Thanks Terry, VHF man,
I've been down this track too and can verify that with the addition of a diode to the CM the treble is indeed slightly softer, and in fact decay is lengthened and improved. There may be a slight improvement in resolution, too.
If the diode (4148) was lapsing into reverse-bias, it was not evident in listening tests, but of course this may not be audible.
I've been down this track too and can verify that with the addition of a diode to the CM the treble is indeed slightly softer, and in fact decay is lengthened and improved. There may be a slight improvement in resolution, too.
If the diode (4148) was lapsing into reverse-bias, it was not evident in listening tests, but of course this may not be audible.
Hugh,
The limitations imposed by the diode would probably never be threatened in the real world. I for one would be more than happy to leave my amplifier configured that way even if I could see potential simulation problems when driven to near maximum ratings. However, the 3 transistor approach wins out sonically (albeit by a very small margin) as well as in the simulations so it's difficult to argue against.
BTW, it goes without saying that low cob and high speed + high beta devices are needed all round for best results. I discovered that years ago when I built my first DIY amps. As stated in my previous posts, my latest DIY amp sounds so far superior to a $6000 commercial Meridian amp it's almost an embarrassment. The Meridian is excellent though at mid and low frequencies but the big differences really exist above 1KHz. You know you're on the right track when the treble is silky smooth and sweet and free of grain and hardness, and devoid that metallic quality that is the signature of many SS amps. Obviously there are many commercial designers that still have much to learn.
Cheers
VM
The limitations imposed by the diode would probably never be threatened in the real world. I for one would be more than happy to leave my amplifier configured that way even if I could see potential simulation problems when driven to near maximum ratings. However, the 3 transistor approach wins out sonically (albeit by a very small margin) as well as in the simulations so it's difficult to argue against.
BTW, it goes without saying that low cob and high speed + high beta devices are needed all round for best results. I discovered that years ago when I built my first DIY amps. As stated in my previous posts, my latest DIY amp sounds so far superior to a $6000 commercial Meridian amp it's almost an embarrassment. The Meridian is excellent though at mid and low frequencies but the big differences really exist above 1KHz. You know you're on the right track when the treble is silky smooth and sweet and free of grain and hardness, and devoid that metallic quality that is the signature of many SS amps. Obviously there are many commercial designers that still have much to learn.
Cheers
VM
VM,
The LTP works at very low output levels in a working gnfb amp anyway, so while the diode bias current is low, around 12uA as Glen noted, it will not vary much since the CCS feeding the LTP stage will add constancy; it may deviate plus or minus a couple of uA, no more.
Thank you for your post, and yes, it's worth doing for the sonic improvement.
Some of the commercial amps are surprisingly lazily designed. I don't see too many of them, but their sonic faults are always evident on a good listen. This comes back to topology, implementation and component choice, and it's hobbled by the economic realities where retail is around 4-5 times factory manufacturing cost.
Hugh
The LTP works at very low output levels in a working gnfb amp anyway, so while the diode bias current is low, around 12uA as Glen noted, it will not vary much since the CCS feeding the LTP stage will add constancy; it may deviate plus or minus a couple of uA, no more.
Thank you for your post, and yes, it's worth doing for the sonic improvement.
Some of the commercial amps are surprisingly lazily designed. I don't see too many of them, but their sonic faults are always evident on a good listen. This comes back to topology, implementation and component choice, and it's hobbled by the economic realities where retail is around 4-5 times factory manufacturing cost.
Hugh
Hi , Dean.. VHF. I see Austrialians really take their amps seriously.
As I am only a day away from actually trying out these "tricks"
on a proven design ,I'm glad to have seen further disscussion/
simulation brought forth.
As a question to the awesome knowledge before me
(I'm not being sarcastic) , how do you select a good
LPT pair (low Cob, enough Vce I know) but all those HFE
groupings (3 or 4 for some trannies)?? Do you hand match
for the CM and LTP (I ordered 30 each)as well?
By the way this is front end of the design I'm using as
the "guinea pig"...
Also , for the AKSA line of amps how /why did you choose
the 2SC5200/2sa1943 's , Do they sound better, or was it
availability/price.(I,m using njw0302/0281,are they worthy?)
