Okay, jerluwoo
That is very good figures.
Looks like I have not improved your original very much.
Maybe only increased the MAX power output a bit,
using 76 Watt Class = 2.0 A at 38 Volt supply.
-------------------------
Anyway, may original version schematic in my last post above:
http://www.diyaudio.com/forums/showthread.php?postid=1036543#post1036543
... had a -3 dB point at like 1.3 - 1.4 MHz!
Having such a fast circuit and attach even a little bit of capacitance at the output ( capacitive load )
is bound to turn the whole thing into an OSCILLATOR.
I have now after much trial and errors made a stable version of
my MJ21194 simple Power Follower.
1. I reduced the -3 dB frequency to 100kHz, using a normal lowpass input filter.
See attached bandwith AC-analysis
2. I added some local negative feedback around transistor
using the normal cap between base and collector method.
This resulted in this new trimmed and compensated version
is STABLE in at least 220nF across 5 Ohm LOAD.
Power output as well as distortion figures in this version
stays very much the same:
--------
0.100% THD: 13.84 Watt RMS in 5 Ohm
0.200% THD: 21.35 Watt RMS
At 22.23 Watt RMS output, the amplifier starts Clipping
because V-peak is above 15 Volt .. and amplifier needs almost 4 Volt margin to put out 3 Ampere ( 75% of theoretical 4 Ampere ).
Each output transistor at current 2A has got 19 Volt across at idle.
So we can say:
If believing simulation figures
this output amplifier never get a chance to get higher distortion
than 0.200%, before going into Clipping.
That is with the ideally suited load of 5 Ohm.
That is very good figures.
Looks like I have not improved your original very much.
Maybe only increased the MAX power output a bit,
using 76 Watt Class = 2.0 A at 38 Volt supply.
-------------------------
Anyway, may original version schematic in my last post above:
http://www.diyaudio.com/forums/showthread.php?postid=1036543#post1036543
... had a -3 dB point at like 1.3 - 1.4 MHz!
Having such a fast circuit and attach even a little bit of capacitance at the output ( capacitive load )
is bound to turn the whole thing into an OSCILLATOR.
I have now after much trial and errors made a stable version of
my MJ21194 simple Power Follower.
1. I reduced the -3 dB frequency to 100kHz, using a normal lowpass input filter.
See attached bandwith AC-analysis
2. I added some local negative feedback around transistor
using the normal cap between base and collector method.
This resulted in this new trimmed and compensated version
is STABLE in at least 220nF across 5 Ohm LOAD.
Power output as well as distortion figures in this version
stays very much the same:
--------
0.100% THD: 13.84 Watt RMS in 5 Ohm
0.200% THD: 21.35 Watt RMS
At 22.23 Watt RMS output, the amplifier starts Clipping
because V-peak is above 15 Volt .. and amplifier needs almost 4 Volt margin to put out 3 Ampere ( 75% of theoretical 4 Ampere ).
Each output transistor at current 2A has got 19 Volt across at idle.
So we can say:
If believing simulation figures
this output amplifier never get a chance to get higher distortion
than 0.200%, before going into Clipping.
That is with the ideally suited load of 5 Ohm.
Attachments
Fourier Analysis of New STABLE VERSION
🙂
Here is the fast fourier analysis of a 1 kHz signal
using this last version, restricted to 100 kHz upper freq limit.
We can see 2nd harmonic at -74dB is totally dominant.
3rd is very low at -113 dB
and 4th is also below -100dB, being at -103dB.
lineup
🙂
Here is the fast fourier analysis of a 1 kHz signal
using this last version, restricted to 100 kHz upper freq limit.
We can see 2nd harmonic at -74dB is totally dominant.
3rd is very low at -113 dB
and 4th is also below -100dB, being at -103dB.
lineup
Attachments
I agree very much with limiting the bandwith below 100k. my studies showed a large increase in gain above 100k to near 1meg, nearly 10db. In sims I used a feedback cap between c and b of input trans. to flaten this down to an even gain out to around 500k with a steep rolloff beyond this. I havn't been able to measure any oscilations with the built circuit but my oscope is kinda old(search eico model 460) and may not be picking up the smallest waveforms.This also made a noticeable reduction in the higher order harmonics but only a small change for lower order.
I'm not sure if my speakers present a heavy reactive load or not, havn't bothered to really delv into that end of the chain yet.
If you are familiar with klipsch type aa crossovers I havn't gone beyond that they work and sound great.
Another sortcomming I have found with this circuit is in trying to reduce power consumption of the input stage has a limiting effect to the available output voltage swing. The voltage stage can only swing negative to the point at which the emitter is sitting. With 100 ohm emmiter resistor its sits around 1v so the available unclipped swing for the output is 1v to supply rather than ground to supply. Increasing the emitter resistor and load resistor to keep the same gain at less current (example 470 emitter 4.7k load has the emitter sitting at 2.5v) reduces the available output swing another 1.5v. This was the cause of my finding earlier that the least distortion was found when vout = 1/2 supply +1v, because the emitter was sitting at 1v making 1/2 point for the swing 1v above acual half supply.
I'm not sure if my speakers present a heavy reactive load or not, havn't bothered to really delv into that end of the chain yet.
If you are familiar with klipsch type aa crossovers I havn't gone beyond that they work and sound great.
Another sortcomming I have found with this circuit is in trying to reduce power consumption of the input stage has a limiting effect to the available output voltage swing. The voltage stage can only swing negative to the point at which the emitter is sitting. With 100 ohm emmiter resistor its sits around 1v so the available unclipped swing for the output is 1v to supply rather than ground to supply. Increasing the emitter resistor and load resistor to keep the same gain at less current (example 470 emitter 4.7k load has the emitter sitting at 2.5v) reduces the available output swing another 1.5v. This was the cause of my finding earlier that the least distortion was found when vout = 1/2 supply +1v, because the emitter was sitting at 1v making 1/2 point for the swing 1v above acual half supply.
Class A Impedance Transformer, 21000:1
.
not much to add to comments
that are to read in my attached schematic
as for me
i consider this project as completed and finished
as is not many % of refinements to get
out of this surprisingly well performing circuit
when we think of it is such a very simple 4 transistor design
maybe you have not reached full understanding of this design
and what good use one can have of a circuit like this
and you might not agree with me
there are probably many that wont agree with me on practically anything
suffering the most from this - i am not sure it is me
😉
October 2006
lineup
Mastering at Lineup Audio Lab
http://lineup.awardspace.com/
.
not much to add to comments
that are to read in my attached schematic
as for me
i consider this project as completed and finished
as is not many % of refinements to get
out of this surprisingly well performing circuit
when we think of it is such a very simple 4 transistor design
maybe you have not reached full understanding of this design
and what good use one can have of a circuit like this
and you might not agree with me
there are probably many that wont agree with me on practically anything
suffering the most from this - i am not sure it is me
😉
October 2006
lineup
Mastering at Lineup Audio Lab
http://lineup.awardspace.com/
Attachments
I have no disagreements with you at all. True I don't fully understand everything about amplifier design but I'm working to better myself. I do think that your design would have better operating point stability. My best instrument I know how to use are my ears, so what I will choose to build and use will be what sounds best to me with the equipment I have to use. So I will have to build one with your implementation and listen to both for some time and see. I don't think the real determination of how good it sounds has anything to do with fractions of numbers in distortion.
I've had this little circuit on the backburner for some time. Finally got around to building a test circuit. Results are quite good. Very stable bias almost 0 drift. Good open sound. It is somewhat unconventional araingment in that the voltage gain is provided by the output stage. It operates true single-ended class a.
