the average power consumption of a ClassA amplifier is more than the average power consumption of a ClassAB amplifier.
The 25W ClassA might be pulling 1A to 2A from the smoothing capacitors.
The 100W ClassAB BJT might be pulling 0.05A to 0.2A from the smoothing capacitors.
The 100W ClassAB mosFET might be pulling 0.2A to 1A from the smoothing capacitors.
The 25W ClassA might be pulling 1A to 2A from the smoothing capacitors.
The 100W ClassAB BJT might be pulling 0.05A to 0.2A from the smoothing capacitors.
The 100W ClassAB mosFET might be pulling 0.2A to 1A from the smoothing capacitors.
The bridge rectifier diodes pass on alternate half cycles, that means each diode is operating on a 50% duty cycle.
If the device passes 2A and drops 0.8Vf, then the dissipation will be 0.8V * 2A = 1.6W, but this reduces due to duty to 0.8W Four diodes inside an integrated bridge would thus dissipate 4*0.8W = 3.2W
But I was having a discussion about this in another Thread.
Today (while answering your question) I realised the diodes are not on a 50% duty cycle.
They pass for about 8% to 20% of the whole cycle and while passing this short term average current is ~3times the whole cycle average.
This may account for our difference of about 2:1 that we were discussing.
It may be that the bridge dissipation could be around half that 3.2W
That will still be pretty warm.
When setting up the output bias of my first big ClassA amplifier I was alarmed at the temperatures of the two bridges. I added a 7.5C/W shared between the two and they ran only slightly warm.
If the device passes 2A and drops 0.8Vf, then the dissipation will be 0.8V * 2A = 1.6W, but this reduces due to duty to 0.8W Four diodes inside an integrated bridge would thus dissipate 4*0.8W = 3.2W
But I was having a discussion about this in another Thread.
Today (while answering your question) I realised the diodes are not on a 50% duty cycle.
They pass for about 8% to 20% of the whole cycle and while passing this short term average current is ~3times the whole cycle average.
This may account for our difference of about 2:1 that we were discussing.
It may be that the bridge dissipation could be around half that 3.2W
That will still be pretty warm.
When setting up the output bias of my first big ClassA amplifier I was alarmed at the temperatures of the two bridges. I added a 7.5C/W shared between the two and they ran only slightly warm.
Last edited:
Sadhill,
You must fold the heat sink underneath the board and make sure it's in physical contact with the base of the amplifier case.
Some builders did not like my mechanical design, and instead mounted the rectifiers on a heatsink. You can get a 5 deg C /W heatsink from RS that will do the job.
However, my sx and nx have the rectifier folded under the PSU board as shown in the write-ups and can deliver their rated powers continuously into 4 or 8 ohms without overheating.
You must fold the heat sink underneath the board and make sure it's in physical contact with the base of the amplifier case.
Some builders did not like my mechanical design, and instead mounted the rectifiers on a heatsink. You can get a 5 deg C /W heatsink from RS that will do the job.
However, my sx and nx have the rectifier folded under the PSU board as shown in the write-ups and can deliver their rated powers continuously into 4 or 8 ohms without overheating.
I did eventually manage to cut the rectifier legs and put the rectifiers against the baseplate. Works fine, but with nothing connected I get +/-26.6V DC on both PSU's, well above the recommended 22V. Is it going to be only a matter of heat dissipation, or may it be a problem for the overall result and for some component values ? My heatsinks are around 0.3° C/W for 1 channel.
(the core transformer is a 2x18V, but my mains voltage is 231V)
(the core transformer is a 2x18V, but my mains voltage is 231V)
Bonsai,
Sorry, I'm lost now. I tried to power on the amps, but inconsistent values for I across Vpos removed fuse (0A on one board, 800mA on the second). I really don't understand how should be R6 : maximum so R21+R6+R20=4.2k ? This seems to be what you said in post 1831. But the pdf still says "minimum" and that's what
I did. Hope this did not harm anything on the pcb...
