I have questions about the output cap.
I am considering Nichicon KG , is there a sonic difference between Type I and Type II of them ? LKG1J682MESBAK Nichicon | Mouser vs LKG1J682MESCBK Nichicon | Mouser ?
Another option would be ALC10S1103DH KEMET | Mouser , kemet alc10s, slit foil capacitors , anyone familiar with this?
Thanks
I am considering Nichicon KG , is there a sonic difference between Type I and Type II of them ? LKG1J682MESBAK Nichicon | Mouser vs LKG1J682MESCBK Nichicon | Mouser ?
Another option would be ALC10S1103DH KEMET | Mouser , kemet alc10s, slit foil capacitors , anyone familiar with this?
Thanks
Some minor power loss with .68 ohm resistors, fractions of decibel since the ear hears logrithmically. .33 Increases risk of thermal runaway. Since base board doesn't have temp sense diodes on the heat sink, probably .33 ohm not a good idea. The thermal trak variant above post 683 a good idea with very expensive parts.
I have thermal sense diodes mounted above the TO3 transistors on a remote heat sink, since the on board TO3 of base layout are located for wimpy degC/Watt rated heat sinks. Thousands of hours on my AX6 with .51 ohm emitter resistors with no problem, but I have fans on the heat sink.
Note to Asus, the 6.8 kohm feedback resistor is to compensate for any "distortion" of the output cap. Build a speaker burner amp if you hallucinate at the thought of caps in the sound path. Then use $12 speaker drivers; they will sound "wonderful" and be disposable to boot. My speaker drivers are about $200 each; produce 2nd order HD 20 db down all frequencies 54 to 14.5 khz @ 1W 1m. Nichicon 3300 uf 80v were out of stock the day I bought. The 3300 united chemicon I use are probably demonstrably worse than any nichicon 3300 made. Oddly, the SP2-XT & the AX6 can sound very like the Steinway piano in the same room.
I have thermal sense diodes mounted above the TO3 transistors on a remote heat sink, since the on board TO3 of base layout are located for wimpy degC/Watt rated heat sinks. Thousands of hours on my AX6 with .51 ohm emitter resistors with no problem, but I have fans on the heat sink.
Note to Asus, the 6.8 kohm feedback resistor is to compensate for any "distortion" of the output cap. Build a speaker burner amp if you hallucinate at the thought of caps in the sound path. Then use $12 speaker drivers; they will sound "wonderful" and be disposable to boot. My speaker drivers are about $200 each; produce 2nd order HD 20 db down all frequencies 54 to 14.5 khz @ 1W 1m. Nichicon 3300 uf 80v were out of stock the day I bought. The 3300 united chemicon I use are probably demonstrably worse than any nichicon 3300 made. Oddly, the SP2-XT & the AX6 can sound very like the Steinway piano in the same room.
Last edited:
I got it figured out. I had some 1R/2W resistors so I just hooked them up in parallel. It is all populated but I will wait until tomorrow to begin the testing. I will have to work up a PSU. I have several PSU panels that I should be able to just ignore the center tap and use the + and - outputs. I have a 22-0-22VAC transformer that should net me about +60V-0VDC output. All I had for output caps was 4700uF/50V so I don't want to push it too hard.
Attachments
50 volt output caps are fine for 60 volt rail. They will see 30-ish volts of DC. Could be up to 60 if the output heads for the rail due to a fault, but a 50 volt cap will take that for a while if there is no ripple current to speak of. 35 volt cap might be iffy of you count on it for speaker DC protection.
Ok I got it on the bench. Testing at +50-0VDC. It will play a sine wave but I can't get the bias to come up at all. It is pure class B right now. Even with the pot adjusted to 0R it still has no bias. I'm not sure if the voltage drop across the two diodes in the Thermal tracks are too much compared to the 4148s. anyone have a thought about that? I didn't try shorting one of them maybe I'll should give that a try.
I've had to use 2 diodes series a schottky to get bias up. Pot parallel one of the diodes. Diodes were salvage rectifiers, generic 600-700 mv test voltage. 1n4148 tend to run more 750 mv. I'm using white box (NTE) mj15003 OT.
You can parallel the extra diode with a pot. If the wiper opens up due to old age, bias goes to zero, no damage.
BTW I like the look of your thermo trak build. Nice transfer of heat to a substantial heat sink with angle. I envisioned the original build with those individual Avid TO3 heat sinks with the 2 cm x 1 mm fins all around.
You can parallel the extra diode with a pot. If the wiper opens up due to old age, bias goes to zero, no damage.
BTW I like the look of your thermo trak build. Nice transfer of heat to a substantial heat sink with angle. I envisioned the original build with those individual Avid TO3 heat sinks with the 2 cm x 1 mm fins all around.
Last edited:
Oposite, bias is maximum with pot adjust to 50R, and will rise with voltage on diodes... If there is no bias with 50R pot use 100R pot.
That is exactly what Ltspice is saying. Needs 90ohm to get bias up. I was doubting that since it seems a few guys have built this amp. How did they get the bias with a 27R resistor?
I got one board working. I have it set to 160mA. The trimmer is set to 113R. Outputs are just warm and it sounds better there. The bass is a little flabby but that may be because I'm feeding it with my phone. Tomorrow I will get the other board sorted and then feed it with a CD player. I put it on the scope. Sine waves look fine but not square waves but again, that is using a wave generator app from my phone.
More to come.
More to come.
No wonder. To get any bias at all, the voltage across the bias diode string has to exceed three Vbe's in this quasi complementary design (both drivers' plus the upper power device's Vbe).
Best regards!
Hey Guys,
I got it working. It needs 200R trimmers in the bias to get it to work. I think they are set and about 120R right now. That gives me 70mV across the pair of 1R resistors in parallel or about 140mA of bias. The outputs run just a little warm. I am running it on +63V-0V. Seems very happy there. Forget what I said about the bass being flabby. It sounds really nice and warm. Cool little amp.
YouTube
Blessings, Terry
I got it working. It needs 200R trimmers in the bias to get it to work. I think they are set and about 120R right now. That gives me 70mV across the pair of 1R resistors in parallel or about 140mA of bias. The outputs run just a little warm. I am running it on +63V-0V. Seems very happy there. Forget what I said about the bass being flabby. It sounds really nice and warm. Cool little amp.
YouTube
Blessings, Terry
Last edited:
