Is this a good way to monitor heatsink temperature when biasing amps? I inserted the probe into the holes for the screws that secures the the covers
Usher R1.5 heatsink temperature monitoring - YouTube
Usher R1.5 heatsink temperature monitoring - YouTube
In theory, yes, but that´s a massive heat sink and probe is in the center, I doubt there is much difference, if at all.
In any case, I´d drill a dedicated extra hole for the probe and use mounting holes forn their intended purpose.
A drop of silicone grease inside the hole would make things even better.
In any case, I´d drill a dedicated extra hole for the probe and use mounting holes forn their intended purpose.
A drop of silicone grease inside the hole would make things even better.
The temperature will not be uniform over the heat sink. The actual temperature of concern is the
device junction temperature, which has to be calculated from the device case temperature,
or from the heat sink, close to the device. Use an IR temperature gun and see.
device junction temperature, which has to be calculated from the device case temperature,
or from the heat sink, close to the device. Use an IR temperature gun and see.
One of the first things I learned upon becoming employed as an engineer was the different "theta"s or thermal resistances. That is, what causes temperature drop between the junction of the device, its case and the cases heatsink mounting.
Since junction temperature is the one that kills devices, that's what's really of importance. Yet, being inside the device, it's hard to measure. So you go for case temperature, as being "once removed" from junction temperature by theta j-c. It follows that heatsink temperature is twice removed, via theta c-a. Therefore a poor estimate of the temperature that really matters, junction.
There's tricks, albeit perhaps impractical. One for an NPN might be to bring the amp up to temperature, turn it off and quickly somehow measure the Vbe, which is a junction that looks like a diode. That voltage will vary with temperature and - once calibrated (the impractical part) - would give the actual temperature of the device die.
Get you probe on the device case; it'll be cooler than the actual junction, but better than the heatsink. I assume you're doing this because you want the amp to survive your biasing effort.
Since junction temperature is the one that kills devices, that's what's really of importance. Yet, being inside the device, it's hard to measure. So you go for case temperature, as being "once removed" from junction temperature by theta j-c. It follows that heatsink temperature is twice removed, via theta c-a. Therefore a poor estimate of the temperature that really matters, junction.
There's tricks, albeit perhaps impractical. One for an NPN might be to bring the amp up to temperature, turn it off and quickly somehow measure the Vbe, which is a junction that looks like a diode. That voltage will vary with temperature and - once calibrated (the impractical part) - would give the actual temperature of the device die.
Get you probe on the device case; it'll be cooler than the actual junction, but better than the heatsink. I assume you're doing this because you want the amp to survive your biasing effort.
Yes, the thermocouple could be directly attached to the TO-3 case, if the DVM being used
is battery powered and floating. That would be much more accurate.
is battery powered and floating. That would be much more accurate.
I am biasing my amp based on Nelson Pass recommendation to set heatsink temp at 50c. I do not know at what temp to set the bias if putting it on the casing of the power transistor.
That's a general suggestion to be safe, but it does not tell you the device junction temperature.
It also assumes good thermal design to start with.
With a poor thermal interface to the sink, the device junction could be too hot.
With rail fuses, you could directly measure and set the bias current, which is the actual parameter involved.
Just sub an open fuse, with leads attached to the ends, for one of the actual rail fuses, and measure
the DC bias current with a DVM. After that is set, then you can check the heat sink temperature, to see if
it is acceptable (usually below 50C). If not, back off the bias.
It also assumes good thermal design to start with.
With a poor thermal interface to the sink, the device junction could be too hot.
With rail fuses, you could directly measure and set the bias current, which is the actual parameter involved.
Just sub an open fuse, with leads attached to the ends, for one of the actual rail fuses, and measure
the DC bias current with a DVM. After that is set, then you can check the heat sink temperature, to see if
it is acceptable (usually below 50C). If not, back off the bias.
Last edited: