Can you guesstimate the degC/W figure for the heatsink by looking others in the catalogue and maybe get an idea of its mass and therefore heat capacity?
Then you could say how fast it should heat up if everything is as it should be.
Those thermal switches are good for shutting things off, but I use some cheap temp controllers from ebay, a thermistor sensor, an LED temp readout and a programmable limit for a relay to switch.
Some amps seem to have heatsinks which are for music, not continuous sinewave testing. I have in front of me a little Creek 4040 which seems to make a good volume of music with a small heatsink.
Then you could say how fast it should heat up if everything is as it should be.
Those thermal switches are good for shutting things off, but I use some cheap temp controllers from ebay, a thermistor sensor, an LED temp readout and a programmable limit for a relay to switch.
Some amps seem to have heatsinks which are for music, not continuous sinewave testing. I have in front of me a little Creek 4040 which seems to make a good volume of music with a small heatsink.
So the emitter resistors are specified as .5 ohm. What is the DC no signal bias voltage set to? Have not seen the value you set in your posts.
Knowing the bias value will let us calculate the transistor power dissipation assuming +- 30 volt supplies.
Are both channels on the same heat sink?
Does it get hot quiescently or do you need to be delivering power to a load?
Knowing the bias value will let us calculate the transistor power dissipation assuming +- 30 volt supplies.
Are both channels on the same heat sink?
Does it get hot quiescently or do you need to be delivering power to a load?
Both channels are mounted to the same heatsink so a little hard to tell, but the output pair that has lower voltage appears to be hotter.
I have set the bias to 20mV.
It only gets hot with load and signal. Seems to MAYBE get into thermal runaway at about 5-7V RMS on the outputs
I have set the bias to 20mV.
It only gets hot with load and signal. Seems to MAYBE get into thermal runaway at about 5-7V RMS on the outputs
The bias goes to maximum if the trimmer wiper arm opens. You may want to replace the trimmer.
Ed
Ed
When it gets hot with signal applied, kill the signal and quickly remeasure the bias to see how it’s changed with temp.
How large is the difference in rail voltages, or more specifically, what are the voltages of the rails for both channels?
How large is the difference in rail voltages, or more specifically, what are the voltages of the rails for both channels?
I've ordered TIP41, 42. They seem like a beefier pair. I will try these other things listed in a bit.
Anyone care to comment on this section of the service manual? I'm trying to understand it and see if it warrants looking into for faults. I've got the TIP41 - 42 pairs in and am going to the tests again.
Maybe I missed it, but seems you don´t mention what´s the DC-offset on the output terminals??
If you maybe have larger offset on one channel, that could be the reason for "one channel hotter"??
Which means, the bias will increase as soon as you connect a load, speakers or dummy.
The bias stays about 20mV when hot and cutting signal.
+/-28V for rail power.
Does it work to measure the voltage drop over the emitter resistors and then calculate that against the RMS value on the dummy load. Those two numbers would give me the efficiency of the amp? And then I would know if it is wasting heat more than it should be for AB?
Your bias may be "set", but is it right? The bias is tricky with the SX-550 and it's important to note that the bias trimmer pots are reversed. Counter Clockwise is maximum and being an emitter follower output they idle pretty warm. I have a 100 Ohm shunt across the bias pots and that tames the maximum bias to about 80 mV which is much higher than needed for proper adjustment and prevents the ridiculous quiescent dissipation the stock circuit allows. The bias is best set with a distortion analyzer and it's quite clear where it belongs judging from the residual. It's also notable that Cdom(miller capacitor) is barely large enough at 22P and needs to be increased if the output transistors are replaced. with anything modern. Mine uses 150P with it's D44/45H11 outputs and MPSA06/56 drivers. It sounds great, runs cool and the distortion at full power is about 0.008% which is a LONG way form the 0.3% the specs suggest.
