Hello everybody,I am just about to finish populating my front board for the AlephX and I wonder which is best value for the resistors to ground (ie R19 and R29 in Grey's schematic) IF I DRIVE THE AMP BALANCED???
Peter,did Nelson made an explanation why 10K would be optimum???
I have surplus of 9610's.I found a pair matched within 8 mV worst case.Do you think it will be OK for DC offset at the output?? (ie differentially 😉
Sorry for the 20 questions.Maybe I should use the 100 K resistors before the capacitors...I opted for 4.7uF value ( full range use I hope it is enough.....)
Best regards to you all
Peter,did Nelson made an explanation why 10K would be optimum???
I have surplus of 9610's.I found a pair matched within 8 mV worst case.Do you think it will be OK for DC offset at the output?? (ie differentially 😉
Sorry for the 20 questions.Maybe I should use the 100 K resistors before the capacitors...I opted for 4.7uF value ( full range use I hope it is enough.....)
Best regards to you all
Nelson said that 68k value for an input resistor seems a bit high and he prefers lower values, like 10K for best performance.
The input Mosfet capacitance cause degradation as the impedances climb.
The production Aleph X is also using 10k input to ground resistors (like the rest of the Alephs actually).
4.7u is more than enough. I recommend to anybody to try some ground resistors before the caps and see the difference. Maybe you'll like it, maybe not.😉
As to 8mV matching on fets, I don't know. Put them in the circuit and see. I can easily adjust for 1mV DC offset on my amp and it stays that way. The same with absolute offset: 40mV is not a problem.
The way I matched my input fets was in a circuit, because although they seemed to be match perfectly outside the circuit, when I placed them in, the voltage drop on drain resistors (392ohm) was different. So I did it in a circuit for the same voltage drop.
The balance betweent the sides for 0V between speaker outputs is adjusted by the trimmer which varies the current through output devices (adjusts the drop on source resistors). The absolute offset is adjusted by the trimmer which sets the current through the differencial pair (I wonder why, BTW). Bith values can be adjusted very easily. Absolut drop changes in the beginning from few volts to 0V, so the best is to check it after few hours of operation. My output to ground resistors are 100ohms and sources to output resistors are 4k7ohm. I suspect that circuit might be stable without 100ohm resistors, I didn't try it yet.😉
The input Mosfet capacitance cause degradation as the impedances climb.
The production Aleph X is also using 10k input to ground resistors (like the rest of the Alephs actually).
4.7u is more than enough. I recommend to anybody to try some ground resistors before the caps and see the difference. Maybe you'll like it, maybe not.😉
As to 8mV matching on fets, I don't know. Put them in the circuit and see. I can easily adjust for 1mV DC offset on my amp and it stays that way. The same with absolute offset: 40mV is not a problem.
The way I matched my input fets was in a circuit, because although they seemed to be match perfectly outside the circuit, when I placed them in, the voltage drop on drain resistors (392ohm) was different. So I did it in a circuit for the same voltage drop.
The balance betweent the sides for 0V between speaker outputs is adjusted by the trimmer which varies the current through output devices (adjusts the drop on source resistors). The absolute offset is adjusted by the trimmer which sets the current through the differencial pair (I wonder why, BTW). Bith values can be adjusted very easily. Absolut drop changes in the beginning from few volts to 0V, so the best is to check it after few hours of operation. My output to ground resistors are 100ohms and sources to output resistors are 4k7ohm. I suspect that circuit might be stable without 100ohm resistors, I didn't try it yet.😉
Ok,so I will go for 10K then. 😉
I planned to use 2k2 as resistors from output to sources.Might give a better control over absolute DC,as my prototype will use 22V rails with 8 IRFP240,4 for outputs and 4 for current sources.I chose 41 ohms as resistors from outputs to ground.I think this way the absolute DC will be quite stable.
Later on if all works I will probably try the 4k7 value from sources to output.Nelson seems to consider it better way to keep good diff.noise rejection,as lowering the value might give some problems(noise+distortion).
I intend to get more output fets +sinks to go from 4,25 A bias to 6,38A.Just to listen if the amp controls load better (but my speakers are mostly high Z,8 ohms or more.
I thought about giving the input resistor a value of 10k,and 50 k for the feedback resistor.my target is a 20dB (= X10) gain,as all other Alephs.26 dB is too much for me 😉
I planned to use 2k2 as resistors from output to sources.Might give a better control over absolute DC,as my prototype will use 22V rails with 8 IRFP240,4 for outputs and 4 for current sources.I chose 41 ohms as resistors from outputs to ground.I think this way the absolute DC will be quite stable.