OS
As I am only a day away from actually trying out these "tricks"
on a proven design ,I'm glad to have seen further disscussion/
simulation brought forth.
As a question to the awesome knowledge before me
(I'm not being sarcastic)
, how do you select a goodLPT pair (low Cob, enough Vce I know) but all those HFE
groupings (3 or 4 for some trannies)?? Do you hand match
for the CM and LTP (I ordered 30 each)as well?
By the way this is front end of the design I'm using as
the "guinea pig"...
An externally hosted image should be here but it was not working when we last tested it.
Also , for the AKSA line of amps how /why did you choose
the 2SC5200/2sa1943 's , Do they sound better, or was it
availability/price.(I,m using njw0302/0281,are they worthy?)
OS
ostripper
Best results will be obtained if the devices are matched for HFE, as well as closely as possible for VBE, then thermally coupled. Some time back there were several posts elsewhere, that suggested that devices with an HFE of around 800 sounded best in the CM. In fact, if the CM uses NPN transistors, the MPSA18 with their high HFE perform very well, and their relatively low voltage ratings don't matter in this application. The LTP will also benefit with devices chosen for an HFE of 300-400, as this will also help reduce DC offset.
SandyK
Best results will be obtained if the devices are matched for HFE, as well as closely as possible for VBE, then thermally coupled. Some time back there were several posts elsewhere, that suggested that devices with an HFE of around 800 sounded best in the CM. In fact, if the CM uses NPN transistors, the MPSA18 with their high HFE perform very well, and their relatively low voltage ratings don't matter in this application. The LTP will also benefit with devices chosen for an HFE of 300-400, as this will also help reduce DC offset.
SandyK
OS,
I would agree completely with Sandy; but aside from careful matching of the LTP devices, wouldn't go overboard. Speed should be at least 150MHz with Cob around 3pF, no more, but easy to find in TO92s.
Ditto for the CM devices, though if using very high mu and low stage current, reverse bias of the 'diode trick' could be an issue, so the transistor approach suggest by GK might be in order.
Hugh
I would agree completely with Sandy; but aside from careful matching of the LTP devices, wouldn't go overboard. Speed should be at least 150MHz with Cob around 3pF, no more, but easy to find in TO92s.
Ditto for the CM devices, though if using very high mu and low stage current, reverse bias of the 'diode trick' could be an issue, so the transistor approach suggest by GK might be in order.
Hugh
AKSA Amp kit
ostripper
If the LTP devices are reasonably matched for HFE, closely matched for VBE, and bonded together, you should be able to get rid of the offset trimmer.
SandyK
P.S. My understanding is that vhfman has used the diode with MPSA18 in his CMs and noticed no ill effects. Incidentally , his preamp sounded considerably better than an earlier model
AU $6,000 Krell when I last heard it.
ostripper
If the LTP devices are reasonably matched for HFE, closely matched for VBE, and bonded together, you should be able to get rid of the offset trimmer.
SandyK
P.S. My understanding is that vhfman has used the diode with MPSA18 in his CMs and noticed no ill effects. Incidentally , his preamp sounded considerably better than an earlier model
AU $6,000 Krell when I last heard it.
HI OS,
There is lots of traffic here on the worthiness or otherwise of various output devices. I believe this is important, but not too important. Topology is far more significant.
I chose the 5200/1943 way back when I designed the AKSA, and bought the devices in bulk. I continue to use them because they are fast (30MHz), with very flat beta around 100 from 100mA to 7A. Cob is a joke on an output anyway, an in any event is not too problematic with strong drivers and in EF configuration. I continue to use them today, and find of 100 devices that I can very closely match around 25 pairs, which is very, very impressive. I match within 1mV for Vbe at the bias current chosen (60mA), and within 3% for beta at this same current. They are robust devices, and moreover, I can buy them inexpensively.
In my view the greatest influences on amp design are the input stage and the voltage amplifier. Output stage does have sonic effect, but once the DEF II is sorted, the contribution is not great.....
Cheers,
Hugh
There is lots of traffic here on the worthiness or otherwise of various output devices. I believe this is important, but not too important. Topology is far more significant.
