JLH Problem
Tim,
I am the other person with the same problem.
I built the JLH for ESL and tried to remove the feed back cap.
Using the 47uF cap and vr3 I got the same problems you described. I even suspcted some RF problems since as soon as I got closer to the Amp the oscilation was worse.
I followed Geoff's advice I removed the 47uF cap and replaced
the original 10k resistor with a jfet. this reduced oscilation dramaticaly but still not eliminating it. I tried replacing the BC212 with MPSA56, with more or less the same results.
I found out that the DC offset was now very sensetive to ambient temprature changes and it is quite impossible to set it in a way
that ensures it does not get to much out of hand when teperature changes (Air conditioning in summer).
After some time I decided to reverse the change in sake of a steady 3-5mV Offset instead of sometimes 60-70mV.
If somebody has any ideas I might try again.
Tim,
I am the other person with the same problem.
I built the JLH for ESL and tried to remove the feed back cap.
Using the 47uF cap and vr3 I got the same problems you described. I even suspcted some RF problems since as soon as I got closer to the Amp the oscilation was worse.
I followed Geoff's advice I removed the 47uF cap and replaced
the original 10k resistor with a jfet. this reduced oscilation dramaticaly but still not eliminating it. I tried replacing the BC212 with MPSA56, with more or less the same results.
I found out that the DC offset was now very sensetive to ambient temprature changes and it is quite impossible to set it in a way
that ensures it does not get to much out of hand when teperature changes (Air conditioning in summer).
After some time I decided to reverse the change in sake of a steady 3-5mV Offset instead of sometimes 60-70mV.
If somebody has any ideas I might try again.
The Class A amplifier site
Ralf,
It is my understanding that Geoff's site may only be temporarily closed. I trust he will not mind a brief word from me on why his site is presently unavailable. I am sure that many readers of this thread are aware of Geoff's desire to explain fully anything that might otherwise be misinterpreted or understood only partially....There is a tendency towards some instability in the JLH (update) circuit, as I understand it this may be due, but possibly only in part, to the removal of the feedback capacitor. At present Geoff is unable to devote the necessary time to include a full (and characteristically thorough) explaination of the problem and a solution on his excellent site. Until he is able to do so, I would refer again to the solution that proved effective in my circuit, which was to use MPSA56 (not pin-compatible with 2SA970) for positions Q5, Q6, Q7 and Q8.
Miki,
From how you describe your oscillation problem it seems you experienced the same problem as I. In my case the remedy mentioned above was sufficient to remove the problem.
I wonder if the large dc offset variations you experienced could have been improved with better heatsinking/ventillation.
Tim.
Ralf,
It is my understanding that Geoff's site may only be temporarily closed. I trust he will not mind a brief word from me on why his site is presently unavailable. I am sure that many readers of this thread are aware of Geoff's desire to explain fully anything that might otherwise be misinterpreted or understood only partially....There is a tendency towards some instability in the JLH (update) circuit, as I understand it this may be due, but possibly only in part, to the removal of the feedback capacitor. At present Geoff is unable to devote the necessary time to include a full (and characteristically thorough) explaination of the problem and a solution on his excellent site. Until he is able to do so, I would refer again to the solution that proved effective in my circuit, which was to use MPSA56 (not pin-compatible with 2SA970) for positions Q5, Q6, Q7 and Q8.
Miki,
From how you describe your oscillation problem it seems you experienced the same problem as I. In my case the remedy mentioned above was sufficient to remove the problem.
I wonder if the large dc offset variations you experienced could have been improved with better heatsinking/ventillation.
Tim.
Not entirely removing C4
As the latest discussions in thsi thread are about the removal of C4 in the ESL configuration I've got some observations from all the posts:
removing C4 introduces LF/RF oscillations (depending on the type of Q8/Q7 and Q4)
measures are proposed against this oscillation.
No observations in change of sound quality are posted regarding changing size of C4; type of capacitor or bypassing C4 with sound-wise decent small caps.
