To get a "harder", clearer, more contoured, defined sound, in bass too, do NOT use TO-3 transes: The smaller the transes the clearer and cleaner and contoured,-)
Have build some JLH - exacter "CLH"-) - with different transes. The TO-3 and TO-3P sound very mild and un-contoured; the bass seems not "black", a little bit in background and separat, discrete.
Have build some JLH - exacter "CLH"-) - with different transes. The TO-3 and TO-3P sound very mild and un-contoured; the bass seems not "black", a little bit in background and separat, discrete.
I did built a higher power PNP version but could not tell the difference between it and the original.
Power dissipation was a bit silly using 42 Volt supply @ 60 watts per channel. I never put it into a nice box so it became my garage amp running from a 30V 10 amp SMPS.
Attachments
I have just been reading the November/December 1989/January 1990 Wireless World. JLH is talking about the evolution of audio amplifiers. When we look at the 1969 JLH design and how he understands the other world of audio it is like our hero just got out of his Spitfire and jumped into any jet you care to mention and could fly it.
JLH makes a very interesting observation which totally rings true. Classical and Jazz music can be referenced to real life. Rock music not. Rock music only becomes real via a hi fi. Often with little trouble we can say good or bad and if we are to live with it only our taste counts. I got around this with 100 dB/W OB speakers that do anything. They don't need much power and are very very dangerous if power is available.They will do 350 watt on transients.
The more I think about it only the power supply is the question with the 1969 amp. The output capacitor is our very best friend as it makes PSU design easier and DC protection is assured. Many high grade caps can be found and some are very good on price ( Panasonic FC if ripple ratings OK ).
Did anyone compare a SMPS and Linear? I did some work with cheap Meanwell PSU's and found them to be not too bad. Not sure about the Hypex one, it was OK, but not special. Whilst I don't want the RF junk in the amplifier I doubt this one would mind very nuch. The Hypex seems imune to it's PSU and itself, I have no reason to think the JLH any more fussy. When Hypex the SMPS sounds brighter. It isn't nasty. If nothing else a SMPS would allow a quick build and less heat.
LRS-100-36 | Mean Well 100.8W Embedded Switch Mode Power Supply SMPS, 0 → 2.8A, 36V dc | Mean Well
JLH makes a very interesting observation which totally rings true. Classical and Jazz music can be referenced to real life. Rock music not. Rock music only becomes real via a hi fi. Often with little trouble we can say good or bad and if we are to live with it only our taste counts. I got around this with 100 dB/W OB speakers that do anything. They don't need much power and are very very dangerous if power is available.They will do 350 watt on transients.
The more I think about it only the power supply is the question with the 1969 amp. The output capacitor is our very best friend as it makes PSU design easier and DC protection is assured. Many high grade caps can be found and some are very good on price ( Panasonic FC if ripple ratings OK ).
Did anyone compare a SMPS and Linear? I did some work with cheap Meanwell PSU's and found them to be not too bad. Not sure about the Hypex one, it was OK, but not special. Whilst I don't want the RF junk in the amplifier I doubt this one would mind very nuch. The Hypex seems imune to it's PSU and itself, I have no reason to think the JLH any more fussy. When Hypex the SMPS sounds brighter. It isn't nasty. If nothing else a SMPS would allow a quick build and less heat.
LRS-100-36 | Mean Well 100.8W Embedded Switch Mode Power Supply SMPS, 0 → 2.8A, 36V dc | Mean Well
I think I have a 48V SMPS which will go down to 42V. I have some Indian 2N3055 and some BC327/337 ( and BD139 ). I also have a giant heat sink and some very standard output caps. I may have some MJ15015 to try. I am home Christmas day so might have a play. I hate SMPS, to have a chance of quick data I will risk it. If it works OK I will post data and RF output to 16MHz.
I have not powered amps with SMPSUs,but they have come a long way in recent years.It is probably a good idea to filter the outputs with high quality caps. (polyprop. ceramic etc.) as some electrolytics in the PSU can appear inductive at the switching frequencies.This can lead to overheating and premature failure. My own preference would be caps.on the output ,then inductors followed by more caps., much like the input filter.Make sure that the outputs are fully isolated from the mains by a PROPER HF transformer.Tread very carefully with plug top PSUs,these do not appear fully isolated,I have lost components,fortunately only LEDs by accidentally earthing part of the circuit.
