Cause of RFI during chiller compressor operation 1

[PQ101]

As soon as the newly installed chiller starts-up the reception on AM radio signal disppears for appx. 5-6 blocks. The 200Ton chiller is equipped with harmonic filter. Equipment manufac. has replaced all circuit boards, the SCR panel, installed isolation transformer, chokes around the cables, installed PVC on incoming power conduit, etc. but made no change.

You can basically hear compressor motor whining-distortion on AM radio anywhere from AM500 thru AM1500.

To start with grounding from 480V Main panel to isolation transformer to chiller starter enclosure to compressor motor was checked for proper installation as per NEC and verified OK. Plan was to set power analyzer to check for waveform distortion including the equipment ground.

As soon as chiller is shutoff the AM radio signal comes back to normal. Appreciate any help to narrow down the cause.

 

[PQ101]

Thanks ozmosis. The unit is sitting outside the buillding in school yard. FYI- The noise is generated during operation of compressor motor, OFF when unit is shutdown. Do you mind elaborating further please.

 

[PQ101]

I was advised the following was implemented in this order:

1) Removed snubbers on drive
2) Grounded chiller with dedicated ground rod
3) Replaced logic and trigger boards
4) Replaced entire drive assembly
5) Replaced inductor
6) Added ferrite beads on incoming and both sides of inductor
7) Added reactor and capacitors on incoming lines
8) Removed line side at chiller and ran unit off generator, this seemed to work
9) Removed reactor and capacitors and added isolation transformer- Problem remain.

Seems like when the chiller was running on generator set the RFI was eliminated completely, indicating suspect cause may be voltage related perhaps.

Providing power quality solutions that empowers you!!!

 

[waross]

Do you mean that you ran the drive on a generator, or just the chiller motor without the drive?
What is the configuration of the isolation transformer?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[PQ101]

I did not, engineers from York did (item#8 above indicate was done sometimes in the past).

The 480V 3-phase from main switchgear lands on islation xfmr primary (left hand side shown on attached picture), incoming ground lands on xfmr chasis & bonded to xfmr neutral, and departs to chiller enclosure (lands on GRD lug). The 3-phase secondary (on right side) departs to chiller and lands on drive 480V disconnect.

Providing power quality solutions that empowers you!!!

 

[PQ101]

Plan was to check the transformer connections, not sure if this is contributing to RFI, but something is not right based on data-voltage ditortions.

 

[Skogsgurra]

Ozmosis has a very good point. If there is nothing to pick up the interference, then you will not know it is there.

Could it be that other units do not work close to where AM radios are used? Or that AM signal strength is much better in other places so that interference from compressors are not noticed?

Again, I do not think that the 'hundreds of other units' are any better than the unit causing problems.

The test with a diesel was probably run with short cables from generator to compressor. That illustrates the importance of cable length. As you know, a conductor starts acting like an antenna when it gets close to 1/4 of the emitted wavelength. At AM 1000, the wavelength is around 1000 feet. So, cables less than around 250 feet will not emit much, while grid cables are much longer and emit much more. So, the test with the generator means that emission is not radiated from the unit, but from the cables.

That makes use of ferrite cores on the incoming (to compressor unit) cables easy and effective.

Why doesn't an isolation transformer work? The reason is that interwinding capacitance is high and lets HF through. If you add a static screen between windings - and connect that screen to a good, low impedance ground and with a short cable - you may also get rid of the interference. But, ferrite cores are much easier to apply.

I think that York should help you with this. They have put the equipment on the market and it is their responsibility to make it clean. FCC 15 is what applies in the US.

I am afraid that Champaign, Illinois is too distant for 'popping in'. I will go directly from Sweden and stay in Buffalo for three days. Going to Illinois and doing some meaningful work will take a couple of days and that would mean money. I think that the York people shall pop in instead.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ozmosis]

PQ101
The elaboration I would make is quite simply that you can spend many thousands of $$'s improving the emission levels of the said equipment but a few $'s improving the immunity might do the trick. You will not eliminate RFI completely, just getting it sitting within the compatibility levels.

I recall a few years ago in the UK we had an installation with 7 x 560kW VFD's on a water pump station in a quite remote location. All VFD's had the (legally) required EMI filters fitted, the installation was top notch and all was running well until some old dear down the road complained that since the new installation her old TV started getting fuzzy lines. Water authority came to us and asked what we could do. In the end it was either make about $100k worth of investment (I've lost my GBP sign now on my keyboard..) to hopefully keep her happy but in the end we bought her a new TV, top of the range for about $1500. She was as happy as larry, the water authority saved themselves a fortune and our reputation remained intact.