As I'm getting ready to build the 2003 ESL version, I'm not keen on removing C4. Based on the theory C4 isolates the signal feedback from the DC feedback.
Soundwise elco's aren't the nicest things you want in a signal path as they are frequency dependant. Usually bypassing with a lower value non-polar cap is suggested.
So: my suggestion would be to replace C4 with a 1uF polyprop cap or, as I plan to do just bypass with a .47uF mkp10.
Let me have your comments please.
As the latest discussions in thsi thread are about the removal of C4 in the ESL configuration I've got some observations from all the posts:
removing C4 introduces LF/RF oscillations (depending on the type of Q8/Q7 and Q4)
measures are proposed against this oscillation.
No observations in change of sound quality are posted regarding changing size of C4; type of capacitor or bypassing C4 with sound-wise decent small caps.
As I'm getting ready to build the 2003 ESL version, I'm not keen on removing C4. Based on the theory C4 isolates the signal feedback from the DC feedback.
Soundwise elco's aren't the nicest things you want in a signal path as they are frequency dependant. Usually bypassing with a lower value non-polar cap is suggested.
So: my suggestion would be to replace C4 with a 1uF polyprop cap or, as I plan to do just bypass with a .47uF mkp10.
Let me have your comments please.
Re: The Class A amplifier site
Tim,
There is one thing I forgot to mention, Using the Jfet reduced Hum
compared to the 47uF cap solution.
I did not notice a significant sonic change comparing to the feed back capacitor being in the circuit. I use a 470uF bypassed with a 0.1uF mkp.
As for heat sinking, the normal temperatur is 49 degC. once it changes to 46 there is a change in the DC offset. I guess that it would happen with any teperature shift.
TimA said:
Tim,
There is one thing I forgot to mention, Using the Jfet reduced Hum
compared to the 47uF cap solution.
I did not notice a significant sonic change comparing to the feed back capacitor being in the circuit. I use a 470uF bypassed with a 0.1uF mkp.
As for heat sinking, the normal temperatur is 49 degC. once it changes to 46 there is a change in the DC offset. I guess that it would happen with any teperature shift.
Re: Not entirely removing C4
Iv'e not posted here before so Hi
having learnt my electronics from studying this design and having tinkered around with it for about 10 years I feel as if I am among friends.
re the question above.
If you replace this cap with 1uF you will have an amp that has no bass response.
If you decide to leave it in, I guess a black gate would be the best way to go perhaps a slightly higher value would give firmer bass.
however I would strongly suggest that it is worth sorting out whatever the oscillation problem is and then just replace this cap with a link.
my own design has gone in quite a different direction so I did not have to deal with this exact problem but leaving out this cap reaped rich sonic rewards for me.
by the way the function of C4 is to roll off the bass response, give the amp a DC gain of 1 and therefore give low DC offset. it is not part of a compensation network.
I hope that helps
mike
dutch diy said:
Iv'e not posted here before so Hi
having learnt my electronics from studying this design and having tinkered around with it for about 10 years I feel as if I am among friends.
re the question above.
If you replace this cap with 1uF you will have an amp that has no bass response.
If you decide to leave it in, I guess a black gate would be the best way to go perhaps a slightly higher value would give firmer bass.
however I would strongly suggest that it is worth sorting out whatever the oscillation problem is and then just replace this cap with a link.
my own design has gone in quite a different direction so I did not have to deal with this exact problem but leaving out this cap reaped rich sonic rewards for me.
by the way the function of C4 is to roll off the bass response, give the amp a DC gain of 1 and therefore give low DC offset. it is not part of a compensation network.
I hope that helps
mike
Re: Not entirely removing C4
Whilst your second suggestion might give a small improvement, your first one will not work. Changing the dc blocking capacitor in the feedback network to 1uF will cause the amp to behave as a high-pass filter with a -3dB frequency of just over 700Hz. The output signal will be 16dB down at 100Hz. This would be fine if the amp is only providing power to a tweeter in a bi- (or multi-) amped active system but is obviously not satisfactory for full-range use.
