VFD power cable shielding (screening) issues
VFD power cable shielding (screening) issues
(OP)
I have been heavily promoting the use of VFD shielded power cable in my installations, and most of the problems I have experienced in the past have been diminishing to near extinction. But today, I went to a site with MULTIPLE motor failures connected to a VFD system that I designed, but was altered before purchase and installed by neanderthals. So in my investigation, I found that ALL of the repeated motor failures are happening in the exact same type of application, wherein they took the one single drive I had put into an MCC, and used it to run 4 smaller motors (12A drive running 4 motors that are 2.7A each) on evaporator coils in a cold room. The motors are supposedly "inverter duty" (but I find that is often about as reliable as a fox telling the farmer that the chickens are safe), I could not see the actual motors. So at first glance it SHOULD have been OK in theory. They did install a Load Reactor on each drive, then they used VFD shielded cable, one that has 3 power leads, 3 symmetrical ground leads nested in them, an overall braided shield over those 6 and a separate EGC (Electrical Grounding Conductor, a term used in the US similar to a "PE" conductor). The typical linear distance from MCC to evaporator set is between 300 and 400 ft, through open cable tray. The VFDs did not show faults in the fault logs, other than Power Loss faults typical of being powered down from opening the MCC cubicle doors.
The only big problem I can find with the installation is that the neanderthal electricians just routed the cables haphazardly (as in a bowl of spaghetti) in cable tray along with other full voltage motor leads and the incoming power leads, and then when they ran the shielded cable to the motors, they cut and taped off the shields at the motor connection box end, so they are shielded only at the VFD end. The motor lead cable goes from the MCC to a fiberglass box on the side of each evaporator, and inside is a local disconnect switch, a power splitter block, and a ground bar, to which the EGC is attached, along with 14ga ground wires from the motors. Motor leads from this box to the motors themselves are NOT shielded any longer, they used what appears to be non-shielded flexible wire through gland fittings on the box. I always tell people to ground both ends of shields on POWER cables, because the intent is to form a sort of Faraday cage. But what does / could happen if only one end is grounded? Could this possibly explain the multiple motor failures?
More relevant info:
(1) The motor re-winder that has repaired these has told the user that the motors are failing on "single phasing", which makes no sense to me. The VFDs in question should trip out on loss (or severe imbalance) of output phase current. I have provided the end user with EASA photos of typical single phase motor failures and typical turn-to-turn insulation failures due to standing waves, they have yet to share that with the re-winder to confirm his findings.
(2) To my horror, the user deliberately CHOSE to remove the separate Manual Motor Starters at the evaporators which were to be protecting the individual motors. So there is NO realistic motor overload protection for these 2.7A motors, other than the OL setting of the VFD, which is set at 10.8A. I have told my contact on the job site that this violates the NEC, but the decision came from a manager, so he said he does not want to challenge him and has asked that my report de-emphasize that issue (not likely). Still, these are evaporator coil fans, it is virtually impossible to overload them, and they are only 3 months old so no time for normal bearing wear. While I was there today the VFDs were at 55-60Hz and the 4 motors were drawing only 4.5 - 4.8A collectively for all 4 motors.
My theories at this point are:
1) The lack of shield grounding at one end rendered the shield ineffective, so standing wave generation caused first-turn insulation failure, which the motor re-winder was unfamiliar with and misidentified as single phase damage.
2) The lack of shielding is causing rapid bearing EDM damage and failure, which causes the motors to overload, which again, the re-winder misidentified as damage from single phasing.
3) None of this has any importance whatsoever, and all of it was just caused by the neanderthals who did the motor hookups likely not making good connections at the splitter blocks. Still, I would have expected the VFD to trip on that though.
Please share your thoughts.
The only big problem I can find with the installation is that the neanderthal electricians just routed the cables haphazardly (as in a bowl of spaghetti) in cable tray along with other full voltage motor leads and the incoming power leads, and then when they ran the shielded cable to the motors, they cut and taped off the shields at the motor connection box end, so they are shielded only at the VFD end. The motor lead cable goes from the MCC to a fiberglass box on the side of each evaporator, and inside is a local disconnect switch, a power splitter block, and a ground bar, to which the EGC is attached, along with 14ga ground wires from the motors. Motor leads from this box to the motors themselves are NOT shielded any longer, they used what appears to be non-shielded flexible wire through gland fittings on the box. I always tell people to ground both ends of shields on POWER cables, because the intent is to form a sort of Faraday cage. But what does / could happen if only one end is grounded? Could this possibly explain the multiple motor failures?
