Anyone have comments about the tube hybrid amp in the September 06 issue of audioXpress? It looks like another variation of the JLH design. www.bdent.com carries most of the parts and the PCB looks simple enough. I'm a tube guy but would like to give silicon a try, maybe use an op-amp driver instead of tubes.
Hi all
I suppose the hybrid amplifier is the one described here: http://www.wdehaan.demon.nl/
The amplifier is not yet described in full detail on the web site.
Torben
I suppose the hybrid amplifier is the one described here: http://www.wdehaan.demon.nl/
The amplifier is not yet described in full detail on the web site.
Torben
Hmmmmm.... there are a couple of problems with that circuit, mostly easily solvable.
1) ECC88 on a 310V supply... hm, marginal, either one or the other triode will be very close to maximum Vak. With ECC82, a rather high input voltage is needed for full swing. Also, ECC99, 6CG7/6FQ7 would perhaps be interesting tubes to try.
2) P1 and Q1 - how many times must it be written?! This way of doing it is DANGEROUS - even just moving the pot might blow up the output stage. You never ever want to lose the base connection on the bias servo transistor, and secondly, you don't want to have a possibility to set infinite bias current, cause you REALLY don't need it! It needs to be a fixed resistor from C to B of Q1, and a pot from B to E, with wiper connected via fixed resistor to either B or E of Q1. Chose resistors to provide a sensible bias current range.
3) P2, R7, R10 (guessing on the last one, it's not really readable) - Not sure if R7 is 10 or 18k - with 10k it is impossible to get zero DC at output because P2+R7 should be about equal to R10 when P2 is at mid position. R7 should be 18k - of course, I may be reading it wrong (it's quite blurry). The bias current and DC offset will be dependant on the +-35V power rail regulation, which is not good but not overly problematic either - a pair of well placed zener diodes would solve this problem. The input impedance will be rather low - at most 10k. This should be as high as possible, and it's not too difficult to do by splitting R7 and R10 (R7 into 10k + 8.2k, R10 into 10k + 10k), and providing a bootstrap via sufficiantly large caps from the output.
4) Why quasicomp? Not really a criticism, but I would like to know
5) It is possible to get the cathode follower to work into a virtual current source by connecting the ground end of R6 to a different place, I'll let readers figure that one out
1) ECC88 on a 310V supply... hm, marginal, either one or the other triode will be very close to maximum Vak. With ECC82, a rather high input voltage is needed for full swing. Also, ECC99, 6CG7/6FQ7 would perhaps be interesting tubes to try.
2) P1 and Q1 - how many times must it be written?! This way of doing it is DANGEROUS - even just moving the pot might blow up the output stage. You never ever want to lose the base connection on the bias servo transistor, and secondly, you don't want to have a possibility to set infinite bias current, cause you REALLY don't need it! It needs to be a fixed resistor from C to B of Q1, and a pot from B to E, with wiper connected via fixed resistor to either B or E of Q1. Chose resistors to provide a sensible bias current range.
3) P2, R7, R10 (guessing on the last one, it's not really readable) - Not sure if R7 is 10 or 18k - with 10k it is impossible to get zero DC at output because P2+R7 should be about equal to R10 when P2 is at mid position. R7 should be 18k - of course, I may be reading it wrong (it's quite blurry). The bias current and DC offset will be dependant on the +-35V power rail regulation, which is not good but not overly problematic either - a pair of well placed zener diodes would solve this problem. The input impedance will be rather low - at most 10k. This should be as high as possible, and it's not too difficult to do by splitting R7 and R10 (R7 into 10k + 8.2k, R10 into 10k + 10k), and providing a bootstrap via sufficiantly large caps from the output.
4) Why quasicomp? Not really a criticism, but I would like to know
5) It is possible to get the cathode follower to work into a virtual current source by connecting the ground end of R6 to a different place, I'll let readers figure that one out
copyrigt issue
I agree, we should respect the copyright issue. The hybrid design is published in latest issue of AudioXpress and some people (like in Europe) didn't even receive it yet and I am one of them! So it will be a bomber to them that it already can be found on some discussion bulletin. Why to have a subscription if you can download it from the net? A magazine like AudioXpress can not survive if people just put part of articles (or schematics) on the net just like that.
About the ECC88 issue. The article mentions the use of an E88CC and according to the datasheet provided by Philips, Telefunken or JJ-Electronic the max. anode voltage of an E88CC is 220V. You might like to check out http://www.jj-electronic.sk/pdf/E88CC.pdf and you can run it even up to 250V if you limit the power dissipation to 0.8 watt. So its no issue to use the E88CC at say 150VDC or so.
About the bias current potentiometer. I do work in a high-tech medical environment and the high-quality 10/15 turns potentiometers are used throughout many high-tech designs like life-support systems and causes no problem what-so-ever, we should not be afraid that one of them as used in the bias circuit will fail.
About why quasicomp? The answer can be found in the article ;-)
About the input impedance will be rather low - at most 10k. The driver is capable of 20Veff to drive the amplifier to full power and this with low distortion. Only disadvantage can be the high value (> 3 uF) of coupling caps. But quality capacitors like ClarityCaps and ICEL PHC types offer good sound with a nice price tag.
Wim de Haan
Author of the hybrid amplifier as mentioned in the AudioXpress September 06 issue.
I agree, we should respect the copyright issue. The hybrid design is published in latest issue of AudioXpress and some people (like in Europe) didn't even receive it yet and I am one of them! So it will be a bomber to them that it already can be found on some discussion bulletin. Why to have a subscription if you can download it from the net? A magazine like AudioXpress can not survive if people just put part of articles (or schematics) on the net just like that.
About the ECC88 issue. The article mentions the use of an E88CC and according to the datasheet provided by Philips, Telefunken or JJ-Electronic the max. anode voltage of an E88CC is 220V. You might like to check out http://www.jj-electronic.sk/pdf/E88CC.pdf and you can run it even up to 250V if you limit the power dissipation to 0.8 watt. So its no issue to use the E88CC at say 150VDC or so.
About the bias current potentiometer. I do work in a high-tech medical environment and the high-quality 10/15 turns potentiometers are used throughout many high-tech designs like life-support systems and causes no problem what-so-ever, we should not be afraid that one of them as used in the bias circuit will fail.
About why quasicomp? The answer can be found in the article ;-)
About the input impedance will be rather low - at most 10k. The driver is capable of 20Veff to drive the amplifier to full power and this with low distortion. Only disadvantage can be the high value (> 3 uF) of coupling caps. But quality capacitors like ClarityCaps and ICEL PHC types offer good sound with a nice price tag.
Wim de Haan
Author of the hybrid amplifier as mentioned in the AudioXpress September 06 issue.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.