How I actually do it, isn't very important. It is the idea of having to use so many regulators, each of which does a different job. The shunt regulator tends to capture any high frequency transient that might get through or even be generated by the first regulator. There are several equally good combinations that should work well.
What is more important is the passive devices and configuration ahead of the first regulator. It is designed as a pi network to remove potential ground contamination from RFI, etc. Soft recovery diodes are used as well, without using a center tap on the secondary of the dual coil power transformer. You now see examples of this in many designs. Years ago, it was considered 'excessive' and it is still somewhat difficult to implement.
What is more important is the passive devices and configuration ahead of the first regulator. It is designed as a pi network to remove potential ground contamination from RFI, etc. Soft recovery diodes are used as well, without using a center tap on the secondary of the dual coil power transformer. You now see examples of this in many designs. Years ago, it was considered 'excessive' and it is still somewhat difficult to implement.
Hi, John
The Toshiba parts you are refering to are 2SJ72 and 2SK147, as an evidence for me... Yeah... Some 2SK147 V can still be found, but no chance for 2SJ72 V (and I tried it very hard, you can trust me !... ). So, perhaps a pair of paralleled 2SJ74 V would do the job too for replacing ONE 2SJ72 V, but didn't try yet !
I was quite puzzled with some choice you made with the Dennesen JC-80 : you used consistently 2SJ75 GR grade against 2SK240 V... (see the attached pic). Not quite in the same Vgs and Idss league, as the J113 and J175... It is VERY interesting because you used all them very soon (J175 and J113 since JC-2 IMO, 2SJ72 and 2SK240 since Dennesen...). The close (ahem ! spreads to x 3 factor !) complementary data Jfets data book claim is quite wrong, and your own EXPERIENCE with them (and not only experiments) show it as an evidence... Without this trick, for sure, I would have gone some cross matching attempts that would fail at most... Here, you are not only saving me money, but lots of matching time ! Really, you are both a great engineer and a great guy, John... Just designing sport engines but also with both hands in the used oil... (LOL ! Somewhat a French expression almost impossible to translate right !).
But returning to the dual LM317 trick, you seem to approve LM317 successive stages in the quite normal use (both series). Wouldn't another approach be far better : the first LM317 in a current source configuration feeding another LM317 in the common series configuration... Would be just a power equivalent, IMO, of a constant current source feeding a precision zener diode or voltage reference... Just like in this attachment from a Motorola note...
Thans a lot, John
The Toshiba parts you are refering to are 2SJ72 and 2SK147, as an evidence for me... Yeah... Some 2SK147 V can still be found, but no chance for 2SJ72 V (and I tried it very hard, you can trust me !... ). So, perhaps a pair of paralleled 2SJ74 V would do the job too for replacing ONE 2SJ72 V, but didn't try yet !
I was quite puzzled with some choice you made with the Dennesen JC-80 : you used consistently 2SJ75 GR grade against 2SK240 V... (see the attached pic). Not quite in the same Vgs and Idss league, as the J113 and J175... It is VERY interesting because you used all them very soon (J175 and J113 since JC-2 IMO, 2SJ72 and 2SK240 since Dennesen...). The close (ahem ! spreads to x 3 factor !) complementary data Jfets data book claim is quite wrong, and your own EXPERIENCE with them (and not only experiments) show it as an evidence... Without this trick, for sure, I would have gone some cross matching attempts that would fail at most... Here, you are not only saving me money, but lots of matching time ! Really, you are both a great engineer and a great guy, John... Just designing sport engines but also with both hands in the used oil... (LOL ! Somewhat a French expression almost impossible to translate right !).
But returning to the dual LM317 trick, you seem to approve LM317 successive stages in the quite normal use (both series). Wouldn't another approach be far better : the first LM317 in a current source configuration feeding another LM317 in the common series configuration... Would be just a power equivalent, IMO, of a constant current source feeding a precision zener diode or voltage reference... Just like in this attachment from a Motorola note...
Thans a lot, John
Attachments
Hi, Darry
I am not so sure that the current source is effective AFTER the LM317 in a regular series regulation, and specially if John runs all audio boards in full Class A (and I bet they do !). The current drawn would be shared between the charge and the amp board, but the whole current taking will remain unchanged, and will be a constant one... So, IMO, to put a LM317 in current source configuration AFTER then LM317 in regular series voltage regulation won't help.
So, I would better advocate for the two cascaded series voltage regulation already proposed. But... My own guess is that a single LM317 in usual series regulation scheme would suffice to have the job done fine... Two cascaded series LM317 wouldn't improve really voltage regulation but perhaps a little reject unwanted rubbish noise... So I will stay with my own power current source feeding a voltage reference analogy. But didn't tested this two configurations yet, and compare them... Has anyone ?
Jbaudiophile
I am not so sure that the current source is effective AFTER the LM317 in a regular series regulation, and specially if John runs all audio boards in full Class A (and I bet they do !). The current drawn would be shared between the charge and the amp board, but the whole current taking will remain unchanged, and will be a constant one... So, IMO, to put a LM317 in current source configuration AFTER then LM317 in regular series voltage regulation won't help.
So, I would better advocate for the two cascaded series voltage regulation already proposed. But... My own guess is that a single LM317 in usual series regulation scheme would suffice to have the job done fine... Two cascaded series LM317 wouldn't improve really voltage regulation but perhaps a little reject unwanted rubbish noise... So I will stay with my own power current source feeding a voltage reference analogy. But didn't tested this two configurations yet, and compare them... Has anyone ?