Probably a good idea to keep the PSU separate from the amp.
Probably a good idea to keep the PSU separate from the amp.
Fast way remembering;-)
JLH without the first transistor. And shortening of some parts. 10 minutes to do. Sounds ... wonderful;-)
You do not need a board! Sounds ugly! Unnecessary materials and horrible solder jounts and resonances and paths ...-(
Now TO-220-transistors.
May be, christmas I will finish.-))))
My experience:
The capacitance. The higher the cleaner and more powerful and "harder" and so on. But not (a) big cap (big ones do scratch very very vwell;-), better a lot of little ones. At ONE circuit;-! And NO little bridging caps!
Shottky-diodes too.
Active psu, of course.
... but all this does not counter TO-3 character enough;-)
Any TO-220 or TO-247 I would use.
LG
A little bit better;-)
JLH without the first transistor. And shortening of some parts. 10 minutes to do. Sounds ... wonderful;-)
You do not need a board! Sounds ugly! Unnecessary materials and horrible solder jounts and resonances and paths ...-(
Now TO-220-transistors.
May be, christmas I will finish.-))))
My experience:
The capacitance. The higher the cleaner and more powerful and "harder" and so on. But not (a) big cap (big ones do scratch very very vwell;-), better a lot of little ones. At ONE circuit;-! And NO little bridging caps!
Shottky-diodes too.
Active psu, of course.
... but all this does not counter TO-3 character enough;-)
Any TO-220 or TO-247 I would use.
LG
A little bit better;-)
Attachments
![IMG_20171219_205056[1].jpg](/community/data/attachments/593/593586-ccf107c8d76871c37ce709097b7a39a4.jpg)
![IMG_20171219_205114[1].jpg](/community/data/attachments/593/593600-900b52f2de06f073ee33ac2b80164cca.jpg)
![IMG_20171219_214636[1].jpg](/community/data/attachments/593/593633-07bdf0252fe05e9419cad70439caceb2.jpg)
The smaller the transistor, usually you get the better specs (lower Cob, higher hfe, etc) and hence the better performance. That's if you need 200mW amp
I have a spectrum analyser so can look at the residuals. I have 48V if I wish so can loose a few volts. A simple RC filter should be OK. 1R + 1000 uF + 1R +1uF should do something ( It should be OK, 10 000uF also ). I suspect 1R is enough to divorce the two sides. One slight problem I have seen with some SMPS ( Switch Mode Power Supply or PSU ) is hard connecting the -ve to the heatsink is not always ideal, I've even known the protection come on. Sometimes a 100 nF -ve to heatsink works. I will have to see how it goes. If I can get this to work it's very low cost.
Most of my beliefs verge on alchemy. Alchemy is science in the making. What people forget is science costs money and usually is not for hi fi. It is very lucky if a bit of new sceince is useful to us. A PSU is something which is the science side of the alchemy border. It should be fun. Being that the JLH is a real class A amplifier anything we see on the analyser should be what we hear. When AB, D , sliding class A it's anyones guess as the PSU + amplifier combination is more like a river meeting the sea with tidal flows. The JLH is a river long before the sea. There is some science to back this up. In the 1980's someone who might have been Martin Colloms took on the Alchemists idea of the class AB amplifier being able to rise above the mains ( 240V ). That is like the sea pushing up a river. It was found to be true and 21 amps peak current measured when 3 amps would be thought likely. The idea was put forward that the energy flows would be like a pendulum, Mains 240V to PSU to Amplifier to Speaker and then back again. Any wobble the pendulum meets is modulated into music. This is further made bad by the fact that briefly the AB crossover point shows very high output impedance for a few uS until the feedback loop is restored. One has to remember unless my old Magnepans the load is complex. If you like the AB amplifier is a hammer in a blacksmiths hands that is passing hand to hand whilst he works the iron when class AB or D. That's fine if a Circus act. The JLH is not a pendulum, It's a waterwheel on a strong flow that can divert power. That blacksmith is using two hands all the time or he could use one if we choose.
