Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

About make an audio power amplifier.

Status
Not open for further replies.
What capacitor should I use as input coupling capacitors? Polypropylene, polyester, silver-mica......??? Or even electrolytic? Too much types and brand: Rubycon, Elna, Nichicon.....:confused::confused::confused:.
 
For audio. Input coupling capacitors. Are you change my question into new thread? If yes, firstly thank you but my question for bigger plan: make an amp.
 
What capacitor should I use as input coupling capacitors? Polypropylene, polyester, silver-mica......??? Or even electrolytic? Too much types and brand: Rubycon, Elna, Nichicon.....:confused::confused::confused:.

Polypropylene by far. They can be hellish expensive if you go for the true audiophile types, like £50 for just one 1uF capacitor. But Wima do a range which are used universally in decent amps and they don't cost the earth. I think they are Wima FKP4 polycarb film type: https://www.wima.com/EN/WIMA_FKP_4.pdf

If you need a capacitor with a value over about 1u5F you will need to go for either aluminum or certain tantalums. Bead tantalums sound OK but don't like much current. Rubicon aluminum electrolytics are generally favored for audio work, and if you shop around they are not too expensive, but never cheap. One dodge with electrolytics is to put a pollyprop or similar in parallel. If possible try to keep electrolytics out of the audio signal path and especially out of the feedback network.

On the transitor equivalent thread I wrote a bit about components which is reproduced here:

General
People often think that a resistor is a resistor and a capacitor is a capacitor, and all the talk about differnt sounds is an illusion. As an engineer that was exactly my view, but I had a rude awakening when I built my first serious audio amp using ceramic capacitors. The next rude awakening was when I heard a good commercial amp with matching quality speakers and audio source.

The other thing is that, if you have ordinary speakers and average signal source you won't notice the different sounds of components. But if, for example, you have a pair of speakers that are revealing analytical and dynamic you certainly will hear a difference. Sometimes it is not imediately obvious but you just get the vague feeling that something is not right. The other point is that you can get listener fatiuge after a couple of hours of continuous listening- not so if you listen to the same music live.

Resistors
Metal foil resistors are the best type for high quality audio work. (inductance and capacitance are not too important). Metal film are the next best and far less expensive than metal foil.

The next quality resistor is tin oxide, but they do not sound as good as metal foil and film. Tin oxide are very stable and rugged and are best for general electronic projects. Because they are so reliable and stable, they are the military's preferred type- they are all I used at work.

Carbon composition resistors are liked by the valve boys, but I have not found them to be brilliant for solid state amps. I think it is because valve circuits have low currents and high voltage, while solid state is the other way around. Carbon film have very little going for them and are best avoided for all work.

There are also wire wound and ceramic composition power resistors (1W upwards). Special low inductance wire wounds will be required for the two low value resistors in the output transistor emitters.

Capacitors
Yes, polypropylene caps are what you want for the solid caps. Polycarbonate are the best, but they are not made anymore. In practical terms polypropylene are just as good

Electrolytic capacitors should be a good quality aluminium type, established as sounding nice, Rubicon and Alps for example.

Cetain tantalum capacitors don't sound bad but, for some reason, not the latest generation of high-value digital decoupling tants. Bead tantalums are quite good and have a sweet pleasing sound, but not as pure as polypropolene.

Don't put ceramic capacitors anywhere near the audio signal path or you will get a brittle disturbing sound. The first half-decent amp I built had ceramics- it sounded awful.

Having said that, there is absolutely no reason why you could not use ceramics to get the amps going. Later you can fit what you want.

Polystyrene also sound good, but you only get them in low capacitance values. Silver mica are very nice, but expensive and also not available in high values

Most run-of-the-mill solid capacitors are polyester. They are cheap and easy to get. They do not sound too bad either. My advice would be to use those if you find polyprops are too expensive or hard to get. Later you can upgrade. Polyester caps are not good for decoupling though.

In the dual-supply version of the amp, only the input capacitor is directly in the audio signal path, so that is the capacitor to concentrate on the most.
 
Last edited:
I used inexpensive Epcos Polyethelene film coupling capacitors in most of my hifi amplifiers and they sound the same as expensive larger polypropylene capacitors.
 

Attachments

  • EPCOS capacitors.jpg
    EPCOS capacitors.jpg
    57.4 KB · Views: 495
I used inexpensive Epcos Polyethelene film coupling capacitors in most of my hifi amplifiers and they sound the same as expensive larger polypropylene capacitors.
Hmm- nice advice. I have never heard those. Epcos components are always pretty good.
 
Issue: 06-01 2016_01_31

Nikolai,

I suggest that you use this schematic as the basis of your power amps. It is a classic Douglas Self design and has been optimized for high quality and simplicity. Make no mistake, it is a nice amp without being overly complex or expensive. Yes, it does use a few more transistors than simpler designs, but what is a few transistors here or there between friends.

The circuit can be inverted if required so that the two input difference transistors would become NPN instead of PNP. The performance of the amp would be exactly the same with either sex configuration.

It would be nice to get rid of C2 and replace it with a short. I will think about doing that.
Likewise it would be a good move to replace the input electrolytic capacitor with with a hifi solid type.
C9 and C10 would be better at 2m2F with pollycarbs across them.

The layout should be as shown with the similar amps in the transistor equivalent thread.

More decoupling also shown on the amp schematics of the transistor equivalent thread would also be good



ETO_self_simple_class_B_power amp_Iss01_2015_11_21.png


VAS= Voltage Amplification Stage
The earth symbols on the schematic are misleading and should be OV supply line
Transistors:
TR1,2,3,5,14: MMBTA06 NBJT SOT23 £0.10 UK
TR4,10,11,12: MMBTA56 PBJT SOT23 £0.10 UK
TR6: MJE243G NBJT TO225 £0.40 UK
TR8: MJE253G PBJT TO225 £0.33 UK
TR7: MJL3281A NBJT TO264 £3.00 UK
TR9: MJL1302A PBJT TO264 £3.00 UK
 
Last edited:
Nice Amp Spec. What happened to the power amp that you invested so much time into for Niko in the other thread. Seems such a shame that you would put in so much time, only to have your work go in vain :banghead:
 
Nice Amp Spec. What happened to the power amp that you invested so much time into for Niko in the other thread. Seems such a shame that you would put in so much time, only to have your work go in vain :banghead:

Hi Mike,

Yes, it is a nice amp. It strikes a good balance between quality, complexity, and cost. It is the Douglas Self Simple Audio Amp from his book, 'Audio Amplifier Design' edition 6, 2013. Reference (1.4) @ https://www.electro-tech-online.com/articles/books-articles-data.758/

My stuff is still all there on the Transistor Equivalent thread and hopefully will be referenced. All my theory is based on the work of Douglas Self and the amp above shows his simple implementation which incorporates nearly all the features that are needed for a good performance with that architecture. All that is missing is the practical side: layout, star points, decoupling, etc and a few twiddly bits like possible class A drivers. But that can also be read across as can the mechanical stuff, if required.

spec
 
Last edited:
Nikolai,

I suggest that you use this schematic as the basis of your power amps. It is a classic Douglas Self design and has been optimized for high quality. Yes, it does use a few more transistors than simpler designs, but what is a few transistors here or there between friends. View attachment 95493

Hola Spec,

Strictly speaking, would you call the input a long tailed pair?
 
I would like to see the amplifier's rated load impedance and rated output power driving it.
I would also like to see its harmonic distortion at its rated power at 20Hz, 1kHz and 20kHz.
IM distortion?
TIM distortion?
S/n ratio?
I assume that its frequency response is flat from 20Hz to 20kHz?
 
Cyril Bates mentioned Epcos metalized polyethylene or polystyrene capacitors but I do not think he measured their distortion.
 
Hola Spec,

Strictly speaking, would you call the input a long tailed pair?
Hola atferrari,

Indeed, that is exactly what the input stage is: a long tailed pair. But it is enhanced by having a constant current generator in the tail which improves the difference function. The difference being between the input signal at the base of TR2 and the feedback signal on the base of TR3. The funny transistor pair (current mirror) in the collectors is also a constant current generator providing a current drive for the next stage (VAS) and a high gain. But it also has the important function of keeping the current through the two difference transistors identical, thus enhancing the difference function even more. The two difference transistors and the current sharing transistors should really be matched pairs and in thermal contact with each other.

You may think that all this is overkill and so it is for an average good audio amp but even so there is more going on inside a BJT than you might imagine. One of the main factors affecting the difference amp is the transistor re. re is the effective resistor in series with the emitter inside the transistor itself. To a first order approximation re = 25/Ic Ohms. Where Ic is in mA.

As you probably well know, a resistor in that position causes negative feedback and not only dictates voltage gain but also base input impedance and other parameters. Bu re varies with collector current so it is a source of distortion. So what the designer tries to do is to make sure that both difference transistors in the long tailed pair have the same Ic, and thus the same re and thus the same characteristics.

Also the designer tries to minimize the changes in Ic for change in input signal. This also improves difference transistor performance and minimizes distortion. How is this done: by providing a constant current load in the collector. Another benefit of this approach is that the voltage gain is enhanced radically and not only that but the linearity of the VAS is improved because the input transistor base is fed from a constant current source rather than a voltage source which is more normal. So it is a kind of win, win, win situation. Of course the downside is complexity.

The type of feedback used on this amp is known as voltage feedback which has advantages and disadvantages. If you had a single transistor at the input and the feedback to the emitter that would be current feedback which has other trade-offs. Current feedback is much better in terms of open loop frequency response.

Audio designers go to extreme lengths to ensure that the input difference function is as perfect as possible. Here is a schematic showing an example of a mid complexity input stage to give you a feel for what is involved:

ETO_amp_complex_input_stage.png
 
Last edited:

Interesting Mike- thanks for links.
 
Surprisingly it has a lot of the characteristics of the Leach Amp. The transistors I recognize.

Yes, the Leach amp was way ahead of its time when it came out and is still highly rated now. One of the Leach amps outstanding characteristics is the output stage with a tipple Darlington arrangement and parallel output transistors. This excess output current approach tends to give an effortless feel with a nice bass slam- sorry for the non scientific terms but that is the only way I can describe it. It also helps the amp control difficult speakers.

Like the Leach, the Self Simple amp has a constant current voltage amplification stage but it only drives a Darlinton pair so the absolute hfe and the hfe consistency with IC and frequency of the transistors very much defines the sound and current capability of the Simple output stage. Nikolai is probably going to use TIP35/TIP36 output transistors on his amp which will be OK, but the Rolls Royce choice would be the fabulous On Semiconductor MJ3281A/MJL1302A, which incidentally are only £3 a shot from RS.

Of course, the other outstanding characteristic of the Leach Amp is the complimentary symmetry used from nose to tail. It is done so elegantly too.

ETO_Leach_audio_amp_approved by KISS.png
 
Last edited:
Nikolai,

I suggest that you use this schematic as the basis of your power amps. It is a classic Douglas Self design and has been optimized for high quality and simplicity. Make no mistake, it is a nice amp without being overly complex or expensive. Yes, it does use a few more transistors than simpler designs, but what is a few transistors here or there between friends.

The circuit can be inverted if required so that the two input difference transistors would become NPN instead of PNP. The performance of the amp would be exactly the same with either sex configuration.

It would be nice to get rid of C2 and replace it with a short. I will think about doing that.
Likewise it would be a good move to replace the input electrolytic capacitor with with a hifi solid type.
C9 and C10 would be better at 2m2F with pollycarbs across them.

The layout should be as shown with the similar amps in the transistor equivalent thread.

More decoupling also shown on the amp schematics of the transistor equivalent thread would also be good



View attachment 97140

VAS= Voltage Amplification Stage
The earth symbols on the schematic are misleading and should be OV supply line​
Oh no! MPSA05-56 not available!
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top