Class A amplifier kit - fixing the distortion problem
Vložit
- čas přidán 9. 07. 2024
- The class A amplifier kit I reviewed in the last video had heavy distortion at 12KHz on up. In this video, I correct the problem but a new one arises.
Error: at 10:46 I said R7 when I meant R5. - Věda a technologie
I met John Linsley Hood on a couple of occasions, His nephew worked in my shop. When I talked to him about distortion he said Transient intermodulation was far more important than people realised. That was the ability if a signal to rise and drop cleanly. Damping factor was critical. He always said use the best quality transistors in matched pairs and ensure you based the amplifier correctly and I should not have a problem. I have several bits of his designs including rare phono preamp with ultra low distortion. It caters for both moving coil and moving magnet inputs.
I still have a 30W+30W RMS amp which sounds so incredibly clean.
I would like to see the circuit diagram of the phono preamp, and make it for my Michell turntable with JLH 40 Watt class A updated amps.Could you possibly share it please?
John Linsley Hood was an instrumentation engineer who worked for a large company.
When I spoke to him about amplifiers he said that Transient Intermodulation Distortion (TID) was more important than harmonic distortion and he did actually design some kits for test equipment at the time he wrote most of his articles. He said most waveform generators could not produce a fit enough rise time to produce accurate measurements.
That's why he designed a lot of test equipment himself.
I also spoke to him about components and he thought that thick film resistors were superior to metal film types because they produced less noise. He also said that a good damping factor in an amplifier was essential because it stops the back EMP from the load on the output.
Keep the videos coming. It harks back to my youth when I made all my hi-fi equipment rather than buying it.
That was a brilliant follow up video, many thanks John. I wish more CZcamsrs took this approach, excellent use of knowledge, well explained and without being verbose. Good out come all round.
I don't know how I missed this channel all these years....Now I am your new sub.
Super analysis and thanks for sharing your debug and fixing ....love it.
Great the way you go through testing trying things and showing results !
Great job John, I have the same circuit board that arrived yesterday and this video will help me to fix this problem too. Thank you so much for this video.
People recovering old amplifier circuits or repairing vintage amplifiers based on 2N3055 (which were still widely used in the 80s in the Soviet Block) discovered that all modern 2N3055s or its substitutes are just troublemakers. That's why they reap old amps and receivers in search of the vintage, original 2N3055 or its analogs. Very good quality represent those made in Hungary by Tungsram, in former Czechoslovakia by Tesla (under name KD502 and KD503 with improved parameters) and in Poland by CEMI as BDP620. The KD502 and 503 are especially good and impossible to kill. There are reports that even in the hardest circumstances in the PA amps you could boil not only water but lead on those transistors and still listen to good sound. They were working even when the smoking holes appeared in their "caps" because of the overload when some drunk comerade has made a mess in the loudspeakers matrix :).
What's wrong with the newer 2N3055's?
Most of these transistors they are selling today, even in the reputable stores, is just fake .
What about the legit new ones? Why are the old ones better?
The old ones are better just because they're not fake. :) The new ones... are they still in production? RCA, Motorola? ;)
mouser still have onsemi 2n3055 in stock. not much but still there
Excellent video! I wish there were videos like this on car amplifiers and home audio receivers.
John, to toggle the beeper on the FY3200, press and hold the rotary encoder when powering the unit on. It will remember the setting between power cycles.
Another excellent diagnostic is two probes in an X-Y configuration on the input and output with a sine wave. Eg a straight line shows linearity, circle/ellipse shows phase shift, bumps in the circle or line show distortion very well
Thanks! That was quite educational. Please keep doing videos like this as I feel that I’m learning a lot. I have a TDA 2030 kit to build and if you’ve done that one let me know.
