That would be fascinating if it worked. No one ever discussed RF and class A. It should hurt class AB more. It's argued that 2V/ uS is the slew rate of real music when a 100 watt amplifier. Conjecture suggests 6V/uS as a sensible minimum iif input RF filtering used. Some then call this the British amplifier sound. What a neurve. It's the sound of real music without unnecessary distortion.
After much reading here I was courius about using smps, I found two large smps, delivering 24 volt at 40 amp max, I'm unsure about connnecting them in series to power my jlh, I added the brochure of the smps here.
Would additional filtering be a good idea ?
Would additional filtering be a good idea ?
I doubt extra filtering will be cost effective.Ypu can try 0R33 and 10 000 if. For common mode LCR ,CF5 6 amp choke or similar at circa 1 mH. . I like Smps closer to the required amperage. That way faulty components often are saved from disaster using the Smps current limiter.
Hi nigel.
Thanks for your suggestion, I will try measure how clean dc they deliver when they arrive, and see if extra filtering is needed, I know they are a tad overkill but as in godfather movie, it was a offer I couldn't refuse🙂
Thanks for your suggestion, I will try measure how clean dc they deliver when they arrive, and see if extra filtering is needed, I know they are a tad overkill but as in godfather movie, it was a offer I couldn't refuse🙂
Just to say it can be very hard to reference the noise to a useful point like ground. Be very happy if you can do that. It is usually dictated by at least one thing connected to the system that it is the automatic reference point.
I usually have parts like that at no cost, hopefully you also. For an initial test expensive parts are not required. When I did this the slightly unusual 75 Hz noise didn't change much. 66kHz noise did and was worth doing. 10 uF polyester I suspect would have worked. That's 1000 times less. A big PSU might have good noise at 50 to 100 Hz or 60/120 due to spare capacity.
I tried a 220 uF 320VDC supply and was certain it would be best. Not really.
If you succeed I suspect it will be the softer hiss you will choose due to filtering. Class AB may not suit simple filtering. JLH should.
I usually have parts like that at no cost, hopefully you also. For an initial test expensive parts are not required. When I did this the slightly unusual 75 Hz noise didn't change much. 66kHz noise did and was worth doing. 10 uF polyester I suspect would have worked. That's 1000 times less. A big PSU might have good noise at 50 to 100 Hz or 60/120 due to spare capacity.
I tried a 220 uF 320VDC supply and was certain it would be best. Not really.
If you succeed I suspect it will be the softer hiss you will choose due to filtering. Class AB may not suit simple filtering. JLH should.
With a serial connection smps, you get 48 volts. This is more than necessary (((
Squarewave overshoot
Hi, I’ve been building and tweaking a JLH for the last 6 months. It’s a pretty much standard 69 design but does have two pairs of output transistors (mj15003) with 0.1 emitter resistors. Single rail, 32V at the moment but I will drop this with some more RC filtering later. 60mV across the emitter resistors when the amp has warmed up for Iq = 1.2A.
I tried BD139-16 drivers initially but have settled on BC337-40 because this gave lower distortion at frequencies above 10kHz in my setup. At 20kHz I measured a 7dB advantage. I’m using a small flag heatsink and it seems fine.
I have 2N3906 input transistors and haven’t been able to improve on that choice so far.
The amp is measuring OK for distortion and sounds good but I’m not sure if the overshoot in the attached needs addressing. This is measured into 8R in parallel with 2uF. There isn’t much ringing but I’m new to audio amps so would really appreciate some expert guidance.
Many thanks
Jon
Hi, I’ve been building and tweaking a JLH for the last 6 months. It’s a pretty much standard 69 design but does have two pairs of output transistors (mj15003) with 0.1 emitter resistors. Single rail, 32V at the moment but I will drop this with some more RC filtering later. 60mV across the emitter resistors when the amp has warmed up for Iq = 1.2A.
I tried BD139-16 drivers initially but have settled on BC337-40 because this gave lower distortion at frequencies above 10kHz in my setup. At 20kHz I measured a 7dB advantage. I’m using a small flag heatsink and it seems fine.
I have 2N3906 input transistors and haven’t been able to improve on that choice so far.
The amp is measuring OK for distortion and sounds good but I’m not sure if the overshoot in the attached needs addressing. This is measured into 8R in parallel with 2uF. There isn’t much ringing but I’m new to audio amps so would really appreciate some expert guidance.
Many thanks
Jon
I'm no expert but I think it looks okay, it's a tough load you tested with, try the same without capacitive load, could be fun to see the difference.
Thanks. Good idea. I just tried it and there is no overshoot at all. Just a slight rounding in both excursion. I'm curious as to what response is good and also trying to understand if there are any instabilities in the amp.
Many thanks
Jon
Many thanks
Jon
Try adjusting bias while monitoring the output with your capacitive load, you might be able to reduce the overshoot a bit, also noise figures,
You can use internal and small power supplys Mindwell SMPS (24V 100W like RS-100-24, or LRS-100-24) and they can be used in parallel with a diode separating outputs or in series. like explain in this manual.
They have an external model too (AS-120P-24-DB15)
Many thanks. I'll take a look at bias with square waves and different loads. All my adjustments so far have focused on sine wave distortion.
Cheers
Jon
Thanks pedro, but I have bought theese two beasts now, build quality looks great, can see used units go for big money, industrial quality.
Member
Joined 2009
Paid Member
If you have a spectrum analyzer, soundcard or other, you should adjust bias so that you get the lowest 3rd harmanonics, second harmonics are preferable, and lowest noise, so a mix of low 3rd and low noise, no need to worry about crossover distortion since its class a.
Play around see what happens.
Play around see what happens.
Member
Joined 2009
Paid Member
really ? Nelson Pass found that around 1/3 people prefer dominant 2nd, 1/3 prefer dominant third harmonic, and the rest had other ideas. Of course, the bootstrap resistor values have quite an influence on the harmonic profile too. Perhaps the bias should be set wherever it produces an acceptable temperature at the heatsink since more is often better the limit is usually thermal.