PS: I had a brief discussion with Geoff Moss about the possibility of building a next generation JLH without global feedback. I hope to generate some debate about this issue on this site. Hence the following quotes of our discussion.
Hello Geoff,
Yesterday Jos de Weerdt has visited my place. He brought his new JLH monoblocks with him and we used the entire afternoon to compare his version with mine. Happily there were barely differences to be noted between the two versions. His version had a slightly firmer grip on the lows, my version sounded more open (only slightly so again).
We noticed much more differences between the JLH's and the Hiraga. Not that we could conclude that the one was better than the other, but there were differences also heard by Jos. The Hiraga also sounds more open, the JLH more shut-in, especially to be heard on piano and female voices.
Jos will make a report on this afternoon on his site.
Following this afternoon, I wondered:
1. Do you know how much dB of negative global feedback has the JLH? I remember the Hiraga has around 16 dB of global feedback. I guess the JLH has more of it?
2. To your knowledge, has anybody experimented with the amount of negative feedback of the JLH? In reality of by simulation? Is it possible to change the R8 resistor of 2k7 to a value of say 100k? Can the JLH be built without (global) negative feedback at all, just for the sake of the experiment?
I recently heard an old Akai AM-U5 amplifier that is one of the very few commercial amplifiers ever made without feedback (so called open loop amplifier). It has something special in its sound. Also the single-ended amplifier I built with the 3055 is open loop and also sounds 'free' and open.
Just some thoughts.
Rudy van Stratum
Hello Geoff,
Yesterday Jos de Weerdt has visited my place. He brought his new JLH monoblocks with him and we used the entire afternoon to compare his version with mine. Happily there were barely differences to be noted between the two versions. His version had a slightly firmer grip on the lows, my version sounded more open (only slightly so again).
We noticed much more differences between the JLH's and the Hiraga. Not that we could conclude that the one was better than the other, but there were differences also heard by Jos. The Hiraga also sounds more open, the JLH more shut-in, especially to be heard on piano and female voices.
Jos will make a report on this afternoon on his site.
Following this afternoon, I wondered:
1. Do you know how much dB of negative global feedback has the JLH? I remember the Hiraga has around 16 dB of global feedback. I guess the JLH has more of it?
2. To your knowledge, has anybody experimented with the amount of negative feedback of the JLH? In reality of by simulation? Is it possible to change the R8 resistor of 2k7 to a value of say 100k? Can the JLH be built without (global) negative feedback at all, just for the sake of the experiment?
I recently heard an old Akai AM-U5 amplifier that is one of the very few commercial amplifiers ever made without feedback (so called open loop amplifier). It has something special in its sound. Also the single-ended amplifier I built with the 3055 is open loop and also sounds 'free' and open.
Just some thoughts.
Rudy van Stratum
reply of geoff (edited)
Hi Rudy
The open-loop gain of the JLH depends on the Hfe of the transistors used and on the load impedance. Usually it will be about 65dB (+/-5dB). The closed-loop gain is 23dB so the global feedback is about 2.5 times that of the Hiraga.
I am not aware of anyone making anything other than small changes to the closed-loop gain. Your suggestion of increasing R8 to 100k will reduce the feedback factor so that it becomes similar to that of the Hiraga but it will also increase the THD and input sensitivity by a factor of about 20. With a 50mV input sensitivity, a lot of attenuation will be needed between the source and the amps and there will be a far greater likelihood of instability due to extraneous coupling between the output and the input.
A more suitable way of lowering the feedback factor would be to reduce the open-loop gain by using low Hfe transistors (though this would increase the
%THD) or by adding emitter degeneration resistors to the input and driver tr ansistors. A quick simulation indicates that this might even improve the THD but I would need to do a lot more work on this before I could be sure that all was well. I will add this to my list of "things to do when time permits".
Geoff
Hi Rudy
The open-loop gain of the JLH depends on the Hfe of the transistors used and on the load impedance. Usually it will be about 65dB (+/-5dB). The closed-loop gain is 23dB so the global feedback is about 2.5 times that of the Hiraga.
I am not aware of anyone making anything other than small changes to the closed-loop gain. Your suggestion of increasing R8 to 100k will reduce the feedback factor so that it becomes similar to that of the Hiraga but it will also increase the THD and input sensitivity by a factor of about 20. With a 50mV input sensitivity, a lot of attenuation will be needed between the source and the amps and there will be a far greater likelihood of instability due to extraneous coupling between the output and the input.
A more suitable way of lowering the feedback factor would be to reduce the open-loop gain by using low Hfe transistors (though this would increase the
%THD) or by adding emitter degeneration resistors to the input and driver tr ansistors. A quick simulation indicates that this might even improve the THD but I would need to do a lot more work on this before I could be sure that all was well. I will add this to my list of "things to do when time permits".
Geoff
reply of millwood in other thread
rmgvs:
since you seem to be interested in making an amp that sounds as good as the JLHs of the world, without global feedback, i would submit to you just the JLH without global feedback, .
it is identical to the JLH1969 (with four transistors, tho. But you can take out either the upper or lower output device if you truly want to make it a three transistor design). the only modifications: take out the lower feedback resistor (220ohm in the JLH design) and upside the DC blocking cap to as large as you can get (3300uf is good enough per my simulation). Here are some performance figures:
================================
JLH1969:
-3db: about 15khz;
gain: 750x;
THD: 0.79%
================================
If you a llow MOSFET, you can replace the output devices (2n3055x2) with MOSFET (I used irf540x2), and run the driver a little bit hotter (the 2.2k resistor on the phase splitter is downsized to 110ohm, as the lower resistor (560ohm down to 110ohm). Here is its performance for the MOSFET version:
================================
-3db: about 25khz;
gain: 2500;
THD: 0.53%
================================
Not bad, isn't?
rmgvs:
since you seem to be interested in making an amp that sounds as good as the JLHs of the world, without global feedback, i would submit to you just the JLH without global feedback, .
it is identical to the JLH1969 (with four transistors, tho. But you can take out either the upper or lower output device if you truly want to make it a three transistor design). the only modifications: take out the lower feedback resistor (220ohm in the JLH design) and upside the DC blocking cap to as large as you can get (3300uf is good enough per my simulation). Here are some performance figures:
================================
JLH1969:
-3db: about 15khz;
gain: 750x;
THD: 0.79%
================================
If you a llow MOSFET, you can replace the output devices (2n3055x2) with MOSFET (I used irf540x2), and run the driver a little bit hotter (the 2.2k resistor on the phase splitter is downsized to 110ohm, as the lower resistor (560ohm down to 110ohm). Here is its performance for the MOSFET version:
================================
-3db: about 25khz;
gain: 2500;
THD: 0.53%
================================
Not bad, isn't?
reply of millwood (2) in other thread
the high open loop gain (750x in my simulation) does need to be address. a few thoughts:
1) lower the collector resistor on the input stage;
2) lower the collector and bootstrap resistors on the phase splitter stage. Be careful however as you may accidentally increase the bias on theoutput stage;
3) use degenerative resistors throughout the amp.
4) use mosfets or jfets as they have lower gains.
5) get rid of the bootstrap on the phase splitter.
6) use a power resistor in serial with the load.
adding attenuation isn't bad idea as long as it isn't too much.
I think it can be made to work. Just how one wants it to be done.
the high open loop gain (750x in my simulation) does need to be address. a few thoughts:
1) lower the collector resistor on the input stage;
2) lower the collector and bootstrap resistors on the phase splitter stage. Be careful however as you may accidentally increase the bias on theoutput stage;
3) use degenerative resistors throughout the amp.
4) use mosfets or jfets as they have lower gains.
5) get rid of the bootstrap on the phase splitter.
6) use a power resistor in serial with the load.
adding attenuation isn't bad idea as long as it isn't too much.
I think it can be made to work. Just how one wants it to be done.
Hi Rudy,
in open loop mode (R8 = endless) the JLH Amp works like a big current controlled current source with a high output impedance.
-> Due to the high output impedance the frequency range of your system will strongly depend on the input impedance of your speaker.
-> The over all feedback converts this amp to a voltage source: T4 work's like a "inverse" voltage follower in the feedback path where the voltage over R6 follows the input woltage on the base of T4.
To get a low nfb amp I propose following aproches to think about:
1- leave R8 and use an output transformer as used in a tube amp: you have to redesign the whole cirquit for a high rail voltage to produce the same output power.
2- reduce the open loop gain of the cirquit: use "low betta" types for T3, decrease R5 (based on the latest amp version). I calculated the open loop gain as you find it in the attached file (latest 10 W version of the amp).
rx = base to emitter resistance of Tx
Bx= Betto of Tx
Rx = Resistor
Regards
Dani66
in open loop mode (R8 = endless) the JLH Amp works like a big current controlled current source with a high output impedance.
-> Due to the high output impedance the frequency range of your system will strongly depend on the input impedance of your speaker.
-> The over all feedback converts this amp to a voltage source: T4 work's like a "inverse" voltage follower in the feedback path where the voltage over R6 follows the input woltage on the base of T4.
To get a low nfb amp I propose following aproches to think about:
1- leave R8 and use an output transformer as used in a tube amp: you have to redesign the whole cirquit for a high rail voltage to produce the same output power.
2- reduce the open loop gain of the cirquit: use "low betta" types for T3, decrease R5 (based on the latest amp version). I calculated the open loop gain as you find it in the attached file (latest 10 W version of the amp).
rx = base to emitter resistance of Tx
Bx= Betto of Tx
Rx = Resistor
Regards
Dani66
Attachments
hi rmgvs
I recently bought Akai AM-U5. It is working but not in good condition. There is a blown HAxxxx integrated 8 pin circuit. It was some kind of a protection (dc protection maybe) i belive..it controls the output relais.
Porhaps you can remember the name or hafve the schematic of this amp???
The operational amplifier is some kind of JRC. Is it original?
Thanks in advance. last hope.
I recently bought Akai AM-U5. It is working but not in good condition. There is a blown HAxxxx integrated 8 pin circuit. It was some kind of a protection (dc protection maybe) i belive..it controls the output relais.
Porhaps you can remember the name or hafve the schematic of this amp???
The operational amplifier is some kind of JRC. Is it original?
Thanks in advance. last hope.
Attachments
John Linsley Hood
I used to work in the audio trade and met John Linsley Hood back in the 1970's ( His nephew worked in my shop in London )
I visited a Hi Fi trade show with John Linsley Hood and his nephew back in the late 1970s.
Being a buyer in the hi-fi trade I was spoilt for choice when it came to hi fi equipment, with many deals and offers to be had with suppliers.
You for example might get a company like Ortofon offering their latest and most exotic MCC but with a price tag of over £200 which was a lot of money back in the 1970/80's. If you could buy one for £45 with a MCC preamp you might just buy one. I would point to that when Moving Coil Cartages were first produced they had such a low output and ultra high compliance they needed and addition preamplifier because you could not plug them straight into a Moving Magnet phono input.
The company I worked for also produced the Jim Rogers JR loudspeakers.
They were a good deal at £30 a pair to company employees. We also made the JR150 sub woofer and distributed products like Koss headphones.
Anyway~ I was also interested in consumer electronics from a constructor / student perspective I needed test equipment.
Because John was primarily an instrumentation designer I built both his high frequency Millivolt Meter design and also his audio signal generator design.
Kits were available.
I also attempted to build his ultra low distortion phono preamp design but the PCB offered by a kit company never worked out so I abandoned the project.
Anyway here I am 40 years on and attempting to revisit my youth and wanting to make something again.
Having worked in a school Design Technology Department for 10 years I learnt how to produce and etch my own Printed Circuit boards.
On the down side I am being hampered by Multiple Sclerosis and having other issues in my life to distract me I don't know how successful I will be in this project.
I can still solder and cut holes in metal with an abundance of workshop tools and equipment.
It's more to do with cognitive issue of motivation. ( Everything is a good idea )
It would be nice to hear from a fellow audio enthusiast in the London area with similar ambitions.
Jeremy Clayton-Travis
I used to work in the audio trade and met John Linsley Hood back in the 1970's ( His nephew worked in my shop in London )
I visited a Hi Fi trade show with John Linsley Hood and his nephew back in the late 1970s.
Being a buyer in the hi-fi trade I was spoilt for choice when it came to hi fi equipment, with many deals and offers to be had with suppliers.
You for example might get a company like Ortofon offering their latest and most exotic MCC but with a price tag of over £200 which was a lot of money back in the 1970/80's. If you could buy one for £45 with a MCC preamp you might just buy one. I would point to that when Moving Coil Cartages were first produced they had such a low output and ultra high compliance they needed and addition preamplifier because you could not plug them straight into a Moving Magnet phono input.
The company I worked for also produced the Jim Rogers JR loudspeakers.
They were a good deal at £30 a pair to company employees. We also made the JR150 sub woofer and distributed products like Koss headphones.
Anyway~ I was also interested in consumer electronics from a constructor / student perspective I needed test equipment.
Because John was primarily an instrumentation designer I built both his high frequency Millivolt Meter design and also his audio signal generator design.
Kits were available.
I also attempted to build his ultra low distortion phono preamp design but the PCB offered by a kit company never worked out so I abandoned the project.
Anyway here I am 40 years on and attempting to revisit my youth and wanting to make something again.
Having worked in a school Design Technology Department for 10 years I learnt how to produce and etch my own Printed Circuit boards.
On the down side I am being hampered by Multiple Sclerosis and having other issues in my life to distract me I don't know how successful I will be in this project.
I can still solder and cut holes in metal with an abundance of workshop tools and equipment.
It's more to do with cognitive issue of motivation. ( Everything is a good idea )
It would be nice to hear from a fellow audio enthusiast in the London area with similar ambitions.
Jeremy Clayton-Travis
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.