More relevant info:
(1) The motor re-winder that has repaired these has told the user that the motors are failing on "single phasing", which makes no sense to me. The VFDs in question should trip out on loss (or severe imbalance) of output phase current. I have provided the end user with EASA photos of typical single phase motor failures and typical turn-to-turn insulation failures due to standing waves, they have yet to share that with the re-winder to confirm his findings.
(2) To my horror, the user deliberately CHOSE to remove the separate Manual Motor Starters at the evaporators which were to be protecting the individual motors. So there is NO realistic motor overload protection for these 2.7A motors, other than the OL setting of the VFD, which is set at 10.8A. I have told my contact on the job site that this violates the NEC, but the decision came from a manager, so he said he does not want to challenge him and has asked that my report de-emphasize that issue (not likely). Still, these are evaporator coil fans, it is virtually impossible to overload them, and they are only 3 months old so no time for normal bearing wear. While I was there today the VFDs were at 55-60Hz and the 4 motors were drawing only 4.5 - 4.8A collectively for all 4 motors.
My theories at this point are:
1) The lack of shield grounding at one end rendered the shield ineffective, so standing wave generation caused first-turn insulation failure, which the motor re-winder was unfamiliar with and misidentified as single phase damage.
2) The lack of shielding is causing rapid bearing EDM damage and failure, which causes the motors to overload, which again, the re-winder misidentified as damage from single phasing.
3) None of this has any importance whatsoever, and all of it was just caused by the neanderthals who did the motor hookups likely not making good connections at the splitter blocks. Still, I would have expected the VFD to trip on that though.
Please share your thoughts.
"Will work for (the memory of) salami"





RE: VFD power cable shielding (screening) issues
Perfect setup for that eventuality with the described crummy setup.
Keith Cress
kcress - http://www.flaminsystems.com
RE: VFD power cable shielding (screening) issues
http://www.epri.com/abstracts/Pages/ProductAbstrac...
So based on sensitivity analysis for the above configuration, according to b), grounding at supply end only would give slower rise time and slightly-reduced level which would make the motor surge environment less severe (compared to grounding at both ends). Whether this conclusion applies to your configuration, I have no idea...
=====================================
(2B)+(2B)' ?
RE: VFD power cable shielding (screening) issues
Any electrician who has installed any amount of 4-20 mA current loops will always insulate the field end of the shield and the shield drain wire.
You don't want the shield to form part of a ground fault circuit.
The Canadian code requires de-rating of power cables when sheath currents may cause overheating. A common and acceptable means to avoid sheath heating and derating is to insulate the sheath, armour or shield at the field end.
If your equipment grounding conductor is outside of the cable the impedance seen by a ground fault may be too high. In the event of a ground fault this may cause the fault current to be too low. We want the ground fault current to be high enough to cause an instantaneous trip. We don't want to wait for an inverse time trip in the event of a ground fault.
Another issue is the ampacity of the shield and the NEC. Shields may not have an ampacity rating. With an external grounding conductor, a shield grounded at both ends may become the lowest impedance ground fault current path and may not have sufficient ampacity.
The interleaved conductors are a different case. If these are grounded at both ends then currents caused by induced high frequency voltages will be dissipated as heat. This will limit the magnitude of the induced high frequency voltages. They will also provide an excellent low impedance ground path.
I doubt that the insulated shields are part of the problem.
However grounding the interleaved conductors at both ends may be a good idea.
I am assuming that the interleaved conductors are the same ampacity as the power conductors.
Other possible reasons for failure?
Loose connections.
If the control system may drive the fans above rated speed they may overload. Check to see that the frequency limit of the VFD is bullet proof.
The reactor may be inadequate.
Get a scope on the motors to check for standing waves.
