I've started a dual F5T V2 and V3 build and I have a few related questions:
1. I assume the load sharing resistors (.5 ohm) were calculated based on an assumed maximum output transistor mismatch. NP says they should be matched to within .1V. I think I can do much better than that, probably as close as .02V. Can I reduce the size of the load sharing resistors a bit to .33 ohms?
2. The current .5 ohm values puts the desired bias of .6 - .48 amps (equivalent to ..3 - ..4 voltage drop) right on the hairy edge of when the bypass diodes start conducting at higher temperatures. If I made the load sharing resistors .33 ohm, the voltage drop across the resistors at the same bias would be .2v to .27v; below the diode conducting point even at high temperature.
3. How much power do the diodes dissipate? Most people seem to attach them to the main heat sinks but with class A the main heat sinks run so hot that they are more of a heat source to the diodes. Could I get by with small to-220 type heat sinks of ~ 15 degC/watt on the diodes and keep them cooler than the main heat sinks?
1. I assume the load sharing resistors (.5 ohm) were calculated based on an assumed maximum output transistor mismatch. NP says they should be matched to within .1V. I think I can do much better than that, probably as close as .02V. Can I reduce the size of the load sharing resistors a bit to .33 ohms?
2. The current .5 ohm values puts the desired bias of .6 - .48 amps (equivalent to ..3 - ..4 voltage drop) right on the hairy edge of when the bypass diodes start conducting at higher temperatures. If I made the load sharing resistors .33 ohm, the voltage drop across the resistors at the same bias would be .2v to .27v; below the diode conducting point even at high temperature.
3. How much power do the diodes dissipate? Most people seem to attach them to the main heat sinks but with class A the main heat sinks run so hot that they are more of a heat source to the diodes. Could I get by with small to-220 type heat sinks of ~ 15 degC/watt on the diodes and keep them cooler than the main heat sinks?
The diodes must heat up with the output stage. When a transistor heats up, the gain goes up and can lead to thermal runaway. Fitting the bias chain diodes to the output transistor heat sink will reduce the voltage drop across the diodes, (as they heat up), reducing the base - emitter voltage on the output transistors which will reduce the quiescent current leading to more thermal stability.
I consider the source resistors as a reduction of current imbalance. NOT as an eliminator of current imbalance. I also consider 0.1Vgs difference to be a poor match for paralleled devices. Use 20mVgs if you can achieve that. Many aim for better than 10mVgs at the operational current, but few try to achieve that with operational voltage and operational heatsink temperature.
I seem to be nearly out on my own in suggesting that the diodes should not turn on at normal bias and maximum operating temperature. Even though I am out on a limb, I do support your idea of arranging the diode Vf at well below turn on voltage.
If you are playing music with an average level of 10dB below the maximum output of the amplifier, that would be very loud, then the diodes should hardly ever reach a 0.5Vf. On the few occasions that they do turn on fully the duty cycle should be very low and thus the heating effect should be very low. I would suggest you try using your 15C/W sink to each bypass diode and measure the device temperature during quiet and medium and loud music playing.
You could try full ClassA power testing temperature measurement and even double power in ClassAB mode temperature measurement and see what you get. I can't recall anyone posting such data.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.