Later on if all works I will probably try the 4k7 value from sources to output.Nelson seems to consider it better way to keep good diff.noise rejection,as lowering the value might give some problems(noise+distortion).
I intend to get more output fets +sinks to go from 4,25 A bias to 6,38A.Just to listen if the amp controls load better (but my speakers are mostly high Z,8 ohms or more.
I thought about giving the input resistor a value of 10k,and 50 k for the feedback resistor.my target is a 20dB (= X10) gain,as all other Alephs.26 dB is too much for me 😉
nar said:
You don't even have to go this low. My components' values are doubled for those locations and the DC offset is perfectly stable, I mean it doesn't change at all, even at start up.😉
I'm using 16 output devices per channel.
Get ready to laugh at me.
What is the absolute DC...???
Is it DC between one output and the ground?
Is it DC between two outputs?
Is it Vgs difference between two MOSFETs?
JH
What is the absolute DC...???
Is it DC between one output and the ground?
Is it DC between two outputs?
Is it Vgs difference between two MOSFETs?
JH
When people say "absolute DC offset" in relation to this amplifier, they mean the common potential between the outputs and ground.
Peter Daniel,
Sometime ago Rtirion mentionned in simulations that the 4k7 would give better results when using more (>=16 output fets),but in my case using 8 I will probably have to try values between 2k2 and 4k7.
Idem for the 100 ohms from outputs to ground.In your case might yeld better results,but still I have 8 output fets I shall try the 41 ohms first. 😉
Do you know a way to calculate output power with +/- 22V rails,gain 20dB balanced and total bias 4,25A?and 6,38A?I'm just curious 🙂
I planned to use 750R value for R 12/ R 34.This value is out from Aleph 3.I am very curious to see if it could help sonics thus we are bridged from the speaker view-point,as each half logically sees half the speaker impedance...mmmhhh.Did you make experiments this way?
Best regards
Sometime ago Rtirion mentionned in simulations that the 4k7 would give better results when using more (>=16 output fets),but in my case using 8 I will probably have to try values between 2k2 and 4k7.
Idem for the 100 ohms from outputs to ground.In your case might yeld better results,but still I have 8 output fets I shall try the 41 ohms first. 😉
Do you know a way to calculate output power with +/- 22V rails,gain 20dB balanced and total bias 4,25A?and 6,38A?I'm just curious 🙂
I planned to use 750R value for R 12/ R 34.This value is out from Aleph 3.I am very curious to see if it could help sonics thus we are bridged from the speaker view-point,as each half logically sees half the speaker impedance...mmmhhh.Did you make experiments this way?
Best regards
Peter, I admit I haven't looked back through the 93 pages, but have you somewhere listed a record of what parts values you've actually used? This would be handy as a baseline at this point.Peter Daniel said:
Thanks a bunch for the construction and listening reports. I think I'm going to have to build one of these.
Thanks,
Paul
Most of the parts values that I used are taken from this schematic: http://www.diyaudio.com/forums/attachment.php?s=&postid=43959
The only difference is using different resistors at the input (33K) source resistor are different for me as well (0.56ohm, although bigger value would be better, my current voltage drop is 0.4V) I used Fred's CCS and schematic is available in my "Not jus another Alpeh" thread. 100ohms at the output and 4k7 source to output stabilising resistors. Feedback cap should be 5pF and I didn't need the other comp caps.
The only difference is using different resistors at the input (33K) source resistor are different for me as well (0.56ohm, although bigger value would be better, my current voltage drop is 0.4V) I used Fred's CCS and schematic is available in my "Not jus another Alpeh" thread. 100ohms at the output and 4k7 source to output stabilising resistors. Feedback cap should be 5pF and I didn't need the other comp caps.
Yes, i used 300ohms, because I thought that higher current might improve the sonics somehow.😉
Nar,
with 22V you can swing aprox 40 Volts (22-2)x2. (=40/2^0.5=28V)
with 4,25 A total bias you have a peak current of 4,25A /3,0A eff.
with 6,38A total bias you have a peak current of 6,38A /4,51A
Power into 8 Ohms would be V^2/R=800/8= 100 watts I= 3,55A eff
For 100 watts 4 Ohm you´ll need 7,1A eff = 10A peak.
This means that with 22V supply the 4,25A bias isn´t enough to get 100 Watt into 8 ohms but only 72. For 4 and 2 ohms 36 and 18watts.