Consider a 1:1 isolating transformer in place of the reactor.
Point out that the installation does not meet the NEC and is not as designed. The warrantee is VOID. The installation must be brought up to spec and up to code before you are legally able to continue to assist. You will not put your ticket at risk by assuming any liability for non-compliant installations.
And;
Evaporators? Water and/or moisture damage is a strong possibility. You may have water dripping on the motors during the defrost cycles.
Bill
--------------------
"Why not the best?"
Jimmy Carter
RE: VFD power cable shielding (screening) issues
Best to you,
Goober Dave
Haven't see the forum policies? Do so now: Forum Policies
RE: VFD power cable shielding (screening) issues
I do lots of measurements on VFD motors and I have seen so many different problems that I sometimes wonder if any laws of Nature apply at all. But, I have decided that at least Ohm's Law, Lenz' Law and Maxwell's Laws are correct and useful. You just have to interpret the installation, the measurements, what people are telling you and - not the least - what you can see for yourself.
A "sloppy" installation by "neanderthals" usually doesn't have much influence. If cables are nicely lined up in a tray or not can have some aesthetical implications, but very seldom any electrical consequences.
A motor reactor usually is a good thing. But what if one of the neanderthals put a small ferrite core there? Seen that. It doesn't work. But I was told that there were motor reactors. Go inspect and see for yourself.
A common mode filter? OK, it definitely reduces voltage stress between Winding and slot wall. But it doesn't do much to the Winding-Winding stress.
Motor voltage measurements need to be done from motor terminal to motor frame. It is very difficult to get a good measurement if you measure terminal-terminal, simply because no (yes, I say no) so-called insulated battery driven portable oscilloscope has the necessary common mode rejection. You need an extremely good isolation amplifier or a differential probe with at least 10 MHz BW. The Fluke DP120 works, it has 20 MHz, but a limited voltage range (1200 V peak).
The best thing is to measure from motor terminals to motor frame and do the math (ch1-ch2, ch2-ch3, ch3-ch1) in the scope. You need to compensate your probes very carefully before you do this measurement, but that is about all you need to care about. Measuring with motor frame as reference potential removes all common mode rejection problems and will tell you the true stress that the motor windings see. Both Winding-Winding and Winding-Frame.
Oops - I got carried away! Sorry, But there's more to say about this. I have seen installations where several motors were connected to one drive and where some reflected waves combined destructively so that the peak voltage was quite low, while other motors did see constructive wave-fronts that took the insulation down. All that differed was the cable length from the distribution boxes along the conveyor system.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
I have to disagree that good probes don't exist. Tek have some excellent probes for use on high(ish) voltage systems. The best ones need a proprietary interface but the universal ones are still good. http://uk.tek.com/datasheet/differential-probe-hig...
Expensive!
RE: VFD power cable shielding (screening) issues
Read again: " It is very difficult to get a good measurement if you measure terminal-terminal, simply because no (yes, I say no) so-called insulated battery driven portable oscilloscope has the necessary common mode rejection".
It is the insulated scopes that don't have the common mode rejection needed. Even the lowly Fluke DP does the job. But it has a limited voltage range. I use a low ohm voltage divider consisting of Three 230 V incandescent lamps to overcome that. A resistive voltage divider limits the BW too much. At least if you want to keep size and dissipation down.
But that is history. I now use a Metrix OX 7204, which has more than ample band-width - and can do the math. Not all scopes can.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
thread238-295115: power system cable shield grounded only one end to reduce surges
=====================================
(2B)+(2B)' ?
RE: VFD power cable shielding (screening) issues
Consider a 1:1 isolating transformer in place of the reactor. NOT! If this were on 50 or 60hz INPUT to vfd side, fine, but NOT on output. You CAN use autofmr on output as we have done may times, but the reason is to change the vfd vs motor voltage - not an issue here.....
But what does / could happen if only one end is grounded? Could this possibly explain the multiple motor failures? Jraef, as you know, but are trying to cover all bases, tying ground at both ends WILL NOT CAUSE FAILURE. The tie both ends is to capture electromagnetic field INSIDE the cable, vs old fashion tie one end only 'at the source' - to prevent outside electrostatic field getting INTO the cables (like signal wires). It sounds like you may be questioning your 20 yrs experience in this: DON'T.