However, if you want an expert on EMI, Gunnar is your man. But don't expect a new TV from him.

 

[waross]

I am wondering about the quality of your grounding. Is the dribbley little #14 wire in the picture the only ground or is there a code compliant ground connection somewhere out of site?
More importantly, I am concerned with the voltage waveforms posted at 21 Nov 09 20:02.
These wave forms are showing an almost textbook case of a neutral shift. A voltage drop due to load usually shows up on two phases. Here we have a voltage drop on only one phase, but with an accompanying phase shift.
Remember the story of the man who lost his keys in a dark alley but looked for then 50 yards away under the light "Because the light was better."
Given the history of trouble-shooting here and the assertion that this is only one out of hundreds with a serious EMI it may be prudent to investigate other possible sources of interference. I am thinking possibly a poor connection on a transformer neutral that tests good with no load but goes open under heavy loading and possibly arcs under normal loading. I would also rigorously check the motor for possible faults.
I submit this suggestion nervously as I feel that I am in the presence of greatness when Skogsgurra and ozmosis are collaborating on a drive problem.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[PQ101]

Thanks waross.

For clarity- York engineer had disconnected equipment ground at both ends (480V Mains & at the isolation XFMR) inserted a new ground rod at foot base of the chiller, connecting ground rod to chiller frame, and bonding via dribbly wire too XFMR ground (thinking this solves the RFI). Grounding system is back to original as per code. The thin copper wire is no longer there as of last Week.

Equipment ground connection between isolation XFMR and chiller- There are 2 pairs of conduit piping departing from isolation xfmr to chiller enclosure carry 3-phase delta and ground (#6AWG) in each pipe. Both ground conductors land on chiller enclosure (GRND lug). I will make sure all connections are tight, but not sure if increasing the size of ground conductor will help the situation. (Attached is as-found with the orginal incoming grounds disconneced-right side is the secondary departing with 3-phase and ground to chiller disconnect).

With respect to voltage sag-drop events during the inrush, it's concern that I need to fix (starting point). Typically the voltage should not decrease but 1-5volts in a good solid system. With load current incresing to 99Amps, voltage delivery decreasing to 250volts, means there is weakness somewhere.

As for the compressor motor, there is no indication of a fault but I plan on running a check to verify OK. Thanks

Providing power quality solutions that empowers you!!!

 

[Skogsgurra]

Bill, I never noticed that ground connection. You may be right that I thought it was brighter somewhere else

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[rasevskii]

For PQ101

What about the 480V 3 phase incoming supply-how is that grounded-is there a neutral (like a 277/440 system), where is that grounded? Is there a dedicated padmount transformer just for this or a pole mounted cluster of 3 units? What else is connected onto the 480 volt system you have? Maybe you have in effect a broadcast antenna there...

Are there any overhead connections to a pole or is it all in grounded conduits...where are the conduits grounded..Is anything connected to a water pipe ground...Are there any non-metallic water pipes in the system, giving an apparent ground but not actually...

Try removing and reconnecting different grounds as the system is running to see what makes a difference.

What about the schoolyard fence? Is that grounded or not, how far away is that...connect and disconnect that...

I was once dealing with a traction substation where we had a 60 to 25hz static converter in operation. Using a clamp-on meter, hundreds of amps were found flowing in fences, railings, pipes and metalwork all over the site..

Just some thoughts..

regards, rasevskii

 

[PQ101]

The 480V Main switchgear is served by 500KVA padmount transfofmer (ABB), 480Y/277V, 3-phase connection. The incoming system neutral (from utility padmoiunt) lands on swgr ground bus, bonded to GRND rod and cold water piping (verified OK). The 480V (chiller dedicated) disconnect switch serves the chiller isolation transformer-added recently. Another 480V disconnect swithch serves a 500KVA padmount xfmr inside the building. Plan was to check loading on both locations to determine cause of voltage sag and get a better idea on loading issue (if related in anyway).

The conduit piping has PVS added to it right before the isolation transformer and as metallic conduit top entry land on the enclosure, (York thinking that would solve the RFI). So, the answer is yes, there are PVCs in the system piping. Plan was to install jumper across all PVC connections, and place back to original.

