Kuei Yang Wang said:
I disagree. Provided you have 2 amps, which are resistent to RF/EMI generated audible components, they do not exhibit cross-over distortion residuals, are fast, stable - the one with lower distortion sounds better. If you ever listened to well recorded philharmonic orchestra - and I guess you did - you must have recognized it. The problem is to compare amp like Z-30 (http://web.telecom.cz/macura/sinclair.jpg) with simple SE amp. Yes, THD says nothing in this case, but we both know we do not speak about simple numbers.
Me, I prefer 4th and 5th harmonic below -100dB and higher order below -120 dB.
All the sweet tube amps etc. colorize badly, just listen orchestra.
Konnichiwa,
The Bias is stablised with a thermistor, not the usual VAS Multiplier, plus diodes, you do need sorta the right kind of components and remember to bias the output stage fairly "warm".
But in the studio the signal was from a Neumann Kondensator Mike, with no limiter or compressor! More difficult than 99% of all CD's.... ;-)
In fact amplifying Vinyl signals well is very difficult, moreso than CD (the reasons will become aparent when considering the real world)....
Originally yes.
Sayonara
PMA said:
The Bias is stablised with a thermistor, not the usual VAS Multiplier, plus diodes, you do need sorta the right kind of components and remember to bias the output stage fairly "warm".
PMA said:
But in the studio the signal was from a Neumann Kondensator Mike, with no limiter or compressor! More difficult than 99% of all CD's.... ;-)
In fact amplifying Vinyl signals well is very difficult, moreso than CD (the reasons will become aparent when considering the real world)....
PMA said:
Originally yes.
Sayonara
Konnichiwa,
I am not sure what a "sweet tube amp" is (has it got a sugar coated chassis?).
However, it is easy to demonstrate that a good SE Amplifier in a suitable system has a drastically lower level of coloration than the best speakers and a much more realistic sound WITH THE FULL ORCHESTRA than pretty much anything Solid State.
If you are in London feel free to drop in for said demo, in the meantime consider taking what I call the "Geddes" test.
Sayonara
PMA said:
I am not sure what a "sweet tube amp" is (has it got a sugar coated chassis?).
However, it is easy to demonstrate that a good SE Amplifier in a suitable system has a drastically lower level of coloration than the best speakers and a much more realistic sound WITH THE FULL ORCHESTRA than pretty much anything Solid State.
If you are in London feel free to drop in for said demo, in the meantime consider taking what I call the "Geddes" test.
Sayonara
Wow, a scoop! Thorsten is not the 100% Englishman, calm and selfcontrolled he appears !!
So this amp is a sweet remind of yours youthness?
Kuei Yang Wang said:
Yes, every speaker has its strong sound color, regardless this fact you can distinguish small differences between amps, as the nature of both distortions is different. (I usually mean the distortion of dynamic spectra).
I have different experience with good SE amplifiers and I am afraid that in this point the discussion starts to be useless, as almost all of them here. It is not about paper persuading, but real test, real listenning - the only prove. We can blame each other we have not heard the "right" SE amp or the "right" another amp, and this is absolutely useless.
Konnichiwa,
Actually, funny thing.
The original Audio Innovations Alto Amplifier from the 90's used more or less the same schematic as well, it measured dreadful by modern standards but consistently did very well in blind group tests in UK HiFi Mags, it also has a great reputation in general in the UK, shame Richer Sounds had the electronic design changed over to "just another generica" Amp when they revived the Amp and CDP (they still look great though)....
Sayonara
plovati said:
Actually, funny thing.
The original Audio Innovations Alto Amplifier from the 90's used more or less the same schematic as well, it measured dreadful by modern standards but consistently did very well in blind group tests in UK HiFi Mags, it also has a great reputation in general in the UK, shame Richer Sounds had the electronic design changed over to "just another generica" Amp when they revived the Amp and CDP (they still look great though)....
Sayonara
PMA said:
Maybe for a professional like You, there is none really new here, but I can assure You that most of the diyaudio thread are full of information for diyer like me.
You and Thorsten are both very well skilled and with lot of experience. Is it common that You have different preferences, and this is a right and a good thing.
Please do not leave this discussion. We're not playing with the aeternal diatriba <solid state vs tube>: once one decided to give a try of a good sounding, simple and cheap 15-20W solid state amplifier what are the schematic he should consider?