Attachments
jerluwoo said:
Basic Idea, Reminds a bit of the Famous DoZ amplifier, by Rod 'ESP' Elliott.
http://sound.westhost.com/project36.htm
An externally hosted image should be here but it was not working when we last tested it.
And you have used some clever ideas. Among these one idea I have had myself ...
And when 2 people, independently, thinks the SAME
it can be a good idea 😎
I download your schematic, jerluwoo.
Will have a closer look, later.
Has got some other things today.
Regards
lineup
I suggest 100 ohm Resistor between Q2 base and emiter for more stable operation.
Q3 and Q6 are darlingtons? Right? Why not single bjt??
cheers!
Q3 and Q6 are darlingtons? Right? Why not single bjt??
cheers!
Hi,
is JLW2 in post 58 the final circuit?
All these extra posts and mods have made it quite confusing to identify the recommended version.
is JLW2 in post 58 the final circuit?
All these extra posts and mods have made it quite confusing to identify the recommended version.
Yes post 58 is the recommended circuit. After all this time since I started this thread my prototype circuit still works quite well.
Class-A Amplifier in Post #58
.
If somebody looks at Schematic in Post #58
it might look as such simple and easy amplifier can never be good.
But I must say, after studying this first Class A tryout by jerluwoo
and running many many simulations
and hearing jerluwoo tell about his very positive real life experiences of this amplifier
.... I can not but say:
😎 This is a Good and really Nice audio amplifier!
Maybe because my friend jerluwoo
is a quick learner and very clever man.
Regards to everybody
halojoy
🙂
.
AndrewT said:
jerluwoo said:
If somebody looks at Schematic in Post #58
it might look as such simple and easy amplifier can never be good.
But I must say, after studying this first Class A tryout by jerluwoo
and running many many simulations
and hearing jerluwoo tell about his very positive real life experiences of this amplifier
.... I can not but say:
😎 This is a Good and really Nice audio amplifier!
Maybe because my friend jerluwoo
is a quick learner and very clever man.
Regards to everybody
halojoy
🙂
jerluwoo said:
By using 2 inverting stages + 1 follower
this makes a NON-INVERTING Power Amplifier.
Intended for lower powers, as the supply is +/-15 Volt.
We could use 4 Ohms LSP to get eventually a bit higher power.
With a good and strong Supply transformer
this could work = low voltage, but higher currents out.
-------------------------------
The input transistor is set up with CCS (constant DC current).
AC Gain is only 1 in this first inverting stage.
Second stage Q4 does the voltage gain.
Here is, the otherwise high gain limited and controlled
by using local AC+DC negative feedback via R2, 27k.
The Third output stage is same follower, that we have seen works very well
by testing and simulation earlier in this topic.
------------------------
🙂 Yes, this one should work well.
With good control of DC operation point.
These type of single ended amplifier stages, like in second VAS Q4,
was very usual in older transistor days.
The DC operaton point at Collector of Q4 keep rather stable
due to the DC negative feedback here.
The input stage and output stage with gain only like ~1
should not be very problematically.
Once they have a reasonable good (optimal) current level to work at.
Not run too hot,
but also not too little current
for to be able to do the job they should.
Drive the next thing
- whether this is another transistor stage or Loudspeaker load.
Good working, jerluwoo ... again!
Regards
lineup - doing many other things now ... like writing nice smart PHP Website Scripts!
😉
i got bored and was searching through the forum for something to build.. well.. i have a bit of a thing for simple class A amps so i thought i would give it a shot..
Have had to modify it all a little to use transistors / resistors that i had lying around but according to ltspice it should still work.. **fingers crossed**
I'm starting off with the basic version (no ccs / bootstrap) and i'll have a play then and see what i think and try out the different versions 🙂
Atached is the schematic of what i have actually built...
Have had to modify it all a little to use transistors / resistors that i had lying around but according to ltspice it should still work.. **fingers crossed**
I'm starting off with the basic version (no ccs / bootstrap) and i'll have a play then and see what i think and try out the different versions 🙂
Atached is the schematic of what i have actually built...
