Asen
I assume your suggestion of using 'two amps in parallel' to obviate the need for an output capacitor refers to using them in a bridging arrangement where the speaker is connected between the two amplifier outputs.
I would strongly advise against using this particular amplifier in a bridging circuit without including the output capacitors. In my experience, the dc offset variations are such that to do so would be asking for trouble. The dc offset variation from a single amp is of concern but if you have two dc offsets, quite possibly of opposite polarity, the problem is considerably greater.
The dc offset drifts as the amp warms up. If set with the amp warm, the initial switch on offset can be quite high. Also bear in mind that the dc offset varies with the quiescent current and the quiescent current varies with the supply rail voltage and with transistor temperatures. The supply rail voltage will vary throughout the day as the mains supply fluctuates (unless you are using a regulated supply) and the transistor temperatures will vary with ambient temperature changes.
I have been seriously considering adding a dc servo to my amps during the current rebuild to overcome this shortcoming.
In my opinion, bridge circuits are really only suitable for amplifiers with good dc offset control, i.e. those with a well balanced differential (long-tailed pair) input or those with a dc servo control.
Geoff
I assume your suggestion of using 'two amps in parallel' to obviate the need for an output capacitor refers to using them in a bridging arrangement where the speaker is connected between the two amplifier outputs.
I would strongly advise against using this particular amplifier in a bridging circuit without including the output capacitors. In my experience, the dc offset variations are such that to do so would be asking for trouble. The dc offset variation from a single amp is of concern but if you have two dc offsets, quite possibly of opposite polarity, the problem is considerably greater.
The dc offset drifts as the amp warms up. If set with the amp warm, the initial switch on offset can be quite high. Also bear in mind that the dc offset varies with the quiescent current and the quiescent current varies with the supply rail voltage and with transistor temperatures. The supply rail voltage will vary throughout the day as the mains supply fluctuates (unless you are using a regulated supply) and the transistor temperatures will vary with ambient temperature changes.
I have been seriously considering adding a dc servo to my amps during the current rebuild to overcome this shortcoming.
In my opinion, bridge circuits are really only suitable for amplifiers with good dc offset control, i.e. those with a well balanced differential (long-tailed pair) input or those with a dc servo control.
Geoff
Geoff,
I was looking at servo circuits myself, take a look at this site:
http://www.danbbs.dk/~ko/dcservo.htm
Since I don't know much about this type of circuit I played around with the program for a while and I was left with mixed feelings about it. Specifically, to get a short time delay you need to compromise the cutoff frequency, I don't even know if I am looking at this right, I am guessing that you want a fast response of the circuit to DC. What do you think? Do you have some idea on how make a better circuit?
I was looking at servo circuits myself, take a look at this site:
http://www.danbbs.dk/~ko/dcservo.htm
Since I don't know much about this type of circuit I played around with the program for a while and I was left with mixed feelings about it. Specifically, to get a short time delay you need to compromise the cutoff frequency, I don't even know if I am looking at this right, I am guessing that you want a fast response of the circuit to DC. What do you think? Do you have some idea on how make a better circuit?
Grataku
Thanks for the URL. I've seen that site before and have a copy of the program. It is a pity that the author didn't include the design criteria and calculation details instead of just a circuit and the calculation results. The circuit seems more complex than others I have seen and I am still trying to work out what certain parts are intended to achieve.
One thing you should be aware of, if you are not already. When dc servos are used with voltage feedback amps (i.e. those with a differential (long-tailed pair) input stage) it is possible to do away with the dc blocking capacitor in the feedback circuit (the one in the leg to ground). In fact, this is one reason why dc servos are used, there are some people who try to avoid the use of capacitors wherever possible. The circuit at the site you posted was developed for this very purpose.
With current feedback amplifiers such as the JLH, the dc blocking capacitor must remain, otherwise it will be impossible to achieve a dc offset anywhere near to zero volts (simulations show an offset approaching +10V if my memory serves me correctly).
However, my intention is to simply replace the 7815 and associated resistors with a circuit, controlled by the dc at the output, that will inject the required current into the emitter of Q4 to maintain zero offset. Initial simulations look promising though I have yet to try anything in practice. I'll let you know of any progress.
Geoff
[Edited by Geoff on 04-22-2001 at 01:29 PM]
Thanks for the URL. I've seen that site before and have a copy of the program. It is a pity that the author didn't include the design criteria and calculation details instead of just a circuit and the calculation results. The circuit seems more complex than others I have seen and I am still trying to work out what certain parts are intended to achieve.
One thing you should be aware of, if you are not already. When dc servos are used with voltage feedback amps (i.e. those with a differential (long-tailed pair) input stage) it is possible to do away with the dc blocking capacitor in the feedback circuit (the one in the leg to ground). In fact, this is one reason why dc servos are used, there are some people who try to avoid the use of capacitors wherever possible. The circuit at the site you posted was developed for this very purpose.
With current feedback amplifiers such as the JLH, the dc blocking capacitor must remain, otherwise it will be impossible to achieve a dc offset anywhere near to zero volts (simulations show an offset approaching +10V if my memory serves me correctly).
However, my intention is to simply replace the 7815 and associated resistors with a circuit, controlled by the dc at the output, that will inject the required current into the emitter of Q4 to maintain zero offset. Initial simulations look promising though I have yet to try anything in practice. I'll let you know of any progress.
