Don't make that mistake again. If you are unsure about something just ask. If you don't understand the answer say so. There are people here that are happy to answer dumb questions. Espescially me.
Last edited:
NYCOne nominated for the Fearless Amp Builder of the month for Feb 2011.
Agree
I posted this in the Cascoded thread too -
I hit myself in the head last night when I realized I could convert the Cascoded board I made (in the Cascoded Thread) to a Balanced Cascoded board witrh a couple of trivial soldering connections. I'm putting the Balanced board on hold for a night to go for the Balanced Cascoded board.
I hope to have sound out of it tonight.
If it goes well, expect a newly designed and etch board over the weekend. I finally believe I'm on the trial of something unique and valuable. I hope reality agrees with expectations...
I hit myself in the head last night when I realized I could convert the Cascoded board I made (in the Cascoded Thread) to a Balanced Cascoded board witrh a couple of trivial soldering connections. I'm putting the Balanced board on hold for a night to go for the Balanced Cascoded board.
I hope to have sound out of it tonight.
If it goes well, expect a newly designed and etch board over the weekend. I finally believe I'm on the trial of something unique and valuable. I hope reality agrees with expectations...
I think your error here is assuming everyone who reads these forums knows as much as the "experts".
For example: I asked very straightforward questions in the forum, often many times, and would get "Read the thread" or "Read a book" answers. I'd read the threads, without a practical example of many of the concepts, the thread conversations went right over my head.
I understand directing someone to a search a thread when a concrete answer is available via that route. However, reading any material fails if a newbie (like me) is asked to draw conclusions from a thread that's on a related, but not exactly the same subject.
The issues with these amp designs can be subtle, it is very easy to draw incorrect conclusions if one is new to the subject matter. In essence, that's what I thought the forums were for - to bounce questions and ideas off each other, not to one-up the next guy with your knowledge.
I received countless PMs urging me to push for some answers from other forum members that were too worried to post questions themselves. That's a shame.
Finally, a shout out to NP. A real gentleman. I suspect he runs circles around everyone in the forums, but makes newbies feel like their questions are reasonable. That's real class.
For example: I asked very straightforward questions in the forum, often many times, and would get "Read the thread" or "Read a book" answers. I'd read the threads, without a practical example of many of the concepts, the thread conversations went right over my head.
I understand directing someone to a search a thread when a concrete answer is available via that route. However, reading any material fails if a newbie (like me) is asked to draw conclusions from a thread that's on a related, but not exactly the same subject.
The issues with these amp designs can be subtle, it is very easy to draw incorrect conclusions if one is new to the subject matter. In essence, that's what I thought the forums were for - to bounce questions and ideas off each other, not to one-up the next guy with your knowledge.
I received countless PMs urging me to push for some answers from other forum members that were too worried to post questions themselves. That's a shame.
Finally, a shout out to NP. A real gentleman. I suspect he runs circles around everyone in the forums, but makes newbies feel like their questions are reasonable. That's real class.
Nycone
Keep going and don't worry about the "experts"
About the conversion I posted a picture on the other tread maybe not that easy to see but
it is easy to convert to the balanced just play about with the 10 homs resistors notice the x ?
(+ 2 capacitors and such)
Just one board mounted 2in on top of the other not tried so far but maybe it works
Keep going and don't worry about the "experts"
About the conversion I posted a picture on the other tread maybe not that easy to see but
it is easy to convert to the balanced just play about with the 10 homs resistors notice the x ?
(+ 2 capacitors and such)
Just one board mounted 2in on top of the other not tried so far but maybe it works
Cascoded F5X
For those who might be interested, some info has already been published before on cascoding my version of the F5X :
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-10.html#post2353070
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-13.html#post2371824
We are following a different approach of cascoding to other threads on the forum, and are using JFETs as cascode devices as per Erno Borbely. The 2SK170/2SJ74 see a (near) constant Vds of about 5V. The cascode JFETs (J111 / J174) are already included in the PCB layout, but can easily be bypassed with a single jumper wire. Simulations showed no sonic benefits. The distortion spectrum remains essentially the same with or without cascode. But tests will be performed for both comfigurations, and we shall only report when we have solid results.
Patrick
For those who might be interested, some info has already been published before on cascoding my version of the F5X :
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-10.html#post2353070
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-13.html#post2371824
We are following a different approach of cascoding to other threads on the forum, and are using JFETs as cascode devices as per Erno Borbely. The 2SK170/2SJ74 see a (near) constant Vds of about 5V. The cascode JFETs (J111 / J174) are already included in the PCB layout, but can easily be bypassed with a single jumper wire. Simulations showed no sonic benefits. The distortion spectrum remains essentially the same with or without cascode. But tests will be performed for both comfigurations, and we shall only report when we have solid results.
