuisance and erroneous battery ground alarms!!!!

If the battery charger which is powering the 125 VDC battery supplying the Mk V turbine control panel has a ground detection circuit you should disconnect it to prevent nuisance ground alarms from occurring. Just disconnect it temporarily and see what happens to the Mk V's ground indication.

If the 125 VDC battery supplying the Mk V turbine control panel is also supplying other equipment--including the DC Emer. Lube Oil Pump--AND it still indicates a ground when the breaker feeding the Mk V turbine control panel (and Generator Control Panel (GCP)) is opened, then that means the ground is NOT necessarily in the Mk V turbine control panel I/O.

If you've effectively disconnected all the 125 VDC I/O (contact inputs and solenoid outputs, including those connected to the <P> core!) and you still have a ground, odds are the ground is NOT in the Mk V I/O.

This author has had numerous, passionate discussions with plant engineers and -management about this issue--just because the Mk V (or the Mk IV or the Mk VI) is indicating a battery ground DOES NOT mean the ground is in some device or circuit directly connected to the Mk V--the ground can be on any device which is also powered by the same battery as the Mk V! Battery grounds have been found in the DC Emer. L.O. Pump motor starter, the DC Emer. L.O. Pump field circuit, the DC Hydraulic Ratchet Pump Motor field circuit, 20FG-1 and 20FL-1 solenoids Which are connected to the <P> core and NOT <QDn>), the fire detection circuit (including the 45FTX and 45FTX-n relay circuits), the EX2000 exciter rectifier bridge cooling fan limit switches, even the Cooling Water Module Fan Vibration switches--none of which were directly connected to the Mk V!!! but were powered via the same 125 VDC battery as the Mk V.

So, you most likely have one problem related to the overall ground detection circuitry in that you appear to have a second circuit whose resistance is in parallel with the Mk V's and will wreak havoc with the Mk Vs and cause nuisance and erroneous Mk V battery ground alarms as well as nuisance and erroneous battery charger ground alarms, and you may indeed have a true ground condition on SOME circuit connected to the Mk V--but not necessarily.

The author has also NEVER perusaded a single plant engineer or manager to disconnect the battery charger's ground detection circuit and rely solely on the Mk V's ground detection circuit. If it was built/provided that way during the original installation, they just INSIST that it MUST remain that way, and those are the same people who PISS AND MOAN the loudest about nuisance and difficult-to-find battery ground alarms!

Usually, the battery charger's ground detection circuit is connected to some DCS (Distributed Control System) and they use this as a primary reason for not disconnecting the circuit. "We'll lose our DCS battery ground alarm," is the rationale they use. But, they won't consider that they've been having nuisance ground alarm indications on BOTH the Mk V and the battery charger/DCS systems for some time, and that by putting the system to the condition it should be in they would be eliminating a source of erroneous and nuisance alarms. (You can lead a horse to water, but it's very difficult to persuade him/her to drink...)

This subject has been discussed several times previously on; if you haven't already, use the search feature to find them. One of the first things one needs to do IMMEDIATELY when a battery ground alarm is annunciated is to ask, "What has changed?" Is the unit being put back together after a maintenance outage and a wire been pinched or improperly terminated? Has there been a recent rainstorm which might have caused water to enter conduits and junction boxes which weren't properly sealed after previous entries or which were not properly sealed during the original construction? Has a high compartment temperature in the Load Compartment or the Turbine Compartment been annunciated or detected recently which might have resulted in melted insulation on conductors in conduits in the compartments?

Most often, ground alarms aren't looked at immediately and they then become harder and more difficult to troubleshoot and resolve--again, refer to previous threads for other's experiences.


Andrew Turner

A full and accurate answer from Markvguy as always... I'd just like to ask.. when you say you powered back up, the ground was still there... how was this annunciated?

If you had disconnected all mk5 cores it wasn't on the MK5... was it on the battery charger?
How did you know the ground came back when you switched the power back on??

In my opinion to determine absolutely that it is not Mk5 then the best way is to measure the voltage at the PD core from one leg to earth. It should be 63V or thereabouts with no load on ie all the outgoing circuits disconnected. It will change as you gradually power up the mk5 but should never go hard one way or the other as your system seems to be saying...

The secret to this as with all trouble shooting is to be logical and systematic.

Good luck.
Ground identified and resolved.

What was found:
It appeared we had multiple grounds, when in reality there was only one.
During this ground search it was discovered that in numerous instances many, many devices were wired using the same and apparently most convenient common (107) wire that could be found, irregardless of the final MKV termination point (i.e., I/O landed on <CD>DTBA may be using the common (107) from the <QD2> DTBB card and so on).
No one on site could remember when the ground first came in. Using the PI/Historian it was determined that the ground initially came in months ago and within the same second we had received a "Main Lube Oil Diff Press Alarm". The differential pressure alarm cleared within seconds but the ground stayed in...
After lifting the wires to 63QQ-21 the ground cleared completely.

Thanks for all of the troubleshooting assistance.

As Mr. Turner says, a logical and systematic approach is required--to any troubleshooting.

It is still believed that if the battery charger is equipped with a ground detection circuit, the Mk V's ground detection circuit is adversely affected by the battery charger's and vice versa. So, that's one issue (which will probably never be resolved). And this condition causes grounds which would otherwise not be detected to be annunciated--in other words, nuisance ground alarms.

One of the really annoying "features" of any Speedtronic heavy-duty gas turbine control system is the way that contact input common wires and solenoid output common wires were used and connected to the Mk V. It makes understanding and troubleshooting grounds very difficult--and because drawings can't be provided in this forum it makes it very difficult to describe.

Every contact input circuit to the Mk V has a common wire on one side of the contact--usually it's wire number 107. To save wire and terminations, only a few (107) wires are run out of the Mk V turbine control panel to various junction boxes on the unit, and in those junction boxes there may two, three, four, or even more devices connected to that wire--and the signal side of the contacts may be connected to <CD>, or <QD1>--regardless of whether or not the (107) wire came from <QD1> or <CD>!

Every solenoid output (except those connected to the <P> core) has a common wire on one side of the coil--usually it's wire number 108. To save wire and terminations, only a few (108) wires are run out of the Mk V turbine control panel to various junction boxes on the unit, and in those junction boxes there may two, three or more coils connected to that wire--and the other side of the coil may be connected to <QD1> or <CD>--regardless of whether the (108) wire came from <CD> or <QD1>. (Very difficult to describe in writing....)

So because of this practice of paralleling common wires it's possible for a ground on a device connected to <QD1> to also cause a ground indication on <CD>.... Are we having fun yet?

All of the recommendations below are made presuming that the turbine and driven device can be completely powered-down for trouble-shooting (which seems to have been done already). That would mean that all the auxiliary motor starter breakers would have to be opened (OFF) and the unit would have to be off cooldown.

One very simple way to isolate the ground to the Mk V panel would be to connect a voltmeter between one leg of the battery charger output and ground--at the battery charger--and then open the breaker that supplies the Mk V turbine control panel (and the GCP). If the battery voltage returns to a nominal 65 VDC with respect to ground, then it's likely the ground is on some device connected to the Mk V or the GCP. If the battery voltage does not return to a nominal 65 VDC when the Mk V and GCP are isolated from the battery/battery charger output then the ground is NOT NECESSARILY in the Mk V. There may be more than one ground...which is very difficult to troubleshoot--not impossible, just difficult.

Most Mk V battery grounds are on circuits, wiring, and devices connected to the contact input terminal boards or to solenoid outputs which are supplied with 125 VDC from the TCPD card in <PD>. The 125 VDC supply to the Mk V is distributed to various locations in the Mk V via the TCPD card and all the cables connected to the TCPD card.

The processor (<C>, <R>, <S>, <T>, and <X>, <Y>, & <Z>) power supplies all take the 125 VDC and convert it to AC, transform it to a lower voltage, then invert it back to the various DC voltages required for the microprocessor and analog I/O (such as +/- 5 VDC, +/- 15 VDC, 24 VDC, etc.). So this provides some "isolation" for the analog I/O and low voltage processor power supplies from the 125 VDC battery supply voltage. This author has seen a couple of occasions where processor power supplies have failed (one <S> and one <Z>) where the failure resulted in a battery ground, but the entire power supply in both cases was failed and replacing the power supply was necessary to get the processor back up and running.

So, presuming the ground is in the contact input wiring or the solenoid output wiring that is connected to the Mk V, the J12x and the J8x cables feed the DTBA/DTBB and DTBC/DTBD cards, respectively, in the digital I/O cores (<CD>, <QD1>, and <QD2>, if so equipped) from the TCPD in <PD>. While measuring one leg or the other of the 125 VDC input voltage to the <PD> with respect to ground (the positive leg of the 125 VDC input voltage should be connected to terminal board at terminals -1 & -2; the negative leg should be connected to terminals -3 & -4) disconnect all the J12x (where "x" stands for A, B, C, etc.) cables from the TCPD one at a time and note any change in voltage. DO NOT RECONNECT any J8x cable before disconnecting the next one, and continue until all J8x cables are disconnected. If the ground still exists, then it's most likely not on any contact input circuit/device.

Next, while still monitoring one leg of the 125 VDC battery input with respect to ground, start disconnecting the J12x cables one at a time and noting any change in voltage. As above, do not reconnect any J12X cable before disconnect the next one, and continue until all the J12x cables are disconnected.

Using this method, when you disconnect a cable (J8x or J12x) and the battery voltage returns to a nominal 65 VDC with respect to ground, you have isolated the ground to a device connected to the cable which was just disconnected. REMEMBER: there may be more than one ground--that's why it's important not to disconnect a cable and then reconnect it if there's no voltage change before disconnecting the next one. You must disconnect a cable and leave it disconnected. Once you have found "the" offending cable/circuit by this method, then you can leave that cable disconnected and start reconnecting each of the other cables one at a time while monitoring the voltage. If the voltage changes when you reconnect one of the other cables, then you have another ground; if the voltage remains unchanged and you have reconnected all the cables but the one, then you have isolated the ground to some device connected to that cable.

If the ground STILL persists--as measured by monitoring the voltage input to the <PD> core as above, the next thing to do would be to start powering-down each of the processor power supplies, one at a time, noting any change in voltage, until they are all powered-down. If the battery ground still persists and the GCP wiring is still connected to the circuit then it would be necessary to disconnect the GCP wiring from the circuit.

There is one more device which is present in most Mk V turbine control panels which isn't very accessible or well-known: the <CPF>, Conditioning Power Filter. It is believed this device, which is usually located at the bottom of the Mk V turbine control panel, at the very back of the bottom of the the Mk V turbine control panel, was provided in order to comply with EU regulations about filtering/conditioning power supplies...and GE just decided to put in all Mk V panels, not just those subject to the rules and regulations of the EU. This author has seen components on this device fail--but usually only after some high-voltage spike such as from a lightning strike--and cause grounds and other unusual conditions.

If you believe you have isolated all the devices powered by the TCPD and the GCP is effectively disconnected and the ground condition still exists only on the 125 VDC supply wiring to the Mk V, then it might be the <CPF>--which is real fun to get to and/or replace!