Trace/Xantrex C40 temperature compensation

[Tom in NH]

...the temperature compensation algorithm used by the C40 charge controller or other Trace/Xantrex units? I've figured out that the temperature sensor is a thermistor with an approximate value of 100 Kohms. I'm wondering how much variation the C40 will tolerate when a homebrew temperature sensor is used.

If any of you have a C-series controller with a temperature sensor attached, I'd be very grateful for a bit of assistance in some research. Would you please disconnect the sensor momentarily to measure the resistance between the two center pins. If you send me the resistance you measured and a temperature reading, I'll compile the results and report back the findings. You could send it to my email address, woods at ncia dot net.

Thanks,

Tom

[RobD]

I'd be interested too. I think you need to get several readings at different temps though so you can find the resistance change of the thermistor as temp varies.(Delta 'T') This way you could find a thermistor to match it.

Is it just a thermistor or is there a circuit involved? I would imagine the circuit is in the inverter and just two wires run from the thermistor.

[richhagen]

I have a C-40 with the xantrex/trace temperature probe on a 48V system.  I would be willing to make some measurements for you, but would need to know what you need.  I'm guessing resistance measurements at various temperatures for the probe, but let me know and I'll take them and post them.  Rich Hagen

[richhagen]

If I recall from the last time I looked at it, the probe has a modulal RJ11 or RJ12 plug on the end with two conductors.  Rich Hagen

[Tom in NH]

The Xantrex temperature sensor I have has four conductors, but they are paired. The left two are the same and the right two are the same. If you would send me temperature and resistance measurments across the center two conductors over a variety of different temperatures, that would be great. The more the better. If you want to get really thorough you could dip the temperature probe in hot water and send temperature and resistance data of the high end as well. If you can get a reading down around 0 deg. F. that would be good too.

I should be able to plot a curve of your readings and compare it to a similar plot of my readings. They will likely be offset a little bit. Of course, readings from two different probes are better than just one. Readings from more than two would be even better still, so if anyone else would like to participate, your help is appreciated.

The key questions I want to answer are how much variation is there between different probes? Is the curve of a 100Kohm thermistor within tolerable limits? If not, what value of thermistor would be?  

--Tom

[richhagen]

I will measure a dataset for you, hopefully tonight and post it.  Now where did I put that mercury thermometer at anyway?  Rich Hagen

[richhagen]

Here's a couple of points to get you started.  Preliminarily I am got wierd results, until I realized that the ohm meter wasn't working right.  Fortunately the volt meter was working.  I had a 3200 Ohm 2% resistor, so I could set up a voltage divider circuit.  I havn't found my thermometer, so I only had a couple of reliable reference points.  Below are the measurments for those points to date. As soon as I locate my good mercury thermometer (and possibly fix my meter or find another one that's working), I will post more data points.  

at 0 deg. Celsius, total voltage 7.23 volts, voltage across 3200 ohms .0834 volts

at 21.7 deg. Celsius, total voltage 7.23 volts, voltage across 3200 ohms .2083 volts

at 100 deg. Celsius...(just kidding, I wasn't sure it would survive)

This yields a measured resistance of approximately 277 Kohms at 0 deg. Celsius and 111 Kohms at 21.7 deg. Celsius.  

Keep having fun, Rich Hagen

[Nando]

TOM:

It is easy to replace the thermistor with a variable potentiometer and vary it until the C40 complains,

You got the trigger point value, just measure the resistance.

Of course a careful data taken with the different temperature values applied to the thermistor you get a chart to determine the type or class -- AND if you send the info to me I will help you to identify what can be used and the many variations for different schemes.

It is worth while to remove the thermistor and measure the connection with a DVM ( Digital Volt Meter)

Regards

Nando

[Tom in NH]

Thanks for your numbers, Rich. I'm putting together a graph that you may find interesting. Nando, would you please tell me what you mean by "when the controller starts to complain." You suggest there might be a trigger point. Maybe there is, but it could also be a gradual compensation rather than all or nothing. What do you think?

I'll be back later with my graph if I can figure out how to post an image.

Thanks for your help and comments. --Tom

[Tom in NH]

Hopefully you can figure out my labeling system. This graph shows the close relationship between Rich Hagen's battery temperature sensor (BTS) and mine. Unfortunately it would be hard for me to remove my BTS and bring it inside to get a reading at 70 degrees F. or so. Guess I'll have to wait til next spring for that reading. There are two different 100K thermistors that I plotted too. They are all very close to each other so I think it is safe to assume that you can go out and spend a couple bucks on parts and roll your own BTS instead of spending $37 on a factory made BTS.

Thanks for your help.

Tom

[richhagen]

Thanks for posting your graph.  It is interesting.  It makes sense that they would design the circuit to use a relatively inexpensive off the shelf thermistor as a sensor for the temperature compensation.  It looks to be well potted in a plastic sleeve with a standard 4 conductor line running from it to the modular plug on the end.  I was wrong about there being two pins on it, there are in fact 4 and it appears to be wired as was commented above.  The only thing that would worry me about it over the long term would that if the C-40 was mounted in an area with high humidity or other harsh conditions. I am not sure if the modular connections would work over the long term, but who knows, time will tell, and so far no problems.  RichHagen

[boB]

Nice Graph..

Yes, it's a 100 K  thermister centered at 25 degrees C.

Can't remember the non-linearity factor though.

Algorithm should be -5 millivolts per degree C per 2 Volt cell, as usual.

i.e. -30 mV per degree C for 24 Volt battery.

Happy Turkey day !

boB

[richhagen]

Tom I'm guessing you put this information to work here:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=41980&item=5939237697&rd=1

"This is a homebrew 100 Kohm temperature sensor. Use it to provide temperature sensing, control, and battery temperature compensation. Ouput resistance of this sensor is very similar to the Xantrex BTS Battery Temperature Sensor used to provide automatic temperature compensation in the Trace/Xantrex C35, C40, or C60 series charge controllers. The unit is pin compatible with the Xantrex BTS. Uses RJ-11 connector. Cable is approximately three feet long and may be extended with extension phone cable.

With temperature compensation, you can protect the batteries in your solar or alternative energy power system by optimising the charge voltage. Reduces excessive gassing at high temps. Boosts charge voltage at low temps. Now you can provide temperature compensation at a reasonable cost. No reserve. Your satisfaction is guaranteed. Try it for 30 days. If you are not satisfied, return it for a cheerful refund of the purchase price (less shipping cost)."

Hope it works out well for you, let me know how it turns out.  Rich Hagen