Load Diversion Problems C35

[herbnz]

I have installed a number of load dump systems using a dump load made of fencing wire.

It takes about 12 Metres og 10g to get 25A at 24volts. This has been buried in the ground or ideally laid out in a concrete pour. I have never had problems using a Morningstar tr45. However recently a neighbour approached me for a system seeing the large price differance to the trace c35 I bought in a C35.

This would flash its lights and do its job up to the point it was necessary to dump then it completely shutdown allowing the battery voltage to climb above setpoint. The only way to restore was to shut down hydro depower C35 and restart, even reset button would not operate.

Have replaced unit with a Tr45, but c35 on a compact nichrome load dump is working fine at my home. My conclusion is the long fencing wire has reactance that the pulses from the trace can not handle ( note this is backed up by occasionaly the orange light appears that means open dump )

Anyone else had problems with c35's? The fact it fails to completly off is a worry I see.

[ETech]

This part of your request is a warning flag to me:

"This has been buried in the ground or ideally laid out in a concrete pour."

The bare fence wire would allow a leakage path to earth ground.

Either of these controller "could" have a problem with this and thier output circuitry if grounded.

I do not have a system, but as a general rule you want to eliminatee any leakage paths as they always cause problems.

You could try using an isolated dump load an see if that makes a difference.

Hope this helps, ETech....

 

[herbnz]

Many thanks for reply.

A did try another length held off the ground, same problem.

Yes I see leakage would be there but this would be small and only look like a parallel load, the battery negative is already earthed.

The resistance of the wire would be 1ohm I would say from experiance trying to get earth resistances down to this level our leakage would be small.

But you raise a good consideration

Herb

[Flux]

I suspect it is inductance that is bothering the C35. These things are assumed to be dumping into resistive loads. Any load resistor covering a large area will be quite inductive. There are no chokes or filters like a mppt controller and I assume the manufacturers expect you to keep the loop to a minimum to minimise rf interference.

Try making the loop as non inductive as possible by running the wires parallel and close and see if that works. R C snubbers across the load may also work but if you have a scope I would suggest you look closely at the load terminal volts. If you get the mosfets into rf oscillation you will have major trouble. If it didn't cut off it would probably cook.

It is an old design and certainly has quirks when used for wind power.

Flux

[scoraigwind]

C-series controllers have a serious time lag which makes them a lot less effective than the tristar.  the voltage tends to surge high in gusts.  Maybe this can cause the controller to cut out?  

I don't think earth leakage through concrete will produce a significant change in resistance for battery-voltage dump loads.  Concrete is a very poor conductor.

I use wirewound resistors with a steel bolt down the middle for mounting, and that must surely have higher inductance than a straight wire?  I see no problems other than a bit of noise.

[herbnz]

Hi Flux

First I never mentioned this is number 2. The first one installed did blow FETs was replaced, it was only when the second one showed similar behaviour I started looking.

It is now obvious that the stray self inductance is the problem and I started to feel a little guilty about getting a replacement from the supplier under warranty. At this moment however I feel to some extent justified by the fact that Xantrex in there spec's or installation instructions mention the level of stray inductance that must be present in any dump resistor. In fact the work I have put into this problem should help them see a problem with their unit.

I believe your ideas of snubber ccts could solve the problem but at this stage I will ever leave the tristar in or replace the dump load with a higher resitivity wire hence a shorter length. Re laying the fencing wire to get less inductance would take away the similistic idea. Also it must be the fact that  steel wire being magnetic is making inductance higher.

As a matter of interest re snubbers, are they ever used with FET's?, Would they not be across the FET as we used to do with SCR's?

Thanks for your effort in replying to this

Herb

[herbnz]

Hi Hugh

Thanks for your reply.

The level of inductance I hit must have been higher, more than likely the steel wire being magnetic the reason.

Flux

I see I made a error in my previous posting the comment re Xantrex mentioning inductance should be "never mention "

How do you edit on this board ?

Herb