Attachments
well.... what can i say.... only problem i have is mains hum. need better power supply.. running +40V And fan cooled it is very nice.. sinks get to 32*C and seem to stay there no matter what i throw at it.. havent managed to blow it up yet which is always promising...
hmmmm...
hmmmm...
DoomPixie said:
DoomPixie said:
DoomPixie
Glad to see you are using my favourite method
= hardwire-ing = direct soldering with minimum use of PCB circuit board.
You mention hum.
Yes, with this simple approach amp,
you will need a very clean and well filtered power supply.
You can have a look at my multi filter Class A supply:
=> C-R-C-R-C-R-C .... using 4 big Electrolytics!!!! in power supply line ...
....
I say, those resistors/caps are NOT there just to impress
... such heavy filtering serves great
to make power supply clean enough for building Simple and Basic Class A amplifiers.
Reduce to very low level any hum-ripple coming from Power Supply.
DoomPixie
Your setup already works
You make some improvements and post back to tell us results
Thanks
I am sure, My other friend, jerluwoo ( topic starter )
would love to hear results from anyone using this kind of amplifier.
This amp is simple & it is very good,
when we see how basic this design is
and it is not very difficult to build
My Power Supply - Heavy RC Filtering
-------------------------------------------------------------
Post #80 ... here above:
http://www.diyaudio.com/forums/showthread.php?postid=1036543#post1036543
My schematic attachment in same post:
http://www.diyaudio.com/forums/attachment.php?s=&postid=1036543&stamp=1161488807
Regards
lineup
Lineup Audio Labs
http://lineup.awardspace.com/
i must say i was very impressed, i am deffinatley going to build a better power supply for this amplifier and post back..
The only other problem i am haveing is that above 1V input signal the amp starts clipping.. I think this is due to the resistor values i substituted into the circuit so i'll modify it and let you know how i get on..
I am a great fan of "point to point" wireing of simple designs, There is absolutley no need to make a pcb for this design so why waste time / money doing so? All the strip of pcb that i used does is serves as a ground plane and makes the curcuit a bit more sturdy.. My JLH is ALL on the main heatsink, no pcb atall, makes for minuimum of connections in the cirtuit which in my oppinion can only be a good thing.
I must say though for such a simple amplifier (i am a great fan of keeping it simple) it deffinatley has good performance and even with the problems i am haveing at the moment i already love the sound more than my JLH.. Lovely tight deep bass with almost valve like mids and crisp highs! The speakers i was tesating with arent great but have 12" bass, 6 1/2" mid and 4" tweeter, a lot of amps have problems driveing them (had to modify JLH to stop it oscilating) but this amp hasnt missed a beat yet! It is a pleasure to work with 🙂
Thanks,
Owen
The only other problem i am haveing is that above 1V input signal the amp starts clipping.. I think this is due to the resistor values i substituted into the circuit so i'll modify it and let you know how i get on..
I am a great fan of "point to point" wireing of simple designs, There is absolutley no need to make a pcb for this design so why waste time / money doing so? All the strip of pcb that i used does is serves as a ground plane and makes the curcuit a bit more sturdy.. My JLH is ALL on the main heatsink, no pcb atall, makes for minuimum of connections in the cirtuit which in my oppinion can only be a good thing.
I must say though for such a simple amplifier (i am a great fan of keeping it simple) it deffinatley has good performance and even with the problems i am haveing at the moment i already love the sound more than my JLH.. Lovely tight deep bass with almost valve like mids and crisp highs! The speakers i was tesating with arent great but have 12" bass, 6 1/2" mid and 4" tweeter, a lot of amps have problems driveing them (had to modify JLH to stop it oscilating) but this amp hasnt missed a beat yet! It is a pleasure to work with 🙂
Thanks,
Owen
- Status
- Not open for further replies.