Hello everyone . Is it a right way to put a miliohm meter in series with one of the + or - DC rails and measuring the idle bias current of the output transistors?
I have an amplifier with 4 pairs of IRFP240 Mosfets. BIAS current for each pair is 50ma (according to designer's suggestion). So, 4 pairs equals 200ma of current draw in idle mode when input is shorted to ground.
I'm a newbie and wanted to ask you guys if you suggest this measuring method?
Or perhaps there is another way to do this.
Thanks.
I have an amplifier with 4 pairs of IRFP240 Mosfets. BIAS current for each pair is 50ma (according to designer's suggestion). So, 4 pairs equals 200ma of current draw in idle mode when input is shorted to ground.
I'm a newbie and wanted to ask you guys if you suggest this measuring method?
Or perhaps there is another way to do this.
Thanks.
Are there any resistor in the FET's source path?
This is usually the way to monitor/set the bias current.
This is usually the way to monitor/set the bias current.
Yes, there should be a .1 or .2 Ohm resistor in series with each tranny. Hafler says to put an ammeter in replacement of a rail fuse, but that sounded like an easy way to slip.
I set my bias with a spectrum analyzer. (many free PC software ones) Run 1K in and tune for minimum distortion. One must make sound card probes to keep from blowing up your sound card. Look at the SoundEasy users guide and several threads on making a sound card interface.
THat way I can leave it on the bench to warm up to normal operating temps.
I set my bias with a spectrum analyzer. (many free PC software ones) Run 1K in and tune for minimum distortion. One must make sound card probes to keep from blowing up your sound card. Look at the SoundEasy users guide and several threads on making a sound card interface.
THat way I can leave it on the bench to warm up to normal operating temps.
This is what I did. the voltage across the resistors is supposed to be around 35-40mv for 200ma of BIAS. But the point is that, each of those 4 resistors show different voltages across themselves!! the voltages are : 8.7 / 11.7 / 21 / 15.9 milivolts. ( I did not increased the bias to the 35mv because the driver transistors were becoming too hot)
The voltages should be equal or vary about 6-7 milivolts at maximum. But in my case the difference is vast!
Is it possible that something is wrong with Mosfets?
The voltages should be equal or vary about 6-7 milivolts at maximum. But in my case the difference is vast!
Is it possible that something is wrong with Mosfets?
The variation could be normal, but you have 8 resistors to measure, not 4. The current through any one of the top FETs doesn't necessarily go down through the one below in your circuit, it all gets mixed up and shared depending on each FET's characteristics. If it's easy, I would try swapping FETs around for a better balance. E.G. try swapping the highest current top FET with the lowest current bottom FET.
This is where you hear about using matched transistors (or FET's) in output stages so that the quiescent and load currents are shared more or less equally between them all.
I would put all of the high values on one "side" (top or bottom) and the low values on the other.
At least then they will be as close a match as possible with what you have available.
Or remove them all and match them in a controlled 50mA circuit.
I would put all of the high values on one "side" (top or bottom) and the low values on the other.
At least then they will be as close a match as possible with what you have available.
Or remove them all and match them in a controlled 50mA circuit.
interesting suggestions... I had never thought to do this. Clever!
I will give it a try tomorrow and report the results. Thank you guys!
I will give it a try tomorrow and report the results. Thank you guys!
Some manufacturers fit an 0R1 resistor in series with the + & - supply to the group of FETs and they measure the total Quiescent current NOT WITH A SHORT OR LOAD TO GROUND on the speaker terminal or you will get strange readings! When recorded and set up, the solder bridge is reinstated.
As this design of boraomega's is a stage amplifier, low distortion is not a priority so bias is specified low, at only 25mA per output pair. With 0R22 resistors, you should only measure around 5-6 mV anyway. Assuming your heatsink is large enough for maximum dissipation at high power and you wait until heatsink temperature is constant before taking measurements, the mismatches should not be too critical.
Are you following reliable advice about using higher bias, the controller and the heatsink requirements for it with this design?
Last edited:
After many hours of shuffling the transistors, I really messed up and lost the count! I couldn't finally arrange them at both sides according to the voltage readings I had made before!!! It was so confusing and made me really tired.
So, I did something silly but believe or not it worked. I started raising BIAS very slowly and at the same time controlling the temperature of the driver transistors with a thermometer.
It took around 25 minutes till the amplifier reached its thermal stability in idle mode. I increased the BIAS till the temperature reached 70 degrees. ( which is almost a bit higher than normal for silicon devices)
I tested the amplifier at maximum volume for an hour and it worked perfectly without ANY distortion or overheating. The sound was so clear. free of any hum or hiss noises. I probably could still increase the BIAS but since the sound was perfect, I thought it's better not to push it higher.
So, I did something silly but believe or not it worked. I started raising BIAS very slowly and at the same time controlling the temperature of the driver transistors with a thermometer.
It took around 25 minutes till the amplifier reached its thermal stability in idle mode. I increased the BIAS till the temperature reached 70 degrees. ( which is almost a bit higher than normal for silicon devices)
I tested the amplifier at maximum volume for an hour and it worked perfectly without ANY distortion or overheating. The sound was so clear. free of any hum or hiss noises. I probably could still increase the BIAS but since the sound was perfect, I thought it's better not to push it higher.
Ian what you are saying is exactly correct. Now I reading something around 6-7 mV across the resistors. I'm using the BIG old computer CPU coolers for this amplifier. After hours of working at maximum volume heat-sinks are only slightly warm.
That's good that you have stabilized the amplifier's bias now, though I think the mass of even big CPU heatsinks (at least the typical Pentium PC types I have seen) is probably not enough for 4-8 pairs of output MOSFETs and the thermal stability will be poor, even with the necessary fan cooling. However, I assume the amplier now stabilizes a lot quicker than 25 mins from power-up!
Anyway, If you have safe, reliable operation at full power and you like the sound quality, you have achieved something good. Nice going. 🙂
- Status
- Not open for further replies.