Here there is the list comes out up to now:
Pavel Makula's error correction amplifier
http://www.diyaudio.com/forums/show...9442#post469442
Mauro Penasa GC variant
http://www.diyaudio.com/forums/showthread.php?s=&threadid=54571
Graham Maynard GEM
http://www.diyaudio.com/forums/showthread.php?s=&threadid=60546
Doug Self Blameless Amplifier
Hiraga 20W class A
http://www.tcaas.btinternet.co.uk/hiraga1.htm
Hiraga 8W Le Monstre
http://www.tcaas.btinternet.co.uk/monster31.htm
Giovanni Stochino Ultra fast amplifier
http://www.soton.ac.uk/~apm3/diyaudio/Stochino_circuit.jpg
East Germany 50 by Thorsten memory
http://img361.imageshack.us/img361/153/hifi500en.gif
Bartolomeo Aloia Volksamplifier
http://www.audiodesignguide.com/ba/volkampl1.jpg
Pavel Dudek PA03 Gainclone variant
http://czechaudio.com/Projects/PA03/PA03_schema.pdf
Nelson Pass Aleph, for example mini Aleph
http://www.diyaudio.com/forums/showthread.php?s=&threadid=2001
All these have advantages and drawbacks, as usual in any field.
In the community opinion what are the sonic and technical character of these? If You were forced to build just one of them, which one?
Upupa Epops said:
I have to support this 😉 . Unfortunately Pavel Dudek is known here by his "gainclone" PA 03. We know its + and - 😉 😉 . As a chipamp, very good one. In absolute measure, nothing special, just easy to build. I would better recommend Pavel's DPA38X series, MOSFET with Hawksford's error correction.
Nice to know this. So we replace Pa03 from the prevous list with the Pave's DPA38X. The schematic of this last entry is posted somewhere?
Regarding the error correction amplifier, for my present understanding layout is not so straightforward.
Do You made it on a PCB, that will be released in a near future?
Regarding the error correction amplifier, for my present understanding layout is not so straightforward.
Do You made it on a PCB, that will be released in a near future?
Konnichiwa,
First, how many Amplifiers have you build and debugged so far? If the answer is "not too many" many of the more advanced discrete designs may simply be "a bridge too far".
If you start with a nice case plus heatsinks and enough space, plus a sensible set of batteries for the supply (there is no real point IMHO in having an AC supply when building a solid state Amplifiers, SLA's are not that expensive and miles better) you can start simple and later simply exchange the audio PCB for a different one.
If you want something SERIOUSLY CHEAP that plays well above expectations I would recommend the "Inverted Gainclone" (IGC from here) with a later addition for unconditional stability as the main circuit. I notice that the IGC is out of favour in some circles, aparently it is too much of a "wire with gain" compared to the Non-Inverted version. Well, it's easy to change around.
Using the Chip inverted kills the early effect induced distortion neatly. Remaining distortion residuals from the LM3875 are all very low order and low in value, in general as good and better than most discrete circuitry. There is much internal circuitry the benefits inherent linearity on a stage for stage level. I am unaware of ANY other Amplifier chip that is as consequent about this and equally of very few discrete circuits.
Anyway, these chips are incredibly hard to blow up and generally get a good rate of success.
By combining a 250K or higher value linear track tandem pot with the 10K input impedance of the IGC you have a perfectly tracking stereo pot generally with much better sonics than ANY "audio taper" one.
The key drawback is that having input stage and output transistors on one chip some thermal feedthrough is unavoidable, HOWEVER as you are looking for an amplifier meant to play at fairly power levels you should be okay, especially if you add some Class A biasing for the output stage.
Class A to 1W implies a current sink for an extra 0.5A (LM317?) and with a +/-24V supply (4 pcs 12V/10AH give around 10 Hours play time without recharging) you dissipate 24W static per channel, of these 12W in the LM317 and 12W in the LM3875, quite doable with modest heatsinking.
Output Power at clipping will be around 25W into 8 Ohm BTW.
If you start with that I'd think you have a rather excellent amplifier as a result and the option to try some more complex and difficult projects.
Sayonara
plovati said:
First, how many Amplifiers have you build and debugged so far? If the answer is "not too many" many of the more advanced discrete designs may simply be "a bridge too far".
If you start with a nice case plus heatsinks and enough space, plus a sensible set of batteries for the supply (there is no real point IMHO in having an AC supply when building a solid state Amplifiers, SLA's are not that expensive and miles better) you can start simple and later simply exchange the audio PCB for a different one.
If you want something SERIOUSLY CHEAP that plays well above expectations I would recommend the "Inverted Gainclone" (IGC from here) with a later addition for unconditional stability as the main circuit. I notice that the IGC is out of favour in some circles, aparently it is too much of a "wire with gain" compared to the Non-Inverted version. Well, it's easy to change around.
Using the Chip inverted kills the early effect induced distortion neatly. Remaining distortion residuals from the LM3875 are all very low order and low in value, in general as good and better than most discrete circuitry. There is much internal circuitry the benefits inherent linearity on a stage for stage level. I am unaware of ANY other Amplifier chip that is as consequent about this and equally of very few discrete circuits.
Anyway, these chips are incredibly hard to blow up and generally get a good rate of success.
By combining a 250K or higher value linear track tandem pot with the 10K input impedance of the IGC you have a perfectly tracking stereo pot generally with much better sonics than ANY "audio taper" one.
