gl said:
My personal preference for amplifiers is low gain 20 dB or less. I think
it's very unfortunate that the customers have been convinced that
they somehow want more gain - all they get for their trouble is more
complex circuits and noise.
PS The 30 dB PL figure is an option - we ship 26 dB as a default.
Jfet option in aleph x amp
I have been following this discussion since it first appeared by Graeme et al. What are the CURRENTLY available jfet options which will successfully work in this amplifier? To continue successfully in this hobby we need to take into account the financial resources and parts sources available to all of us.
If at all possible reference to an easily sourced item would be appreciated. The level of achievement on this particular endeavor to date is exceptional. Please continue on. Tad
I have been following this discussion since it first appeared by Graeme et al. What are the CURRENTLY available jfet options which will successfully work in this amplifier? To continue successfully in this hobby we need to take into account the financial resources and parts sources available to all of us.
If at all possible reference to an easily sourced item would be appreciated. The level of achievement on this particular endeavor to date is exceptional. Please continue on. Tad
Here is one of the modules under test.
I have Graeme's feedback and input divider values installed.
I had a quick listen to it tonight. Some details tomorrow when I get the other channel up.
For the numbers people, offset at start up was 7 volts from cold and dropped quickly to a few hundred mv.
Relative Offset is about 30mv.
Ian
I have Graeme's feedback and input divider values installed.
I had a quick listen to it tonight. Some details tomorrow when I get the other channel up.
For the numbers people, offset at start up was 7 volts from cold and dropped quickly to a few hundred mv.
Relative Offset is about 30mv.
Ian
Attachments
I measured the actual voltage gain tonight.
These are the measurements;
Input voltage sine wave 100 hertz
34mv negative to ground
34 mv positive to ground
64 mv differental
Output voltage
640mv to negative to ground
640mv to positive to ground
1280 mv differential
That suggest to me 20 x gain
I am bit tired so perhaps someone can express that at db voltage gain.
Ian
These are the measurements;
Input voltage sine wave 100 hertz
34mv negative to ground
34 mv positive to ground
64 mv differental
Output voltage
640mv to negative to ground
640mv to positive to ground
1280 mv differential
That suggest to me 20 x gain
I am bit tired so perhaps someone can express that at db voltage gain.
Ian
Ian,
Excellent work there! And thank you for the pictures. Pictures are always encouraging. That said, I'll see what I can do about posting one or two here in the near future.
Tad, thank you for the feedback. Sometimes you wonder if anyone is interested in this thread other than the small number of people who post. On the other hand the number of views is getting pretty high and the thread does seem to stay on page 1 a lot of the time. Anyway, it's good to receive an encouraging word.
The 2SJ74 and 2SK170 are still available and probably will be for a few years.
Cheers,
Graeme
Excellent work there! And thank you for the pictures. Pictures are always encouraging. That said, I'll see what I can do about posting one or two here in the near future.
Tad, thank you for the feedback. Sometimes you wonder if anyone is interested in this thread other than the small number of people who post. On the other hand the number of views is getting pretty high and the thread does seem to stay on page 1 a lot of the time. Anyway, it's good to receive an encouraging word.
The 2SJ74 and 2SK170 are still available and probably will be for a few years.
Cheers,
Graeme
Hi Graeme,
I should have both channels ready to do some evaluations with different gain settings tonight.
In comparison to the X250.5 the X Jfet Aleph has noticably less gain. I guess the X250.5 must be set up for 30 db of gain.
It may only seem marginally more but on the volume control settings its a lot more gain and there is more source noise.
My preference is certainly for lower gain with high sensitivity loudspeakers.
Ian
I should have both channels ready to do some evaluations with different gain settings tonight.
In comparison to the X250.5 the X Jfet Aleph has noticably less gain. I guess the X250.5 must be set up for 30 db of gain.
It may only seem marginally more but on the volume control settings its a lot more gain and there is more source noise.
My preference is certainly for lower gain with high sensitivity loudspeakers.
Ian
Graeme
One note. Do you think the actual thickness of the individual heatsinks you used made that much of a difference overall in the efficiency of there cooling ability. I would guess that square foot area to the atmosphere would be the most important thing. I was going to replicate this chassis design with 1/16 inch aluminum plate salvaged from old street signs. After I finish my A-75/300 Pass/Thagard amp your amp will be next.
