Extrapolating Trip time curves on a "Thermal only breaker"

[ELA]

My question is how would you extrapolate the trip time curve for a TA35 breaker in the spec sheet here on pg3 for a 50 -60 amp current lasting only 2 - 3 cycles?
http://www.schurter.ch/pdf/english/typ_TA35_Wippe_1Pol.pdf

Here is some background:
I am wanting to use a 3-5 amp thermal breaker to protect a piece if equipment that normally draws less than 1 amp running. Use the 3 amp , 1 pole, for this discussion. ((although I will most likely use the 5A for increased ability to handle the inrush.))

My concern is inrush current. The Switch Mode Power Supply inside of the equipment states a worst case inrush current of 50 -60amps. Unfortunately they do not state a time duration.
** I measured one example unit and only observed a 2-3 cycle worst case inrush duration.
The amplitude was normally less than 10 amps at 120Vac. I am not sure under what conditions they predict 50 -60 amps (unit can be run at 240V so of course it would be higher there) . I want to design around this worst case value to be safe.

NOTE that there are many different specifications given such as::
Overload: min 40 cycles at 6x In.
Conditional short circuit current : 2000A
And the trip time curves.

From these specifications it can be seen that multiple overloads cause degradation of the unit thus limiting its life.

My question is actually two fold:
1) How many cycles do you think this breaker will hold for at 50-60 amps.
2) How much of a concern for the life of the breaker should there be if powering up inrush current is 50 -60 amps lasting for only 2-3 cycles?

 

[Cold Fusion]

I highly suspect you already know every thing ai have to say. So take this as maybe useful.

My question is how would you extrapolate the trip time curve for a TA35 breaker in the spec sheet here on pg3 for a 50 -60 amp current lasting only 2 - 3 cycles? ...

Same as you - No clue. 60A for a 3A rated CB is 20x. That's off the chart. Maybe the curve extrapolates or maybe it drops off to an instantaneous time because of the low thermal mass.

...
1) How many cycles do you think this breaker will hold for at 50-60 amps.
2) How much of a concern for the life of the breaker should there be if powering up inrush current is 50 -60 amps lasting for only 2-3 cycles?

1) If the curve extrapolates - 150 to 250 ms, say 10 to 15 cycles

2) None.

Now that your questions are addressed, some comments and other questions:

...From these specifications it can be seen that multiple overloads cause degradation of the unit thus limiting its life. ...

I didn't see that. Where did that come from?

...I am wanting to use a 3-5 amp thermal breaker to protect a piece if equipment that normally draws less than 1 amp running. ...

What are you trying to protect or protect from?

Here's what I'm thinking (again, I pretty sure you already knew this):
If you are trying to protect the power supply from overload, then fast OCP (quick fuses) on the output would make more sense.

If the idea is to protect the power supply from an internal fault, then the unit is already dead there is nothing to protect. You just want the fire out fairly quickly. The incoming power OCP just needs to trip inside the cord damage curve.

Just wondering

cf

 

[gar]

100605-1828 EST

ELA:

Consider a Heinemann breaker. These are magnetic-hydraulic and some are good for rather large inrush. Combine this with a negative temperature coefficient thermistor.

http://02d14e9.netsolhost.com/acrobat/cirguide.pdf
http://02d14e9.netsolhost.com/acrobat/ams-cata.pdf

I have used a Keystone CL-70 NTC in the past. You would probably want a somewhat lower resistance than the CL-70.

I tried to search for Keystone Carbon Co. They may have been bought by another company.

An NTC with a high inrush magnetic breaker is a very good solution.

.

 

[ELA]

Thanks for your input Cold Fusion,

The breaker is rated for 50,000 cycles but when exposed to a 6x overload it goes down to min of 40 cycles.

I am only really needing to protect the internal 16 awg interconnecting conductors. NOTE: This breaker also serves as an on-off switch for the device.
I agree ... if it were totally up to me I would use fuses + a switch but the customer does not want fuses. They require a resettable device (don't want field personnel to have to hunt for replacement fuses).