I wish I had seen this video, before I replied on the last. Your reverse engineering answered many of the questions I had about the schematic of this amplifier: It is indeed boot strapped, and the large 2.200 µF capacitor is used as the output capacitor, which makes more sense to me. With the new transistors, I think the boot strap impedance is just simply too high, to fully drive the new transistors, which lead to less pull up current on the output. You would indeed need to reduce the value of the 2.2kΩ resistor across the base-emitter of the low side 2N3773 transistor, and reduce the value of the resistor connecting the boot strap capacitor to the base of the high side transistor. The problem is, there would be more current across the 2N1711, running it hotter. Apparently, the 2N3055 transistors have a higher hfe,(higher ß)than the 2N3773's(more current gain)allowing a higher impedance driver circuit.
When you held the capacitor in your hand in parallel with the feed back resistor, your hand itself, introduced a stray capacitance, similar to touching the open input. When you installed the capacitor, the stray capacitance was removed, allowing a well damped output with the square wave input. I agree, you could downsize the capacitor, maybe 220 pF, but 330 pF would be safer. I would use the smallest value I could get away with. Nice video, nice fix!
The 2N3773's also have a lower transistion frequency of 200kHz. Could also try replacing the 2N1711 driver transistor with a BD139-16 or 2SC3421 to handle slightly more bias current power. There's maybe a suffixed version of the BC141 as well, BC 141-16 or the ZTX 653 for a direct drop in replacement. There is the connection terminal position difference to sort out as well with the first 2 replacement driver transistor suggestions.
Nice work John.. Thank you.
Two things:
(1) When you have asymmetric ringing, it suggests that the ac current path to ground directly is different from the ac current path to ground via the supply rail. In other words, the supply decoupling is insufficient at these frequencies, which is hardly surprising given that big electrolytics will have considerable impedance from their parasitic inductance at 10KHz+. So one of the first things to try is to decouple the supply with a decent 100nF decoupling capacitor in parallel with the electrolytic.
(2) The compensation capacitor you introduced across the feedback resistor will indeed improve the stability of the amplifier by increasing the phase margin at higher frequencies. Its direct affect will be to roll off the hf response of the amplifier at a frequency that you can calculate: f = 1/2πCR. A 500pF cap across a 2K7 resistor rolls off at around 120KHz, so well beyond the audio range, which suggests that you could use a higher value capacitor if needed.
There's no point in worrying about "slow" rise times on a 20KHz square wave for audio use as I've yet to meet anyone can hear the difference between a 20KHz square wave and a 20Khz sine wave. After all, the difference lies principally in the third harmonic at 60KHz and you'd have to be a dog to hear that. Who said "on the internet, nobody can tell you're a dog"? Well, testing the difference between a 20KHz square wave and a 20Khz sine wave would be a sure-fire way to decide.
When you hold the cap in your hand instead of tacking it, you change the caps value because of the capacitance of your body.
There are two possible answers to your question: 1. The information is not available online, then I can not find it. 2. The information is available online, then you can find it with the same effort as I can.
No, far more likely to be oxides on both components leads. The body capacitance would have very little effect on the 500nA cap
There is another possibility, the information on line is wrong as in this case. I like John prefer to call on many years of experience. If you concentrate off line you to can learn this discipline, basically less time here talking nonsense. You must let go of Googles apron strings.@@1959Berre
Hello John thanks a lot for the great video. I've noticed two little trimpots on the board; what are they for?
I wonder how this amp would sound with the suggested transistor substitutes.
What about the Class A circuit that was used for the color gun amps one for each of the colors R,G,B of those color TV sets that use a CRT, that unit is simply not wide band.
Nice job! Thank you.
I read the paper of J. L. Linsley Hood about his very interesting design. One problem of this design is, that it is not possible to adjust the bias current and the offset voltage independently by use of two trimmers resistors. You have to recalculate the whole circuit if you need to go to other (lower) speaker impedances. His original design was dimensioned for a speaker impedance of 16 Ohm, which you will not find today.
Just use two 8 ohm speakers in series, that would fix that problem..
