Hi Ian and Dave,
Both amplifiers are now converted to JFET inputs. The second amp shows 55mv of relative offset that remains unchanged from turn-on to full warm-up. One amp has 5 hrs of run-in time on it and the other has 2.5 so both are still burning in and the sound is continually changing. They need a week or two before I can really determine the extent of what I've done. For the moment though I will say that I agree with Nelson - they are definitely better with parallel JFET inputs. I think it's time to take a picture or two and start another thread.
Dave I'm glad you solved your transformer noise issue.
So Jeremy - how much are you asking for your Lono's?
Cheers,
Graeme
Both amplifiers are now converted to JFET inputs. The second amp shows 55mv of relative offset that remains unchanged from turn-on to full warm-up. One amp has 5 hrs of run-in time on it and the other has 2.5 so both are still burning in and the sound is continually changing. They need a week or two before I can really determine the extent of what I've done. For the moment though I will say that I agree with Nelson - they are definitely better with parallel JFET inputs. I think it's time to take a picture or two and start another thread.
Dave I'm glad you solved your transformer noise issue.
So Jeremy - how much are you asking for your Lono's?
Cheers,
Graeme
Graeme,
Send me email. It was kind of an off-the-cuff remark. Before I would decide to get rid of them, I'll have to try dual standard 1kVA transformers and be happy with the result. That could take some time. And then there's that SOZ project if I have extra transformers lying around.
Jeremy
Send me email. It was kind of an off-the-cuff remark. Before I would decide to get rid of them, I'll have to try dual standard 1kVA transformers and be happy with the result. That could take some time. And then there's that SOZ project if I have extra transformers lying around.
Jeremy
aleph 1.2 to xa100
GL and Kropf, thanks for the replies, I guess we wait and see how GL's front end turns out. I will have to get some Jfets from somewhere to experiment.
GL -did you use the source resistors on the diff pair to get the low relative dc offset or did you find matched pairs?
I reread some of Erno Borbely's articles and looked at his schematics for his class A amps and he does the same thing by using cascodes on the diff pair.
I am all ears about how your project turns out.
Kropf, my amps are over 100lbs each and I had to put them on dollies to move them around. I can understand what you are talking about. dave
GL and Kropf, thanks for the replies, I guess we wait and see how GL's front end turns out. I will have to get some Jfets from somewhere to experiment.
GL -did you use the source resistors on the diff pair to get the low relative dc offset or did you find matched pairs?
I reread some of Erno Borbely's articles and looked at his schematics for his class A amps and he does the same thing by using cascodes on the diff pair.
I am all ears about how your project turns out.
Kropf, my amps are over 100lbs each and I had to put them on dollies to move them around. I can understand what you are talking about. dave
Dave,
the J109 is a matched pair although there a tolerance of 3% on it. I used 2 J109's per amp but did not match those to each other. I am going to try and match my remaining parts and substitute matched ones if possible. We'll see how that goes. It may take me months to get there. The Source resistors don't seem to affect the relative offset. One amp has the Source resistors but they are temporarily shorted out. The other amp doesn't have them. Their presence doesn't seem to affect the sound but I didn't do any exhaustive comparison. I've left them in the one amp so that I can do some additional measurements.
I re-read wuffwaff's Aleph-JX thread this weekend and realized I had forgotten how much excellent work he and Patrick(EUVL) had done modifying their amps to have JFET front ends. If I had re-read that thread five months ago I would have gotten discouraged and not done my upgrade. I have not had any of the problems they did. I haven't done any waveform measurements or analysis but I can't hear any distortion at all even at high levels. I am using Magnepan 3.6's which are 4 ohm loads and which suck current like hogs. In fact the AX100J would probably be considered too anemic to drive them if you posed the question to the folks over on the planar forum at audioasylum. So it's not like the amps are being treated gently. In addition I found that the bass performance has improved with AX100J. It's cleaner, more solid and more articulate. I don't know why. This is in contrast to what wuffwaff reported. Maybe he found and fixed that problem. It would be nice to know.
Graeme
the J109 is a matched pair although there a tolerance of 3% on it. I used 2 J109's per amp but did not match those to each other. I am going to try and match my remaining parts and substitute matched ones if possible. We'll see how that goes. It may take me months to get there. The Source resistors don't seem to affect the relative offset. One amp has the Source resistors but they are temporarily shorted out. The other amp doesn't have them. Their presence doesn't seem to affect the sound but I didn't do any exhaustive comparison. I've left them in the one amp so that I can do some additional measurements.
I re-read wuffwaff's Aleph-JX thread this weekend and realized I had forgotten how much excellent work he and Patrick(EUVL) had done modifying their amps to have JFET front ends. If I had re-read that thread five months ago I would have gotten discouraged and not done my upgrade. I have not had any of the problems they did. I haven't done any waveform measurements or analysis but I can't hear any distortion at all even at high levels. I am using Magnepan 3.6's which are 4 ohm loads and which suck current like hogs. In fact the AX100J would probably be considered too anemic to drive them if you posed the question to the folks over on the planar forum at audioasylum. So it's not like the amps are being treated gently. In addition I found that the bass performance has improved with AX100J. It's cleaner, more solid and more articulate. I don't know why. This is in contrast to what wuffwaff reported. Maybe he found and fixed that problem. It would be nice to know.