Patrick
2SK3497 / 2SJ618 GB
I just sent off to the GB participants details for payment. Pls check your email.
Message to ichiban
I got an error message saying hotmail won't deliver emails from my mail service provider. Please check whether you received a email from me titled 2SK3497..GB (i). If not I need to send you via other email accounts, which is a pain. Would be great if you have another email address that is not at hotmail.
Thanks to all,
Patrick
I just sent off to the GB participants details for payment. Pls check your email.
Message to ichiban
I got an error message saying hotmail won't deliver emails from my mail service provider. Please check whether you received a email from me titled 2SK3497..GB (i). If not I need to send you via other email accounts, which is a pain. Would be great if you have another email address that is not at hotmail.
Thanks to all,
Patrick
One of your questioned how I arrived at the 2.50 Euro per FET.
Special one-time discount price for 100 MOSFETs USD200.
10% for 2x Paypal (you to me, and me to dealer)
19% for UK import tax
14% for matching rejects (22 pairs from 50 FETs)
10% postage & packing, etc
Euro to USD 1.3
=> per FET 2.36 Euros.
I rounded off to 2.50 Euros for ease of calculation for now.
This is estimated costs only, and the exact sum will be calculated later on when all cost have been accounted for. The excess amount will be returned to you / donated to the forum / contribute to pay for prototyping the F5X PCBs / ..... Whatever you wish to decide.
Regards,
Patrick
Special one-time discount price for 100 MOSFETs USD200.
10% for 2x Paypal (you to me, and me to dealer)
19% for UK import tax
14% for matching rejects (22 pairs from 50 FETs)
10% postage & packing, etc
Euro to USD 1.3
=> per FET 2.36 Euros.
I rounded off to 2.50 Euros for ease of calculation for now.
This is estimated costs only, and the exact sum will be calculated later on when all cost have been accounted for. The excess amount will be returned to you / donated to the forum / contribute to pay for prototyping the F5X PCBs / ..... Whatever you wish to decide.
Regards,
Patrick
I just caught this post. I'm sorry, I don't check this thread as often as I did in the past.
Can you enlighten us about your cascoding thoughts? Trial and errors?
I built the Borbely All JFET no feedback Preamp (EB2000-402). I thought Borbely's ideas were fabulous. I had some questions, but Erno was great and jumped right in to help me with my build.
I have some 2SK246BLs left over as well as 2SK170 and 2SJ74s from that build. I cascoded the Balanced F5 with BC546 and BC556 mostly because am not confident of my efforts and I hate to waste (relatively) expensive transistors if I'm wrong - it happens lots of times.
I thought about your comments regarding cascoding and then about Nelson Pass' - on the surface they seem at odds. I'd completely overlooked the idea of cascoding and was going to leave it alone until Nelson reminded me of his paper on cascoding (indirectly in a post). I have no training in engineering or electronics, so I may be interpreting it incorrectly, but it seems to show cascoding always helps. As I understand it, you stay in the "sweet spot" with cascoding (though you need to find it first).
I would love to understand your efforts, you're a very sharp fellow. Unfortunately, you seem to keep so many interesting aspects of your build to the few you've invited to help you. I'd asked to be part of that effort, but was rebuffed - not experienced enough. I can't argue with that point.
None-the-less, there's no commercial angle here, no awards, can you share some of your learnings with the forum? Even if they will be more complete in the future? Aren't we always more knowledgeable as we accrue more experience?
Can you enlighten us about your cascoding thoughts? Trial and errors?
I built the Borbely All JFET no feedback Preamp (EB2000-402). I thought Borbely's ideas were fabulous. I had some questions, but Erno was great and jumped right in to help me with my build.
I have some 2SK246BLs left over as well as 2SK170 and 2SJ74s from that build. I cascoded the Balanced F5 with BC546 and BC556 mostly because am not confident of my efforts and I hate to waste (relatively) expensive transistors if I'm wrong - it happens lots of times.
I thought about your comments regarding cascoding and then about Nelson Pass' - on the surface they seem at odds. I'd completely overlooked the idea of cascoding and was going to leave it alone until Nelson reminded me of his paper on cascoding (indirectly in a post). I have no training in engineering or electronics, so I may be interpreting it incorrectly, but it seems to show cascoding always helps. As I understand it, you stay in the "sweet spot" with cascoding (though you need to find it first).
