EDM - Breakdown Voltage on bearings 3

[Marke]

Hi

In a situation of severe abnormal bearing loading due to mechanical misalignment and vibration, is the potential breakdown voltage of a bearing reduced?
Logic would suggest that as the thickness of the insulating lubrication layer would be affected, the breakdown voltage would also be affected.
I have looked at a situation where there has been premature bearing failure which is being blamed on EDM. There are alignment issues and possibly major vibrational problems also which alone would cause premature bearing failure. The reported failure is EDM. I have not seen any failed bearings, but the shaft voltage is low.

Any thoughts or experiences?

Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[Skogsgurra]

Right. Oil film thickness is reduced by high temperature, high load, speed and also dependent on oil quality. Mineral oil produces thicker films than synthetic oils do. (Note: the lubricant in grease is still oil. The grease is a result of (lithium) soap mixed with oil).

Vibrations are not good and bad alignment neither. Does bad alignment still exist? Arent everyone using laser and such things?

High total capacitance can be a problem and cause severe damage to bearings even at low voltages below 5 volts.

More details in answer to your e-mail.


Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Skogsgurra]

...reduced by LOW speed... of course

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Skogsgurra]

I made a presentation in Schenectady last week. The material is here.

http://www.gke.org/presentationer/files/GE Bearings and their electric environment w O and E.pdf

It is a short version of the seminar we are going to run in Sweden in April. Short but comprehensive.


Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[MagicSmoker]

That was a fascinating presentation, Gunnar, and worthy of the coveted magenta star!

I design definite-purpose traction drives and have been slowly getting up to speed (pardon the bad pun) over the last few years on EMC compliance, but many of my customers don't really care if a drive I design radiates less noise (that is to say, they only care about EMC compliance if regulations force them to care). However, it seems that the same solution to reducing EMI on the motor side of the drive - ie, a common mode choke - can also be touted as a means to reduce bearing erosion from EDM (and could also be used by the original poster here - just slip a toroid of appropriate size and core material over all of the motor cables). Assuming this is a valid proposition, how do you optimize the common mode choke impedance to reduce bearing voltage?

 

[Marke]

Hello MagicSmoker

EMC is an issue with any electronic drive system and you are correct, the common mode choke will have an impact on both EDM and EMI.
The other critical thing to get right, is to provide a very good low impedance return path from the motor frame to the DC bus. This requires good low impedance decoupling from the DC bus to the chassis and a large surface area on the return path. These paths must be a low impedance at high frequencies, (100KHz + ) This requires short leads on bypass capacitors, no pigtails anywhere and the large surface area return is easily achieved by the use of a screened cable properly terminated at both ends. see

http://www.vfd-emc.co.nz

The type of core and the cross sectional area of the core of the common mode choke have a major impact on it's effectiveness. Not just a case of using the closest ferrite toroid!! I have found good effects on existing installations by using high permeability ferrite U cores designed for LF switch mode applications. You can glue them together over the existing cables after holding them in place to check if they actually work.

Gunnar's presentation is fantastic as usual, he is the guru and needs to come to NZ to teach us over here!

Mark Empson
Advanced Motor Control Ltd

 

[electricpete]

LPS to Gunnar for the presentation – I’m going to study that. Good info from Marke as well.

What popped into my mind in response to the original question is that we need to be aware of EHD (elastohydrodynamic) lubrication principles.

While a sleeve bearing (hydrodynamic lubrication) has very simple relationship where film thickness decreases in response to load, rolling bearing (elastohydrodynamic = EHD lubrication) is different. The rolling bearing deforms as approximated by hertzian theory. This particular factor (increase in contact area under load) actually would make the contact resistance of a given element decrease as load increases.... oppposite the direction we’re thinking about for the film.

What happens to film thickness under load under EHD conditions: To my simplistic thinking, under EHD lubrication, things like surface roughness and thickness of tightly adherent molecular layers dominate the picture, such that film thickness does not change much with load.

