Originally posted by Nelson Pass
Let's see....What is my position on Bandwidth?
Oh, I remember. Bandwidth is good unless it's not. By that I
mean that giving consumers a power amplifier with gain and a
very wide bandwith is an invitation to trouble because many
of them (consumers) can't put batteries in a flashlight, much less
do a proper job of grounding in their systems.
Also, bandwidth is good if I get it without having to make a
special effort. As a result my stuff usually gets up to 100 KHz
without any extra parts, so I don't monkey with it.
I have an abstract interest in the acceptance of the F4, but it's
bandwidth is not the part that will influence that - it is the total
lack of voltage gain. It's a surprisingly good sounding piece,
but very few consumers will find that out.
Lastly, since I switched over to FETs, I have not encountered an
instance of RF pickup except for the owner of an Aleph-Ono who
lived in the Empire State Building under the antennae.
1. ... is an inviatation to truble ...
yes, we can read issues/problems in topics all day if wish at diyaudio.com
2. ... 100 kHz without extra parts ...
yes, keeping amplifiers simple & good we often can end up
with an upper bandwidth 5-10 times higher than any man can hear
Your thinking on this is good & down to earth.
And this is one reason people build your amplifiers.
I know this is your philosophy when comes to Pass DIY designs.
Probably you have several exceptions in some of your experiments and commercial designs.
But a few exceptions to a rule, does not make the basic rule worse.
Exceptions does confirm that there is a General Rule.
Lineup
regards
Let's see....What is my position on Bandwidth?
Oh, I remember. Bandwidth is good unless it's not. By that I
mean that giving consumers a power amplifier with gain and a
very wide bandwith is an invitation to trouble because many
of them (consumers) can't put batteries in a flashlight, much less
do a proper job of grounding in their systems.
Also, bandwidth is good if I get it without having to make a
special effort. As a result my stuff usually gets up to 100 KHz
without any extra parts, so I don't monkey with it.
I have an abstract interest in the acceptance of the F4, but it's
bandwidth is not the part that will influence that - it is the total
lack of voltage gain. It's a surprisingly good sounding piece,
but very few consumers will find that out.
Lastly, since I switched over to FETs, I have not encountered an
instance of RF pickup except for the owner of an Aleph-Ono who
lived in the Empire State Building under the antennae.
1. ... is an inviatation to truble ...
yes, we can read issues/problems in topics all day if wish at diyaudio.com
2. ... 100 kHz without extra parts ...
yes, keeping amplifiers simple & good we often can end up
with an upper bandwidth 5-10 times higher than any man can hear
Your thinking on this is good & down to earth.
And this is one reason people build your amplifiers.
I know this is your philosophy when comes to Pass DIY designs.
Probably you have several exceptions in some of your experiments and commercial designs.
But a few exceptions to a rule, does not make the basic rule worse.
Exceptions does confirm that there is a General Rule.
Lineup
regardsF4 and Aikido
CVILLER
I want to build a F4 NP amplifier according to the schema and the PCB layout that you brought in post #163. But as a first step I want to use Aikido. This will give me a kind of a “Moskido” amplifier, very much identical to what Marsupialx suggest in the Moskido thread.
Have I understod this right that with an Aikido as a first step, I have to leave out Q101/102. R101/102/103/104 and C103/104 from the PCB? The signal out from Aikido shall then go directly to the “meeting point” of C101/102.
The PCB you suggest, can I buy it from anyone?
Sincerely
Eivind Stillingen, Norway
CVILLER
I want to build a F4 NP amplifier according to the schema and the PCB layout that you brought in post #163. But as a first step I want to use Aikido. This will give me a kind of a “Moskido” amplifier, very much identical to what Marsupialx suggest in the Moskido thread.
Have I understod this right that with an Aikido as a first step, I have to leave out Q101/102. R101/102/103/104 and C103/104 from the PCB? The signal out from Aikido shall then go directly to the “meeting point” of C101/102.
The PCB you suggest, can I buy it from anyone?