I was unsure of extrapolating since the curve is not linear. I called Schurter and they did not have a definitive answer for me. I did an estimate using an I^2 *t approximation and came up with numbers in the same area as yours.


Gar,
As I mentioned to Cold Fusion I had wanted to use fuses but the customer over ruled that. The issue with a magnetic is the cost.
This thermal breaker is only $8. The customer is driving for cost savings.


I suspect the breaker will be fine.
I am going to set up a test on Monday to test the breaker with a ~60 amp peak using a pulsed solid state relay for a varying number of cycles. I will do this just to be sure I am not overlooking anything.
Will post the results if anyone might be interested.

 

[gar]

100605-2032 EST

ELA:

Also you can use a thermistor with a thermal breaker and greatly reduce the stress on the breaker.

.

 

[Cold Fusion]

...The breaker is rated for 50,000 cycles but when exposed to a 6x overload it goes down to min of 40 cycles. ....

Okay I didn't translate right. "When exposed to a 6X load that trips it, the life decreases. Yes I agree. I suspect the life goes to 1 cycle if the equipment does indeed develop a fault and the CB opens.

...I am only really needing to protect the internal 16 awg interconnecting conductors. NOTE: This breaker also serves as an on-off switch for the device. ...

Maybe a 10A device is the right one. That will protect the #16 conductors from catching fire and inrush is 6X and the trip time is .9 to 2 seconds.

... if it were totally up to me I would use fuses + a switch but the customer does not want fuses. They require a resettable device (don't want field personnel to have to hunt for replacement fuses) ...

The customer requests sound good to me - their reasoning is good. The power supply might not even need output fuses if the output folds back. A lot of current (no-pun) designs do.

I was unsure of extrapolating since the curve is not linear. I called Schurter and they did not have a definitive answer for me. I did an estimate using an I^2 *t approximation and came up with numbers in the same area as yours. ...

Schurter's response in not surprising - unless you are planning on buying 10,000 or 100,000 units. I didn't get as technical as you did. I extended the existing curve slopes by eyeball and pen and marked of a 20X current, by using the 1 - 2 spacing. Your guess sounds a lot better than mine.

...Will post the results if anyone might be interested.

Sure

cf

 

[ELA]

Gar,
I had considered doing that and would probably add one if I thought it were absolutely necessary.

I have used NTC thermistors from Ametherm in the past for nuisance tripping of Magnetic branch circuit breakers when a servo drive pulled 260 amps inrush. Back in those days I worked for a larger company and designs were not so cost sensitive. Packaging was an issue since this was an add-on external to the design.

The main issues with adding one is again the added cost and where to mount the device (insulate it etc). If testing shows a concern with the breaker life I will still consider it but would most likely add it to a pcb layout.

 

[ELA]

A follow up

A follow up

I was able to setup a 44 amp load ( closest value high power resistor I had) and pulse the SSR for a varying number of cycles. It quickly became obvious that I needed to wait a good time period between tests to allow the breaker to cool between tests so as not to skew the results.

I found the 3A breaker tripped after 30 cycles at 44 amps. This is right in line with an extrapolation using (I^2 * t) that predicted 28 cycles for that current.

So this breaker should be fine ( I will be using the 5 amp version for added margine) I could use the 10A to be even more safe (if only considering wire protection) but I do like the idea of keeping the breaker closer to the demand for dedicated protection.

This was an interesting exercise as you consider that at 44amps the breaker is dissipating 116 watts, for that short 30 cycle period. And of course it did not take long for the load resistor to get warm either.

 

[Cold Fusion]

...This was an interesting exercise as you consider that at 44amps the breaker is dissipating 116 watts, for that short 30 cycle period. ...

The CB has a 2.6V drop at 44A?

cf

 

[ELA]

Yes that should be accurate.
The breaker resistance is 0.06 ohms.

