O.k. there are plenty of headamps and plenty of f5 Variants too...
But I wanted a headamp and I like Pass Amps, who does not?
After fiddling around with several variants of the Aleph-J scaled for headphones in simulation, I tried out a f5 and scaled it down for my needs.
The gain was too high for my needs, I wanted around 10dB. Increasing the feedback did lead into instability, so I reduced the gain of the first stage like Nelson does in the BA-3.
Looks like this
Not much parts, easy to build e.g. on stripeboard. My version looks like this:
But of course it could be wired point to point or one could use a BA-3 board from the forum store (and a litte fantasy).
And it works like a charm. Even without optimising P3, its set to the middle position at the moment.
Bias is approx 200mA and offset below 1mV. But it takes some time till the amp reaches its values. Perhaps I should try the NTC à la F5 to speed up warming up a little and give some more temperature stability. But I have no idea how to size it right now. Any hints?
Best regards
Floric
But I wanted a headamp and I like Pass Amps, who does not?
After fiddling around with several variants of the Aleph-J scaled for headphones in simulation, I tried out a f5 and scaled it down for my needs.
The gain was too high for my needs, I wanted around 10dB. Increasing the feedback did lead into instability, so I reduced the gain of the first stage like Nelson does in the BA-3.
Looks like this
Not much parts, easy to build e.g. on stripeboard. My version looks like this:
But of course it could be wired point to point or one could use a BA-3 board from the forum store (and a litte fantasy).
And it works like a charm. Even without optimising P3, its set to the middle position at the moment.
Bias is approx 200mA and offset below 1mV. But it takes some time till the amp reaches its values. Perhaps I should try the NTC à la F5 to speed up warming up a little and give some more temperature stability. But I have no idea how to size it right now. Any hints?
Best regards
Floric
Attachments
Addendum: If someone else did anything like this and posted it already - sorry, I did not find it.
Changed the feedback to 330R. I wanted a little more gain because some of my sources have a really low output.
The supply is still +- 12V supplied by lm317/lm337 at the moment. O.k that's the lazy solution but it's clean enough and provides current limitation in the case of a fault. The bias is set to 180mA now, this value was randomly chosen.
I made some experiments with the setting of P3 and had a look at the distortion. Don't want to say that I measured it because it was viewed using Mulitisine and Audioanalyser with the analog outputs and inputs of my soundcard (ESI Juli@) at 44,1 kHz sampling rate and I'm not sure if this setup could stand quantitative proofs. But you can see the effect quite well.
In the attechments you see the display of the distortion spectrum of a 1kHz sine with an output of 0,5V rms into a resistive load of 33R. Fist image shows the "miniumum" setting where 2nd and 3rd harmonics are approximately equal. The second image shows the setting that I prefered while listening with dominant 2nd harmonics and relatively low higher order ones.
For listening I use a Beyerdynamic DT880pro with 250R, with this load the distortion is lower.
Have fun
Flo
The supply is still +- 12V supplied by lm317/lm337 at the moment. O.k that's the lazy solution but it's clean enough and provides current limitation in the case of a fault. The bias is set to 180mA now, this value was randomly chosen.
I made some experiments with the setting of P3 and had a look at the distortion. Don't want to say that I measured it because it was viewed using Mulitisine and Audioanalyser with the analog outputs and inputs of my soundcard (ESI Juli@) at 44,1 kHz sampling rate and I'm not sure if this setup could stand quantitative proofs. But you can see the effect quite well.
In the attechments you see the display of the distortion spectrum of a 1kHz sine with an output of 0,5V rms into a resistive load of 33R. Fist image shows the "miniumum" setting where 2nd and 3rd harmonics are approximately equal. The second image shows the setting that I prefered while listening with dominant 2nd harmonics and relatively low higher order ones.
For listening I use a Beyerdynamic DT880pro with 250R, with this load the distortion is lower.
Have fun
Flo
Attachments
I don't know, what went wrong the last time. The last days I have been fiddling around with p3 a little bit. Suddenly when I wanted to take a new measurement, the higher order hamonics vanished. They must have been an error in my measurement setup. Now my favourite harmonic distortion looks like attached.
It seems funny: Settings with lower THD are possible with this amp, but I prefer some distortion over my headphones. Even if its as low as 0,03%, it seems more musical to my ears.
But even without those settings, I think this is a really good little amp. It sounds good and it' thermally stable.
Best regards
Flo
It seems funny: Settings with lower THD are possible with this amp, but I prefer some distortion over my headphones. Even if its as low as 0,03%, it seems more musical to my ears.
But even without those settings, I think this is a really good little amp. It sounds good and it' thermally stable.
Best regards
Flo
Attachments
Just to document the whole thing:
Concerning the thermal stability of this amp, I was fooled by the thermal characteristics of my transistors: I adjusted bias and dc offset when the case was open. After that I used the amp with an open case for a wile, made some measurements and tests (e.g. the graphs above) with the open case (on my desk that does nor matter).
After a while of using the amp, just after my last post, I wanted to find a place for it and closed the lid (only little ventilation). I was sure that the temperature in the case would rise when it's closed so I controlled the dc-offset regularly, it only changed by a mV. So I thought that the amp was thermally stable. But then I recognized, that the whole case got a little too hot for my feeling. I opened the case and measured: In the "hot" case, I have a bias of more than 300mA even though I adjusted it to 200mA in the open case!
In my opinion, it's a matter of chance that the offset remained almost stable with this change of the bias.
To be sure that the values won't change too much in the future even if the sun shines on the case, I turn on the heating, a magazine is placed on the case etc., I implemented two thermistors like Nelson did it in the f5-turbo. After that the adjustment of the amp got a little more difficult but now it remains relatively stable over a range of temperatures (~10°C to ~30°C) if the lid is closed or not. The dc offset varies 1-2mV and the bias changes some mA, nothing to worry about.
So I strongly recommend to use the schema below to anyone who wants to build this amp and couple the thermistors thermally to the output transistors.
The theristor are standard 4,7k NTC types. I don't notice a change of the sound, still very good.
Best regards
Flo
Concerning the thermal stability of this amp, I was fooled by the thermal characteristics of my transistors: I adjusted bias and dc offset when the case was open. After that I used the amp with an open case for a wile, made some measurements and tests (e.g. the graphs above) with the open case (on my desk that does nor matter).
After a while of using the amp, just after my last post, I wanted to find a place for it and closed the lid (only little ventilation). I was sure that the temperature in the case would rise when it's closed so I controlled the dc-offset regularly, it only changed by a mV. So I thought that the amp was thermally stable. But then I recognized, that the whole case got a little too hot for my feeling. I opened the case and measured: In the "hot" case, I have a bias of more than 300mA even though I adjusted it to 200mA in the open case!
In my opinion, it's a matter of chance that the offset remained almost stable with this change of the bias.
To be sure that the values won't change too much in the future even if the sun shines on the case, I turn on the heating, a magazine is placed on the case etc., I implemented two thermistors like Nelson did it in the f5-turbo. After that the adjustment of the amp got a little more difficult but now it remains relatively stable over a range of temperatures (~10°C to ~30°C) if the lid is closed or not. The dc offset varies 1-2mV and the bias changes some mA, nothing to worry about.
So I strongly recommend to use the schema below to anyone who wants to build this amp and couple the thermistors thermally to the output transistors.
The theristor are standard 4,7k NTC types. I don't notice a change of the sound, still very good.
Best regards
Flo
Attachments
I did another thing during the last weeks: being offended by the "too much in head" sound of some recordings, I made some experiments with crossfeed circuits. I came to the conclusion, that the best compromise for my ears is the Meier-crossfeed. The other ones I tried were different, the one Rod Elliott published on his site e.g. is a little better concerning the "out of head" localization but it changes the sound too much.
The one I finally implemented was the Meier-crossfeed with buffers before and after it. The buffers are JFET common drain with css, a little like the DC-B1 but using the j310 JFETs with reduced bias. See the schema below.
In my implementation, the crossfeed-Resistor (R28 2,2k in the schema) can be switched between 2,2k, 6,8k and infinity (open connection) with a rotary switch, so I can adjust the stereo image from "wide and in my head" (open connection) to "narrow and before my head" (2,2k) with one step between according to the music I listen to.
In "open" position, the crossfeed does not change the sound compared to the amp without crossfeed.
I like it.
best regards
Flo
The one I finally implemented was the Meier-crossfeed with buffers before and after it. The buffers are JFET common drain with css, a little like the DC-B1 but using the j310 JFETs with reduced bias. See the schema below.
In my implementation, the crossfeed-Resistor (R28 2,2k in the schema) can be switched between 2,2k, 6,8k and infinity (open connection) with a rotary switch, so I can adjust the stereo image from "wide and in my head" (open connection) to "narrow and before my head" (2,2k) with one step between according to the music I listen to.
In "open" position, the crossfeed does not change the sound compared to the amp without crossfeed.
I like it.
best regards
Flo
Attachments
Adjusting P3 changes not only the harmonic balance, but also the output voltage offset.
Did your preferred setting result in higher or lower offset?
/R
P.S. I like the buffered crossfeed. After all, this kind of application is what buffers live for, might as well make good use of them.
Did your preferred setting result in higher or lower offset?
/R
P.S. I like the buffered crossfeed. After all, this kind of application is what buffers live for, might as well make good use of them.
Hello Richard,
I always adjusted the offset after turning P3, so the offset was around 1mV in every case.
They do the job really well and you can place them where you want. I already built active crossovers like that (passive eq between buffers). This version with the j310 is even cheap.
Flo
I always adjusted the offset after turning P3, so the offset was around 1mV in every case.
They do the job really well and you can place them where you want. I already built active crossovers like that (passive eq between buffers). This version with the j310 is even cheap.
Flo
- Status
- Not open for further replies.