C3
220uF is a nearly universal value for this capacitor. Sometimes you will see 330uF. If the value is increased there is little benefit The importance of the value is that it should be high enough that the AD resistance is very very low. This is to avoid introduction of dsischarge noise typival of elevtolytics getting into the signal. Were it not for that a much lower value might be acceptable. I've used a smaller values on SPICE simulations where the assumption was that it was a theoritically perfect capactor. In that case it's fuction was simply a DC blocker.
I suspect attaching q9 to the main heatsink or piggybacking it to one of the output transistors might be a good idea. I have built one of these and noted when setting the quiesent bias that the voltage read oscillates a bit. It was nearly impossible to set it so that it stays put. This doesn't seem to have an impact on actual performance, but it manages to bug me. It certainly can't hurt.
220uF is a nearly universal value for this capacitor. Sometimes you will see 330uF. If the value is increased there is little benefit The importance of the value is that it should be high enough that the AD resistance is very very low. This is to avoid introduction of dsischarge noise typival of elevtolytics getting into the signal. Were it not for that a much lower value might be acceptable. I've used a smaller values on SPICE simulations where the assumption was that it was a theoritically perfect capactor. In that case it's fuction was simply a DC blocker.
I suspect attaching q9 to the main heatsink or piggybacking it to one of the output transistors might be a good idea. I have built one of these and noted when setting the quiesent bias that the voltage read oscillates a bit. It was nearly impossible to set it so that it stays put. This doesn't seem to have an impact on actual performance, but it manages to bug me. It certainly can't hurt.
C3
220uF is a nearly universal value for this capacitor. Sometimes you will see 330uF. If the value is increased there is little benefit The importance of the value is that it should be high enough that the AD resistance is very very low. This is to avoid introduction of dsischarge noise typival of elevtolytics getting into the signal. Were it not for that a much lower value might be acceptable. I've used a smaller values on SPICE simulations where the assumption was that it was a theoritically perfect capactor. In that case it's fuction was simply a DC blocker.
I suspect attaching q9 to the main heatsink or piggybacking it to one of the output transistors might be a good idea. I have built one of these and noted when setting the quiesent bias that the voltage read oscillates a bit. It was nearly impossible to set it so that it stays put. This doesn't seem to have an impact on actual performance, but it manages to bug me. It certainly can't hurt.
I you are exporing modifcations to the P3A, you may want to consider setting it up in a SICE model. Rod recommends the freeware version os SIMetrics. I've been checking it out for a few weeks and it seens very user friendly -- besides the price it right.
220uF is a nearly universal value for this capacitor. Sometimes you will see 330uF. If the value is increased there is little benefit The importance of the value is that it should be high enough that the AD resistance is very very low. This is to avoid introduction of dsischarge noise typival of elevtolytics getting into the signal. Were it not for that a much lower value might be acceptable. I've used a smaller values on SPICE simulations where the assumption was that it was a theoritically perfect capactor. In that case it's fuction was simply a DC blocker.
I suspect attaching q9 to the main heatsink or piggybacking it to one of the output transistors might be a good idea. I have built one of these and noted when setting the quiesent bias that the voltage read oscillates a bit. It was nearly impossible to set it so that it stays put. This doesn't seem to have an impact on actual performance, but it manages to bug me. It certainly can't hurt.
I you are exporing modifcations to the P3A, you may want to consider setting it up in a SICE model. Rod recommends the freeware version os SIMetrics. I've been checking it out for a few weeks and it seens very user friendly -- besides the price it right.
be careful ....
jojo:
I would very strongly suggest paying attention to andrea's comment about looking at the 300W subwoofer project to parallel the output devices appropriately. A power amp of this design having 1. no emitter ballast resistors, 2. high voltage rails and 3. decent value power supply caps can make for some unpleasant and possibly spectacular results.
Not to mention the waste of some innocent and perhaps expensive, difficult to acquire output devices. Actually, the drivers may decide to check out as well ...
And by the way, make sure the drivers are biased "appropriately" to drive all those output devices "appropriately" (load impedance, frequency range, etc.)
mlloyd1
who has been there, long ago...
jojo:
I would very strongly suggest paying attention to andrea's comment about looking at the 300W subwoofer project to parallel the output devices appropriately. A power amp of this design having 1. no emitter ballast resistors, 2. high voltage rails and 3. decent value power supply caps can make for some unpleasant and possibly spectacular results.
Not to mention the waste of some innocent and perhaps expensive, difficult to acquire output devices. Actually, the drivers may decide to check out as well ...
And by the way, make sure the drivers are biased "appropriately" to drive all those output devices "appropriately" (load impedance, frequency range, etc.)
mlloyd1
who has been there, long ago...
Andypairo said:
darkm4n said:
if Onsemi is available in your part of the world but you don't know which ones to get, I have used the MJLs and MJWs with good luck (21193/21194 or 3281/1302 for example).
If onsemi isn't available, you can try 3281/1302 from the Japanese or their newer variants.
darkm4n said:
the P3A from Elliot is copyrighted, I am sure. you can either do a hand layout or use veroboards: I used a veroboard myself.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.