TimA has given an indication as to why my website has been temporarily removed. The reason for my inability to update the information on the site is ill-health, both personal and that of my elderly mother whom I look after. I hope to have at least part of the site back on the air (with a new URL) within the next week or two. I will post the new address as soon as it is working.
My apologies to those who have tried to access my website during the past few weeks only to receive an error message.
Geoff
dutch diy said:
Whilst your second suggestion might give a small improvement, your first one will not work. Changing the dc blocking capacitor in the feedback network to 1uF will cause the amp to behave as a high-pass filter with a -3dB frequency of just over 700Hz. The output signal will be 16dB down at 100Hz. This would be fine if the amp is only providing power to a tweeter in a bi- (or multi-) amped active system but is obviously not satisfactory for full-range use.
TimA has given an indication as to why my website has been temporarily removed. The reason for my inability to update the information on the site is ill-health, both personal and that of my elderly mother whom I look after. I hope to have at least part of the site back on the air (with a new URL) within the next week or two. I will post the new address as soon as it is working.
My apologies to those who have tried to access my website during the past few weeks only to receive an error message.
Geoff
Trafos, heatsink and output transistors.
Hello
i have a few qustions.
Im planing to build a JLH updated ESL- version. 30 watt per channel. (for over a year now) whit regulated supply. but i do not know what trafo to buy. what sek. voltage and how many VA. will be good for one channel?
will 2 0,3 K/W heatsink per channel keep is runing under 55 C.? and how many watt of heat will be dissipationt per channel?
Can i use MJL3281A for output transistors since a have some already.
Hello
i have a few qustions.
Im planing to build a JLH updated ESL- version. 30 watt per channel. (for over a year now) whit regulated supply. but i do not know what trafo to buy. what sek. voltage and how many VA. will be good for one channel?
will 2 0,3 K/W heatsink per channel keep is runing under 55 C.? and how many watt of heat will be dissipationt per channel?
Can i use MJL3281A for output transistors since a have some already.
Re: Trafos, heatsink and output transistors.
I use these and am pleased with results but they are quite fast so you may need some HF compensation to make it stable.
JLH suggests an arrangement for this in one of the early articles. It would probably work in your implementation.
mike
Calebay said:
I use these and am pleased with results but they are quite fast so you may need some HF compensation to make it stable.
JLH suggests an arrangement for this in one of the early articles. It would probably work in your implementation.
mike
Re: Trafos, heatsink and output transistors.
I have used 300VA 22 0 22 V trafos with a circuit from Geoff for the P.S, Thanks Geoff.
It works fine giving 25v rails with a 3.5A current.
I am using the MJ15003 though. My heat sinks are rated 0.7 C/W per output device but I built the case from Aluminum so it serves as additional heat sink. the AMP is usualy running at 49 C. depending on ambient temp of coarse.
running at 25V with 3.5A will give you about 88W per pair of devices with a 0.3 heatsink you get 26.25 C + 25C ambient will give 51 C so you should be ok.
you should take into account the dissipation from the power supply devices (transistors / regulators).
besides that the amps are great, stop planning start building
Calebay said:
I have used 300VA 22 0 22 V trafos with a circuit from Geoff for the P.S, Thanks Geoff.
It works fine giving 25v rails with a 3.5A current.
I am using the MJ15003 though. My heat sinks are rated 0.7 C/W per output device but I built the case from Aluminum so it serves as additional heat sink. the AMP is usualy running at 49 C. depending on ambient temp of coarse.
running at 25V with 3.5A will give you about 88W per pair of devices with a 0.3 heatsink you get 26.25 C + 25C ambient will give 51 C so you should be ok.
you should take into account the dissipation from the power supply devices (transistors / regulators).
besides that the amps are great, stop planning start building
hello
so the MJL3281A will not be the best choice i see. so ill think ill try 2SC2922 instead.