What about the PSU's 22ohm resistor between 0V and chassis : needed or not ?
Sorry, I'm lost now. I tried to power on the amps, but inconsistent values for I across Vpos removed fuse (0A on one board, 800mA on the second). I really don't understand how should be R6 : maximum so R21+R6+R20=4.2k ? This seems to be what you said in post 1831. But the pdf still says "minimum" and that's what
I did. Hope this did not harm anything on the pcb...
What about the PSU's 22ohm resistor between 0V and chassis : needed or not ?
In fact, I discovered that the internal fuse of my DVM has blown up ! thus the 0A measure. with R6 at the max position, the current rises to 2.2 amps, and blows the 1A fuse on the Vneg input... there is apparently no short, I have to investigate. All the parts are the recommended ones, transistors included. All my 4 pcb's behave the same.
Looking at the circuit diagram (page 14 of the V2.10 document)
- lift one end of R20
The total supply current should drop to 500-600mA (amp actually works very well with 600 mA bias BTW)
If it does, check the value of R20 (should be 2.2k) - if its lower than this - it could explain your problem.
Check R21 - should be 1k. If it is higher than this, it would explain your problem.
Make sure R6 (the 1k pot) is actually 1k
If this all checks out ok, you need to look elsewhere - but check the above out first and if no luck, come back.
- lift one end of R20
The total supply current should drop to 500-600mA (amp actually works very well with 600 mA bias BTW)
If it does, check the value of R20 (should be 2.2k) - if its lower than this - it could explain your problem.
Check R21 - should be 1k. If it is higher than this, it would explain your problem.
Make sure R6 (the 1k pot) is actually 1k
If this all checks out ok, you need to look elsewhere - but check the above out first and if no luck, come back.
thank you Bonsai.
Just to make sure : R6 must be adjusted to 1kohm, that is a total of 2kohm between BC547 (in fact I use BC550) base and NJW3281 emitter ?
By the way, I checked all the components on the pcb's, no misplacement, correct values, zeners working, 1N4007 put the right way, no solder bridge...
We tried to have it work with an electronician friend (very skilled in A-class amps, part of his business) this morning. He used a variac to feed the PSU, there was no way to go above 0.2V for the Q4 emitter voltage, which occured at roughly 135V mains for 0.3 amp overall current. Any attempt to push the mains higher resulted in Q4 emitter voltage dropping whereas it was raising on the Q7 emitter.
I'll try your solution once you confirm for R6.
Just to make sure : R6 must be adjusted to 1kohm, that is a total of 2kohm between BC547 (in fact I use BC550) base and NJW3281 emitter ?
By the way, I checked all the components on the pcb's, no misplacement, correct values, zeners working, 1N4007 put the right way, no solder bridge...
We tried to have it work with an electronician friend (very skilled in A-class amps, part of his business) this morning. He used a variac to feed the PSU, there was no way to go above 0.2V for the Q4 emitter voltage, which occured at roughly 135V mains for 0.3 amp overall current. Any attempt to push the mains higher resulted in Q4 emitter voltage dropping whereas it was raising on the Q7 emitter.
I'll try your solution once you confirm for R6.
Bonsai,
After R20 disconnection and R6 adjusted to 1k, I get 647 mA across Vpos (with Vpos 26.5V)
R20 is 2.19k, R21 is 0.997k
I saw RBalu had the same issue and speaks of the "pcb error of Q3 base connection" in post 2147 solved "with jumpers". But I don't find anything wrong with Q3 and my pcb's here...
What should be the next step ?
After R20 disconnection and R6 adjusted to 1k, I get 647 mA across Vpos (with Vpos 26.5V)
R20 is 2.19k, R21 is 0.997k
I saw RBalu had the same issue and speaks of the "pcb error of Q3 base connection" in post 2147 solved "with jumpers". But I don't find anything wrong with Q3 and my pcb's here...
What should be the next step ?