Google shows me this which perhaps supports my simplistic view:

http://bura.brunel.ac.uk/bitstream/2438/5077/1/FulltextThesis.pdf

Equation 55 suggests that for a cylindrical roller bearing, the film thickness h ~ W^(-1/8)
Thus thickness decreases as load increases, but only very slightly (a very weak dependence).

I’m sure someone somewhere has made same types of estimation for ball bearing, but google does not take me there quickly.

Does the increase in contact area or reduction in thickness win out? I'd guess contact area but my guess is no better than anyone else'. I guess the main point to make is it's not as simple as hydrodynamic lubrication and difficult to predict without finding something written by someone who has studied it very carefully.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Sorry, I was framing my discussion in terms of contact resistance, which is affected by both contact area and film thickness.

You asked about breakdown voltage, that logically would not be affected by contact area... only by film thickness (which changes weakly with load).


=====================================
(2B)+(2B)' ?

 

[Skogsgurra]

Thanks!

Pete, I think that you will appreciate Annette Mütze's thesis. It is here:

http://www.ew.tu-darmstadt.de/media/ew/dissertationen/dissannette.pdf

Look for Herzian contact area and oil film thickness. Lots of data there.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[electricpete]

Thanks, it is a good reference although I’ve never given it the time it deserves to read. I see from flipping to Appendix D, equation 2 that it gives the minimum film thickness (which I interpret to mean the film thickness at location of minimum, ... NOT minimum required) as a function of load (w’~Q = load), and the dependence is:
hmin~w’^(-0.13) which is practically the same as the previously-linked reference which had exponent –1/8 = -0.125. They don’t come out and say whether it’s ball or roller, but one of the parameters v2 mentions surface contact at point or line which suggests maybe the equation applies to both.

If we take that number as representative, then doubling load would change film thickness to 91%. Not a huge change. They also show strong dependence with speed (low speed gives smaller film thickness) and temperature (higher temperature gives smaller film thickness). .


=====================================
(2B)+(2B)' ?

 

[electricpete]

ok, now I see that equation D1 is for point contact (ball bearings) and equation D2 is for line contact (cylindrical rollers).

D1 for ball beairngs gives an exponent -0.073, which is even lower than previously discussed
D2 for cylindrical roller gives the exponent -0.13... (makes sense this is the one that matches the other reference which was talking about cylindrical roller bearings)

=====================================
(2B)+(2B)' ?

 

[MagicSmoker]

Marke - as the OP here, have you already employed filter/shielding measures on the motor cables to reduce the level of impressed voltage on the motor? Also, how are you evaluating the level of impressed voltage? It seems to me that getting an accurate measurement of the impressed voltage would be especially difficult, whereas measuring the common mode current would be trivially easy (by throwing an amp clamp or amp-flex coil around all of the motor cables). That said, Gunnar's method of using an incandescent lamp to check for impressed voltage was brilliant (even though he seems a bit ashamed of it in his presentation... )!

 

[Skogsgurra]

You may be surprised to learn that a perfect low-impedance symmetric and shielded cable is the worst thing you can have if your load is floating (fan coupled directly to motor shaft).
I am going to a Siemens factory tomorrow where that is the case. I expect some resistance from the guys I'm going to talk to. But I have armoured myself with convincing arguments and will also do measurements to prove it.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Marke]

Both EDM and EMC benefit from common mode filters, but when it comes to the low impedance return path, that improves EMC dramatically, but can make EDM much worse as the reduced HF impedance can increase the current flow through the bearings.
I commonly see the "EMC solution" promoted as the "EDM solution" also, but the "EMC solution" is generally about screened cables and input filters, not common mode filters.
Some drives include common mode filters as standard, others have common mode filters as an option and others just do not offer at all.

From an EMC perspective, the low impedance return path and internal decoupling are paramount. The common mode filter will further improve EMC and also improve EDM, so it is a good idea to do as standard to reduce the risks of problems.


Mark Empson
Advanced Motor Control Ltd