Thermistor help

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello, need a little help. Been searching and I think I've over searched and confused my brain. I'm new at this so keep that in mind. I have a Fosgate amp I've been working on. In trying to diganose 1 problem, I created another. I broke what appears to be a thermistor. I looked at the schematic from RF and it from what my brain tells me is a ntc thermistor. The schematic shows 10k@25c 1k@90c which would be ntc. The colors on the good one on other side are brown,black,orange, gold. My brain says 10k +_ 5%? My searches lead me to believe it is a Murata possibly. I've found 3 part numbers, but not sure which is right, also having hard time sourcing them. I've come up with NTSCA0XV103(F or E) 1BO, or NSDS0XV103(F or E) 1BO, or NTSD1XV103FPB(30,40, or 50). They all seem to be the same from what I can see. Any Help would be great. Thank you.
 
No ideas on this? The ones I have listed seem to be bare lead, the one I'm looking for has shielded leads. Also as far as source goes, looking for someplace I don't have to order 100 of them. A cross reference would be good if someone knows it and a place to get only a few pieces. Should I post a pic of the schematic or of the good one or ? Thanks.
 
I'd be temped to remove the one from the good side and measure it with an ohmmeter. Warm it up with a hot air gun or something, and see if it's really a thermister, or if somebody built the board with a regular resistor. If it is a thermister, sub in whatever you can get that meets the specs and fits. Add your own insulation to the leads if necessary. My experience has been that getting small quantities of odd parts is difficult. Can you get it as a service part from the manufacturer?
 
Unless you get lucky and get a part from the manufacturer, or can find out exactly which thermistor was used and then get lucky-enough to find good data on it, your best bet might be as Conrad suggested, i.e. remove the remaining good one and measure its resistance at several temperatures.

You can use boiling water, and ice water, to get very close to 100 degC and 0 degC, without a thermometer. (You might want to have the thermistor and leads in a plastic bag that can withstand the water temperatures.)

You can calculate the Beta constant, which is one common spec for thermistors, with:

Beta = (T x T0)/(T0-T) x LN(RT/RT0)

where T0 and T are in degK (i.e. degc + 273.15), and RT0 and RT are the resistances at those temps. T0 is usually the 'nominal' temp.

For your schematic-given resistances for 25C and 90C, the above equation would give Beta(25C/90C) = 3835.5. However, you would want to measure at the two temps that are commonly used for the published Beta constants for thermistors, which vary by manufacturer, etc. You can just check the catalogs or suppliers that you have available, and get measurements that match how they specify them. For example, the Xicon NTC thermistors at mouser.com are specified with Beta(25C/50C).

On the other hand, if it looks like a 5% resistor, i.e. cylindrical with color-code stripes, then maybe you'll get lucky and find out that that's all it is.

You could also just pop in a 10K resistor, there, run the unit, and see how hot that resistor gets, under various conditions. If you get lucky and it always stays near-enough to 25C, then maybe there's not much of a problem with just using a plain 10K resistor.

However, if it does turn out to be a thermistor, and has to remain a thermistor, and you can get the resistance measurements (from the remaining good one) for at least three temperatures (and I'd probably try to get it for at least four temps, i.e. 0C, 25C, 50C, and 100C, and probably also 75C while I was at it), then you could also download the huge EPCOS library of spice thermistor models, and download LTspice (from Linear.com) , and could hope to find a pretty good match, by plotting the various Epcos 10K thermistors' resistances for the temperature points at which you measured, and comparing (which would use a trivial LTspice circuit, and procedure). Once you found an Epcos match, you could either find the Epcos part, or cross-reference it, or just look for the same specs. (If all else failed, you could probably also just alter the response of an Epcos part by using parallel and series resistors, until it matched your data points closely-enough.)

You could also do more-or-less the same thing manually, by solving the equation above for RT, and calculating what an available thermistor's resistance should be for the temps at which you measured yours, using the published Beta constant.

i.e. RT = RT0 x exp[Beta(T0-T)/(T0 x T)]
 
Note that the "Beta constant" equations given in my last post are fairly-gross approximations, for thermistors, and that using the Epcos spice models would be much more accurate, for finding a matching thermistor. They use the now-standard polynomial-type models, probably using the 'Steinhart-Hart coefficients' thermistor modeling method, and also even include self-heating effects.
 
Thank you for the replies. I ended up pulling one from another amp, same brand and using it to get this one fixed. I will pull the remaining good one from the other amp and take the equations and see what I come up with. I'm pretty sure it's a thermistor because it get's clamped to the heatsink with the transistors. I Have to think it's thermal shutdown protection. Like it was mentioned, it's hard to find someplace to buy only a few of something like this so I'll have to find something to make due. Thanks again for the replies and the formula.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.