I chose the 5200/1943 way back when I designed the AKSA, and bought the devices in bulk. I continue to use them because they are fast (30MHz), with very flat beta around 100 from 100mA to 7A. Cob is a joke on an output anyway, an in any event is not too problematic with strong drivers and in EF configuration. I continue to use them today, and find of 100 devices that I can very closely match around 25 pairs, which is very, very impressive. I match within 1mV for Vbe at the bias current chosen (60mA), and within 3% for beta at this same current. They are robust devices, and moreover, I can buy them inexpensively.
In my view the greatest influences on amp design are the input stage and the voltage amplifier. Output stage does have sonic effect, but once the DEF II is sorted, the contribution is not great.....
Cheers,
Hugh
ostripper
you can have a read in this topic
there are several telling about experiences matching pairs
Matching transistors - Measuring hfe
http://www.diyaudio.com/forums/showthread.php?threadid=93898
My own experience, while measuring and matching BC54x BC55x BC56x, is that if you buy 100 from same batch
then those with same VBE are mostly very close in HFE gain, too.
'Same batch' means that these were made by same machine on the same day.
That is they are made from the same silicon piece of raw material.
If we want to match, we should not buy 10 pieces at one time and 10 pieces another time.
We should buy and ask for to get 100 or even more from the same batch, big package.
you can have a read in this topic
there are several telling about experiences matching pairs
Matching transistors - Measuring hfe
http://www.diyaudio.com/forums/showthread.php?threadid=93898
My own experience, while measuring and matching BC54x BC55x BC56x, is that if you buy 100 from same batch
then those with same VBE are mostly very close in HFE gain, too.
'Same batch' means that these were made by same machine on the same day.
That is they are made from the same silicon piece of raw material.
If we want to match, we should not buy 10 pieces at one time and 10 pieces another time.
We should buy and ask for to get 100 or even more from the same batch, big package.
Thanks for the input ,Andrew T., that 15k gives me 1.5 ma per
transistor. At 10k, I calculated I'd have 2.2 ma. The quasi
line of amps (which is a similar self design) uses the 15k,
but I always wondered ..why the low current.This amp
has a more robust CCS so forcing more current through the
tail would not be an issue.
I'm most likely going to drop that 3'rd tranny that mr.
klienschmidt discussed in my CM with an SMT under the board
(it is used in most precision op-ampsas the CM),
and with my "same batch" tape reel devices I might be able to drop the trimmer.
Ahh, the box of parts is at the door...
OS
Many years I have been looking for the best audio circuit.I'm getting too old and exhausted in internetting only to find out that AKSA is a copyright. Well, what can I say, "I'm Frustrated"
I am not a leech nor a blood-sucking vampire to suck what you've got and make an audio factory to earn a living and I not assuming too much for it is only that I could not afford to buy such kit because we poor people in Philippines could not get such thing due to the meager of the income. We are born reader and not a designer that you have study in school with your money. You should be proud that you are gifted and has the ability to draw such circuit with your brain.
AND You're right..You have the right to remain silent. keep the circuit secret
and Go ahead, keep on saying your AKSA sounded best or great..That's the reason why I bumped here because of reading too much of insights of AKSA yet I and others could not get and learn.
I am only searching for the best audio for my front seat in my living room and be proud to bring the system (If I completely assembled it) outside the manila street to compete with the music loving Filipinos who kept on listening and singing without jeopardizing other brands as you dont know.
You are talking about the circuits or diodes which we dont have here in forum, only to understand by yourself because each of you have your own kit and own diagram in your hands. You know that when you discuss something, it is necessary to see some of the details discussed which i could not see and at least investigate what you are talking about.
How could you say that Hugh has contributted to the DIY "Community" without giving any references and basic
schematic values to follow thru. Is there another ultralinear solid state invention here. what's new in this circuit ? Is there a name in Hugh's patent like "unity coupled circuit" or other?.. Hafler, Keroes, Williamson, Mcintosh, Nelson Pass, Daniel Meyer and others have their circuit published and earn their reputation and fame. Where is Hugh here?