The large drop only occurs during the inrush period.

 

[kingpb]

You cannot extrapolate a protection curve. It is what it is.

The curve is indicating that up to 10X the 2-pole will trip starting at 1 sec. The 3-pole shows around 0.4 secs. Thta tells me this is the capability of the device. It is thermal only, so for currents higher than 10X there is not a instantaneous trip. A good guess would be that the contacts are going to weld shut, or at least be damaged with currents greater than that.

A tripping time of 3 cycles is very fast, and in general, breakers cannot trip this quickly, usually 7 cycles is more realistic.

BTW: This is an IEC type rated device. Not sure if your looking at 50Hz system, or not, the OP did not say. If this is in the US, I would be looking for non-IEC components.

I would think fuses would fit better with what you are trying to achieve.

 

[ELA]

Kingpb,
Thanks for your definitive answer. Did you happen to read my post?

At first I was concerned about whether or not the curve would extrapolate and how. Testing has confirmed that it does indeed extrapolate just fine.
No problems with welding contacts at all.

 

[winnie]

[...] NOTE: This breaker also serves as an on-off switch for the device.[...]

Is this going to be a problem? The device will be _switching_ the inrush current of the power supplies. Presumably switching _on_ is easier on the contacts than switching _off_, but we are still talking about high current cycles.

-Jon

 

[kingpb]

Kingpb,
Thanks for your definitive answer. Did you happen to read my post?

At first I was concerned about whether or not the curve would extrapolate and how. Testing has confirmed that it does indeed extrapolate just fine.
No problems with welding contacts at all.

I see where you did do some experimentation, read through kind of fast. My feeling is that because it worked for a given case, can you repeat the experiment multiple times without failure, or determine degradation. Only time will tell, especially if it is a cycling effect.

But if the device is not either rated for your conditions, or the manufacturer cannot confirm the usage, I might suggest your look for a listed device with the ratings you need, especially if the equipment is critical to the operation.

 

[ELA]

winnie, kingpb,

Thanks for your inputs.

I was concerned initially primarily about the switch duty of the breaker and inrush current more so than I was about nuisance trips. I figured nuisance trips would not be an issue since the inrush was so short of a duration.
The extrapolation of the curve exercise was just to be sure I was not making a bad assumption.

The device life does degrade with over currents and it lists a short circuit current of 2000A being worst case.

My initial concern was over degradation and possible welded contacts from short duration inrushes from a SMPS load.
In testing the normal inrush currents of the SMPS were much smaller ( less than 10A) than the published maximums.

In my initial testing I switched (4) of the power supplies in parallel on one switch (manually) with a resultant 20A inrush combined (normally only one supply is present). I did this 50 times with no issues. Just to be sure I then ran the test with a pulsed SSR providing both 44 amps and 70 amps. I experienced no issues when two different samples were tested from a 3 cycle duration all the way up to the 30 cycles that caused the expected trip.

One breaker was exposed to a total of over 90 interations during all of the testing combined. It still functions fine and I have taken it apart and inspected the contacts. They are slightly pitted and I think that is probably mostly from the 30 cycle trip testing at 44 amps.

I was testing at the higher currents for a worst case scenario. If I felt the switch would need to operate hundreds of times into those higher currents I would select a different device. This switch will seldom be used in the life of the product (maybe 20 to 40 times in the life of the device).

As I mentioned earlier in this post cost is the driving factor and the customer will not allow the use of a fuse. I am used to "over designing" on the side of reliability when not driven to contain costs.

The breaker is being used well within the devices "normal operation" ratings. I am just looking at a worst case design scenario.
Sometimes a manufacturers specifications do not include all possible data collection scenarios and you then have to make a judgement call.

I contacted the manufacturer and they suggested testing since they could not provide the data. Rather than rely solely on a judgement call I elected to perform at least a minimum sample size test.
May not be 100% but I feel I have done my due diligence.

Thanks for your comments as they keep me thinking.