Aleph-X builder's thread.

[Netlist]

Terry,
William is correct. If R1/4 got hot, there is either DC or AC running.
Could you update what you measure over these R's? Both scope and DMM would be nice.

/Hugo

[still4given]

Originally posted by wuffwaff
Hi Terry,

an oscillating amp can´t sound very well...... You still have the problem of R1/4 being very hot wich means that you´ve got a problem here. Try to put the scope on these and see what waveform you have.

This maybe is also the cause for your transformer running hot.

Try upping the C´s in the active current source from 3n3 to 6n8 or more. Maybe the bogger input capacitance of the fets you use has an influence here.

William

For some reason I used 7pF for C9 and C10. I don't remember now where I came up with that value. Where is it stated what value that should be? I looked through the wiki and I didn't see it.

I'll try something different there.

Bessings, Terry

[wuffwaff]

Terry,

then try 4n7.

William

P.S. 7pF is a very strange value. Are you shure?

[still4given]

Hi William,


Yeah they were 7 +/- .5pf 500VEM. I bought them thinking I needed 5-10pf. I figured they fit right in there.

I didn't have any 4n7's handy but I did have some 3n3's so I put those in both channels. I'll pick up some 4n7's this week. I also changed out all of the 0r15 resistors for 0r22 on one channel only so I could test the difference. The channel with the 0r22's sounds more muffled than the other. A little less presence in the upper mids and treble. This was not an improvement in sound to my ears.

The channel with the 0r22's had to be biased at .530v on the source resistors in order to get the offset right. If I set it at .5v the offset would go to about 200mv. With them at .530 the offset is about 72mV. That's with the absolute offset at around +/-5mV.

Oh yeah, I also changed the input caps to 12uf orange drops in bothe channels. That's as close as I could find locally.

Well, I don't know how to measure output but here's what I tried.

I fed a 440hz sine wave into both channels one at a time by using one audio feed and moving it from channel to channel. I measured AC voltage across the + and - outputs. I had a pair of 4ohm speakers plugged into the speaker jacks, one each side.

The channel with the 0r22 source resistors measured higher. I then hooked up the scope and hooked one probe each to the + and - outputs, set it to inv and add, and set the waveform to just touch two lines on the scale. Then I switched channels and they were almost identical in strength. If this is a dumb test, please let me know so I won't keep doing it.

OK on another note, after about an hour and a half, the heatsink on the channel with the 0r22 resistors is at 51c. The other is at 60c.

Oh yeah, I changed the resistors in one channel of the PSU to a pair of 0r22 25w resistors per rail. This is the same channel that got the other changes. I'm reading a drop across those of 0.570VDC at a rail voltage of 14.75VDC. The channel with the 0r15 source resistors still has a pair of 0r1 5w resistors per side and I'm reading 0.356VDC drop across them with a 14.6VDC rail.

I guess that's enough rambling for now. Still looking for suggestions.

Thanks, Terry

[still4given]

Just a small update. I switched peakers and found that the speakers had more of an affect on the sound than the amp. Reversing the speakers revealed that the channel with the 0r22 resistors doesn't sound more muffled. That was coming from the speakers. Probably more placement than anything but still, I wanted to make that clear. The two channels sound almost identical.

I'm beginning to warm up to the sound of this amp. It seems to have it's own charm.

I believe I'll change out the other channel's resistors too. I like the lower heat.

Blessings, Terry

[Adams_Leo]

Originally posted by GRollins
Mark,
Choose the resistor value according to how much bias you want and how much heat you can safely dissipate. Lower resistor values (all things being equal, meaning you don't fiddle the current source, etc.) will, of course, lead to higher bias current. Look at your speaker load impedance and decide what its minimum value is. Work backwards from there, then throw in a bit of extra bias, just for spice.
5 1/2 digits will do fine for matching. What tolerance are the Vishays? 1%? I wouldn't even bother matching them, myself, but then I'm short on time. Someone with more time on their hands might feel that tweaking the R values would be worthwhile.

