Voltage spikes on ground wire.

[BigAl]

I was called out to look at a machine, where the customer was complaining about voltage spikes on the ground wire going to an ampliphier card. I thought it would be an easy trip, I expected to see a clutterd panel, with no shielding, and wires runnining every which way. This turned out not to be the case, and it turned out to be more than one machine. The customer has been burning up the ampliphier cards, about every 3 to 4 months. The manufacturer of the card evaluates why the card has failed, and has indicated that it is due to voltage on the ground terminal. The card is fed with 24VDC, to power the card, and receives a 0 to 10VDC signal voltage from an Allen Bradley analog output card.
Measuring between OV, and ground you can see votage spikes, somtimes as high as 16VDC, when the E-Stop is actuated, and when the machine is first powered up. The machines all have about 15 to 20 DC relays, and some DC safety relays. There is only one main DC power supply, and the OV return wire is bonded to ground. The customer has tried powering the analog module, off the backplane of the PLC, and externally, and it doesn't seem to make a difference. We are measuring these spike right at the terminal blocks which feed the analog signal to the card. The wiring in the panel, is great, everything is shielded, and high, and low voltage is seperated.

Any ideas, or thoughts would be greatly appreciated.

Thank you.

 

[rsdoran]

I have to ask a couple of questions.
1. Is there just one 24vdc supply?
2. Is this a floating supply...ie common not bonded to ground.

From your statements it sounds like a floating DC supply that has a short to ground.

 

[DonsDaMan]

I'd run a new wire from the 0V back to the ground pit, bypassing the ground you're now using.

My guess is that the ground at this plant has been "polluted" (my word for it) by too many people equating ground with neutral with zero volts with anything they see as having "no voltage."

Ground is a safety. It is for cases and housings. You don't want current in a ground circuit, unless a fault occurs. Then you want this current to trip your safety, alerting you to the fault without killing anyone.

Neutral is a reference where all of your vectors add to zero. We tie points that SHOULD be zero to ground to INSURE they are zero, but this introduces a current to ground when that point is not balanced to zero. Usually three phase systems are at fault here.

Zero volts is an arbitrary point that we use for our reference. Wherever you put your black probe when making a measurement... that's zero for that measurement. Zero volts is not necessarily ground. For example, when I measure V A-B, B phase is my "zero" for that measurement. No, I'm not going to touch it, since my feet would be "zero" for THAT measurement, causing a certain unpleasantness for me, and anyone in the vicinity who doesn't like loud noises and curse words.

I'm guessing you have some ungrounded three phase with a "ground fault" introducing a voltage to this overloaded, polluted ground. It's going to take a very thorough examination to clear up if this is the case.

And if that is that case, these amplifier cards are just the first identified symptom of a much bigger problem.

I don't envy you, unless you're being paid hourly (or daily)....

Good luck. I hope I'm wrong.

 

[Mickey]

Sounds like you are getting a voltage spike from relay coils.I would install
"Free Wheeling Diodes" acrossed every relay coil.

Do a search on this site for more info on free wheeling diodes.

 

[DickDV]

It sounds like your DC power supply has its negative terminal connected to earth ground thru a high resistance. This is commonly done to pull the negative pole to ground potential without tying it tightly. This permits some "float" under shortcircuit conditions and when multiple power supplies with different references are mixed.

Earth ground, unfortunately, is not necessary at zero potential or noise-free especially when comparing ground at one point to ground at another point, say, thirty feet away.

If these 16V pulses are fast, I would be tempted to put a little capacitance between earth ground and power supply negative so the power supply has a tighter AC connection to earth ground. This won't change the DC ground connection and just might swamp the 16V pulse just enough to avoid further damage.

The disadvantage to doing this is that some of the noise on earth ground may be coupled into your power supply leads. This may or may not cause other problems so you should be on the lookout for that.

 

[Pierre]

I'd run a new wire from the 0V back to the ground pit, bypassing the ground you're now using.
...
Ground is a safety...
...
I'm guessing you have some ungrounded three phase with a "ground fault" introducing a voltage to this overloaded, polluted ground. It's going to take a very thorough examination to clear up if this is the case.
...

Good points Don!

I have encountered similar problems in the past.

When check your panel Ground to the REAL ground with a wire, be carefull. Electrocution is a definite possibility.

First Check de potential with a good Fluke set at 1000V and then lower the scale for readings.

Normaly you could have some current flowing but almost no voltage. So insert an Amprobe on this wire to check for current.

I have seen in the past some pretty intense current flow INDUCED by long motor power cable.

The current will flow through the least resitive path. (Hey guys please dont argue just for the sake of it)

Soooo. The spikes could come from afarr. Could even be a welding machine grounded close to a distant motor which has a GOOD ground.

Cables grounded on the panel side will make a good path for induced voltage from any magnetic field.

I saw Ground current detectors on Main switches set at 400A. Almost no voltage where present but God where there ever current!

Again, good luck. But BE CAREFULL. Ground conductors can kill you as much as live ones.

 

[mordred]

I would also check at the main transformers for poor grounding.

 

[rsdoran]

I think many missed a point in the original post, the system and problems seem to be with the 24vdc system. I even missed the fact that there is just ONE DC power supply.

If the DC power supply common is connected to ground if the +24vdc goes to ground then it "should" short out and blow a fuse. Since he is seeing as much as +16vdc on ground then I suspect the DC supply is floating and there is some form of a short.

Recently I had to deal with a Minster stamping press that had a new feeder added that used an Indramat drive, the drive failed and was replaced a total of 5 times in approximately a year. On the 5th install I observed the system using an oscilloscope and determined I had an AC voltage riding on the 24vdc control voltage. I determined that the drive itself was addding this AC voltage to the DC and thru trial and error determined poor/improper grounding was causing or allowing it to happen.

Sometimes problems like this are not easy to locate but I would definitely try to determine if any devices could be shorting to ground...especially if the power supply is not common/bonded to ground. Measuring a DC value to ground especially with the ESTOP actuated makes me think loose wiring and/or faulty DC device.

I am not positive it matters if the existing ground is good or not as far as the DC voltage is concerned. If the DC common (-) is bonded to ground then a DC + voltage should blow a fuse if a + voltage goes to ground and its bonded to the DC common.

Another issue to look at is the meter you are using, sometimes Fluke (and other brands) digital meters can/will detect leakage current (voltage) that isnt a true reading. Try using an analog meter or better yet an oscilloscope to troubleshoot the problem. The false reading thing may not be an issue because you are burning up cards but never hurts to verify.

 

[elevmike]

Floating potential voltage on Ground Wire??? Hmmmm... Dos'nt sound like a Grounded Ground wire...

Ok, so going along with Dick some. Bonding alone may not provide for a very good ground. Maybe a bigger ground wire and better overall grounding would be an answer to handel the spikes. Ive been on jobs where we had to run ground wires to all the equipment, and back to the main power supply, then back to the ground rod.

Are all the ground wires connected at one lug, or are the various lugs spotted around the panels?

The other thing is all thoes DC relays are great spike generators. Especially if you have a bunch of them dropping all at once. Are all the coils connected to one common thats well grounded? Are they equipted with blocking/clamping diodes across the coils??

 

[mordred]

There is one other remote possibility. That being power harmonics.

This is an unlikely scenario however poor harmonics can affect neutral conductors. I will include a Fluke pdf on harmonic troubleshooting.

This pdf is handy as it shows how to troubleshoot harmonics with either a DMM or scopemeter.

 

[DickDV]

Really, elevmike, if you compare two different earth ground points located, say 50 feet apart, in an industrial facility, you will often find measureable differences in potential between them.

I'm not talking about two different points on a copper ground conductor, now. I'm talking about the earth---dirt--- being at different potentials.

So, if you have a +24v power supply grounded at the first location and another grounded at the second location, you may find that the voltages on the plus sides don't match. If you check their ground sides, that voltage doesn't match either due to the difference in ground potentials between the two points.

It can cause all kinds of chaos in systems spread over significant distances. Of course, the solution would be a nice low-resistance copper ground between all points. If that's not practical, optical isolation and floating power supplies become the next option.

Ask any telephone technician about the unpredictability of ground potentials. In my early days in that field I saw as much as 40VDC difference in 5 miles.

 

[BigAl]

Thanks for the great ideas.

Thank you for all the great ideas. It has definitely got me thinking. I did check all of the grounding, and this customer went well above and beyond what I usually see. All of the paint is removed, where the ground bars are attached to the sub panels, and they used external star washers, and even added extra bonding jumpers from each ground bar, to seperate locations on the panel. The primary transformers are grounded correctly. I also thought there would probably be a difference between the machine ground and the building ground, because I have also experienced this at other facilities. No such luck here. The same customer, is experiencing the exact same problem, at a sister plant, in a different State.
I will definitely explore the use of diodes on all of the DC relays.



Thanks again.

Big Al

 

[elevmike]

Re: Thanks for the great ideas.

I will definitely explore the use of diodes on all of the DC relays.

No claming?? personally I consider this a "almost must". I don't know how many DC relays you have, or how big the coils are and so on, but sure wouldnt hurt. Just connect a scope accros the coil of a DC relay, and toggle it a few times and you'll see what I mean. The cathode (striped side) of the diode goes on the positive side of the coil, and the anode goes on the negitive side of the coil.

 

[rsdoran]

I am going to withhold judgement on that part, it was stated:

Measuring between OV, and ground you can see votage spikes, somtimes as high as 16VDC, when the E-Stop is actuated, and when the machine is first powered up.

If that is the only time or the most noticeable times the spikes are seen then the relays have no bearing.

I stand by my idea of an intermittent short to ground from a 24vdc device, more than likely a wire with bare spot going to ground.

The strange part of this are the cards bubrning, a short shouldnt have that much effect on other devices except the fuse or power supply. Low voltage from a short should stop the cards from working.

I would test all wiring from the effected cards to ground.

Usually when cards fry there is an external voltage (AC) being applied when it shouldnt be.

PLEASE explain what you mean by amplifier card.

 

[Kim Gold]

We had a problem in our area a few years ago with computer boards burning up. They were Compaq computers (not that that makes a difference) and the board burning up was called a riser board. They would last approx. 3 or 4 months.

It turned out to be a problem with the grounding system in the building (Large maintenance shop). It took a while to find it. At first we thought it was just a bad batch of computers. Once we identified the problem we had a company come in and they ran large bonding conductors throughout the building to eliminate any differences in ground potential. They then drove ground rods.
Once this was completed we have had not problems since. It has been approx. 2 years.

Kim