The failures are from something else, not from NOT tying motor end of motor cable shield to ground.
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
The point was, they did NOT tie the motor end to ground, it was taped off, not connected to anything. The shield was only connected to ground at the VFD end.
I was not given a chance to connect a scope yesterday. I am not allowed per company policy because of Arc Flash safety requirements, only certified technicians are now allowed to handle tools. Technically, I was not allowed to open the MCC cubicles either because the bus was still energized. But I digress...
More data emerged today in the form of photos taken by the customer after he went up on a lift to where the evaporator coils are. Turns out that the cable they told the contractor to use is apparently NOT the cable they had used, the electrical contractor did his own thing. So it does NOT have symmetrical grounds nested in with the conductors, it ONLY has the EGC and an aluminum foil shield. Actually, it looks almost like it is a Mylar shield, very flimsy.
I also got a photo of the motor windings from the re-winder, it looks a lot more like an insulation failure at the first turn than a single phasing failure to me.
I'll upload these to my photobucket site at home tonight and post them here, I am blocked from photobucket and all other photo upload sites I have tried from my work PC.
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
Another revolting development, 5 more motors failed today...
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
RE: VFD power cable shielding (screening) issues
Its not the physical distance from the VFD to the motor.
length from vfd to most of the motors = 100
measured length = 100 x 5 = 500
so if max is below 500 then you would have problems.
Used to do multiple motors on one vfd and this was a rule you had to pay attention to so your motors did not fail.
I thought some vfd manufacturers had this note in it. Cant remember which one had a detailed explanation of this issue?
RE: VFD power cable shielding (screening) issues
RE: VFD power cable shielding (screening) issues
The total cable length is more a question of how tough it will be for the drive to handle cable capacitances and does not primarily influence the overvoltage at the motor end.
But, if there are several cables with differing lengths connected and the PWM edges reflected waves combine constructively, you will have more overvoltage than the standard Udc*1.8 formula says. If they combine destructively*, you get a lower, and broader, peak voltage.
Two examples are shown in http://gke.org/rapporter/files/PWM%20transients%20... **
The installation has one VFD for each conveyor section. The VFD supplies junction boxes at the conveyor. The motors are connected to these junction boxes with separate cables with differing lengths.
* Or, more precisely, "unconstructively" due to different cable lengths.
** Measurements taken from motor terminals to motor frame.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
I would not expect the screen connection or lack of to have any impact on the reliability of the motors. I certainly will have an impact on the electrical noise and stray voltage around the area.
The standing waves that cause first turn failure are present in both screened cable and unscreened cable, but the velocity factor of the cable changes, so the voltage rise per meter is higher for screened cable than unscreened.
A properly bonded screened cable between the motor frame and the VFD frame will reduce the transient noise voltage on the frame of the motor, but in doing so will increase the transient voltage between the stator windings and the frame. It will also increase the discharge current through the bearings for capacitively coupled voltages as the high frequency impedance will be lower.
So, in answer to your question, I do not believe that the lack of bonding on both ends of the screen will contribute to motor failure unless there is some form of cable resonance/ringing, but, depending on the drive, it is possible that the output stage of the drive is being overloaded by the capacitance of the cable as the total length probably exceeds the maximum length for the drive. In this case, one or more outputs may be "desaturating". This would normally cause an earth fault or similar, but it may be disabled or malfunctioning.
Many gate drives check that the device is fully on a very short time after the gate drive is applied. If the capacitance is too high, it can take time for the device to fully saturate and if the device is not in saturation, the gate drive is removed.
There could be a mechanism along these lines that is causing the output waveform to be assymetric and damage the stator windings.
I would suggest that as a minimum, you may need an output reactor, or better, a sinusoidal filter between the drive and the motor cables. Sometimes, just a simple "ferrite" common mode ring will make a significant improvement.
Best regards,
Mark
Mark Empson
Advanced Motor Control Ltd
RE: VFD power cable shielding (screening) issues
"...or better, a sinusoidal filter between the drive and the motor cables"
I had an interesting case in Eindhoven, NL. There, a cooling system with Ziehl-Abeg fans had been installed. The motor supplier said that the fans needed sine filters and there were such filters.