Sincerely
Eivind Stillingen, Norway
12sx7 and 6sn7 are variants of the same beast-imho the part count of a B1 and F5 is so much lower than an Aikido and F4 that they should have an advantage,and i think they do, on solo voices with one instrument it's a toss up but the more complex it gets the F5 just shows more detail and better placement---and on stuff that isn't "audiophile grade" the F5 shines a little more
F4/F5/Aikido
Thanks to Bubba 177, Jameshillj and tms for your answers. Your comments was interesting, even if they did not “hit” exact what I was asking for: What components in F4 do I have to drop out if I want Aikido to replace the first gain stage.
Bubba177, you have had the opportunity to compare F4 and F5. Your conclusion: the more complex it get-F5 give more details and better placement. Its a conclusion that, as far as can see, very much fits well to what NP said when he was asked the same question.
This thoughts have “given birth” to that I might reconsider F4 and go for F5 instead, but still with the Aikido as the first gain stage. Is it a good idea?? And again, what components shall I
“force out”, and where do I connect the signal from Aikido?
I have to mention that the signal to this mix of Aikido and F5 (or F4) will come from a Copland 306 pre, then into a very modded DCX 2496. My loudspeaker are Magnepan 1.6. From 28 to 60 I use NAD 208. From 60 to 600 I intend to use This F4/F5. The top range is taken care of with a modded Music Angel 845. My music taste is very wide, but very much into classics.
Any good advices from all the expert on this forum??
Sincerely
Eivind Stillingen, Norway
Thanks to Bubba 177, Jameshillj and tms for your answers. Your comments was interesting, even if they did not “hit” exact what I was asking for: What components in F4 do I have to drop out if I want Aikido to replace the first gain stage.
Bubba177, you have had the opportunity to compare F4 and F5. Your conclusion: the more complex it get-F5 give more details and better placement. Its a conclusion that, as far as can see, very much fits well to what NP said when he was asked the same question.
This thoughts have “given birth” to that I might reconsider F4 and go for F5 instead, but still with the Aikido as the first gain stage. Is it a good idea?? And again, what components shall I
“force out”, and where do I connect the signal from Aikido?
I have to mention that the signal to this mix of Aikido and F5 (or F4) will come from a Copland 306 pre, then into a very modded DCX 2496. My loudspeaker are Magnepan 1.6. From 28 to 60 I use NAD 208. From 60 to 600 I intend to use This F4/F5. The top range is taken care of with a modded Music Angel 845. My music taste is very wide, but very much into classics.
Any good advices from all the expert on this forum??
Sincerely
Eivind Stillingen, Norway
When i was building mine, i had the same issue with one source-resistor. It was because of improber-matching i think... I just swapped the IRF/IRFP, with another having the same Vgs, from my matching.
You say, that the difference is 15-20mV, which is not very much i guess. - I donno through.
Jesper.
I have two questions for the F4 experts here.
1) Which version of the F4 sounds best: a) the initial one (schematic in the manual), or b) the second one with the 220uf/1K networks connected to the outputs?
2) In the second version - the one with the 220uf/1K networks connected to the output - how high can the voltage rails safely get before the input JFET's are at risk of smoking? The transformers I have produce +-32VDC rails.
Thank you.
As background information I came home from BAF with a number of LONO transformers, courtesy of Nelson, with a 28V-0-28V @ 22A secondary and dual 0-100V-120V primaries. I would like to build two amplifiers - each with four F4 channels. These would be run as two channels of balanced in and out drive a pair of Magnepans bi-amped. There would be an electronic crossover in front of the amps.
Again, thanks to everyone.
Graeme
1) Which version of the F4 sounds best: a) the initial one (schematic in the manual), or b) the second one with the 220uf/1K networks connected to the outputs?
2) In the second version - the one with the 220uf/1K networks connected to the output - how high can the voltage rails safely get before the input JFET's are at risk of smoking? The transformers I have produce +-32VDC rails.
Thank you.
As background information I came home from BAF with a number of LONO transformers, courtesy of Nelson, with a 28V-0-28V @ 22A secondary and dual 0-100V-120V primaries. I would like to build two amplifiers - each with four F4 channels. These would be run as two channels of balanced in and out drive a pair of Magnepans bi-amped. There would be an electronic crossover in front of the amps.
Again, thanks to everyone.
Graeme
Hi Nelson,
Thank you. I realized last night that you are probably the only person in the world who has heard both versions.
I would like to preserve as much voltage swing as possible considering a) the hassle of adding heatsink to deal with extra dissipation, and b) wanting to take as much advantage as possible of these great BAF transformers. The preamp/crossover has 32V rails and is balanced out, not that I need to use all the voltage output swing available.
What voltage would you recommend I not exceed at the JFET drains?
Thank you again.
Graeme
Thank you. I realized last night that you are probably the only person in the world who has heard both versions.
I would like to preserve as much voltage swing as possible considering a) the hassle of adding heatsink to deal with extra dissipation, and b) wanting to take as much advantage as possible of these great BAF transformers. The preamp/crossover has 32V rails and is balanced out, not that I need to use all the voltage output swing available.
What voltage would you recommend I not exceed at the JFET drains?
Thank you again.
Graeme
Mr Pass
I would like to ask you about the single ended complementary follower 2sk170BL/2sj74BL, which you have been using in the F4 and F5 designs
I looked into the data sheets and 2sj74 has really diffrend transconductance curve then 2sk170 device.
I remember , in other famous thread someone stated :
“ .............. At a bias current of say 6-8mA, 2SK170V has a typical transconductance of 33mS, and 2SJ74 40mS. So ..you are correct in that the current contributions of the K device and the J device can be balanced by using a source resistor for the 2SJ74 which is 5 ohm higher than that of the 2SK170. But that does not mean that this guarantees optimal cancellation of (even order) distortions.
And of course, the Idss of the devices should be so chosen to give the same bias current with their corresponding source resistors. This means that their Idss should be different by a defined amount, depending on the value of the source resistors chosen.
So a simple 2-transistor circuit is not that simple after all..." ;
Mr J.Curl said about the sk170/sj74 follower :
„............ Well, now you know why I tend to use the matched pair N channel follower.........
The automatic DC offset tracking allowed input and output connections without coupling caps.
It might be possible to slightly offset the Idss's of the N and P channel parts to get good DC performance with identical source resistors, or perhaps find a slightly different value for the N device and the P device, that 'might' give both low distortion and low offset..........”
Based on those a.m. two statements , I would like to ask You what is the optimum Idss value of the sk170 / sj74 devices used in the input stage of the F4/F5 amps considering that the are BL grades and the source resistors are this same value (Rs= 10 R in yours designs )
Is that truth , sj74 device should have little higher Idss parameter compared to the sk170 device , or it is not so important issue in those designs ?
I would like to ask you about the single ended complementary follower 2sk170BL/2sj74BL, which you have been using in the F4 and F5 designs
I looked into the data sheets and 2sj74 has really diffrend transconductance curve then 2sk170 device.
I remember , in other famous thread someone stated :
“ .............. At a bias current of say 6-8mA, 2SK170V has a typical transconductance of 33mS, and 2SJ74 40mS. So ..you are correct in that the current contributions of the K device and the J device can be balanced by using a source resistor for the 2SJ74 which is 5 ohm higher than that of the 2SK170. But that does not mean that this guarantees optimal cancellation of (even order) distortions.
And of course, the Idss of the devices should be so chosen to give the same bias current with their corresponding source resistors. This means that their Idss should be different by a defined amount, depending on the value of the source resistors chosen.
So a simple 2-transistor circuit is not that simple after all..." ;
Mr J.Curl said about the sk170/sj74 follower :
„............ Well, now you know why I tend to use the matched pair N channel follower.........
The automatic DC offset tracking allowed input and output connections without coupling caps.
It might be possible to slightly offset the Idss's of the N and P channel parts to get good DC performance with identical source resistors, or perhaps find a slightly different value for the N device and the P device, that 'might' give both low distortion and low offset..........”
Based on those a.m. two statements , I would like to ask You what is the optimum Idss value of the sk170 / sj74 devices used in the input stage of the F4/F5 amps considering that the are BL grades and the source resistors are this same value (Rs= 10 R in yours designs )
Is that truth , sj74 device should have little higher Idss parameter compared to the sk170 device , or it is not so important issue in those designs ?