Contactor pole frequency limits 4

[thinker]

We plan to install AC contactor between VFD inverter and PM motor.This contactor would be opened only under no-load conditions, so we have no concerns regarding arc breaking issues. However, when contactor is closed and the system is running, the inverter output frequency could reach 800 Hz (fundamental). The contactor pole specification (Telemecanique, type LC1F630, 630 A) limits the frequency to 200 Hz. Is this limitation based on additional losses in closed contacts? If we operate above the limit of frequency, is there a chance of contacts welding? Please advise.

 

[electricpete]

The contactor pole specification (Telemecanique, type LC1F630, 630 A) limits the frequency to 200 Hz. Is this limitation based on additional losses in closed contacts?

It could be skin effect losses. That seems like a straightforward and logical reason.

Just thinking out loud, it seems that another possibility would be based interrupting capability. Interruption of can involve a race between recovery of dielectric strength and recovery of voltage accross the contact (after interruption at current zero). Normally the transient recovery voltage of concern is oscillatory associated with capacitance/inductance of the system. But perhaps if the line frequency gets high enough, there is not enough time for the dielectric strength to recover before the fundamental voltage reaches a level that will cause breakdown again.

Maybe someone else can provide more definitive answer.


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

 

[ScottyUK]

Why do you intend to put the contactor in the VFD output circuit? Almost every VFD manufacturer strongly advises against this type of design.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[thinker]

Scotty, you are right, generally speaking we would not do that. However, this a special test stand for HEV motor testing, and the end user wants from time to time (during durability test cycle) to remove power from the motor under test, then spin that motor from a dynamometer and measure back EMF values at open stator windings (they name it a sanity check). Again, the contactor will be turned off without load. The concern is its closed position performance.

 

[Mobius44]

It's a very interesting question, but I think it can only be answered by your specific contactor supplier.

Google found me some information from Rockwell showing derating multipliers ranging from 0.7 to 0.3 for various contactors switching 1kHz load. Since you're not switching under load I would expect those derates are worst-case for you.

The Rockwell info also included this explanation:

The skin effect produces an increased resistance of the current path with increasing frequency. In addition, magnetic induction in adjacent metal parts will cause increased hysteresis and eddy-current losses. Especially steel parts (quenching devices, screws, magnets, base plates) may be heated to levels in excess of permissible temperatures.
Since the cross section of the current path as well as the type and distance of adjacent metal parts can vary, the total heat generation and the local overtemperatures depend on the contactor type.

 

[thinker]

Thanks, Mobius44. Could you please send a link to that Rockwell materials? I could not successfully google it so far. Also, I was under impression (may be wrong one) that the skin effect would not be such important factor at relatively low frequencies.

 

[jraef]

I used to work for Siemens, we had a de-rate factor for frequencies up to 1kHz. Over 1kHz there are issues with heating of ferro-magnetic parts so they stop there. But for the contact paths, the skin effect heating resulted in a non-linear increased heating effect. The paper I have says the following de-rate factors:

For 100Hz: 6.7%
For 200Hz: 12.9%
For 300Hz: 16.4%
For 400Hz: 18.8%
For 700Hz: 23.2%
For 1000Hz: 26%

Plot those out and you can interpolate your de-rate for 800Hz, or I'd just use 25% and call it good. However only Schneider can answer to whether or not this is valid for their products.

By the way, no decrease in switching (making - breaking) capacity at increased frequencies. Just the steady state current carrying capacity.

But there were interesting debates internally as to whether or not you should use DC switching ratings when used on the load side of a VFD. I was in the not-an-issue camp because even though the PWM consists of DC pulses, they do change direction and will therefore have the same arc interruption effect as AC from that standpoint. The other side felt that was OK at 50/60Hz, but what about if the output was 10Hz or 5Hz? The reversal is happening a lot slower at that point. I'm not convinced that's an issue.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[thinker]

Thank you, Jraef, very useful table. BTW, I just received response from Schneider Technical Support: for 800 Hz the de-rating factor is 32%.

 

[electricpete]

It sounds like the original question was well answered (skin effect)

Jeff – good discussion and I’m glad you responded to my line of discussion regarding interruption.

There is no reason in physics I can come up with that your colleagues would ever suspect loss of interrupting capability at low frequency. Low frequency certainly does not share the unique dc feature of having no current zero's. Can you articulate what possible reservation anyone could have regarding interruption of low frequency current?

In contrast, there is imo strong reason to suspect reduction of interrupting capability as line frequency increases. The basic interrupting mechanism involves a race between recovery of dielectric strength and voltage accross the open contact after interruption. Whenever the actual voltage accross the open contact overtakes the dielectric strength, the arc reestablishes. If this occurs very shortly after reignition during the period of high frequency L-C ringing, it is termed “reignition”. If this occurs more than one quarter cycle after interruption in response to power frequency voltage, it is termed “restrike”. See slide 4 here:

http://www.ece.mtu.edu/faculty/bamork/ee5220/5220_L19.pdf

If we compare 60hz to 800hz interruption, there seems no doubt that the dielectric strength recovers at the same rate in both cases, but the power frequency voltage accross the open contact returns faster for 800hz, so it seems inevitable that restrike is more likely at 800hz than at 60hz.

Restrike does not necessarily lead directly to failure to interrupt, but is certainly not a favorable factor for contact life or for successful interuption.

To what extent restrike is a concern for low voltage contactors, and how high frequency would have to get before it becomes a problem, I’m not sure.


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

 

[Mobius44]

I'm glad your supplier was able to help. While the different supplier's factors are very similar, it's always good to have "official" numbers.