I was waiting for quite awhile for Nelson to publish any corrections to Grey's original schematic and or the current model schematic. Since Nelson came out with the AX.5 series does not the AX original become old stuff. Just wishing here I know the folks at Pass Labs have to eat and pay bills. They already do far more than all the other designers combined. Thanks all Tad
One note. Do you think the actual thickness of the individual heatsinks you used made that much of a difference overall in the efficiency of there cooling ability. I would guess that square foot area to the atmosphere would be the most important thing. I was going to replicate this chassis design with 1/16 inch aluminum plate salvaged from old street signs. After I finish my A-75/300 Pass/Thagard amp your amp will be next.
I was waiting for quite awhile for Nelson to publish any corrections to Grey's original schematic and or the current model schematic. Since Nelson came out with the AX.5 series does not the AX original become old stuff. Just wishing here I know the folks at Pass Labs have to eat and pay bills. They already do far more than all the other designers combined. Thanks all Tad
Tad,
The material I used is .187 (3/16") thick. As they stand, the sinks I built do the job. They don't rise above 45-50C even in summer. But they're not efficient because they're too thin. The heat doesn't spread out enough so a great deal of the sink plates aren't being utilized maximally. Each plate is hottest right next to the transistor(s) as you would expect. The heat then radiates outward and because of convection the heat pattern tends upward toward the top of the plate. So imagine a parabolic heat pattern around each transistor. The ends of the plate are not hot at all. The reason it works is that the total surface area of the sink is probably double or triple (probably lots more) that of the factory version.
I placed the transistors laterally at the 33% and 66% points on the plate. If I did it again I would place them at the 25% and 75% points.
I would also use .25" plates instead of .187. That would spread the heat better and the plates could be made smaller. As it is the plates I have could probably be reduced an inch in height and two inches in width without causing much of a thermal change overall.
1/16" material is not going to work. I don't like raining on anyone's parade but I wouldn't recommend it.
Nelson talked about publishing an upgrade to the original Aleph-X design but he has been silent on this for quite a while. I don't believe that he will ever publish the factory XA schematics. That said he has said on several occasions that original Aleph-X is very, very close to the factory version. And that was before the refinement process set in. So I think it's safe to say that the Aleph-X still holds a place of serious value-for-effort-expended in the DIY world.
Cheers,
Graeme
P.S. Ian, yes I believe the the 250.5 is in the 30dB group.
The material I used is .187 (3/16") thick. As they stand, the sinks I built do the job. They don't rise above 45-50C even in summer. But they're not efficient because they're too thin. The heat doesn't spread out enough so a great deal of the sink plates aren't being utilized maximally. Each plate is hottest right next to the transistor(s) as you would expect. The heat then radiates outward and because of convection the heat pattern tends upward toward the top of the plate. So imagine a parabolic heat pattern around each transistor. The ends of the plate are not hot at all. The reason it works is that the total surface area of the sink is probably double or triple (probably lots more) that of the factory version.
I placed the transistors laterally at the 33% and 66% points on the plate. If I did it again I would place them at the 25% and 75% points.
I would also use .25" plates instead of .187. That would spread the heat better and the plates could be made smaller. As it is the plates I have could probably be reduced an inch in height and two inches in width without causing much of a thermal change overall.
1/16" material is not going to work. I don't like raining on anyone's parade but I wouldn't recommend it.
Nelson talked about publishing an upgrade to the original Aleph-X design but he has been silent on this for quite a while. I don't believe that he will ever publish the factory XA schematics. That said he has said on several occasions that original Aleph-X is very, very close to the factory version. And that was before the refinement process set in. So I think it's safe to say that the Aleph-X still holds a place of serious value-for-effort-expended in the DIY world.
Cheers,
Graeme
P.S. Ian, yes I believe the the 250.5 is in the 30dB group.
Erno Borbely Audio is always quoted as a source. You can Google for that. Then there's mcmelectronics.com. They haven't had the 74's for a while. But here you go for 170's:
http://www.mcmelectronics.com/product.aspx?productid=2SK170
And there are other places. Search this site or Google the net. Don't give up. People on this site offer these parts for sale or trade once in a while too.
GL
http://www.mcmelectronics.com/product.aspx?productid=2SK170
And there are other places. Search this site or Google the net. Don't give up. People on this site offer these parts for sale or trade once in a while too.
GL
1sk170
I just checked and ampslab sells the jfets in limited quantities. Just google ampslab.com and go to components and jfets.
Graeme, So it seems that as usual thick heavy aluminum is the material of choice for the sinks. Mo money!~!! I like this hobby it gives me direction for my working efforts. I guess I could take up golf or building motorcycles. There is always something to buy. And I do love spending money. Just can not seem to find an unlimited source of it. Maybe melting a few hundred soda cans and molding some diy heatsinks would be a worthwhile effort. Again thanks Tad
P.S. Where is the next great amp project going to come from?