I would love to understand your efforts, you're a very sharp fellow. Unfortunately, you seem to keep so many interesting aspects of your build to the few you've invited to help you. I'd asked to be part of that effort, but was rebuffed - not experienced enough. I can't argue with that point.
None-the-less, there's no commercial angle here, no awards, can you share some of your learnings with the forum? Even if they will be more complete in the future? Aren't we always more knowledgeable as we accrue more experience?
NYCO
Euvl has built this amplifier for himself.
He succumbed to persuasion and released just sufficient details for the Members to decide if they wanted to join him.
The promise has always been that once the copies are up and running and the build details sorted/developed then he plans to make all the information available to all the builders.
There's no secrecy in that.
In the meantime, there is sufficient information from Euvl for any Memeber to build his own "clone" of the balanced F5 in various forms.
This is definitely not a Secret Society within DIYaudio.
BTW,
you post in a thread and then abandon it !
Euvl has built this amplifier for himself.
He succumbed to persuasion and released just sufficient details for the Members to decide if they wanted to join him.
The promise has always been that once the copies are up and running and the build details sorted/developed then he plans to make all the information available to all the builders.
There's no secrecy in that.
In the meantime, there is sufficient information from Euvl for any Memeber to build his own "clone" of the balanced F5 in various forms.
This is definitely not a Secret Society within DIYaudio.
BTW,
you post in a thread and then abandon it !
> Any thoughts on how these will perform as the amplifier FETs?
Such UHC-MOSFETs have been very popular amongst Japanese DIYers. You can find, e.g. circuits from Kaneda using these FETs at least 4 or 5 years back. However, there are a few aspects to consider in using the 2SK3497 / 2SJ618 in the F5.
They have high transconductances. This means higher open loop gain, hence higher negative feedback, and lower distortion. They also have much higher capacitances. Thus the first corner frequency of the open loop response will be lower, which might be compensated by the higher NFB. The sonic signature will certainly change. The question is whether it is to your liking, which is of course subjective.
They are not truly complementary devices as the 2SK1530 / 2SJ201. So you need to change other things in the circuit to make the distortion cancellation (e.g. playing with values of the source resistor). They also have a very high tempco. It is probably essential to include the thermistor in the biasing circuit.
The Spice models are openly available. If you have time, I suggest you spice the single-ended circuit, using 16V rails, 0R22 source resistors (as a start), and 2A bias. If you could twist the circuit until you get near zero 3rd and 5th harmonics, then we have a good basis to experiment from. The rest would have to come from tests.
> Can you enlighten us about your cascoding thoughts?
Cascoding is a very complicated subject that I'd rather leave for Nelson to explain. There are at least 4 reasons to cascode – to sustain higher voltage and/or dissipation, to reduce Miller Effect, and to provide extra isolation from external influences, and to improve linearity. I can also think of at least 3 methods of cascoding – self biasing (Borbely), fixed voltage (Pass F5), signal modulated (Pass ZV9).
If we eventually determine that cascoding is beneficial, we shall publish the results and explain why. If not, then I rather not spend the effort.
> Unfortunately, you seem to keep so many interesting aspects of your build to the few you've invited to help you.
The people in the team are all seasoned DIYers who know their basic electronics, many of them more skilled than I in PCB design & production, in debugging and testing, ….
I have already explained to you in a PM in great length a few months back why I cannot afford the time to explain everything in such detail that every beginner with no electronic background will understand. This task has been kindly alleviated from me by the two gentlemen who will build, test, write-up and answer questions. In selecting the team members, the one single criterion is what their contribution would be, and whether this contribution will save me time or not.
If you have a real interest in understanding circuits, please take my advice and read a book from Cordell, or Self, or Sloan, …. They have answers to a lot of your questions. As I have also explained to you in the PM, if you just want a definitive answer to a specific question, Nelson is the best person to ask.
> Euvl has built this amplifier for himself.
> He succumbed to persuasion and released just sufficient details for the Members to decide if they wanted to join him.
> The promise has always been that once the copies are up and running and the build details sorted/developed then he plans to make all the information available to all the builders.
> There's no secrecy in that.
> In the meantime, there is sufficient information from Euvl for any Member to build his own "clone" of the balanced F5 in various forms.
Thank you for re-summarising for me.
> Sorry to ask this but could you direct me to where the power supply cct for this amp is buried - I just can't find it although I think you did post it on the F5 amplifier thread some time ago.
No problem to ask. We have not published one as yet, but it is similar to the one for Zen Version 3. We will do so at the end of February when the PCB work is finished.