In this installation, there was around one inverter failure per month. Inverter failure, not motor failure.
I have measured funny current and voltage waveforms at certain frequencies and it seems that the filter's resonance frequency and a fourth of the carrier frequency coincide so that there are large curren transients between VFD and filter. Simulations confirm that this can happen and the Q of the sine filters has now been reduced (a few ohms added in series with C). We are watching what happens. It seems to help, but we do not know for sure. Needs a few months evaluation.
So, filters can be a problem, too.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
They were also rather small. And there were lots of them on every inverter. The manufacturer says that those motors need sine filters. Without a sine filter, the motor will fail, he says.
I find that quite remarkable. Why deliver motors that cannot tolerate PWM? For applications where PWM is the rule.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
Yes, been there, done that, got the tee shirt etc.
Those motors are a high slip motor designed for use with variable voltage (not inverter) control. They are typically an integral part of the fan and commonly inside out where the rotor is on the outside.
Because the rotor is very high resistance, the carrier frequency creates high slip losses in the rotor which overheats and cooks everything else!
With sinewave filters, there are two types. The cheap ones are a simple series reactor into a capacitor bank and while they give a reasonable sinewave phase to phase, they have a lousy waveform to earth. Lots of common mode through a low pass filter. If you run a long screened cable on those, you get a lot of earth current at lower frequencies. I lit up a 230V 100W light bulb with the earth currents on a three hundred meter length of cable. And to make life more interesting, the capacitors between the DC bus and the chassis are too low in value to provide a return path for the common mode current, so it has to go back to the transformer and come in via the three phase to the rectifier. We had 150V of noise at a relatively low frequency superimposed on the phase to neutral voltage and blew lamps and switchmode power supplies.
The better sinewave filters include sizable capacitors from the filter three phase output back to the DC bus on the drive to get rid of the common mode rubbish.
Most of the submersible pumps here, run on VFDs, yet they still make them with a maximum voltage of 800 volts peak. Still havbe not moved with the times!!
Best regards,
Mark.
Mark Empson
Advanced Motor Control Ltd
RE: VFD power cable shielding (screening) issues
I can't wait to hear from Jeff what make his motors are.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
Have you ever tried it? I only did limited testing, but the results I got showed that a delta-wye transformer will provide a sinewave output voltage without the issues a sinewave output filter can introduce as mentioned by Marke. Also, it's much simpler then a sinewave filter with no capacitors so it's less likely to fail.
RE: VFD power cable shielding (screening) issues
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
Why not? Stay within the V/Hz ratio of the transformer and it will work fine as the frequency varies. You need to ensure that any voltage boost (either due to the fixed V/Hz drive programming or the automatic flux vector operation) doesn't saturate the transformer.
RE: VFD power cable shielding (screening) issues
Overall MCC installation.
One of the buckets with a single VFD feeding 4 motors. Notice that shield is terminated separately from EGC.
Terminations removed to show more detail. the cable has a braided shiled in addition to the foil, but they cut the braid, twisted the foil and terminated it...
Cable routing to tray. These are not "VFD" cables, they are just shielded tray cable.
Spaghetti bowl routing through a wall penetration. No attention to noise abatement procedures (not to mention professionalism).
Junction box / common disconnect at Evaporator end (where the motors are). Shield is just cut off at this end, not connected to anything. EGC terminates on the ground lug next to the disconnect, but the remainder of the ground wires are only connected to it via the enclosure back panel.
Detail showing that the cables exiting this box going to the individual motors are just using flex cord through cord grips, no more shielding.
Photo from the motor shop, showing what they claim is 'single phasing" damage to the winding. I believe this looks more like damage due to standing wave insulation breakdown, at roughly about 2 o'clock.
They are claiming that the while taped device in this photo is the embedded thermal protector for the motor, hence they do not need individual motor overload relays. If it really is a 3 pole "klixon" device opening the Wye point of the motor windings, then they would be correct in that code assessment, but I'm not convinced until I see data sheets on the motor. It's possible, but it's also possible that this is just a thermistor connection, and they ASSUMED it would be OK, but in fact is completely useless without being connected to a Thermistor Relay somewhere that would then tell the VFD to shut down. Still waiting on details of that.