One day Rufus a 19 year old friend discovered the Quad 303. After a short time the 303 rejected him. Usually when this happens the set 67 VDC drops to 25V. His went the other way and gave 89V + hum. When repaired he was very dissapointed as the amplifier lost it's sense of power. I heard this also so don't say he was wrong. It seems Quads 2N3055 type devices could take 100V, however the specs were set when 67V was the safe maximum. The 303 has a remarkable PSU and makes this AB amp work almost exactly as a JLH 10 watt. The current gain of the 303 output stage is vast and needs almost no bias to work ( highly linear also ), we mostly hear the PSU when the 303. JLH pointed out that not everyone would like this as a fine tuned AB amp with fine tuned unregulated PSU will sound more dynamic. The reason is simple, the speakers are the cause. Even a speaker box is the first problem. A baffle is better. My own are flat to 40Hz and have useful output at 30Hz ( - 15 dB ). Plenty of bass EQ and power used to have that. They are 2x4 feet each with 15 inch bass, 12 inch 100 to 6000 Hz ( 8 kHz - 6dB ) and tweeter above. The sound very fast and not too coloured.
Just to say a little of Alchemy. A friend asked my to design some cables. He didn't ask me to doubt him as part of the deal. I had met his Russian friend who insisted a 4% gold 96% silver alloy works, he seemed a very sensible person not given to weird ideas. I then stopped to think. Take lead and take tin, when mixed they melt about 20C lower than their two metal when ideal. Hi fi seems to like Oxygen Free Copper. It is about 102% the conductivity of ETP copper. OFC is not expensive so I asked myself why? Industries other than hi fi use it. One use when very pure is to contain nuclear waste. I notice some talk about capacitors sounding good or bad ( Read Cyril Bateman ). Often they do not sound as bad as they should and measure very well ( better than - 100 dB ). All the same where is the harm to try some. With the JLH there are so few you might for once be able to hear how the work, mains class X2 can be excellent and cheap. X2 types follow the same rules as audiophile types. The JLH is an amplifer that could be made from what some would think junk and sound better than most Super Fi.
If I get to listening which means unpacking the hi fi I will use a NAD 3020 as preamp. I have a DL110 PU which will just about give enough.
Cyril Bateman's Capacitor Sound articles | Linear Audio NL
it's funny, I have never liked the sound of quad 303, I find it too limited, not open enough.
Yet, years ago, I went to dismantle a super-charged quad / thorens installation td124 + SME3012 / arena speakers.
the deceased gentleman had built his house around his hi-fi and his daughter found this set unsightly and wanted a smaller system.
maybe I should listen to one again.
As far as capacitors are concerned, I think the big difference comes from the schema first and foremost.
the more simple it is, the more the difference will be heard, especially in class A.
I stopped trying to understand for a long time, now I try and I keep what I like
Yet, years ago, I went to dismantle a super-charged quad / thorens installation td124 + SME3012 / arena speakers.
the deceased gentleman had built his house around his hi-fi and his daughter found this set unsightly and wanted a smaller system.
maybe I should listen to one again.
As far as capacitors are concerned, I think the big difference comes from the schema first and foremost.
the more simple it is, the more the difference will be heard, especially in class A.