In the meantime I hope someone could reach Variac (or Jason) for me, as Marc has not answered my emails for months, and the forum has to decide in the next week or two how to proceed with the PCBs once all Gerber files are ready for prototyping. If the forum no longer wants to organise this, then some of you would have to take over as a GB or whatever else, as I really have no time to do so for you.
Regards,
Patrick
.
Such UHC-MOSFETs have been very popular amongst Japanese DIYers. You can find, e.g. circuits from Kaneda using these FETs at least 4 or 5 years back. However, there are a few aspects to consider in using the 2SK3497 / 2SJ618 in the F5.
They have high transconductances. This means higher open loop gain, hence higher negative feedback, and lower distortion. They also have much higher capacitances. Thus the first corner frequency of the open loop response will be lower, which might be compensated by the higher NFB. The sonic signature will certainly change. The question is whether it is to your liking, which is of course subjective.
They are not truly complementary devices as the 2SK1530 / 2SJ201. So you need to change other things in the circuit to make the distortion cancellation (e.g. playing with values of the source resistor). They also have a very high tempco. It is probably essential to include the thermistor in the biasing circuit.
The Spice models are openly available. If you have time, I suggest you spice the single-ended circuit, using 16V rails, 0R22 source resistors (as a start), and 2A bias. If you could twist the circuit until you get near zero 3rd and 5th harmonics, then we have a good basis to experiment from. The rest would have to come from tests.
> Can you enlighten us about your cascoding thoughts?
Cascoding is a very complicated subject that I'd rather leave for Nelson to explain. There are at least 4 reasons to cascode – to sustain higher voltage and/or dissipation, to reduce Miller Effect, and to provide extra isolation from external influences, and to improve linearity. I can also think of at least 3 methods of cascoding – self biasing (Borbely), fixed voltage (Pass F5), signal modulated (Pass ZV9).
If we eventually determine that cascoding is beneficial, we shall publish the results and explain why. If not, then I rather not spend the effort.
> Unfortunately, you seem to keep so many interesting aspects of your build to the few you've invited to help you.
The people in the team are all seasoned DIYers who know their basic electronics, many of them more skilled than I in PCB design & production, in debugging and testing, ….
I have already explained to you in a PM in great length a few months back why I cannot afford the time to explain everything in such detail that every beginner with no electronic background will understand. This task has been kindly alleviated from me by the two gentlemen who will build, test, write-up and answer questions. In selecting the team members, the one single criterion is what their contribution would be, and whether this contribution will save me time or not.
If you have a real interest in understanding circuits, please take my advice and read a book from Cordell, or Self, or Sloan, …. They have answers to a lot of your questions. As I have also explained to you in the PM, if you just want a definitive answer to a specific question, Nelson is the best person to ask.
> Euvl has built this amplifier for himself.
> He succumbed to persuasion and released just sufficient details for the Members to decide if they wanted to join him.
> The promise has always been that once the copies are up and running and the build details sorted/developed then he plans to make all the information available to all the builders.
> There's no secrecy in that.
> In the meantime, there is sufficient information from Euvl for any Member to build his own "clone" of the balanced F5 in various forms.
Thank you for re-summarising for me.
> Sorry to ask this but could you direct me to where the power supply cct for this amp is buried - I just can't find it although I think you did post it on the F5 amplifier thread some time ago.
No problem to ask. We have not published one as yet, but it is similar to the one for Zen Version 3. We will do so at the end of February when the PCB work is finished.
In the meantime I hope someone could reach Variac (or Jason) for me, as Marc has not answered my emails for months, and the forum has to decide in the next week or two how to proceed with the PCBs once all Gerber files are ready for prototyping. If the forum no longer wants to organise this, then some of you would have to take over as a GB or whatever else, as I really have no time to do so for you.
Regards,
Patrick
.
Last edited:
Some examples of UHC-MOSFET amps :
2SJ554/2SK2955 classAB SEPP power amplifier
audio: ‹à“cŽ®•—–¡�@ƒVƒ“ƒvƒ‹�Eƒpƒ��[ƒAƒ“ƒv‚Ì�§�ì�iNO.192) �i‚P�j
‰�”N‚Ì‚c‚bƒAƒ“ƒvƒp�[ƒc�iƒpƒ��[�E‚e‚d‚s�j
Patrick
2SJ554/2SK2955 classAB SEPP power amplifier
audio: ‹à“cŽ®•—–¡�@ƒVƒ“ƒvƒ‹�Eƒpƒ��[ƒAƒ“ƒv‚Ì�§�ì�iNO.192) �i‚P�j
‰�”N‚Ì‚c‚bƒAƒ“ƒvƒp�[ƒc�iƒpƒ��[�E‚e‚d‚s�j
Patrick
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.