I just checked and ampslab sells the jfets in limited quantities. Just google ampslab.com and go to components and jfets.
Graeme, So it seems that as usual thick heavy aluminum is the material of choice for the sinks. Mo money!~!! I like this hobby it gives me direction for my working efforts. I guess I could take up golf or building motorcycles. There is always something to buy. And I do love spending money. Just can not seem to find an unlimited source of it. Maybe melting a few hundred soda cans and molding some diy heatsinks would be a worthwhile effort. Again thanks Tad
P.S. Where is the next great amp project going to come from?
I managed to get both channels running tonight with the Jfet front end as per Graemes modifications.
There are two differences: feedback bypass caps are 4.7 pf and I run 4 x IRFP240 and 0.75R sources resisters.
Subjectively its always are hard call but I would describe the result was silky smooth, perhaps a tad plump in the bass.
I certainly prefer the 26 db gain but would like to lower the gain to 20 db.
Perhaps Nelson can suggets the best approach?
Ian
There are two differences: feedback bypass caps are 4.7 pf and I run 4 x IRFP240 and 0.75R sources resisters.
Subjectively its always are hard call but I would describe the result was silky smooth, perhaps a tad plump in the bass.
I certainly prefer the 26 db gain but would like to lower the gain to 20 db.
Perhaps Nelson can suggets the best approach?
Ian
I have tried halving the value of the feedback resisters from 220k to 110K as this was quite simple to implement .
Input voltage sine wave 100 hertz
I measured the gain at 9.6 times
Input voltage differentail 140mv
Output voltage differential 1345mv
Using the above equation gives 19.64 db gain.
So what happened subjectively?
It made a very noticable difference.
I would describe the result as less powdery (as in soft), with far more prominant extension of bandwidth and clarity and the bass was truer, more firm.
I dont think there is a right or a wrong answer to any of this as it also depends on one's own system and environment.
Perhap spmewhere in between the two gain settings would be interesting.
Ian
Input voltage sine wave 100 hertz
I measured the gain at 9.6 times
Input voltage differentail 140mv
Output voltage differential 1345mv
Using the above equation gives 19.64 db gain.
So what happened subjectively?
It made a very noticable difference.
I would describe the result as less powdery (as in soft), with far more prominant extension of bandwidth and clarity and the bass was truer, more firm.
I dont think there is a right or a wrong answer to any of this as it also depends on one's own system and environment.
Perhap spmewhere in between the two gain settings would be interesting.
Ian
Attachments
Tad,
Yes, thick and heavy is the rule for heatsinks - particularly for class A applications. The trick is to get the heat away from the transistors as fast as possible. That means thick hear the transistors at least, lots of radiating area, dark color, as few thermal interfaces as possible in the heat path, etc, etc, ... . The bulk of the cost for Al heatsinks is the cost of the raw material. There's just no getting around it. If you find surplus heatsinks or raw Al you'll be looking at similar numbers. It's like a law of physics. And yes, fun hobbies also seem to be the expensive ones. Another law of physics maybe.
As for the next great amp project - we'll all just have to see what comes along. My preference is for something that has 100W to 200W per channel. I need that as a minimum for my speakers. The project would also have to have demonstrably better sound than the AX100/AX100J project. I watch the Firstwatt developments with interest. It is tempting to try scaling one of them up to higher power output. For now it's fun just listening to what I have. But I know that sooner or later I will be smitten by "Power-amp AcquiSition Syndrome" or PASS.
Graeme
Yes, thick and heavy is the rule for heatsinks - particularly for class A applications. The trick is to get the heat away from the transistors as fast as possible. That means thick hear the transistors at least, lots of radiating area, dark color, as few thermal interfaces as possible in the heat path, etc, etc, ... . The bulk of the cost for Al heatsinks is the cost of the raw material. There's just no getting around it. If you find surplus heatsinks or raw Al you'll be looking at similar numbers. It's like a law of physics. And yes, fun hobbies also seem to be the expensive ones. Another law of physics maybe.
As for the next great amp project - we'll all just have to see what comes along. My preference is for something that has 100W to 200W per channel. I need that as a minimum for my speakers. The project would also have to have demonstrably better sound than the AX100/AX100J project. I watch the Firstwatt developments with interest. It is tempting to try scaling one of them up to higher power output. For now it's fun just listening to what I have. But I know that sooner or later I will be smitten by "Power-amp AcquiSition Syndrome" or PASS.
Graeme
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.