88W per pair of devices, so will that be 176 W per channel + the waste from LM338K?
how many watt will the amp put into 8 Ohm whit 25volt rails and 3.5A current.?
and yeah i will try building soon.
so the MJL3281A will not be the best choice i see. so ill think ill try 2SC2922 instead.
88W per pair of devices, so will that be 176 W per channel + the waste from LM338K?
how many watt will the amp put into 8 Ohm whit 25volt rails and 3.5A current.?
and yeah i will try building soon.
Assuming 6V drop with 3.5A the power supply devices should dissipate around 21W per rail
with an 1.5 to 2C/W heat sink the temperature should be OK
remeber to check that the devices you plan to use can handle the the internal temp.
I used BD250 / BD249 and a 2C/W per device
I used the smaller sinks at the back of the AMP for the PS and the Mj15003 sinks are at the sides.
with an 1.5 to 2C/W heat sink the temperature should be OK
remeber to check that the devices you plan to use can handle the the internal temp.
I used BD250 / BD249 and a 2C/W per device
I used the smaller sinks at the back of the AMP for the PS and the Mj15003 sinks are at the sides.
Re: Oscillation Problem
1. I forgot to mention that my JLH updated version is the one shown in Fig. 3- JLH for ESL circuit but with C4 removed.
2. I got no oscillation at all, I run it at +-27v with quiescent current of 3.2A. The only drawback is the higher DC offset range of about 250mv(may somebody not accept it) from cold start to being warm up so I adjusted it at +180mv cold start, I am not sure whether the problem is due to the temperature drift of Q7 and Q8. Good ventilation may help but I've not yet tried it.
3. Regarding the oscillation problem. It is better to find out the cause of instability, perhaps it may come from improper wirings or components placing. I had experience in connecting power supply to main PCB with long wires causing supply instability and hum occurred. In my case, I avoid long wirings, I incorporated the Cap mult circuitry in the same PCB of the main circuitry but the wirings from PCB to the power transistors remained inevitably with the wires kept as short as possible(within 6 inches). I am amateurish in PCB design but with thinking in a more logical way of placing the components and running the traces such that output components be placed not closely to the input components(I am ashamed as my PCB's components placing were still very congested due to lack of space(like apartments in HK). Not good for air ventilation and driver transistors heat sinks placing, design fault!). I haven't tried faster transistors, still using Mj15003 and 2SC3621 in place of 2SC3421, other remained the same. I hope my little experience could help and wish all of you would solve the problems and enjoy the musics.
Best regards
John
1. I forgot to mention that my JLH updated version is the one shown in Fig. 3- JLH for ESL circuit but with C4 removed.
2. I got no oscillation at all, I run it at +-27v with quiescent current of 3.2A. The only drawback is the higher DC offset range of about 250mv(may somebody not accept it) from cold start to being warm up so I adjusted it at +180mv cold start, I am not sure whether the problem is due to the temperature drift of Q7 and Q8. Good ventilation may help but I've not yet tried it.
3. Regarding the oscillation problem. It is better to find out the cause of instability, perhaps it may come from improper wirings or components placing. I had experience in connecting power supply to main PCB with long wires causing supply instability and hum occurred. In my case, I avoid long wirings, I incorporated the Cap mult circuitry in the same PCB of the main circuitry but the wirings from PCB to the power transistors remained inevitably with the wires kept as short as possible(within 6 inches). I am amateurish in PCB design but with thinking in a more logical way of placing the components and running the traces such that output components be placed not closely to the input components(I am ashamed as my PCB's components placing were still very congested due to lack of space(like apartments in HK). Not good for air ventilation and driver transistors heat sinks placing, design fault!). I haven't tried faster transistors, still using Mj15003 and 2SC3621 in place of 2SC3421, other remained the same. I hope my little experience could help and wish all of you would solve the problems and enjoy the musics.
Best regards
John