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
any more motors fail?
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
No restrictions when it comes to PWM? Like for the Ziehl-Abegg ones?
I still haven't got a clear picture of the cabling. Are all motor cables run from the inverter? With one cable for each motor?
Or is there one cable running from inverter to junction box and then separate cables running from junction box to motors? See 17 Jun 14 23:44.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
The length of the cables to each motor 300ft to 400ft is around the length that could cause transient voltage amplitudes to double.
The interesting thing to find out, is if the breakdown is first turn phase to phase, or phase to earth, or center winding shorted turn or similar.
Do you have a maximum insulation voltage rating for the motors?
We commonly use a single VFD, around 4KW, to supply up to 4 motors on the output with neutral screened cable and average cable length of 60M per motor. We add a manual motor starter at the VFD end of each cable to provide the individual motor protection. - never had a problem.
As a thought, I wonder what would happen if the thermal switch opened a few times while it was running on the output of the VFD and that length of cable. Would it produce some very high inter winding voltages at the switch end? I would expect so. Once again, an indication of where the breakdown was may give a clue, but here, it is very hard to get sensible information from the motor rewinders. I have to go and look myself because every failure of a motor on a drive is reported as "first turn failure" even when the windings are 100%.
Best regards,
Mark.
Mark Empson
Advanced Motor Control Ltd
RE: VFD power cable shielding (screening) issues
The more I think about your last statement, the more it makes sense. Whenever I tell people not to open a contactor down stream of a VFD, they ask why and I explain how the arc formed in the separating contacts acts as a capacitor tank circuit tht increases the voltage until the dielectric of the gap finally stops the current flow, and that dV/dt will eventually damage the transistors. But I never thought through the effect that might have on the motor windings and although I have never seen that, it may be because the transistors fail first. In this case it may be that because the VFD is protected from that dV/dt by the load reactor the effect is moved down to the windings, or at least maybe exacerbating any reflected wave spike action also taking place.
New info that may also be interesting. My employer recently bought a small company that makes power quality equipment, most famous for having a voltage sag ride-through device that does not rely on batteries or capacitors. As part of their sales efforts, they created an on-line power quality meter that, if enabled, allows users to access its records of line power anomalies. As it so happens, there is a nearby customer (less than a half mile away) that has one installed and is web enabled, so on Monday my product manager for that system is going to check the reports for the window of time in which the latest failures happened, since we know they happened between when I left and the following morning. That may be interesting as well, either to see if there was something, or even to eliminate that possibility.
I'm writing my report this weekend, in which I am going to tell them that without much more thorough analysis by someone who can spend sigificant time on site and work live with proper metering equipment, the best I can offer up are theories and possibilities. But I'm going to lean on this being insulation failure due to high voltage spikes caused most likely by the reflected wave phenomenon exacerbated by installation problems and possibly other factors. I'll wait for that other report before I send it of course, but I just can't help but stay focussed on this as the issue. I think it is too many failures to be loose connections, nobody is THAT bad, and the damage evidence I see in that motor photo does not point in that direction, or to sustained overload conditions.
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
It appears that the common denominator seems to be time related. If I understand you correctly, you seem to get multiple failures rather than just one randomly.
If it is a winding to ground fault, then is it possible that there is a three phase supply problem where the three phases are moving together relative to ground and transient voltage exceeding the insulation to ground? This could be a combination of transient voltages from the VFD coinciding with transient voltages on the supply.
Many years ago, I found a problem, always the same your of the night, causing a bunch of VFDs to shut down on instantaneous overcurrent.
This was tracked down to power factor switching on the 33KV causing common mode transient voltages on the 33KV which were then capacitively coupled through the supply transformer and shifted the whole supply relative to ground. The capacitance to ground on the output of the VFD (Motor and cables) cause a current transient at the output of the VFD.
Fixed by replacing the corroded earth strap on the star of the transformer.