Hello guys, please advise which of those three 220Mkf capacitors (Black ones) is responsible for reversed Negative feedback (C3 on the drawing) ? I want to replace it with a silmic2 470mkf. I heard this would enhance the sound. Thanks
Is there any chance that this distortions affect a computer ?
Good educational videos! The thing with your distortion measuring is very high noise floor of the scope, so you don't know what's happening below 0.1-0.2%. All of these kits and chipamps sound different with good speakers and source. You could do better fft measurements with a cheap soundcard and free software, like Arta. In testings like this you dont see the difference between various harmonics, (below the clipping, where you actually listen to amplifier) and that is important because it effects the sound.
is there a zobel network across the output ?
i have a question about this amplifier i saw the circuit the output is coupling with capacitor. My question is during turn on the power will the speaker have a pop sound? will it become problem? or the pop sound is loud?
IMHO distortion at 12 KHz is a non-issue, unless you plan to play music for dogs. A distorted 12 KHz signal will have some energy at 24 KHz and higher. Ugly on the scope, yes, but inaudible. I really don't see this as a problem for all but the placebo audio bunch. There is I admit, the sole argument of making it better... because we CAN. So, OK, carry on!
Nice follow up!
500pF parallel R5 corect ceramic or polyster ?
Hi dear,
I am looking for a Class A/AB High-end amplifier to play bass-guitar (with a pre-amp plugged before it).
please wich one do you recomend?
The complet produt with all the boards, power supply, case-box, etc..
best regards,
Gabriel.
Very good improvements. I see that you have replaced the 3773 I think by another ones with lower caps, I don't have catched the references that you have used, old 2n3055? The differences were huge! Very good explanation.
In a well made conventional amplifier (3 stages architecture for example: input stage (input: signal source and negative feedback, responsible to correct the output, output: current), volt-amplifier stage (input: current, responsible for the open-loop gain, output: voltage), and output stage (give the current to drive the load, gain nearly unity)), there is a current source (in the order of mA) in the input stage, and in the volt-amplifier stage, and their value affect slew-rate (higher current = better, but too much = other drawbacks).
But, in this design, it is difficult to tell what is the main thing that affect slew-rate: not an optimal design. Also, the linearity of the output stage (class A) isn't fully exploited: no differential pair with current mirror and tail current (poor distortion reduction), and probably not enough open-loop gain to provide better linearity. For a little bigger board, it should be possible to add other components, and the small extra cost will be worth it for the distortion performance gain.
There is a possibility that the substitute transistors are the counterfeits. They were sent to you for free along with the kit, so you really don't know their "pedigree".The 3055's were probably recycled from all the junk we send to Chin, that's why they look so beat up. (A very common practice over there)Besides, 3055's are so generic, commonplace, and cheap - I don't think anyone would even bother to counterfeit them. Actually relabeled 3055's are frequently passed off as high end audio power transistors.
Not for anything but do you think when you threw those transistors across the room it was a good thing?
What distinguishes this piece from others?
According to his book the design uses mj480 transistors not 2n3055s. I lost my copy of his original Wireless World article but I remember he did write a modification to it about a year later.
thank you for this video. I just ordered this kit some time ago and got it assembled yesterday. Mine came with 2d hand 2N3055. However, I have some other NPN transistors in my bin which could give better result. I think 12 volts is not enough to power this circuit but no datasheet came with mine and I don't know how high I can power the circuit without frying it.
Nice job
very good review /could you expertise a naim NAP140 kit/in advance thanx
4:40 - Looks like an RC circuit problem, but only in the positive direction!
I dunno that board is crammed. Looks complex.
How hard can you drive an amp into a static load? I have a 1000W 8 Ohm load but i'm worried to use it because once i lent a 2600W amplifier to a scientist for an experiment with mice triggering a device or a computer network. There was a huge 4 Ohm coil but when we drove it with the amp it saw it more like a short circuit and i blew the amp to smitheries.
nice work good job
I think it's the leads' inductances that caused ringing when touching the cap. after soldering there was no lead inductance hence no ringing.
