| Nelson Pass |
Previously I have mentioned that we have put thermistors
in parallel in CRC type power supply filters.
Initially there was some concern about current hogging, as
any mismatch between thermistors would cause the lower
resistance one to take more current and get hotter and ad
infinitum. When we built a few of these, we didn't encounter
this, but it has become clear that it's a real problem and we
were just lucky the first time.
Any mismatch or puff of air will start current hogging, and you
will quickly find yourself with one thermistor taking most of the
load. This is not necessarily a reliability problem, but the circuit
is not working as intended. If you want to use thermistors
here, I recommend using just one or having some resistive
ballast in series with each one.
If you don't need a thermistor in this application, just use
resistors. In fact, I think I recommend a big fat coil instead.
If you are concerned about LC resonance, then place a resistor
or thermistor in parallel with the coil.
:cool:
You don't like these ideas? I've got others.... |
|
|
| Peter Daniel |
I also used thermistors in my Aleph X, but since the current draw is less I had to use only 2 per rail.
The way I used them is not parallel them directly, but feeding from the same capacitor bank ea. of the two is connected to their own capacitor bank on left and right heatsink. This way, the possible problems of mismatch could be avoided. |
|
|
| Mark A. Gulbrandsen |
Nelson,
Peter suggested epoxying a single one to a heatsink which I suppose is ok to do, but I think if one will be using them in parallel you may want to epoxy them all in a block, or perhaps ceramic encapsulate them all together. That way they'll all be at the same temp. I measured my two in parallel again and there is a big temperature difference now..after about 2 hours operation. One is a bit over 100 deg C, and the other about 88 deg C. So you are definately right.
I think matching is important with these if one is going to use them in parallel, but so is keeping them all at the same temp....somehow....Perhaps that Red Silicon stuff from the auto parts store would make a good encapsulating material......it has an extremely high temp rating.
Mark |
|
|
| Nelson Pass |
Keeping them at the same temp is mostly a matter of
aesthetics. If they do the job and are not too hot, it
is not important per se. In your case, I wouldn't
get very excited, unless that sort of imperfection keeps
you awake at night.
If you are a true audiophile, it probably does...
;) |
|
|
| Peter Daniel |
| quote: | Originally posted by Mark A. Gulbrandsen
Peter suggested epoxying a single one to a heatsink which I suppose is ok to do |
Also, when you decide to put a heat sink on one of those, don't make it too big, otherwise thermistor might not get hot enough.;) |
|
|
| Mark A. Gulbrandsen |
Very good point Peter!! I'm still thinking about going the encapsulation method. This way a poof of wind or something won't cool down one thermistor more than the other......
I ordered extras so I may try this over the weekend. Will get back about the results.
Mark |
|
|
| Peter Daniel |
| And I think that encapsulating will probably solve the problem . |
|
|
| Nelson Pass |
Putting a small ballast resistance in series with each
thermistor, say .5 ohm or so, will solve the problem and
is easier. |
|
|
| Mark A. Gulbrandsen |
Thanks Nelson, will try that first.....finishing up both chassis this weekend first.......then on to the thermisistor problem. I'd like to stay away from any relays in there too.......keeping it simple like the circuit. I think the chassis has been the most complex problem so far but they are comming along nicly.
Mark |
|
|
| wessol |
Dumb newbe question...
In regards to the power supply RC filter configuration;
If the CRCRC filter was setup in a lowpass filter configuration
what would be a wise choice for the cutoff freq.?
Would a person look at the audio spectrum or
try to filter out just below the AC mains freq. (50 or 60 Hz)? |
|
|
| Edwin Dorre |
Hi Nelson,
what do you use with your production XA's at the moment when the 4 paralallel CL-60's gave you problems? I hope you do not mind I am asking you this...
As I have just finished power supply and then hit this thread!
Many thanks,
Edwin |
|
|
| Nelson Pass |
| We have been using coils and resistors. The coils alone would be preferred due to lower dissipation, but the resistors are physically quieter, and in some units I can afford to to throw away a couple watts while we work through the coil vibration issue, which was a consequence of PC mounting creating a sounding board. |
|
|
| fdegrove |
Hi,
Match your thermistors, create a virtual midpoint between hot an neutral...bingo.( can use the earth).
If you can't match, treat the thermies like as if they'd just be a regular case of caps or resistors in //, i.e. use a balancing network if you must...as N.P suggest if I read correctly.
All pretty straightforward if I've read this correctly.
Cheers,;)
P.S. When a PTC/NTC goes south it's to be replaced..just like a fuse, difference is you're not always going to have a visual clue.
Just my measle two cents,;) |
|
|
| wuffwaff |
.......or use a thermistor that´s big enough to dissipate all of the power. Thi biggest can dissipate 6,3 watts and that should be enough for this application.
william |
|
|
| Nelson Pass |
Certainly you can use the highest current thermistors and run
them in series with no issues. The biggest one in my Digikey
catalog is a KC024L-ND, and if you don't like the price on it,
use the KC022L-ND.
:wiz: |
|
|
| promitheus |
Has anybody used only one big thermistor?
It sounds like a very good idea to me.
The cost is not so big, at the end you will be using only 1 instead of 4 thermistor per power supply rail. |
|
|
| SI |
| quote: | Originally posted by Mark A. Gulbrandsen
Very good point Peter!! I'm still thinking about going the encapsulation method. This way a poof of wind or something won't cool down one thermistor more than the other......
I ordered extras so I may try this over the weekend. Will get back about the results.
Mark |
Mark any updates on the thermistor issue.
Regards. |
|
|
|
