I powered up a recently completed Balanced Line Stage built on power supply and amp PCBs from Kristijan (http://web.vip.hr/pcb-design.vip/bls.html). The power supply PCB was built first and tested fine. The two channels on the amp PCB were built at the same time, part for part since I was matching resistors.
The problem: upon powering up, I noticed that the power resistors R7 and R8, acting as the current sources at the drain of the MOSFETs, for one channel were getting much hotter than the other channel's.
I followed the tips from the Balanced Line Stage manual to measure voltage at the gate, drain, and source of each transistor. The one channel that had the cooler resistors had voltages that were as expected.
But the other channel had the left IRF610 with a drain voltage that was of the correct magnitude, but the wrong sign. Also the voltage at the source was the same as the voltage at the drain (it was tested with the amp powered down and it is not a short). The same problem occurs with the right-side IRF610, except the drain and source voltages are more than double the magnitude that they should be.
All resistor values of both channels were re-checked and all were the same (between both channels).
Any ideas?
The problem: upon powering up, I noticed that the power resistors R7 and R8, acting as the current sources at the drain of the MOSFETs, for one channel were getting much hotter than the other channel's.
I followed the tips from the Balanced Line Stage manual to measure voltage at the gate, drain, and source of each transistor. The one channel that had the cooler resistors had voltages that were as expected.
But the other channel had the left IRF610 with a drain voltage that was of the correct magnitude, but the wrong sign. Also the voltage at the source was the same as the voltage at the drain (it was tested with the amp powered down and it is not a short). The same problem occurs with the right-side IRF610, except the drain and source voltages are more than double the magnitude that they should be.
All resistor values of both channels were re-checked and all were the same (between both channels).
Any ideas?
Ah, nevermind. It turns out my mica insulators with thermal paste on one side were somehow having a resistance between the tabs (drain) of the IRF610 and the big heatsink that all transistors are mounted to.
Are there any household items that could act as an insulator with sufficient thermal properties for the IRF610's in the Balanced Line Stage? How about a piece of paper? 🙁
Are there any household items that could act as an insulator with sufficient thermal properties for the IRF610's in the Balanced Line Stage? How about a piece of paper? 🙁
To be sure it wasn't the fasteners, I had switched to a shorter screw, so all I have now is a threaded nylon spacer that's pushing down on the transistor's tab.
I'm a bit afraid of the setup now. I re-measured voltages on the IRF610's, and they are now correct. However, I'm seeing a voltage on my heatsink compared to points on the amp circuit, even though I thought the heatsink was electrically isolated from the circuit. All standoffs are nylon spacers. All transistors either have an insulating washer and mica insulator (from a TO-220 mounting kit), or just a mica insulator without a mounting screw. Measuring resistance between the tabs and heatsink gave some odd reading for one power supply MOSFET - the reading was a negative resistance? How does this happen in a digital multimeter?
I'll recheck that mosfet's insulation.
EDIT: Something was definitely wrong with the insulation somewhere. I just pulled everything off the heatsink and ran it - all voltages were fine, and there weren't any "fishy" voltages between grounds, tabs, heatsinks, etc.
I'll recheck that mosfet's insulation.
EDIT: Something was definitely wrong with the insulation somewhere. I just pulled everything off the heatsink and ran it - all voltages were fine, and there weren't any "fishy" voltages between grounds, tabs, heatsinks, etc.
Possum said:
Classical one, this! You turn off the amp, start measuring resistance which really is sending a current out and measuring a voltage drop. But those caps in the circuit are still (partially) charged, and they will send out their own current when you touch points in the circuit. This current may swamp the measuring current and can be reverse, so you get very odd results.
Telltales are: reading continuously changes (caps are discharging) and reversing the meter probes gives completely different results.
Cure: discharge caps (especially the power supply reservoirs) and/or wait 10 minutes.
Jan Didden
Jan,
Thanks for the explanation. I noticed that there was no resistance between the transistor tabs and the heatsink when I went back and measured a while later (but something was still in contact with the heatsink, causing the problem I was having).
I ended up using a strip of printer paper over the mica pads to help ensure the isolation between transistors and heatsink. The BLS sounds great. 🙂 If using paper won't have any adverse effects in the future, I think I'll leave it in there. But if it's a bad idea, is there any place to buy large strips of a proper insulator? Say 3"x1" and 6"x1"?
-Greg
Thanks for the explanation. I noticed that there was no resistance between the transistor tabs and the heatsink when I went back and measured a while later (but something was still in contact with the heatsink, causing the problem I was having).
I ended up using a strip of printer paper over the mica pads to help ensure the isolation between transistors and heatsink. The BLS sounds great. 🙂 If using paper won't have any adverse effects in the future, I think I'll leave it in there. But if it's a bad idea, is there any place to buy large strips of a proper insulator? Say 3"x1" and 6"x1"?
-Greg
Insulating tabs
I don't know where to buy strips, but I would not keep the paper in there too long.
Probably not good enough as electrical isolator, and too good as thermal isolator. All of which is most probably varying all over the place with temperature, humidity and time of day.
Jan Didden
I don't know where to buy strips, but I would not keep the paper in there too long.
Probably not good enough as electrical isolator, and too good as thermal isolator. All of which is most probably varying all over the place with temperature, humidity and time of day.
Jan Didden
- Status
- Not open for further replies.