I hope you can return to your class A amplifier type video.
I am particularly interested in the transformer output design.
Since the 2N3055 can operate at a collector to emitter voltage of up to 60,could you try using a much higher VCC?
This would more closely mimic a valve amp output,requiring less quiescent current with less heat providing the transformer primary D.C. resistance is lowish.
Would this work and sound any good?
I have never tried it.
my amplifier just got full distorted like blown , bass is terrible.. it is blown? what should i check :((((((
A good follow up John. I could not believe the original Hood amplifier could be so bad. Many people built this in the past and it had a very good reputation. Over the years the circuit was up dated and power increased. Does anyone know if the updated version is available as a kit?
I've seen a kit for the updated version on ebay but best to buy the bare boards, and source better transistors you can trust,use good outputs.not too fast, say 4 meg max ( 2n5886 in my last because I had them) and don't mount them on the boards.
I've found better stability with leads about 3-4 inches to BIG heatsinks.
Well made, it's still superb, read the original articles, they are still valid.
Output transistor match is critical for best performance.
It's not the original. How can you use different transistors and call it original?
Also the operating Voltage.
Is that ringing bcz different hfe value
Josh, I have a pair of bx5a's studio monitors by M-Audio that have a filter on the amp to cut off anything below 50hz. What can I do to change that so I can get 30hz?
This doesn't seem particularly advisable. Monitors in this class don't generally extend too far below 100 Hz anyway (for any sensible level spec such as -3 dB), so the main purpose of this highpass is keeping low frequencies off that would mainly make the woofer pump air and generate IMD due to high excursion. Thus level handling is improved at generally negligible cost.
Thank you for the response. I took your advice and left them alone and added a TRUAUDIO sub to my studio instead of going crazy. It sounds amazing by the way.
You should do a video on how you setup to measure distortion.
That was shown in the previous video, he uses a spectrum analyzer.
Can you suggest a good amplifier kit because im new in the audio stuff i have been following you for like a year now and i want to build a nice amp around 50 or 100 W
The TDA7293 kit I reviewed is great.
If you would like to build a good class A amp skip the class D’s amps. Go to diyaudio.org you’ll find amp kits with circuit boards in a community that’s willing to guide you a little bit with very good visual photographs on step-by-step assembly.
Interesting. The thing I notice about the circuit is that it has "active" drive for the negative swings, but just "resistor" drive for the positive swing. Kind of like the difference between RLT and TTL digital logic. As a result, I think it's going to be pretty sensitive to what it takes get the top transistor turned on by (from the resistors) as well as how long it takes the bottom transistor to "leak down" until it turns off.
In this design, the lower transistor is signal driven and the upper transistor is a current source. The current source gets modulation from the signal for better efficiency, but is still considered class A.
Something else that I'd check - the voltage at the point between R8 and C4. It seems to be the supply for turning on the top transistor. Perhaps C4 needs to be made bigger.
C4 appears to be a bootstrap. If it were too small, low frequency drive would suffer.
From where is the kit?
Electrolytic capacitors are terrible for audio especially at high frequencies, unfortunately they are necessary because of the size needed for the output capacitor. It may help to use a high quality part with low ESR, and also to bypass it with a ceramic capacitor in parallel to give a better high frequency path
God did we make money in the capacitor plague aftermath.
Molson, that is correct. It had a major impact on computers back then
I have the same JLH 1969 kit with a tube preamp module driving it. I replaced the 2N1711 driver transistor with a transistor socket into which I plugged in a 2N3866 UHF/VHF RF transistor and it improved the sound - so using RF transistors might be a way to solve the issue of slow rise.