Geoff
[Edited by Geoff on 04-22-2001 at 01:29 PM]
Single ended or not?
There seems to be some discussion (for a longer while) about the true nature of the Linsley-Hood amp, is it single-ended or not? A true single-ended amp always works in class-A and therefore can not draw more current than the current that is idling. So is it true in the case of the JLH that if you make the amplifier draw a current of say 1 Ampere the maximum power is limited to a few watts even if you have plenty of voltage available in your power supply (say 40 Volts or more in case of the old 10 Watter)? Is there no chance that the JLH can work in class-AB?
Someone out there can tell the answer, of simulate the thing in case of a moderate current of 1 Ampere or less.
There seems to be some discussion (for a longer while) about the true nature of the Linsley-Hood amp, is it single-ended or not? A true single-ended amp always works in class-A and therefore can not draw more current than the current that is idling. So is it true in the case of the JLH that if you make the amplifier draw a current of say 1 Ampere the maximum power is limited to a few watts even if you have plenty of voltage available in your power supply (say 40 Volts or more in case of the old 10 Watter)? Is there no chance that the JLH can work in class-AB?
Someone out there can tell the answer, of simulate the thing in case of a moderate current of 1 Ampere or less.
The JLH is a single-ended Class-A amp only in the sense that it cannot significantly increase its current output 'on demand', for example when presented with a low impedance load. The output stage operates in push-pull with the 2N1711 acting as a phase splitter to feed the 2N3055s.
From my (extensive) simulations of this circuit, it would seem that it can provide a peak current into the load of just over 1.4 times the quiescent current. This figure is confirmed by Doug Self in his Audio Power Amplifier Design Handbook in which he states when referring to the JLH amp "the effective quiescent current varies by more than 40% over the cycle".
It is impossible for the JLH amp to operate in Class-AB.
Geoff
From my (extensive) simulations of this circuit, it would seem that it can provide a peak current into the load of just over 1.4 times the quiescent current. This figure is confirmed by Doug Self in his Audio Power Amplifier Design Handbook in which he states when referring to the JLH amp "the effective quiescent current varies by more than 40% over the cycle".
It is impossible for the JLH amp to operate in Class-AB.
Geoff
The initial idea of this thread was the PSU of the amp.
When I built it I found out that the PSU is not of great importance given that your speakers are not too sensitive.
I myself have got Odin loudspeakers (90db/w/m) and use regular PSU that includes two 150VA trafos, two 35A bridge rectificators and two 33000uF caps (1 per channel). There's absolutely no noise. Here I must add that I've built the old version of the amp - 1969.
I intend to make further experiments with the PSU but at this stage I don't need that.
It's a nice sounding amp.
Regards
Asen
When I built it I found out that the PSU is not of great importance given that your speakers are not too sensitive.
I myself have got Odin loudspeakers (90db/w/m) and use regular PSU that includes two 150VA trafos, two 35A bridge rectificators and two 33000uF caps (1 per channel). There's absolutely no noise. Here I must add that I've built the old version of the amp - 1969.
I intend to make further experiments with the PSU but at this stage I don't need that.
It's a nice sounding amp.
Regards
Asen
Hi there, I am also building this amp, just finished the amp boards obtained from Hart electronics. All my components I have sourced locally. Output transistors are MJL3281A. I am not able to obtain the MJE371 here so have opted for the BD440. Geoff Moss has been very hepful. like you I have wondered about which power supply to use. I will probably go for the updated regulated supply but I am looking around for any other suggestions.
Phil
Phil
JLH in progress
Asen,
I'm currently building a dual-mono setup of the 1996 version with the fully regulated power supply.
Find more details at my amp-site http://www.knoware.nl/users/cfw/amp.
It's nearing completion.
I have a progress report on my site.
I'm interested to know who you're progressing.
Jos
E-mail: cfw@knoware.nl
[Edited by dutch diy on 10-08-2001 at 02:54 PM]
Asen,
I'm currently building a dual-mono setup of the 1996 version with the fully regulated power supply.
Find more details at my amp-site http://www.knoware.nl/users/cfw/amp.
It's nearing completion.
I have a progress report on my site.
I'm interested to know who you're progressing.
Jos
E-mail: cfw@knoware.nl
[Edited by dutch diy on 10-08-2001 at 02:54 PM]
Joe
The only thing I've changed is that I've added some 10000uF per channel. Now there are 43000uF per channel.
I'm a bit worried (see Geoff's page) about the 30MHz transistors I use - Toshiba 2SC3280 - but my scope doesn't show any problems and I cannot hear any either.
May be it will be of interest that I prefer the amp with a tube pre in front of it despite it doesn't need the gain. That combination gives a sweeter sound. Of course that depends on the whole system one will use. I have a DIY ECC88 SRPP preamp.
Regards
Asen
[Edited by Asen on 10-08-2001 at 04:18 PM]
The only thing I've changed is that I've added some 10000uF per channel. Now there are 43000uF per channel.
I'm a bit worried (see Geoff's page) about the 30MHz transistors I use - Toshiba 2SC3280 - but my scope doesn't show any problems and I cannot hear any either.
May be it will be of interest that I prefer the amp with a tube pre in front of it despite it doesn't need the gain. That combination gives a sweeter sound. Of course that depends on the whole system one will use. I have a DIY ECC88 SRPP preamp.
Regards
Asen
[Edited by Asen on 10-08-2001 at 04:18 PM]
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.