Inrush current limiting for inductive load

[littlej040]

I am using an isolation transformer in a VFD circuit to prevent a GFCI from tripping (due to harmonic noise getting back on the line.) This solved my GFCI problem, but brought about another issue due to the inrush current the transformer created on this circuit.

Due to UL requirements, i placed a 10A (K curve for increased holding current due to inductive loads) in series with the transformer primary. Occassionally if the power is interrupted then recovered (during a storm etc.) The breaker is tripping due to the high inrush current.

My resolution was to install a NTC thermistor in series with the breaker that is capable of handling almost twice the rated current of the circuit. While this solved my issue, I feel their may be a more elegant solution that will remove this Hot component from my circuitry. I considered placing a PTC device in series with the transformer primary (and removing the 10 A breaker), but it appears these devices are not intended for high voltage high current situations (115 V, 8 A in this situation.)

Does anyone have any suggestions?

Thanks,
Jeff

 

[littlej040]

Any thoughts?

 

[DragonPetter]

Couldn't you put capacitance on the primary input to supply the inrush current?

 

[nsaspook]

Any thoughts?

Use the correct breaker or fuse type.

http://www.cooperindustries.com/content/dam/public/bussmann/Electrical/Resources/Solution%20Center/technical_library/BUS_Ele_Tech_Lib_Transformer_Protection.pdf

 

[littlej040]

Gents, thanks for the replies. The breaker that i am using is an ABB S201-K10, and FLA of the transformer is 8A. To my understanding I have specified the breaker appropriately for this. ABBs K-curve is designed to hold 10-15 times rated current for 1 second, which would be well within the inrush possible. Is there a more suitable breaker curve available?

 

[Carl Pugh]

Different brands of transformers have different inrush current. Could you try several brands of transformers?

Do you have room for a larger transformer? A transformer rated for twice the primary and secondary voltage and the same current should have a lower inrush current.

Maybe add a resistor in series with the transformer primary. Resistor to have maybe twice the resistance of the transformer primary.

 

[littlej040]

The footprint I have available is very limited for mounting this transformer. I would like to move away from using a resistance in series with the primary due to the heat dissipation. I am currently using the Thermistor in the same manner, and was hoping to find an alternative method. Should i try to measure the inrush current to find a breaker with a more suitable curve?

 

[jim hardy]

Some thoughts:


if you can use the transformer at less than rated voltage it'll increase its volt-second tolerance and stop those transients. Does it have perhaps a set of 150 volt taps?

I have used Sola constant voltage transformer at 120V into 208 volt tap, that works very well. Just you don't get to use the transformer's full VA capability.

As you did, I have also used NTC thermistors with excellent results, mine were Keystone Carbon brand.

experiment with the transformer
you'll find the inrush happens only about one in ten closures where the closure falls near sinewave zero crossing. Current spikes are short but high peak. You can observe them with a meter shunt in neutral and storage oscilloscope. Mine were 20X full load. As Carl suggested, a surprisingly small amount of series resistance will limit the peak current but at 8 amps you'll make some heat... You'll find using it at less than rated voltage both reduces those inrush peaks and makes the core run cooler.

A slower breaker, as suggested by NSAsp'k, would also solve it provided it co-ordinates with your source. (I had to co-ordinate with a flimsy source that had almost no overcurrent capacity)

lastly can you use one of those solid state relays? They are available in a variety that closes at sinewave peak instead of zero crossing , specifically for transformer loads.

old jim

 

[littlej040]

Jim and all, thank you for your responses.

I was able to measure the peak current this morning and saw 14X FLA. I was a bit surprised in this but it is in tune with what everyone has mentioned.

One thought I had was that potentially the circuit breaker was tripping magnetically and not thermally due to the spike. I replaced the 10A breaker with a 10 A slow blow fuse, and repeatedly turned on and off the circuit with no issues. The NTC thermistor was removed at this point so it appears that this would work as well.

Have any of you seen the same characteristics of a inrush current causing a breaker to trip and not the fuse?

I have actually used a SSR in another design to switch on and off the same VFD circuit with mixed results. The SSR was a puck style zero crossing model, that occassionaly would get latched on due to the noise on the line. I resolved this issue by placing an RC network in parallel with line and load, but again this was just a bandaid to the real issue. These noise issues are always a blast to work with!

Any additional comments are greatly appreciated!

Jeff H

 

[Antiphon]

The better way is to keep using the resistor but after a very short time you bypass the resistor with a relay. The relay would be a time-delay type and would require no special external circuitry on your part to do its job.

You can also use a much larger resistor with this method without any heating whatsoever.

 

[jim hardy]

yes i have seen 'instantaneous' magnetic breakers tripped by inrush.
i found a breaker with 'inrush tolerance' which was a small flywheel attached to the magnetic mechanism, increasing inertia to give the breaker modest inrush tolerance.
That in combination with a small resistor to limit those occasional peaks fixed my system. But it took considerable experimentation to get right. Best i recall it was an Airpax UPL 10 amp breaker.

a Zero Crossing SSR would guarantee a large inrush peak nearly every time. You'd need one that closes on peak instead. Most manufacturers make both, read catalog carefully.

Since you have abiity to measure peaks you can experiment with breakers and find one that works. ABB should have a sales engineer who's happy to help pick out one with more inrush tolerance.

old jim

 

[sophiecentaur]

The humble washing machine machine motor controller uses a thyristor (or something like!) to give a soft turn on for a fairly powerful motor. This would be better than a simple resistor. Thyristors are incredibly robust and tolerant to abuse so it might do what you need.

 

[NascentOxygen]

Is there a reason why a small-valued series inductor could not do the job? While bulkier than a resistor, it would have no associated power wastage, and the operating voltage loss could (in theory) be compensated by a tap on the transformer. Air cored would ensure it was non-saturable.