Jbaudiophile
Hi, John
Thanks a lot for the input ! The Dennesen JC-80 is crowded (ahem... Just a few pairs...) with such strange things...I understand now that your own schematic didn't showed unprobable matches like these ones, and that only relies to Jim Dennesen... First, I thought that you had perhaps found that the P-channel high Idss class just matched the N-channel low end of the class...
Now, I understand what happened with your own designs a lot... And much more than audio specialized press related it, making your glory and then wiping you out of the market... They resulted from parts changes WITHOUT your own advice but on YOUR schematics, at an evidence to cut costs, by your employers at the time... A shame, as they knew little (or no? !) electronics at the time : Mark Levinson, a former musician, Jim Dennesen (I was heard he was first selling ice creams... Really ?).
But here in Europe as well as in most countries, employers HAVE THE INTELLECTUAL PROPERTY RIGHTS...
but if you have a registered patent... I am fighting against that and found some ways that seems valid (I am at first a lawyer, as I studied for but choosed another professional way). Anyone interested in full explanations of how itwould work, please drop me a line !
Jbaudiophile
Thanks a lot for the input ! The Dennesen JC-80 is crowded (ahem... Just a few pairs...) with such strange things...I understand now that your own schematic didn't showed unprobable matches like these ones, and that only relies to Jim Dennesen... First, I thought that you had perhaps found that the P-channel high Idss class just matched the N-channel low end of the class...
Now, I understand what happened with your own designs a lot... And much more than audio specialized press related it, making your glory and then wiping you out of the market... They resulted from parts changes WITHOUT your own advice but on YOUR schematics, at an evidence to cut costs, by your employers at the time... A shame, as they knew little (or no? !) electronics at the time : Mark Levinson, a former musician, Jim Dennesen (I was heard he was first selling ice creams... Really ?).
But here in Europe as well as in most countries, employers HAVE THE INTELLECTUAL PROPERTY RIGHTS...
but if you have a registered patent... I am fighting against that and found some ways that seems valid (I am at first a lawyer, as I studied for but choosed another professional way). Anyone interested in full explanations of how itwould work, please drop me a line !
Jbaudiophile
Here is a first schematic for a global PSU to start learning with and to see progressively what’s important and of interest in audio power supplies, as well as what could be in the wrong way.
I didn’t represent the dual coil power transformer and the soft recovery diodes bridges.
I am not very experienced in shunt regs…
PSU - Global
I didn’t represent the dual coil power transformer and the soft recovery diodes bridges.
I am not very experienced in shunt regs…
PSU - Global
Hi, Callmedad,
WOW ! A big effort indeed ... Close to my own thinking of the Blowtorch PSU... John said that the Xformer secondaries and rectyfying were taken apart from each other... And on the Blowtorch board, we can only see ONE CMC, not two... But perhaps you are not on the Blowtorch idea.. If such a case, I beg your pardon....
John didn't say that the 3 terminals stage was only made with ONE LM317... Two, IMO... But I already posted a thread about it... To be discussed !
Your proposed shunt regulation is interesting, as the first stage is a current source as I guess it (but I think that it would be far more manageable to do that with a single LM317... Just see my previous post in answer to Darry's proposal).
Also, the last stage is VERY interesreting to me, and I will study this one a lot (after some sleeping time !). But IMO, this one woud be much more at ease on the Blowtorch Line amplifier, in lieu of a local regulatiion (when John Curl speaks, I just listen... then wonder why he said so... NO parables from him... But to be thought of several times !).
Seems that the Jfet constant current trick I suggested to you is now in all of your designs... A good thing, really !
I am not sure, but on the Blowtorch PSUs, all Mosfets seem to be driven from an op-amp, and NOT through a Jfet... To be discussed further... But wait a minute.. I notice now that you are also in France... Could drop me a message through the "Direct line" (out of the forum)... Perhap that we can cross some guess and data to share with other guys... But anyway, just an interesting last stage for me...
All the best, Callmedad
Jbaudiophile
WOW ! A big effort indeed ... Close to my own thinking of the Blowtorch PSU... John said that the Xformer secondaries and rectyfying were taken apart from each other... And on the Blowtorch board, we can only see ONE CMC, not two... But perhaps you are not on the Blowtorch idea.. If such a case, I beg your pardon....
John didn't say that the 3 terminals stage was only made with ONE LM317... Two, IMO... But I already posted a thread about it... To be discussed !
Your proposed shunt regulation is interesting, as the first stage is a current source as I guess it (but I think that it would be far more manageable to do that with a single LM317... Just see my previous post in answer to Darry's proposal).
Also, the last stage is VERY interesreting to me, and I will study this one a lot (after some sleeping time !). But IMO, this one woud be much more at ease on the Blowtorch Line amplifier, in lieu of a local regulatiion (when John Curl speaks, I just listen... then wonder why he said so... NO parables from him... But to be thought of several times !).
Seems that the Jfet constant current trick I suggested to you is now in all of your designs... A good thing, really !
I am not sure, but on the Blowtorch PSUs, all Mosfets seem to be driven from an op-amp, and NOT through a Jfet... To be discussed further... But wait a minute.. I notice now that you are also in France... Could drop me a message through the "Direct line" (out of the forum)... Perhap that we can cross some guess and data to share with other guys... But anyway, just an interesting last stage for me...
All the best, Callmedad
Jbaudiophile
- Status
- Not open for further replies.