Hi, I like to leave my amplifiers "always on"! Is this possible or should I always power it down? Was thinking of silent cpu fans to keep temperature steady. May I ask, does it continue to heat or does it need to be held below a limit? I turned mine off at 40°c. Thanks
@@mike.a.irving8227 You can leave the JLH 1969 on as long as the main output transistors (2N3055) are adequately heatsinked. I just use the chassis (part of the enclosure box) as a heatsink without a fan and kept it on for 150 consecutive hours to give the unit a burn in without any problems. Check out www.alvenh.com/misc-projects/tubeamp for details.
Nice build in the link.@@alvenhchanne
did i understand it right?: i only have to put on another capacitor and then it works fine?
The parts they supply in these Chinese kits can vary so you have to run tests.
I NEED THAT TOY!!!
I think most people wouldn't really notice the distortion at those kind of frequencies anyway, so it wouldn't have mattered too much. Still a good fix though. I also say "bachree" instead of "battery" but then that's the way everyone else round here says it.
Did you tried to use 2s5200 instead of those 3055?
Edit: *2SC5200*
I've used some 5200 and are really good!
And, at 18-20 volts use 1-1.2 amps... from all A that i have used, this one are from farest possible way the most "efficient" and powerfull amp XD
I actually just got one in the mail that uses 3 C5200 Toshiba transistors and I haven't tested it yet but what voltage and amperage do you recommend I use for setting the biased and output ? And them what kind of voltage and amperage do you recommend I use for running this thing at 4 ohms ?
During the same year in 1969, J. L. Hood published another article (Wireless World October 69) which explains the same phenomenon, and workaround. If anyone would like to learn more about, and fix it ultimately, I recommend to read the article. The introduced board by John is far from the engineering excellence as he pointed from many angles. Unfortunately the old 3055 with low < 1MHz current gain bandwidht or a better PCB layout do not solve the oscillation by one simple magic hit. The circuit needs some additional RC network here and there. Happy reading.
Szatmari Mihaly Cheap 1969 amp from China are pretty good.
top job
From what I see concerning the polarization of the output stage, it seems that this amplifier, although it is said to be class A, it works as class B.
What's your opinion?
The output transistors are on for the full cycle of the waveform, therefore it is operating in class A
@@JohnAudioTech Correct, it seems that both transistors conduct in the active region all the time and there is an overlap of the signal generated by both at the output. So there is a reinforcement of both.
Great Video. I need to open up some (remove the top if possible) 2N3055's from China without destroying the transistor. Is there any chance that cases or portions of cases are made of Beryllium and therefore toxic if inhaled?
I don't think so mate, the casing on those is just metal.
Beryllium Oxide was used almost exclusively on RF power transistors
@@Dazzwidd That's just berylliunt.
Really enjoy the videos. Wish you were a guitar player and was into experimenting with tube and SS guitar amp circuits.
Have you ever considered making a video about headphone amplifiers? And btw, I hope you dont mind me copying your home-made pre-amp design. Already ordered all the parts
Yeah, headphone amps, especially good sounding ones for higher (like 300 Ohm) impedance cans.
I have a amplifier with a stk4141 ii it has distortion i think i need to replace it bcuz itz a Ic and it has the output transistors inside
And my amplifier is a kenwood ka35 not much inside lol everything is driven by that big ic
And everything elese looks fine
Its like notting on the board
My transistor guitar amp is hissing at every note played, but silence when im not playing anything, can anyone help suggest a solution? This is way too technical for me to understand, thank you
Probably has a noise gate so that when you don't play anything the noise is below threshold and so the output is suppressed. When you start playing and the threshold is met, the gate lets go and you can hear the noise. Hiss as opposed to buzzing/humming is typical of amplification noise, especially in cheaper designs. If it's humming/buzzing, it's probably mains related and can definitely be reduced based on the setup!
Thanks, it's interesting. What a new one problem arises?
If you have issues with that beeper being too loud just put some tape on its went if that makes sense and it will be slightly less annoying.