At the isolation transformer, 3-phase comes in with the ground conductor bonded to xfmr neutral, and departs from secondary to the chiller. There are no additional ground rods in the system-all removed. There are no over-head cluster-mount 3 pot transformers anywhere nearby, and the school fence is about 500 yards away.

Thanks for your input raseskii. I appreciate that.


Providing power quality solutions that empowers you!!!

 

[rasevskii]

For PQ101

Thanks the photo. This may sound ridiculous, but if the building has a metal roof, that may be the antenna... If so, it is probably bonded somewhere to the electrical panels or conduits. If so try breaking that. The roof, if so, has to have grounding at some points legally, I believe. But that can be elsewhere in the building complex.

regards, rasevskii

 

[waross]

I have been on many projects where the refrigeration techs were not allowed to do the power connections to their machines. Been on a few projects where the electrical crew corrected the power wiring done by the refrigeration guys. That ground connection is a good example why.
But now that it has been repaired, enough said.
Gunnar, will you take a second look at the voltage waveforms and comment please? It looks to me as if a dirty connection on a transformer primary wye point may be going open under load. I don't see that waveform as causing the EMI, but whatever is causing that waveform may also be causing EMI.
The changes in voltage are too abrupt and too regular to be a reflection of starting current, and it is always only one phase.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[PQ101]

Thanks waross, it's affecting all 3-phase.

Providing power quality solutions that empowers you!!!

 

[Skogsgurra]

I had a look at the Dranview recordings. They do not look right. But, they also do not look like a bad connection in the power circuits. Such a bad connection would result in heavy arcing and the voltage would, because of the arcing, not change as rapidly as you can see in these recordings - the saw-tooth waveform at the end of disturbance in the blue phase is a very clear example. It looks, at least to me, like the connections between grid and Dranetz had a problem. Not between grid and compressor.

Also, the influence on AM radios is not something you get from loose connections. But you do get it from a PWM inverter. It outputs a continuous carrier frequency that is more intense than the AM broadcast stations and that is what kills the radio stations. Intermittents do not do that.

I need to repeat that recordings made with a milliseconds timebase never can show details that interfere in the MHz band. To see those, you need a microsecond timebase and corresponding BW. None of the Dranetz boxes have that.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[waross]

Thanks Gunnar. I was wondering if a loose connection ahead of the VFD may be arcing under load and acting as a spark gap x-mitter.
Yours
Bill

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[PQ101]

Thanks skogsgurra.

As you can see the line voltage on VFD disconnect decreased to 250volts (when load increased to 99Amps) indicating the existing 500KVA padmount xfmr is unable to support the voltage during the inrush. Padmount (attached picture) may need to be upsized to support the load during inrush as well as under continueous load. I was wondering if undersize xfmr could be a suspect cause-in any shape.

As for arcing, I believe all connections are verified-tight (I will recheck again though). Comments are very helpful.

Providing power quality solutions that empowers you!!!

 

[PQ101]

Based on monitoring: looking at phasor diagram it appears there is something wrong with supply voltage- or perhaps neutral-bonding connection somewhere before it gets to chiller disconnect switch, either at isolation xfmr or 480V main panel.

Data shows minimum load, and yet there is voltage diturbabce (taking chiller out of picture for a moment). Seems like I have to fix this situation first- making sure supply voltage-voltage delovery is good to start with.

Providing power quality solutions that empowers you!!!

 

[waross]

Full load rating on a 500 KVA X-former at 480;277 is 600 amps.
Even if it was already 100% loaded, it should sustain an additional 99 amps.
Assume the regulation at 5% (conservative) and full load will be expected to drop the voltage 24 Volts. The addition of a 99 Amp load may be expected to drop the voltage by about 4 volts. A rigorous solution will take power factors and phase angles into consideration, but transformer voltage drop can only explain a voltage drop of around 4 Volts, not a drop of over 200 Volts.
Is it possible that there is a bad connection somewhere on the supply side that is emitting EMI, but notenough to overcome local radio stations? The addition of the chiller load may be increasing the amplitude of such interference to the extent that it does overcome the local station signals.
Did you take your voltage readings ahead of the VFD on the incoming supply? Did you change connections between starts? It is strange that two phases stay healthy but a different phase shows a drop on each trace.
Have you walked around the area with a small radio to find the area of greatest interference? Try near the VFD drive with the cover open and closed.

Bill
--------------------
"Why not the best?"
Jimmy Carter