A severe brown out situation could cause current to rise and overloads to operate as another scenario, but that should be more obvious.
Best regards,
Mark.
Mark Empson
Advanced Motor Control Ltd
RE: VFD power cable shielding (screening) issues
Jraef,I was sent to one of the large last remaining steel mills in Philli on such a case. For a month previous, the local motor repair folks, the OEM motor mfgr sent engineers, and no one found why 460v motors were failing anywhere from 5 MINUTES TO 3 HOURS - even under NO LOAD. About 20 motors failed during that time. Never once did the drive fail
I told the customer he had to have a loose wire on one of the motor phases. In my diagnostic process, I first added a motor inductor - no help; they lost 2 more motors in 1 day. Next we had the motors vacuum impregnated as we KNEW the fault was ozone eating the varnish causing shorts. No fix.
I finally went to the customer site. I found a loose screw on one motor phase at a terminal strip below this huge machine. Simply a loose screw. Result was in excess of 3000v spikes. Scoping the normal motor spikes showed about 1700v spikes; my motor designer assured me the insulation was good to 2500v; above that ozone would build up and in short order eat the varnish and short windings.
2 turns tightening on that screw and the motor has run ever since my trip in about 2000.
Don;t discount a possible loose connection in all those cables.
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
Loose screws usually show in motor current and performance. And it shall be a hell of a loose screw to be able to produce even 1 kV peaks. Remember that a loose screw isn't the same as an air gap.
Also, if there is any current in a loose connection, you will have heat.
And, finally, there should (yes, I know about should and reality) be maintenance looking after loose screws. IR inspection is used for that.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
That said, one can google 'how fast does ozone breakdown polymers' and find various studies that prove it can cause issues in much less than 1 hour. For example, take this study I just googled:
http://books.google.com/books?id=EV6vG-LfwWsC&...
You can see in figure 1.11 that ozone in such small concentrations as 3ppm can cause cracks in polymers (typical wire varnish) in less than 2 hours; 23 minutes for some materials per figure 1.12.
But what about near 100% concentration? How long would THAT take to crack the insulation?
But irregardless of cracking, what if a wire is ALREADY NICKED from being hammered into its slot? It does not require this 23 minutes to crack; so if already exposed, is it not READY to short as soon as ozone builds up?
What is the resistance of ozone? I did not google it as my experience with DC motor commutators flashing over already proved to me many many years ago that ozone build up around a commutator makes for ZERO ohms and thus instant flames; I may be wrong but I doubt it.
I must admit that attempts to vacuum impregnate the stators DID extend the life of these motors from this one loose phase connection: from 5min or 2 hours - to a day or two. Reason seemed obvious at the time: No air pockets between windings inside the stator meant no O2 to chemically change to O3, so it took longer for the ozone to build up and cause cracks in the insulation.
I believe ozone can be miade from air in very short order. I did not research this, but assume from empirical data of seeing atom bombs split atoms in useconds it does happen quickly.
So as strange as it sounds on the surface, I stand by my empirical data that tells me a loose connection can cause motor insulation failure in as little as a few minutes.
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
RE: VFD power cable shielding (screening) issues
I think that by the time I was done I used pretty much everything mentioned in here to assign blame, then I might have ended up getting a bit carried away by including President Obama (a popular sport now), but recommending that they hire someone with REAL tools to go spend some time on site and investigate this more deeply.
We'll see what happens next.
** For those too young to recognize it, that's an excerpt from the song "Alice's Restaurant".
"Will work for (the memory of) salami"
RE: VFD power cable shielding (screening) issues
Arlo (still touring - saw him last year) would be so proud of you!
www.KilroyWasHere<dot>com
RE: VFD power cable shielding (screening) issues
Guthrie's autobiographical incident that inspired "Alice's Restaurant" occurred in 1965 in western Massachusetts (anything west of Dedham is the frontier to a blueblooded Bostonian and Stockbridge is nearly on New York border) . Ten years later, in 1975, having been the discharged from long-suffering 'service' in the US army, I returned from Germany and landed in, what to me, was a foreign land, eastern Massachusetts (I'm a mid-western by birth and worldview).