The key drawback is that having input stage and output transistors on one chip some thermal feedthrough is unavoidable, HOWEVER as you are looking for an amplifier meant to play at fairly power levels you should be okay, especially if you add some Class A biasing for the output stage.
Class A to 1W implies a current sink for an extra 0.5A (LM317?) and with a +/-24V supply (4 pcs 12V/10AH give around 10 Hours play time without recharging) you dissipate 24W static per channel, of these 12W in the LM317 and 12W in the LM3875, quite doable with modest heatsinking.
Output Power at clipping will be around 25W into 8 Ohm BTW.
If you start with that I'd think you have a rather excellent amplifier as a result and the option to try some more complex and difficult projects.
Sayonara
Good point by Thorsten, more complex discrete amps, especially those with error correction, need to set operating points to give their best results, especially if the components are changed from original.
And I would always recommend to every DIYer to measure his electronic baby. An oscilloscope and signal generator is a must, IMHO. Otherwise it is difficult to tell any judgements.
And I would always recommend to every DIYer to measure his electronic baby. An oscilloscope and signal generator is a must, IMHO. Otherwise it is difficult to tell any judgements.
Konnichiwa,
Yup, plus most amplifiers do not have enough PSRR to make the power supply inaudible, or suppressed to below the noisefloor at least, hence often a very poor circuit with a good powersupply provides better sound than the reverse.
Yes, also even fairly basic Soundcards can manage FFT with a resolution to better than -80db, add a seperate filter for the fundamental and a suitable scaling amplifier (both nice little eductaional projects) and looking down as far as -120db using a standard PC or Notebook is not that difficult.
IMD test with at least 8 discrete tones can be very telling, but most PC based soundcards have insufficient bandwidth to make this useful, however 500/600/700/800/900/1K/1.1K/1.2K at equal levels and well below clipping can still be instructive even on AC97 sound systems, in this case the old "Null" test is good, subtratc input from output and FFT the residue with sufficient amplification for scaling.
Nice what sort of tools modern technolgy gives us, when I think back to what we had in the 80's in Eastern Europe.....
Sayonara
PMA said:
Yup, plus most amplifiers do not have enough PSRR to make the power supply inaudible, or suppressed to below the noisefloor at least, hence often a very poor circuit with a good powersupply provides better sound than the reverse.
PMA said:
Yes, also even fairly basic Soundcards can manage FFT with a resolution to better than -80db, add a seperate filter for the fundamental and a suitable scaling amplifier (both nice little eductaional projects) and looking down as far as -120db using a standard PC or Notebook is not that difficult.
IMD test with at least 8 discrete tones can be very telling, but most PC based soundcards have insufficient bandwidth to make this useful, however 500/600/700/800/900/1K/1.1K/1.2K at equal levels and well below clipping can still be instructive even on AC97 sound systems, in this case the old "Null" test is good, subtratc input from output and FFT the residue with sufficient amplification for scaling.
Nice what sort of tools modern technolgy gives us, when I think back to what we had in the 80's in Eastern Europe.....
Sayonara
Komban wa Thorsten,
yes, I support the use of PC cards, and would like to add one possible problem: in case of amplifier oscillations, that are often above 1MHz, the soundcard in the scope mode would not find them, for the low sampling frequency. For this reason, I would support an idea of an oscilloscope, even a cheaper type.
Oiasumi nasai 😉
yes, I support the use of PC cards, and would like to add one possible problem: in case of amplifier oscillations, that are often above 1MHz, the soundcard in the scope mode would not find them, for the low sampling frequency. For this reason, I would support an idea of an oscilloscope, even a cheaper type.
Oiasumi nasai 😉
Kuei Yang Wang said:
has anyone tried to conduct an experiment to see how "audible" a PS is? I am curious as to real, factual basis of the above statement.
Yes it is, and the "wiring" of the amplifier as well. You can make great mistakes by improper leading of 0V (analog ground and supply ground). Also screening/shielding effect of the case is audible, I am often disappointed when I see horrible DIY wired amplifiers.
Konnichiwa,
The o'scope is a troble-shooting tool, IMHO, the Analyser is an Evaluation tool.
Both are essential. Soundcard 'scopes are near useless....
Sayonara
PMA said:
The o'scope is a troble-shooting tool, IMHO, the Analyser is an Evaluation tool.
Both are essential. Soundcard 'scopes are near useless....
Sayonara
Thorsten,
going back to the title of the thread, which I thought was about making a 300B like amp with ss devices (silly me), would it be possible to take a high voltage mosfet load it with an output transformer and make a SE amp out of that?
going back to the title of the thread, which I thought was about making a 300B like amp with ss devices (silly me), would it be possible to take a high voltage mosfet load it with an output transformer and make a SE amp out of that?
grataku said:
If I may reply, I was interested in the result of the same idea some 7 years ago. I made quite exactly the what you are describing. For me, the sonic result was not only not good, but quite horrible.
- Status
- Not open for further replies.