Graeme
Hi,
somewhere in the thread you mentioned above the bass problem was solved by changing the value of the McMillan resistors. I changed them from 10k to 20k and then even higher.
This just compensates for the much lower bias and the bigger drain resistors.
I can´t hear any distortion, I can only see it on my scope at high currents and high frequencies. The sound as such is just fine and better than it was with the 9610´s.
When I´ve got a bit of time I will try two things:
1. Put zeners (5,7V) parallel to the bootstrap caps in the active current sources
2. play with some NTC´s in the current source for the diff pair to change the absolute offset behaviour
3. maybe try paralleling the Sj74´s
But time is a thing that I don´have too much at the moment.
Nice greetings from a not so very cold Sweden,
William
William
somewhere in the thread you mentioned above the bass problem was solved by changing the value of the McMillan resistors. I changed them from 10k to 20k and then even higher.
This just compensates for the much lower bias and the bigger drain resistors.
I can´t hear any distortion, I can only see it on my scope at high currents and high frequencies. The sound as such is just fine and better than it was with the 9610´s.
When I´ve got a bit of time I will try two things:
1. Put zeners (5,7V) parallel to the bootstrap caps in the active current sources
2. play with some NTC´s in the current source for the diff pair to change the absolute offset behaviour
3. maybe try paralleling the Sj74´s
But time is a thing that I don´have too much at the moment.
Nice greetings from a not so very cold Sweden,
William
William
Hi William,
Thank you for the update. It answers my my questions. Am I correct in thinking that you had the same distortion issue with the MOSFET front end?
Adding the zeners to the bootstrap caps is on my list too. I would also like to replace the resistor feeding the CCS zener with a constant current diode. On my amps I would also replace the resistor feeding the cascode reference zener. I have a bag of 3.3ma units so this would be quick and easy.
Graeme
Thank you for the update. It answers my my questions. Am I correct in thinking that you had the same distortion issue with the MOSFET front end?
Adding the zeners to the bootstrap caps is on my list too. I would also like to replace the resistor feeding the CCS zener with a constant current diode. On my amps I would also replace the resistor feeding the cascode reference zener. I have a bag of 3.3ma units so this would be quick and easy.
Graeme
wuffwaff said:
I couldn't resist this opening.
This simple circuit compensates for all manner of temperature drift in the Aleph-X and it is easy to plug into the HiFiZen board. The control voltage is at the collector of transistor Q1. R1 parallel to the NTC thermistor, R2, linearizes the response of R2.
Jeremy
Attachments
How does it work? Pretty well. For the best results, the thermistor should be in good thermal contact with the heatsinks.
Some tweaking is certainly necessary to temperature compensate your particular amplifier. Adjusting the CCS current via R6 or the linearization resistor R1 will adjust the range and magnitude of the response. The usual differential pair CCS adjustments can then be used to adjust the absolute offset after the amplifier is fully warmed up.
You may have some motorboating problems during a soft start, so be careful and don't blow your speakers.
Jeremy
Some tweaking is certainly necessary to temperature compensate your particular amplifier. Adjusting the CCS current via R6 or the linearization resistor R1 will adjust the range and magnitude of the response. The usual differential pair CCS adjustments can then be used to adjust the absolute offset after the amplifier is fully warmed up.
You may have some motorboating problems during a soft start, so be careful and don't blow your speakers.
Jeremy
Attachments
Hi Ian,
you have to measure Idss and Vp. Then you can plot the Id-Vgs diagrams and pick the ones that are closest at the bias (5mA) you want to use.
I got within ,2mA but still had more offset than with my closely matched IRF9610´s
Another possibility would be to match them at 5mA and measure Vgs
William
you have to measure Idss and Vp. Then you can plot the Id-Vgs diagrams and pick the ones that are closest at the bias (5mA) you want to use.
I got within ,2mA but still had more offset than with my closely matched IRF9610´s
Another possibility would be to match them at 5mA and measure Vgs
William
Hi Ian,
That's good news regarding the JFETs. I agree with William on the matching. I would go with matching Vgs at the Id you are planning to use. Beyond that it's a matter of trying different pairs in the circuit until you find a pair you like best. But that's a lot of trouble.
The ZTX550's aren't all that critical. I would say use anything with an hfe of 100 or more, Vce of 60V or more, and Pd of 600mW or more - although my preference on this last would be 1000mW.
The amps continue to perform really well. The one big improvement I note is their ability to unravel complex musical textures much better than before.
Cheers,
Graeme
That's good news regarding the JFETs. I agree with William on the matching. I would go with matching Vgs at the Id you are planning to use. Beyond that it's a matter of trying different pairs in the circuit until you find a pair you like best. But that's a lot of trouble.