This is advice from someone that repairs APC units and their beepers are annoying as hell.
Cheers
Thanks great video
Thanks
Saludos. Este. Amplificador. Es. Estéreo. Y. Vatije. Es. Gracias
Teşekkürler
Usually John throws these amps in the ash heap of history, but for some crazy reason I thought this Class A would bug him into fixing it. Kinda strange I thought that. So John, did it eat at you for a while?
Yes! I wasn't sure I really wanted to "reverse engineer" the board for the schematic to get me on the way.
don't bother to fix the ringing its 200KHz, it is not audible. Anyway if you want to fix it put a 100 pF parallel to R4.
how is the audio quality?
Sounds okay to me.
Increase the size of the capacitor
Does any body know where she is? What is her name? Watching her video after 6 years and hopefully see her if those bright eyes are not taken yet!
Kiss Analog needs to see this video...
Ha. BATry is British, blessings John!
1:24 - It does not 'beg the question', it raises the question!
My amplifier works well, but at high volume, bass more, i hear distortion. 😥 It is an Alto mac 2.3
my only class A amplifier is modded Apex FA9.
not a good square wave for the amp
Great solution 👍
And ,I don't like the duck , lol
10:57 for solution!
sawtooth wave
Nice!! Chick poops...yummy....LOL
You need to download the JLH 1969 Design to compare. This clone is crap should get bandwith 10Hz to 100KHz and THD < 0.1% at 10W. Also you need to change component values for different loads.
Ver very classic troll amplifier.
Diagram
sound.whsites.net/jll_hood.htm
I rather enjoyed the sleuthing there John...
1:25 - I'm sorry but it does not "beg the question", it raises the question. :-)
I'm from India.
I want class a amplifier. Details send me 🎉
Cheap badly made capacitors have a tendency to have a poor bandwidth from their reactive corner frequency to the high corner where they tend to drop off quite early and often less then 20kHz making them not audio grade.
ничего же не понятно... можно на русском?)
Why bother
Aussies pronounce battery "bat tree".
Steve Smith No we don't. Never heard it here in 66 years.
QuadMaestro do live in Perth ?
Bat tree is a yobbo thing..
Steve Smith No, Melbourne. Not that it's relevant. Your claim is that all Aussies do it, and I am an Aussie and not only don't do it but have never heard it, which is refutation by counterexample.
Bat tree came from Sunnyvale NSW lol.
not working FIX capacitor + ALPS 10kx2 distortion sound bad. Sound hissing low volume .
wut
Sorry John, but that's a class B amp, maybe AB, but certainly not class A.
Oh yeah? I await your circuit analysis of why you think it is not class A.
Simple, really two output transistors Q1 and Q2 , driven in anti-phase by Q3, the junction capacitively coupled to ground, classic class B. if it was class A then there would be either a resistive load to one output transistor, or perhaps a constant current load, or even a transformer to couple the output, the giveaway is the two output transistors being driven, although the current through them is excessive.
Nope. A push-pull amplifier can still be class A - the definition of which merely states that the output device(s) have to be active 100% of the time. In fact, one would not even use an output stage like this in an AB amp, since it performs poorly as such.
Conversely, any kind of B/AB topology can also run in class A if beefed up accordingly.
It may look class B but it is biased hard to keep both outputs on for the full waveform cycle which is the definition of class A. Constant current drawn off the supply with load or without, with or without a signal is also indicative of class A operation.
It's still class B if the output consists of two output devices are actively used to source and sink the output, even if the circuit is so poorly designed as to draw a large standing current. Class A has only one driven output, You are confusing this with a class A amp which has a constant current load. This circuit was very common in the 60's and 70's , before we had complimentary transistors, and I built many of them, in fact I used a very similar design, but with Q3 capacitively coupled and biased to get a proper quiescent current in the 2N3055's.
Please stop saying "acrost". There is no such word.