With 6,38A bias you´ll have 100 watts into 8 ohms and 81 watts into 2 ohms.
These figures are true for 50% ac current gain but can be raised easy when you change this to 60-65%.
william
with 22V you can swing aprox 40 Volts (22-2)x2. (=40/2^0.5=28V)
with 4,25 A total bias you have a peak current of 4,25A /3,0A eff.
with 6,38A total bias you have a peak current of 6,38A /4,51A
Power into 8 Ohms would be V^2/R=800/8= 100 watts I= 3,55A eff
For 100 watts 4 Ohm you´ll need 7,1A eff = 10A peak.
This means that with 22V supply the 4,25A bias isn´t enough to get 100 Watt into 8 ohms but only 72. For 4 and 2 ohms 36 and 18watts.
With 6,38A bias you´ll have 100 watts into 8 ohms and 81 watts into 2 ohms.
These figures are true for 50% ac current gain but can be raised easy when you change this to 60-65%.
william
That is true (providing your calculations are correct😉 )
I was initially planning for 20V rails and 7A total current. It came out, that under the load my rails dropped to 15V and the total current is about 6A. Should my rails be higher, they wouldn't match the available current and something would go to waste. So I'm actually glad my rails are only 15V. The heatsinks I'm using, although large, cannot handle more power dissipation at 55 deg C(2 heat sinks per channel size 11"x11"x2", 23 fins ea.) They were perfect size for A75 mono.
I was initially planning for 20V rails and 7A total current. It came out, that under the load my rails dropped to 15V and the total current is about 6A. Should my rails be higher, they wouldn't match the available current and something would go to waste. So I'm actually glad my rails are only 15V. The heatsinks I'm using, although large, cannot handle more power dissipation at 55 deg C(2 heat sinks per channel size 11"x11"x2", 23 fins ea.) They were perfect size for A75 mono.
Peter Daniel said:
Fans, brother. <a href="http://www.siliconacoustics.com/pabst84122ngl.html">Fans.</a>
Just a sidebar on fan noise. Most computer style fans are designed to run a bit fast on the tip speed and consequently are also a bit noisy. Throttle them back just 10% and you'll be amazed at the difference in sound. Just put a bit of resistance in line.
Sorry dude fans are wimpy measures for wimpy diyers,
if you can't take the heat stay out of the kitchen! 😉
Audio manifacturers stay as far away as possible from moving parts because
-they are noisy
-unreliable
-they give a sense of sloppiness and poor built quality to the whole piece
Nothing like a hunk of metal brushed and anodized to give substance, and justify a 10k price sticker. I am all for emulating that in my diy projects.
We have maybe 250 fans in my lab working 24/7, they slowly become noisier and noisier, and they breakdown continuously. We have tech staff that go around and change a couple of fans on any given month.
if you can't take the heat stay out of the kitchen! 😉
Audio manifacturers stay as far away as possible from moving parts because
-they are noisy
-unreliable
-they give a sense of sloppiness and poor built quality to the whole piece
Nothing like a hunk of metal brushed and anodized to give substance, and justify a 10k price sticker. I am all for emulating that in my diy projects.
We have maybe 250 fans in my lab working 24/7, they slowly become noisier and noisier, and they breakdown continuously. We have tech staff that go around and change a couple of fans on any given month.
fans in an audio amp
Once I made some tests with one of the lowest noise Papst fans (60x60mm) that are avaible. I ran it without mounting in free air on a silicon pad to avoid resonance . Mounting fans in whatever type of housing always make resonance things worse compared to free air (not to mention to mount them in way that they do not rattle).
Even this way (without mounting in an enclosure) I was able to hear the fan in a quiet room in 3m distance clearly. Decreasing the fans speed helped much about noise but it was not possible to lower it under a level that I could not have recognised it, without making the fan almost useless regarding air-flow.
I`d NOT recommend using any kind of fans in an audio amp.
Once I made some tests with one of the lowest noise Papst fans (60x60mm) that are avaible. I ran it without mounting in free air on a silicon pad to avoid resonance . Mounting fans in whatever type of housing always make resonance things worse compared to free air (not to mention to mount them in way that they do not rattle).
Even this way (without mounting in an enclosure) I was able to hear the fan in a quiet room in 3m distance clearly. Decreasing the fans speed helped much about noise but it was not possible to lower it under a level that I could not have recognised it, without making the fan almost useless regarding air-flow.
Does it mean you never played music?😉 With the music I'm listening too, fans don't bother me.😉