The ZTX550's aren't all that critical. I would say use anything with an hfe of 100 or more, Vce of 60V or more, and Pd of 600mW or more - although my preference on this last would be 1000mW.
The amps continue to perform really well. The one big improvement I note is their ability to unravel complex musical textures much better than before.
Cheers,
Graeme
aleph 1.2 to xa100
GL, what is your listening expeirence been with your modified amps in the low bass region. I am using mine as treble/mid range amps where they sound excellent. I only used them for a short time full range on some Watkins WE-1's when testing. These speakers have three 8" woofers. One of the three acts as a subwoofer. These are acoustic suspension types and need a lot of current to set off the bass. The XA100 reminded me more of the Aleph's in pushing bass drivers. The 350.5 and my A75 were much better in the low bass region.
The reason I ask is that you use yours full range. thanks, dave
GL, what is your listening expeirence been with your modified amps in the low bass region. I am using mine as treble/mid range amps where they sound excellent. I only used them for a short time full range on some Watkins WE-1's when testing. These speakers have three 8" woofers. One of the three acts as a subwoofer. These are acoustic suspension types and need a lot of current to set off the bass. The XA100 reminded me more of the Aleph's in pushing bass drivers. The 350.5 and my A75 were much better in the low bass region.
The reason I ask is that you use yours full range. thanks, dave
Hi Dave,
My listening impressions to date are:
1) The musical lines are definitely cleaner. The micro-detail level has increased so that I can hear more of the low level stuff that's going on. The word "unravel" comes to mind regarding the musical textures.
2) It's as if the light level on the sound stage has been turned up a notch. There are fewer shadows. I don't believe that this has anything to do with tipping the tonal balance toward treble or bass or any of that old yin and yang stuff people talk about with the Alephs. This effect has improved my listening enjoyment a lot.
3) Bass is cleaner, more solid and more articulate. This improved still more going from 10K to 22K with the MacMillan resistors.
OK. Enough of the Stereophile style narrative. All of these effects are modest. So if I was you I wouldn't get obsessed about upgrading. I have kept my old MOSFET boards and I may switch back in a few months to see what I think. I have never heard an Aleph so I can't comment on the comparison of the bass performance.
At this point I've moved on to creating an active crossover for the MG3.6's. I've come up with two versions of a UGS type circuit to provide gain and balanced output. One of these is now in the system as a temporary line amp, for listening tests and tune up.
Cheers,
Graeme
My listening impressions to date are:
1) The musical lines are definitely cleaner. The micro-detail level has increased so that I can hear more of the low level stuff that's going on. The word "unravel" comes to mind regarding the musical textures.
2) It's as if the light level on the sound stage has been turned up a notch. There are fewer shadows. I don't believe that this has anything to do with tipping the tonal balance toward treble or bass or any of that old yin and yang stuff people talk about with the Alephs. This effect has improved my listening enjoyment a lot.
3) Bass is cleaner, more solid and more articulate. This improved still more going from 10K to 22K with the MacMillan resistors.
OK. Enough of the Stereophile style narrative. All of these effects are modest. So if I was you I wouldn't get obsessed about upgrading. I have kept my old MOSFET boards and I may switch back in a few months to see what I think. I have never heard an Aleph so I can't comment on the comparison of the bass performance.
At this point I've moved on to creating an active crossover for the MG3.6's. I've come up with two versions of a UGS type circuit to provide gain and balanced output. One of these is now in the system as a temporary line amp, for listening tests and tune up.
Cheers,
Graeme
I will start a new thread when I've gotten a bit farther along. On page 1 of the "UGS adventures" thread there is link to a preliminary schematic of one of the versions. It contains several mistakes. As you can see, this project dates back a long way and is a graphic demonstration of how long I take to get things done.
I have made different design decisions from some of those in the UGS Adventures thread.
The project is essentially a line amp and active crossover combined into one unit.
Graeme
I have made different design decisions from some of those in the UGS Adventures thread.
The project is essentially a line amp and active crossover combined into one unit.
Graeme
Hi Graeme,
I just want to confirm your current setup in terms of McMillian resisters, the input to ground resisters and any other tweaks so i can compare notes. Can you also please confirn the zenor orientation.
On the input to ground resisters my understanding is 22K reduces the closed loop gain or increases the feeback. Perhaps this is why the offset is lower? I just wonder what the closed loop gain is compared to 10Kinput to ground resisters? 6db.
I have made a small daughter vero board for the dual matched quad 2SJ74's!
imac
I just want to confirm your current setup in terms of McMillian resisters, the input to ground resisters and any other tweaks so i can compare notes. Can you also please confirn the zenor orientation.
On the input to ground resisters my understanding is 22K reduces the closed loop gain or increases the feeback. Perhaps this is why the offset is lower? I just wonder what the closed loop gain is compared to 10Kinput to ground resisters? 6db.
I have made a small daughter vero board for the dual matched quad 2SJ74's!
imac
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.