I, like Mr. Guthrie, was charged with violating the sanctity of the town dump; specfically, illegally removing property, to wit, a Chevy 305 engine complete with bell housing. He was dumping, I was scavenging, both illegal in the Commonwealth.
I worked the incident to great advantage, being 20-something at the time and trolling the bars for feminine company. Any 20-something knew the Alice's Restaurant in 1975 and could sing the refrain; some lovely ladies could even in harmonize.
Dan
RE: VFD power cable shielding (screening) issues
Home of Museum of Bad Art - Art, too bad to be ignored!
Been there twice. Love the Place.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
5 months later, they repaired all of the shield connections, made up missing ground connections, cleaned up the wiring, everything I told them to do. We sent a field tech out to the site who scoped everything, no problems found, no voltage spikes seen, good ground continuity throughout. Clean, normal, good.
20 more motors have failed... I feel like I'm in the twilight zone here.
The motors, Baldor "Inverter Duty", have been sent to Baldor corporate (as opposed to local shops), they came back and said the failures are consistent with first turn insulation breakdown due to reflected wave phenomenon...
Per my suggestion, the user installed what we call "Motor Terminators", basically a self-contained RC snubber system that you mount near the motor to absorb the spikes if they happen. Since then, no more motor failures, but it hasn't been long yet, only about a week for the last installs. But... now the Load Reactors in the VFD cubicles are overheating and the insulation on THEM is failing, but only on the units in which the distance from drive to load is right around 250 ft!
The consensus now, from a large sampling of power quality gurus within my company, is that it might be a rarely seen phenomenon in which the reactors act in resonance with the cable capacitance to actually create reflected wave voltages, and since nobody has ever added the terminator down stream of a load reactor (because there would be no point), the phenomenon is exacerbated but the effect is only seen at the reactor end. Can't explain exactly why, but we're going to tell them to remove (jump out) the reactors on those units and see if the motor failures cease.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
RE: VFD power cable shielding (screening) issues
I think that the whole situation could benefit from real world measurements instead of consensus in a "large sampling of power quality gurus within your company".
Excuse me, but PQ gurus seldom have a clue about what can happen between motor, cable, VFD and other components that have been installed or added. Have those gurus produced anything else than consensus? A failure model with verified data so you can do some basic calculations?
Have any measurements been made? With calibrated and compensated probes that 1) can withstand the voltage and 2) whose band-width isn't degraded by a "protection resistor" that some of the very large companies tell their field guys to use. And, further, have those measurements been made directly between motor terminals (which is a bad idea because of the capacitance from scope GND to real world GND that exists even in the best battery operated scopes) or have the measurements been made terminal to GND in differential mode (the only reliable way to do that measurement if an insulation amplifier isn't used)?
The fact that the load reactors are overheating points to resonance between reactor and cable and that resonance often is in the 10+ kHz range, which turns the coil into an inductive oven that heats the core. Are there any damping resistors parallel to the reactors? If not, you need to add them.
The terminators (RC combinations) that I have tested usually never worked. My recipe is thus:
1. Find out what the motor terminal voltage actually is and compare rise-time and peak voltage to the manufacturer's limit.
2. Check the reactor/cable system for ringing. Add damping resistors parallel to the reactors until the ringing is gone.
3. If necessary: Add common-mode filters. Not measly ferrite cores, they are not very useful. Use nanocrystalline cores. They are, in my experience, the only ones that help.
Excuse my rant, but when I hear about 20 motors that have failed - I can't shut up. This is, after all, what I do most of my time.
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
RE: VFD power cable shielding (screening) issues
Gunnar Englund
www.gke.org
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
RE: VFD power cable shielding (screening) issues
Bridging out the reactors could cause issues with the capacitive current overloading the IGBTs and causing VFD failures.
I think Gunnar is probably right (as usual) and there will be some high frequency ringing. I have seen it a number of times, and also been told that there is none because the scope has too low a bandwidth, or the operator is not looking at the HF components. Power quality type scopes usually just don't cut it for this type of investigation.
Becoming an expensive exercise, glad I am not in the firing line for paying the costs!!
Best regards,
Mark.
Mark Empson
Advanced Motor Control Ltd