Hello
After some schematics essay and chating with a good friend, I've redone and simplified a single-ended amp using a standard LTP input type topology with a boot strap and also a phase lead at the VAS.
Single-ended amp compared to push-pull amp do have two big advantages, there is no cross-over distortions and no drivers/output transistors switching noises.
But to work correctly a single-ended amp need to work in class A, so it's mean a very high bias current and lot of heat. This is because the current is limited by the current source Q3 and if you have to low current you will not reach enough voltage for the output transistor and it will clip before reaching the max power that the rails voltage should permit.
But single-ended amp can sound excellent with a superb sound-stage.
So, for those who want to try this type of amp, here is my 35 watt single-ended amp, in my amp I've done a current source of about 2.7 amp, that should be quite enough for a load down to 7 ohm.
I should say that it's an amp for experiments diyer.
Since the bias are fixed by the current source, there is no need for a VBE multiplier.
You will need a really very BIG heat-sink.
Don't try to get more output power by raising the rails voltage, at higher power you will end-up with an horrible clipping sounds.
The total gain of that amp are arround 37x
It would sound better if you matched the two LTP input transistors for HFE and VBE.
You need to use a low cob and fast speed for Q5 the vas transistor, at least 120 mhz with a cob lower than 10 pF.
Here is the schematic.
Bye
Gaetan
After some schematics essay and chating with a good friend, I've redone and simplified a single-ended amp using a standard LTP input type topology with a boot strap and also a phase lead at the VAS.
Single-ended amp compared to push-pull amp do have two big advantages, there is no cross-over distortions and no drivers/output transistors switching noises.
But to work correctly a single-ended amp need to work in class A, so it's mean a very high bias current and lot of heat. This is because the current is limited by the current source Q3 and if you have to low current you will not reach enough voltage for the output transistor and it will clip before reaching the max power that the rails voltage should permit.
But single-ended amp can sound excellent with a superb sound-stage.
So, for those who want to try this type of amp, here is my 35 watt single-ended amp, in my amp I've done a current source of about 2.7 amp, that should be quite enough for a load down to 7 ohm.
I should say that it's an amp for experiments diyer.
Since the bias are fixed by the current source, there is no need for a VBE multiplier.
You will need a really very BIG heat-sink.
Don't try to get more output power by raising the rails voltage, at higher power you will end-up with an horrible clipping sounds.
The total gain of that amp are arround 37x
It would sound better if you matched the two LTP input transistors for HFE and VBE.
You need to use a low cob and fast speed for Q5 the vas transistor, at least 120 mhz with a cob lower than 10 pF.
Here is the schematic.
Bye
Gaetan
Attachments
Last edited:
Nice circuit, Gaetan. I think the sound of these amps is amazing.
I once built a water cooled SE amp with tube front end which ran at 9A and 70V - my wonderful daughter failed to turn on the water one day when she wanted to play it to her friends, you can imagine what happened.........
I think of this amp as the Manhattan Project!!
Enjoy the music, Gaetan, these things sound fantastic.
Hugh
I once built a water cooled SE amp with tube front end which ran at 9A and 70V - my wonderful daughter failed to turn on the water one day when she wanted to play it to her friends, you can imagine what happened.........
I think of this amp as the Manhattan Project!!
Enjoy the music, Gaetan, these things sound fantastic.
Hugh
Most people would not need that much gain.
The rule is to have no more gain than absolutely needed.
If we have sources with low signal output voltage, we build us a small nice preamplifier.
Electrically there is nothing that will improve with higher gain.
A bunch things in fact will be worse.
Now, people have different sources. So, some may want some extra gain. I know.
With supply 24 Volt, you can count on max 22 Vpeak out.
1 Vrms in is a good sensitivity for most modern sources.
This is 1.414 Vpeak
I need for my CD-player: 22/1.414 = Gain x15.5
R21 = 47k .... divided with 14.5 = 3.24k for R24
We can make R24 like 3.3k or 3k
The good thing is we can reduce C12 (33uF) electrolytic.
We can make it 10 uF, and still have same low freq roll-off.
Which means we can start use film capacitors.
Either one 10uF, or parallell two 4.7uF, which gives C12 = 9.4uF
Thanks for publishing you nice idea.
It is straight forward and Class A is still the King of Performance!
35W into 7r0 requires 3.16Apk and 22.1Vpk.
2.7A bias is not enough for a 7ohm load.
Most people would not need that much gain.
The rule is to have no more gain than absolutely needed.
If we have sources with low signal output voltage, we build us a small nice preamplifier.
Electrically there is nothing that will improve with higher gain.
A bunch things in fact will be worse.
Now, people have different sources. So, some may want some extra gain. I know.
Thanks for publishing you nice idea.
It is straight forward and Class A is still the King of Performance!
Hello
I've experimented myself that the less global negative feedback you have the better the sound are, it would not allways be big differences but there is a little one for most amps topologies. So I alway give less global negative feedback in my amps, and yes it's give more gain.
Most of the times I was fixing amps gain at arround 30 db, but now I go a bit higher and wen you reduce the global negative feedback you can reduce a bit the value of the cdom cap.
Thank
Bye
Gaetan
This is precisely why JLH set many of his amplifiers to what in modern terms are very high gains.
Leach also alludes to this for better accuracy/sound quality.
Hello Andrew
Yes, very low GNFB reduce the spray of odd high harmonics thd in the spectrum and making a lower nfb will need lower cdom cap for the VAS and it's do a better Slew Rate.
Btw, you was right, my current source was too low for a 7 ohm load, I did not calculate it for PP output, that was my error.
Thank
Bye
Gaetan
Yes, very low GNFB reduce the spray of odd high harmonics thd in the spectrum and making a lower nfb will need lower cdom cap for the VAS and it's do a better Slew Rate.
Btw, you was right, my current source was too low for a 7 ohm load, I did not calculate it for PP output, that was my error.
Thank
Bye
Gaetan
Last edited:
That is why all my mongrels are set at 38DB ! Both my sound blaster and my auzentech sound cards can barely muster 1V OP, so I might take lineups advice and build a small preamp with a gain of 2 (adjustable of course).
OS
we can go both ways:
1. use one preamplifier with hardly any distortion at all (OPA134)
and use a power amplifier with sensitivity 1 Vrms for max nominal output
2. go without preamplifier
and use power amplifier with much higher sensitivity, say 0.2 Vrms
in these 2 cases power amplifier in case 2 needs five times more voltage gain
with probably more transistors for added gain
which maybe could render like 5 times less bandwidth, if we take mathematically
and no doubt also will give more distortion
my way, and the normal way, would be solution 1.
it is like you should amplify phono signal 400 times
and use one OP-Amp x400
or use two OP-Amps in series x20 x20,
the difference in outcome would be like a mountain!
1. use one preamplifier with hardly any distortion at all (OPA134)
and use a power amplifier with sensitivity 1 Vrms for max nominal output
2. go without preamplifier
and use power amplifier with much higher sensitivity, say 0.2 Vrms
in these 2 cases power amplifier in case 2 needs five times more voltage gain
with probably more transistors for added gain
which maybe could render like 5 times less bandwidth, if we take mathematically
and no doubt also will give more distortion
my way, and the normal way, would be solution 1.
it is like you should amplify phono signal 400 times
and use one OP-Amp x400
or use two OP-Amps in series x20 x20,
the difference in outcome would be like a mountain!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.