Grey

Grey,I have an other question,I want to build hight power verions aleph-x with 100W power output,but I can not get the 0.47Ohm resistor for A-X,so I plan use 2pcs 1ohm parallel connection,1ohm//1ohm=0.5ohm to replace the all 0.47Ohm resister in attached schematic.Can it work well and steady?thanks
Leo

[GRollins]

That will do just fine. All that matters is the total resistance and the ability to dissipate the heat. Anything else (wirewound vs. metal film vs. carbon comp. vs. etc.) is just icing on the cake.

Grey

[BobEllis]

Originally posted by still4given
Hi William,


... Oh yeah, I also changed the input caps to 12uf orange drops in bothe channels. That's as close as I could find locally.

... OK on another note, after about an hour and a half, the heatsink on the channel with the 0r22 resistors is at 51c. The other is at 60c.

Oh yeah, I changed the resistors in one channel of the PSU to a pair of 0r22 25w resistors per rail. This is the same channel that got the other changes. I'm reading a drop across those of 0.570VDC at a rail voltage of 14.75VDC. The channel with the 0r15 source resistors still has a pair of 0r1 5w resistors per side and I'm reading 0.356VDC drop across them with a 14.6VDC rail.

I guess that's enough rambling for now. Still looking for suggestions.

Thanks, Terry

The orange drops are OK - did you notice an improvement in the sound when you got rid of the electros?

Increasig the source resistors reduced the bias current, resulting in less heat. Looks like 0R22 is right for those heat sinks.

How is the psu ripple voltage and no signal output trace on the changed channel? Your bias there is about 5.5A, which would show reduced ripple, (measure AC on the rails) and your increased R in the CRC woulod further decrease the ripple.

Are your R1 and R4 still getting hot? Did you souble check their value? I can't tell you how many times I have typed the wrong value into my mouser order (6.8R meaning 68, etc.)

How warm is your transformer with the reduced load?

[still4given]

Originally posted by BobEllis


The orange drops are OK - did you notice an improvement in the sound when you got rid of the electros?

Increasing the source resistors reduced the bias current, resulting in less heat. Looks like 0R22 is right for those heat sinks.

How is the psu ripple voltage and no signal output trace on the changed channel? Your bias there is about 5.5A, which would show reduced ripple, (measure AC on the rails) and your increased R in the CRC woulod further decrease the ripple.

Are your R1 and R4 still getting hot? Did you double check their value? I can't tell you how many times I have typed the wrong value into my mouser order (6.8R meaning 68, etc.)

How warm is your transformer with the reduced load?

Hi Bob,

I really couldn't hear any difference. Perhaps if I did them one channel at a time and compared I could have, but I changed them all at once. My memory is not good enough to tell you if there was much difference.

The AC voltage is reduced with the pair of 0r22 resistors in the power supply. I read only 24mV of AC after the filters. If I lower the scope enough I can see a small ripple on the scope but I don't think it is hurting anything.

R1/R4 are not hot as long as I keep the absolute offset low. This is like chasing your tail though. Everytime I check the absolute offset it is different. I will get it settled down and then check it a half and hour later and it will be back to a volt or two +/-. Even so, at a couple of volts, R1/R4 won't be very warm. When I first start the amp cold the absolute offset might be close to rail voltage. As it warms up it drops. Very weird! Relative offset is more stable. I have about 70mv offset on the channel with the 0r22 source resistors. Probably because I didn't take the time to match them. When I matched the 0r15's originally they were all very, very close. I'm using Dale 1%. I will match the rest before I put them in the other channel. If it looks like I need to, I will pull these and match them. From what I've read, 70mV is OK.

I listened to the amp for a couple of hours last night through my nearfields. It sounds quite nice through them. I don't know if I'm missing anything yet. I will do more comparisons once I get the case all put back together, and get it off of my bench. I'll let it be my listening amp for a while. I still need to finish up my second KSA50. The AX will be what I listen to while I work on that. By the end of that, I should know if it's a keeper or not. It might end up in the master bedroom as a space heater.

Blessings, Terry

[still4given]

Well, I got all of the resistors changed to 0r22. Sounds nice and is much cooler running. I bumped the bias up to .530v across the source resistors. Still plenty cool. I'm going to just listen to it for a while.

Blessings, Terry