molded-case cb inst trip response: peak, rms, or other` 3

[electricpete]

Molded case circuit breaker instantaneous trip settings are set/checked by injecting sinusoidal current.

We specify the rms value of that test current, but of course we also know it's peak. RMS/Peak ratio for sinusoid test current is 0.707.

Now, if I measure a motor starting current with an oscilloscope, it of course has a sinusoidal component plus a decaying dc component which varies based on closing angle. This waveform has a theoretical worst case peak 2.81*LRC (assuming worst case closing angle and negligible decay). More importantly, this waveform has a rms/peak factor of as low as 61% (much lower than the sinusoidal test waveform).

If I wanted to determine whether I expect the breaker to trip for one particular measured current waveform (and determine how much margin to trip on that particular start), there are at least two ways I might think about comparing?

Method A - Compare the peak of the measured waveform to the calculated peak of the sinusoidal test waveform (ie compare to 1.41*setting).

Method B - Compare the rms value of the measured waveform (over the interval before the first current zero) to the rms value of the test waveform.

Questions:
Q1 - Which of the above would make more "sense" from a physical standpoint?

Q2 - Would you consider either of the above approaches is "bounding" in a certain direction? i.e. perhaps 1 always conservatively predicts trip even though trip may not occur.


Q3 - Perhaps neither of the above approaches is a fair comparison. It may not only be a matter of a measure of the magnitude. For example, the time before the first current zero could theoretically approach one full cycle for the motor starting waveform, but is limited to 1/2 cycle for the sinusoidal waveform. Maybe this has an effect as well?

Note, I am not really looking for recommendations on how to set an instantaneous breaker. I am aware of NEC and other guidance for breaker setting. I would really like to understand the fundamental theory governing whether the breaker magnetic trip on a given phase would trip for a given waveform on that phase...if there is such a thing.

An interesting note - the rms value of the fully offset waveform with no decay is in fact sqrt(3) * LRC - a number sometimes quoted in instantaneous breaker setting recommendations.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[alehman]

Q1: I think of A and B, A would be better. Many breakers will unlatch in well under half a cycle, so RMS measured over half a cycle is probably not the best approach.

My understanding is that the magnetics unlatch the breaker based on the instantaneous current, so it seems like the best approach would be to establish what that value is for a given breaker and compare your waveform. I'm probably oversimplifying.

Maybe someone here is more familiar with the physics involved.

 

[electricpete]

Note that if we neglect the decay, the waveform looks something like:
i(t) = sqrt(2)*ILRCrms * [1 -cos(w*t)]
peak value is: 2*sqrt(2) * ILRCrms = twice the peak of sinusoidal LRC

RMS value is calculated as follows:
i(t)^2 = <sqrt(2)*ILRCrms * [1 -cos(w*t)]>^2
i(t)^2 = (2)*ILRCrms^2 * [1 -2*cos(w*t) +cos^2(w*t)]
<i(t)^2> = (2)*ILRCrms^2 * <1 + -2*cos(w*t) +cos^2(w*t)>
<i(t)^2> = (2)*ILRCrms^2 * <1 + 0 +0.5>
<i(t)^2> = (2)*ILRCrms^2 * <3/2> = 3*LRCrms^2
irms = sqrt{<i(t)^2>} = sqrt(3)*ILRCrms = sqrt(3) times rms of sinusoidal LRC


Therefore,
method A (examining peaks) would lead toward setting instantaneous at 2*LRC (plus allowance to provide margin for breaker variability)

method B (examining rms) would lead toward setting instantaneous at sqrt(3)*LRC (plus allowance to provide margin for breaker variability)

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[waross]

A magnetic trip will need a threshold current to establish a strong enough magnetic field to start the trip mechanism moving. Once the current is above the threshold you will need enough I^2*T to supply the energy to complete the trip initiation. The wave form will be important if the trip coil has enough inductance to noticeably increase the impedance at your test frequency.

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

 

[dpc]

The magnetic trip unit responds to peak current, but you are dealing with a device that is not very accurate or repeatable to begin with. The tolerance band for these breakers is pretty wide.

Just to add to the fun, at higher fault current levels, newer MCCBs will actually open their contacts before the breaker trip mechanism trips, due to the blow-apart contact design.

 

[electricpete]

dpc - You are correct that the breaker tolerances are an important part of the equation and something we have been thinking about those very carefully also. Again, I am not trying to develop a method for establishing setpoint. The question arises during investigation of a very intermittent instantaneous trip during starting. We tested the circuit breaker many and we think we understand the variability on that particular test for that particular breaker. We were hoping to develop a rational basis to compare those test results against our multitude of measured starting waveforms on this particular motor. There are other issues to consider: breaker test method - pulse method vs ramp method, voltage at time of start, effect of contact bounce on starting current etc.

Thanks waross. That makes good sense. There is an aspect of peak/threshhold behavior to begin the movement, and then an aspect of integrated behavior to move the trip element far enough. So the interval for the I^2*t would begin at some threshold and end either at same threshhold or at the next current 0?

I have attached comparison of a motor current waveform (with decay neglected) and sinusoidal test waveform of the same peak. You can see that for any threshhold current, , the motor waveform spends more time above the threshhold than the sinusoidal curent that has the same peak. From this simple analysis, we might conclude that even method A (comparing peaks) might not necessarily be conservative. i.e. It is perhaps possible that breaker might be expected to trip even though the peak of the measured current waveform is less than the peak of the sinusoidal test current used for breaker setting? (at least when we neglect the decay of the offset).


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[dpc]

The mag trip is generally a solenoid - one per phase and these actuate a common trip bar. Typically, only one phase will see the peak transient inrush, but there will be substantial (locked-rotor) current in the other two poles. There can be variations between poles on the same breaker.

For adjustable instantaneous trips, the allowable tolerance for trip current is like 80% to 120% (UL).

Also, if these are Motor Circuit Protectors, there can be some slight differences with standard thermal-mag breakers. I know that some HMCPs from Cutler-Hammer used to have and may still have some type of "transient inrush suppression" to help prevent tripping on starting. I assume this is some type of dc filtering.

 

[electricpete]

Yes, we test injecting current in one phase at a time to test each phase.

I know that some HMCPs from Cutler-Hammer used to have and may still have some type of "transient inrush suppression" to help prevent tripping on starting. I assume this is some type of dc filtering.

I haven't heard of that. I did find this which sounds similar:

http://www.clayeng.net/Training/myweb6/Module08/Output/MotorCircuitProtectors.html

Certain HMCPs also have a transient inrush trip suppression device. This allows the startup of energy efficient motors without nuisance tripping the sensitive short circuit protection of the current sensing coil.

A tuned spring introduces a time delay of approximately 8 ms into the trip sequence under normal conditions. It allows the HMCP to ignore the initial high inrush current during the first half-cycle of start-up. A true fault current would supply a magnetic force to override the spring action and provide instantaneous tripping of the device

However, I could not find a single mention of this in the Cutler Hammer literature. Does anyone know how to tell whether a breaker has this? For example I am interested in two motors: one with HMCP100R3 and one with HMCP150T4

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

One more piece of information possibly related to this mysterious damping feature:

NEC 2005 430.52(C)(3) FPN: "For the purpose of this article, instantaneous trip circuit breakers may include damping means to accomodate a transient motor inrush current without nuisance tripping of the circuit breaker".

Again, if anyone can help me figure out how to determine whether a given HMCP circuit breaker has this feature, I would be very very interested.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[dpc]

I guess the obvious idea is to ask Cutler-Hammer,but you've probably already tried that. Dave Roybal used to be a good resource on C-H circuit breakers. I'm not sure if he is still there or not.

I have an old (4th Edition) "Breaker Basics" Manual from Westinghouse that states that all Frame F (3 - 150 A) HMCP breakers have transient inrush trip suppressor as standard.

Regards,

Dave

 

[dpc]

electricpete,

I found one additional brief reference in the Breaker Basics book:

"The HMCP, with the unique transient inrush trip suppressor, or electronic time delay, permits a ride through of the first half cycle peak without sacrificing sensitive short circuit protection."

 

[dpc]

electricpete,

Here's a copy of a drawing from the Westinghouse book. Looks like a spring on the plunger provides the "transient suppression" magic.

 

[electricpete]

Thanks dpc. I did contact Cutler Hammer technical support today and they told me that their HMCP does not have this transient suppression feature.

It is still a very interesting piece of information. We are troubleshooting some intermittent trips that have occurred after we changed out our old ITE breakers with new CH HMCP breakers due to obsolescene. We had many years trouble free years with the old breakrs. I will try to see if those breakers had this transient suppression feature... may be an important piece of the puzzle.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[dpc]

Keep us posted - I'm interested in learning what you find out.

Dave

 

[electricpete]

Cutler-Hammer called me back with a "correction". Their HMCP does indeed have the mechanically-implemented transient suprression feature also (frame sizes 3 - 150A). I verified this by looking in a breaker.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.