For reference the Rockwell information google found is linked here:

http://tinyurl.com/6zmfa6d

and the Rockwell derating table is found here:

http://tinyurl.com/6lczxan

 

[electricpete]

Hmmm. Just found this in Mobius' link, which seems to contradict all my conclusions:

2.7.2 Effect of line frequency on switching capacity
When alternating current is switched off, the switching path must be sufficiently deionized during the zero passage to prevent
reignition of the arc when the next half-wave appears.
At higher frequencies, the regeneration of the voltage after the zero passage is usually accomplished faster. However, the duration of
the arc and thus of the ionization phase is shorter. This is why contactors have practically the same switching capacity at 400Hz as at
50/60Hz.
Difficulties can arise at lower frequencies, as the effect of strong ionization by the prolonged arc becomes dominant. At lower
frequencies, the switching capacity is reduced and becomes very dependent on the voltage and the inductance of the load.

Let me think on that.

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

 

[electricpete]

It makes sense. There are 2 halves of the race: the actual (recovery) voltage and the dielectric withstand.

I focused on the actual (recovery) voltage, which suggested more severe duty at high frequency.

I assumed no change in dielectric withstand with frequency, but in fact it is apparently the dominant effect. After the contact parts, an arc occurs for duration of up to 0.5 cycles before the first natural current zero occurs. The time duration of that 0.5 cycles is much longer for the low-frequency case and therefore causes more ionization... and it will take longer for the dielectric strength to recover from that more-ionized state.

Thanks Mobius and Jeff for explaining that.



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

 

[jraef]

Well, that interestingly is the gist of the argument my colleagues had actually. Like I said I don't agree based on the specifics of that application being on the PWM output of VFDs, because technically, the individual pulses are going to zero in between each one. They don't CROSS zero and go the other way, but to me, that's an irrelevant distinction. Zero is zero. This argument may be valid if you actually had real sine waves at lower frequencies, such as 25Hz and 16.67Hz as they used to years ago.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[electricpete]

Jeff - I'd like to propose as a simplification that you consider the motor to be a low pass filter, so the only aspect of the voltage that matters is the fundamental. Now, the PWM acts just like a variable frequency sinusoidal source for our discussion. The nature of individual pulses and when they are zero is not important. What is important is the fundamental current and voltage. With that in mind, re-read the link or my most recent post (ignore the earlier one). Wouldn't you conclude that the low frequency causes more severe interrupting duty based on the longer arc before the first interruption.... resulting in more ionization and slower dielectric strength recovery.

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

 

[jraef]

Hmmm... low pass filter. Makes sense now that you point that out. I was thinking just of the PWM signal itself, but you're right, you have to consider the entire circuit because the return path of each phase is through that low pass filter.

Of course, none of the participants posed that as the issue at the time, and being biased to the position I had already taken I failed to consider it either.

(I know, I know, bad engineering practice but hey, we're all human...)

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[electricpete]

On thinking some more, it is not a trivial assumption that the motor acts as a low-pass filter with respect to preventing high-freq PWM pulses from appearing accross an opening contact.

I can justify it to myself if I consider the contact to be a resistance.. then the voltage accross that resistance is proportional to the curent.. and the current does not change in response to those high-freq PWM pulses.

But it is probably not fair to consider the contact as a resistance during the timeframe of interest when current is not flowing.... if we model instead as an open circuit than we might conclude full PWM voltage is seen accross the contact. Or maybe neither of these simple models is enough. Hmm.

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

 

[electricpete]

I guess another way to look at is this: contactor interrupting capability is not usually derated for PWM operation at normal power frequency (60hz), are they? I think the answer is no, which suggests that PWM voltage probably does not appear accross the opening contacts. So motor and cables etc probably does like a low pass filter

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

 

[ScottyUK]

jraef / ePete,

Arc interruption requires a current zero, not a voltage zero. A PWM output current will either flow in the power switch when 'on' or the body diode (or another switch) when 'off'. The current zeros will follow the fundamental output frequency, not the PWM carrier frequency. I agree about the low pass fileter effect and it certainly happens with the current because of the load inductance.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[electricpete]

Scotty - what leads you to believe I thought interruption depends on voltage zero rather than current zero (quote please). I have discussed the role of current zero many times: "After the contact parts, an arc occurs for duration of up to 0.5 cycles before the first natural current zero occurs."

The voltage is important because it is an element of the race between recovery voltage and dielectric strength.

It is easy to see the motor acts as a low pass filter for current, but it's not as easy imo to see it acts as a low-pass filter with respect to voltage appearing accross the just-opened contacts (see my post 11 May 11 20:12). In the end I concluded it does act like a low pass filter simply because we do not derate the interrupting capability for PWM operation at normal frequency (if PWM spikes voltage did appear accross opening contacts, we would have to derate regardless of frequency). But I am also still interested if there is a simple explanation as to how the motor/system acts as low pass filter with respect to voltage appearing accross the just-opened contacts during the critical period after interuption.

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

 

[electricpete]

So in summary, I agree with what you said Scotty. I believe it is consistent with what I had said (sorry if I sounded argumentative). I am still left searching for a more complete explanation of why the motor acts as a low-pass filter in the particular case of voltage accross open contacts. A scenario of interest would be when phase A has mechanically opened and electrically opened (current stopped flowing after current zero) and phases B and C remain conducting.... can we show whether the PWM creates any higher frequency components accross the electrically open A phase contact. My simple inclination is to reduce it to a single-phase circuit with source, inductance, and open contact... then no current is flowing in inductance and source voltage appears accross the contact. That doesn’t lead to the answer we expected and must be wrong (since we don’t derate interrupting capability for PWM at normal frequency).. The right explanation as to why voltage doesn’t appear across the contact must involve either the additional phases or the capacitive elements. Just a curiosity if anyone can come up with simple explanation. (Otherwise I will stop monopolizing this thread.)

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