I stopped trying to understand for a long time, now I try and I keep what I like
303 with the right speakers can be very good. Strangely Magnapan SMGa is one. The very simple current limiters actually work. If you clone the 303 output stage look at the diodes. It's so easy to dial up what you need. They sound like a valve design that just gives up when clipping. It's hard clipping that sounds OK. It's not subtle. My tests suggest standard current limiting doesn't work. Far better like the Naim NAP 250 it's in the PSU. The NAP250 is said to be the amplifer Quad didn't make. It nearly clones a 303 in distortion cuves. Tantalum capacitors are not it's best thing. Alan Morningon West was with both companies. Some say it was the 303 replacement that the 405 prevented. The 33 premap can work ( 34 and 44 less so ). It has a very narrow window of OK performance. The Phono stage is antique. The spare radio input is very high impedance, that can be very helpful. The sound of that input can be dreadful or excellent. If dreadful you didn't respect that small window of possible performance. When I say dreadful read " like it is underwater ". Many valve designs are the same. Leak Varislope has an anode follower output. It has 30 cm of very nasty cable. It is 75 ohms cable and needs > 470K to work into. When you do it is very wonderful. Buy some silly money cable and it will sound underwater if 300pF/metre. My Pro Audio friends call all silly money cables " Yaksahumi ". I invented the term as I like to pretend cables are unimportant when with them. It's far better than a whole evening of making me the spokesman for the hi fi industry. If anyone says cable capacitances is unimportant, think again. I am not convinced my Pro Audio friends could work outside of their world. I almost think they would see a vitamin pill and KFC as being no different to a healthy meal. Everything in Pro Audio works because the standards were set perhaps in 1940. The joy of hi fi is mostly it's wrong and can be made right.
Here is the power supply I shall use for the data. Any idea what current I should set the output stage to? 42V is about where the lowest ouput voltage sits ( with RC filtering 40V ? ). My heatsink at a guess is 0.25C/W. It takes 8 T03. If 1amp and stereo I guess it would sit at 50C max. I have two heat sinks if required. My friend has a Krell and BBC LS5/9. She won't mind if the JLH is better. That's the plan. The bigger plan is to use the cheap switchmode SMPS. I have data somewhere on that PSU. For very little money I made it work well. Alas I have no idea where I put that data ( file name ). It was a true class B working at constant power after me being fed up with class D repairability. Using the SMPS for the first time in my life I got full class B efficiency. The transistors TIP35/36. The thing I didn't think of was the power supply was at high current all the time just like class A. That's all the data I need. It was a very reactive load. It was great to see the SMPS have a think about the load, the little protection LED slightly dims on starting. When the linear supply goodness knows what the starting current was ( far less than 1R initial load ). The TIP35/36 are tough. All the same I prefer the junctons don't need to prove it. Sorry to talk about the 303 so much. Often the 303 and JLH are the benchmarks of the era.
LRS-100-48 | Mean Well 110.4W Embedded Switch Mode Power Supply SMPS, 0 → 2.3A, 48V dc | Mean Well
Here is the power supply I shall use for the data. Any idea what current I should set the output stage to? 42V is about where the lowest ouput voltage sits ( with RC filtering 40V ? ). My heatsink at a guess is 0.25C/W. It takes 8 T03. If 1amp and stereo I guess it would sit at 50C max. I have two heat sinks if required. My friend has a Krell and BBC LS5/9. She won't mind if the JLH is better. That's the plan. The bigger plan is to use the cheap switchmode SMPS. I have data somewhere on that PSU. For very little money I made it work well. Alas I have no idea where I put that data ( file name ). It was a true class B working at constant power after me being fed up with class D repairability. Using the SMPS for the first time in my life I got full class B efficiency. The transistors TIP35/36. The thing I didn't think of was the power supply was at high current all the time just like class A. That's all the data I need. It was a very reactive load. It was great to see the SMPS have a think about the load, the little protection LED slightly dims on starting. When the linear supply goodness knows what the starting current was ( far less than 1R initial load ). The TIP35/36 are tough. All the same I prefer the junctons don't need to prove it. Sorry to talk about the 303 so much. Often the 303 and JLH are the benchmarks of the era.
LRS-100-48 | Mean Well 110.4W Embedded Switch Mode Power Supply SMPS, 0 → 2.3A, 48V dc | Mean Well
This is from TubeCad I think ? This will be my starting point as the starting values are useful. I suspect the current limiter of the SMPS will keep me out of danger. For many reasons Christmas day is not the best for me. I will cook a meal and the usual. Starting this project will be an excellent distraction. Any circuit observations?
I might do a PSU test first if I can find a large resistor for a load. If I can I will post noise to 400 kHz and ripple. I am going to try a very JLH type voltage regulator idea. I only have 2V to play with so wish me luck. I suspect it will have to be the amplifier itself. 14R 60 watt.
I'm sure you do
you just have to think outside the box on what your points of reference are.