Truly I admire Mr. Hugh's work based on what is advertised and promoted here, however, my soldering iron is waiting. DIY I understand is do-it-yourself , not buy and do-it-yourself after the kit has arrived. For people like us who can't buy, what should i do. "I LIKE AKSA" you know, but it's too expensive if compare to
assemble parts here.
Now it's a waste of time hanging in here. Perhaps it's about time to quit early and change my mind in the eagerness to have AKSA or ASKA.
Yes I post a less than 10 post, but I've spent most of my time reading your discussions but could not comprehend what you're up to without any reference. It's so unfair that Moderator had removed my post for those are the words suited and i feel is the best comment.
May I suggest to put it this way. Why dont you make a forum for AKSA owners group, just like Atma-Sphere Owners Group(ASOG) where they discussed all details separately who has their schematics in their room while discussing.
I pity to all of you for God will remove your wisdom at the end of this days (I Cor 1:19)
Forgive me for what I am saying here.
Poor Bala, 777, Mr.Bean which they didn't get anything here. I sympathized them. they are like me who is eager to have your Best.
If i offend anyone here and angry with me, remove again this post or e-mail me.
I am not a leech nor a blood-sucking vampire to suck what you've got and make an audio factory to earn a living and I not assuming too much for it is only that I could not afford to buy such kit because we poor people in Philippines could not get such thing due to the meager of the income. We are born reader and not a designer that you have study in school with your money. You should be proud that you are gifted and has the ability to draw such circuit with your brain.
AND You're right..You have the right to remain silent. keep the circuit secret
and Go ahead, keep on saying your AKSA sounded best or great..That's the reason why I bumped here because of reading too much of insights of AKSA yet I and others could not get and learn.
I am only searching for the best audio for my front seat in my living room and be proud to bring the system (If I completely assembled it) outside the manila street to compete with the music loving Filipinos who kept on listening and singing without jeopardizing other brands as you dont know.
You are talking about the circuits or diodes which we dont have here in forum, only to understand by yourself because each of you have your own kit and own diagram in your hands. You know that when you discuss something, it is necessary to see some of the details discussed which i could not see and at least investigate what you are talking about.
How could you say that Hugh has contributted to the DIY "Community" without giving any references and basic
schematic values to follow thru. Is there another ultralinear solid state invention here. what's new in this circuit ? Is there a name in Hugh's patent like "unity coupled circuit" or other?.. Hafler, Keroes, Williamson, Mcintosh, Nelson Pass, Daniel Meyer and others have their circuit published and earn their reputation and fame. Where is Hugh here?
Truly I admire Mr. Hugh's work based on what is advertised and promoted here, however, my soldering iron is waiting. DIY I understand is do-it-yourself , not buy and do-it-yourself after the kit has arrived. For people like us who can't buy, what should i do. "I LIKE AKSA" you know, but it's too expensive if compare to
assemble parts here.
Now it's a waste of time hanging in here. Perhaps it's about time to quit early and change my mind in the eagerness to have AKSA or ASKA.
Yes I post a less than 10 post, but I've spent most of my time reading your discussions but could not comprehend what you're up to without any reference. It's so unfair that Moderator had removed my post for those are the words suited and i feel is the best comment.
May I suggest to put it this way. Why dont you make a forum for AKSA owners group, just like Atma-Sphere Owners Group(ASOG) where they discussed all details separately who has their schematics in their room while discussing.
I pity to all of you for God will remove your wisdom at the end of this days (I Cor 1:19)
Forgive me for what I am saying here.
Poor Bala, 777, Mr.Bean which they didn't get anything here. I sympathized them. they are like me who is eager to have your Best.
If i offend anyone here and angry with me, remove again this post or e-mail me.
read through the DX series of threads.
These are a very basic kit and/or PCB that needs much more input from you and as a result is considerably cheaper.
Many of the design features in the original AKSA55 are replicated in the DX and developed a bit farther/wider in the other models that followed.
These are a very basic kit and/or PCB that needs much more input from you and as a result is considerably cheaper.
Many of the design features in the original AKSA55 are replicated in the DX and developed a bit farther/wider in the other models that followed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.