Dual Mono means 2 channels in one case, but with separate transformers and power supplies. So we have not changed the situation since then.
A single channel of F5X consumes 32Vx4A = 128W electrical power, and dissipate all of that into the heat sink when not playing music. So a heat sink for 75W is certainly under-dimensioned. The MF35-151.5 will just take that much power without excessive junction temperature at the MOSFETs. So I am not prepared to advise you to put out more heat into it. Of course I cannot stop you from doing so either.
🙂
Hope this answers your question,
Patrick
.
A single channel of F5X consumes 32Vx4A = 128W electrical power, and dissipate all of that into the heat sink when not playing music. So a heat sink for 75W is certainly under-dimensioned. The MF35-151.5 will just take that much power without excessive junction temperature at the MOSFETs. So I am not prepared to advise you to put out more heat into it. Of course I cannot stop you from doing so either.
🙂
Hope this answers your question,
Patrick
.
Last edited:
F5 Versions supported by the Main Amplifier PCB
I was asked by PM as to whether one can use our PCB for SE F5, why not bipolar cascode, how about all the other F5 variants, ....., etc.
I try to explain it simply.
The circuitry (independent of component values) of the PCB is identical to the standard F5 as published by Nelson a few year ago, including the over-current protection, but with the following exceptions :
1) It contains 2 Single Ended F5's in (95%) mirror image. The difference to using 2 separate F5 SE PCBs (e.g. from Peter Daniel) is that the left and right halves of our PCB share a common single-point star ground, a common connection to +Vs and another to -Vs in the middle. This is a key layout requirement for the balanced circuit. But still, you may, if you so wish, use one PCB to build 2 channels of SE F5's without problems, even though you might get a bit of power-supply, and to a lesser extent signal, coupling between the left and right channels.
2) It supports both grounded F5X and floating F5X.
3) It supports optional JFET cascode for the first stage (using J111/J174 as per Borberly).
4) It does not include thermistors, but there are enough connection points available if you wish to use a thermistor.
5) It supports the use of 2SK389 / 2SJ109 dual JFETs. This puts quite some constrains on the PCB layout.
You can pretty much use any value of components, and any devices you want, as long as they are TO92/TO247 (TO3P), and pin compatible.
There are provisions to allow the use of MPC74 or TO220 type source resistors.
Patrick
.
I was asked by PM as to whether one can use our PCB for SE F5, why not bipolar cascode, how about all the other F5 variants, ....., etc.
I try to explain it simply.
The circuitry (independent of component values) of the PCB is identical to the standard F5 as published by Nelson a few year ago, including the over-current protection, but with the following exceptions :
1) It contains 2 Single Ended F5's in (95%) mirror image. The difference to using 2 separate F5 SE PCBs (e.g. from Peter Daniel) is that the left and right halves of our PCB share a common single-point star ground, a common connection to +Vs and another to -Vs in the middle. This is a key layout requirement for the balanced circuit. But still, you may, if you so wish, use one PCB to build 2 channels of SE F5's without problems, even though you might get a bit of power-supply, and to a lesser extent signal, coupling between the left and right channels.
2) It supports both grounded F5X and floating F5X.
3) It supports optional JFET cascode for the first stage (using J111/J174 as per Borberly).
4) It does not include thermistors, but there are enough connection points available if you wish to use a thermistor.
5) It supports the use of 2SK389 / 2SJ109 dual JFETs. This puts quite some constrains on the PCB layout.
You can pretty much use any value of components, and any devices you want, as long as they are TO92/TO247 (TO3P), and pin compatible.
There are provisions to allow the use of MPC74 or TO220 type source resistors.
Patrick
.
Last edited:
Cascode using Bipolar Transistors
Another specific question was "Why not bipolar cascode".
This is a good question, and there are indeed some technical merits in using BJT cascodes. BJTs have a higher gm than JFETs like J111/J174 or 2SK246/2SJ103, and hence can reduce Miller effect of the first stage JFETs (2SK170/2SJ74) better, at the expense of 6 more components per channel.
I shall talk to fitzfish about this and see whether we can offer this option on the PCB as well.
Whether we shall also test this particular version will be decide between the test team, as they not only carry the work load, but also the expenses.
Hope I have answered some of your questions sufficiently,
Patrick
.
Another specific question was "Why not bipolar cascode".
This is a good question, and there are indeed some technical merits in using BJT cascodes. BJTs have a higher gm than JFETs like J111/J174 or 2SK246/2SJ103, and hence can reduce Miller effect of the first stage JFETs (2SK170/2SJ74) better, at the expense of 6 more components per channel.
I shall talk to fitzfish about this and see whether we can offer this option on the PCB as well.
Whether we shall also test this particular version will be decide between the test team, as they not only carry the work load, but also the expenses.
Hope I have answered some of your questions sufficiently,
Patrick
.
Last edited:
thx
so If I do not get it wrong again, only output devices are on MF35 heatsink and we will have another heatsink for regulator per channel
what Is dissipation for regulator board
so If I do not get it wrong again, only output devices are on MF35 heatsink and we will have another heatsink for regulator per channel
what Is dissipation for regulator board
Dual Mono means 2 channels in one case, but with separate transformers and power supplies. So we have not changed the situation since then.
A single channel of F5X consumes 32Vx4A = 128W electrical power, and dissipate all of that into the heat sink when not playing music. So a heat sink for 75W is certainly under-dimensioned. The MF35-151.5 will just take that much power without excessive junction temperature at the MOSFETs. So I am not prepared to advise you to put out more heat into it. Of course I cannot stop you from doing so either.
🙂
Hope this answers your question,
Patrick
.
> only output devices are on MF35 heatsink and we will have another heatsink for regulator per channel
Yes, that is the plan.
Each of the two regulators (one for each channel) sits on a Fischer SK88-125 heat sink, so do the rectifying diodes. Per channel, assuming 16Vac secondaries from your transformer, the dissipating for rectifier and regulator will be 26.5W in total. The SK88 heatsinks sit inside the case, along the centre line and close to the top cover. The top cover has very generous openings protected by a wire mesh for ventilation.
Patrick
Yes, that is the plan.
Each of the two regulators (one for each channel) sits on a Fischer SK88-125 heat sink, so do the rectifying diodes. Per channel, assuming 16Vac secondaries from your transformer, the dissipating for rectifier and regulator will be 26.5W in total. The SK88 heatsinks sit inside the case, along the centre line and close to the top cover. The top cover has very generous openings protected by a wire mesh for ventilation.
Patrick
Those who are interested in the discussion why to have 16V 2A per FET for the 4-MOSFET balanced circuit can have a look at :
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-21.html#post2474669
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-22.html#post2475929
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-22.html#post2475623
Hope it helps to clear rather than to confuse,
Patrick
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-21.html#post2474669
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-22.html#post2475929
http://www.diyaudio.com/forums/pass-labs/172770-balanced-f5-question-22.html#post2475623
Hope it helps to clear rather than to confuse,
Patrick
Is 16V rail voltage too low ?
See :
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads.html#post2475952
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads.html#post2475958
I find that I do most of the listening at no more than 4V pk-pk, which means in the F5X (balanced) the Vds across each FET is no less than (Rail Voltage - 2.5V). The extra 0.5V deduction comes from the source resistors of the MOSFETs.
Patrick
See :
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads.html#post2475952
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads.html#post2475958
I find that I do most of the listening at no more than 4V pk-pk, which means in the F5X (balanced) the Vds across each FET is no less than (Rail Voltage - 2.5V). The extra 0.5V deduction comes from the source resistors of the MOSFETs.
Patrick
Comments about paralleling output devices
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads-2.html#post2476282
Patrick
http://www.diyaudio.com/forums/pass-labs/182966-f5-low-z-loads-2.html#post2476282
Patrick
I have revisited the cascode options in the last couple of days, and read some of Borbely's paper again. He has abandoned the use of bipolar cascode in favour of JFETs in all his new designs, with reasons.
So I have come up with a new JFET cascode which will give better performance than the J111/J174 and yet not more complex than the bipolar cascode. And we shall include that on the PCB layout, together with the J111/J174 option. The amplifier will then remain an all-FET design.
Again, this is only for experimentation purposes. Only actual test will decide whether it brings anything.
Patrick
.
So I have come up with a new JFET cascode which will give better performance than the J111/J174 and yet not more complex than the bipolar cascode. And we shall include that on the PCB layout, together with the J111/J174 option. The amplifier will then remain an all-FET design.
Again, this is only for experimentation purposes. Only actual test will decide whether it brings anything.
Patrick
.
Last edited:
Hi,
I have just found this new thread.
Hopefully this will stop all the complainers, moaners, etc from distracting the interested from their goal.
I have just found this new thread.
Hopefully this will stop all the complainers, moaners, etc from distracting the interested from their goal.
Last edited:
I have revisited the cascode options in the last couple of days, and read some of Borbely's paper again. He has abandoned the use of bipolar cascode in favour of JFETs in all his new designs, with reasons.
Patrick
.
hi Patric
sorry to bother you - take your time to reply
what is the reason for Borbely to switch to FET cascodes?
thanks
Kannan
well i have received no reply from fitzfish on the transformer query. perhaps he will see here. perhaps i am not worthy? 😀 anyone else in the build team i can contact who might know?
perhaps the switch to fets also was congruent with his advertising byline creating 'all fet' designs hehe
perhaps the switch to fets also was congruent with his advertising byline creating 'all fet' designs hehe
Here is what the invoice said:
RC0600 915 5 banded transformer
My invoice number is 5149.
Hope this will help.
RC0600 915 5 banded transformer
My invoice number is 5149.
Hope this will help.
it has come to my attention that my post above if taken seriously could be taken the wrong way. my post
all sorted thanks to pchw anyway, so as you were
was more me thumbing my nose at the sometimes deserving attitude that seems to be going around lately. it was not meant to be taken seriously thus the smiley, as i know others like myself have their own lives they must get on with. so Dave if you read it like this my apologies, i was not being serious, but i guess its still a bit raw for some people at the moment.well i have received no reply from fitzfish on the transformer query. perhaps he will see here. perhaps i am not worthy?
all sorted thanks to pchw anyway, so as you were

I can give you a small piece of good news.
We have been in touch with the forum and are discussing details of the PCB prototyping run.
So we are back on schedule.
🙂
Patrick
We have been in touch with the forum and are discussing details of the PCB prototyping run.
So we are back on schedule.
🙂
Patrick
- Home
- Amplifiers
- Pass Labs
- F5X -- the EUVL Approach