Then how to match them?
Andrew if you have a better way please show us, with schematic of course.
You know that my words understanding is not good
But usually I use DMM only& the result quite good
There is a Web Site belonging to Rod Elliot that discusses and shows how to match Devices. The article is titled "Matching Power and Driver Transistors". I suggest you have a look there for Elementary instructions. You can find other Articles there that you may find useful in Many ways.
The simple thing to do to establish the basics of good amplifier design is read Douglas Self's book. A few pointers:
Use a degenerated input stage with a current mirror. The mirror doubles input stage transadmittance while ensuring push-pull action which cancels even order harmonics. The degeneration allows raw gain to be traded for linearity of the input stage and permits the value of the compensation capacitor to be reduced which increases slew.
Elliots design uses a 100p compensation with no degeneration for the input stage. This is terrible as the slew induced distortion in his design is likely to be high; the slew rate is likely to be mediocre at best.
Moreover, to maximise forward path, it is prudent to use a triple emitter follower for the output stage as this minimises the loading on the second stage.
Last edited:
Thanks for that AndrewT
Here is Rod's method, I will use that one:
Matching Power and Driver Transistors
Thanks so much, never visited Self's site before.
New things to read, new things to learn, that's always good
Poor design?
michaelkiwanuka - I don't know who you are, nor do I know why you seem to be so vitriolic about my designs or PCB layouts, but your comments are way out of line with reality and are offensive and defamatory (and you spelled my name incorrectly). You obtained PCB layouts from where exactly?
Thousands of my PCBs for a great many projects have been sold all over the world, and I have never had a complaint from a customer about poor performance or the design not meeting any claims I may have made for it.
I design for DIY people, some of whom are newbies, others are very experienced and know what they are doing, how to take detailed measurements, etc. Neither newbies or experienced constructors have had any complaints - quite the reverse in fact.
The designs I produce are specifically intended to be easy to build and to have no bad habits. I don't design for DC to daylight frequency response or sub 0.001% THD, but without exception they perform exactly as claimed.
I see from your profile that you have made many posts, but have no friends - the reason for that seems pretty clear.
michaelkiwanuka - I don't know who you are, nor do I know why you seem to be so vitriolic about my designs or PCB layouts, but your comments are way out of line with reality and are offensive and defamatory (and you spelled my name incorrectly). You obtained PCB layouts from where exactly?
Thousands of my PCBs for a great many projects have been sold all over the world, and I have never had a complaint from a customer about poor performance or the design not meeting any claims I may have made for it.
I design for DIY people, some of whom are newbies, others are very experienced and know what they are doing, how to take detailed measurements, etc. Neither newbies or experienced constructors have had any complaints - quite the reverse in fact.
The designs I produce are specifically intended to be easy to build and to have no bad habits. I don't design for DC to daylight frequency response or sub 0.001% THD, but without exception they perform exactly as claimed.
I see from your profile that you have made many posts, but have no friends - the reason for that seems pretty clear.
Rod, I don't think you help yourself here. mk has provided no explanations on what he finds wrong, and he's been called out on it. He'll either dish up an actual explanation (which you can take issue with, if you like, on a technical level) or he won't, and his comments can be judged and discounted in that context. No need to take this to the level of personal insults.
I may not be 'helping myself', but MK's comments are in fact defamatory, and as such should be dealt with accordingly.
His remarks have the real potential to harm my business, and as such should not be tolerated by DIY Audio.
As you probably know, I rarely post here (for a variety of reasons), but when I am defamed there has to be a response.
His remarks have the real potential to harm my business, and as such should not be tolerated by DIY Audio.
As you probably know, I rarely post here (for a variety of reasons), but when I am defamed there has to be a response.
how will newbies be able to compare degenerated input LTP against one that is not if you ask them not to try it?
Generally, I prefer the sound of current feedback amps. But, recently I had a chance to examine schematics of some old VFB amplifiers from the seventies and early eighties that sounded good to me. Amps like old heavyweight Pioneers, Technics, Musical Fidelity A1. One thing in common was that neither of these amps had current mirror for the input pair and some do not have even active CCS. It is foolish to think that Japanese circuit designer is not well educated in electronics, that he is not aware that such enhancements exist. It is even less possible that manufacturer is saving on components because there are two huge transformers inside, huge hetsinks, expensive pot, etc. Additional cost of a couple small signal transistors is negligible. My explanation is that Japanese engineer had eight hours each working day to design amps and that he actually built prototypes and judged how they sound. Very probably he felt that additional enhancements, although improve lab performance, do not sound better! I think the same. So, Rod's decision to simplify circuit is not bad decision. Especially in the diy context, where each additional transistor in the circuit could be the potential source of some troubles, like oscillation.
Later came specs race and since transistors are cheap designers lost sense of limits. But as specs race unfold quality of sound went rapidly down!
Later came specs race and since transistors are cheap designers lost sense of limits. But as specs race unfold quality of sound went rapidly down!
No-one has "defamed" you. Your circuit was criticized, no reason given. Like I said, if mk leaves it at that, no-one can possibly take that criticism seriously. If he actually has some technical argument, address it.
Clearly, the comments come from the SOTA engineering camp which is largely engaged in commercial and scientific development. This is hardly in the scope of newb DIY activity which is a very high proportion of members' interests here.
The thread is about PCB layout and someone elects instead, to rubbish the design on which it is based. Irrelevant, insensitive trolling, I'd say.
'Just my 5c
The thread is about PCB layout and someone elects instead, to rubbish the design on which it is based. Irrelevant, insensitive trolling, I'd say.
'Just my 5c
On the contrary Mr Elliot, I gave very sound reasons why your designs are sub-standard here:
http://www.diyaudio.com/forums/solid-state/201446-rod-elliot-p3a-layout-critics-21.html#post3494036
To add to those valid criticisms, your second stage, the transimpedance stage, uses a single transistor. This is a poor decision because this transistor's non-linear base-collector parasitic capacitance is exposed to the full voltage swing at the collector and varies with it. This creates a significant source of unnecessary distortion. Ideally the second stage should be a cascode.
Back to topic
Hi michaelkiwanuka, let's go back to topic
don't critics other design if you can not make a better one,
if you wish just make a new thread okay.
So here's my PCB design that I keep & now I need some critics
So anyone feel free to critics me
there is two red & orange and I don't sure which one is better.
Thanks to Rod Elliot who give us free schematic
of course if you guys want the real P3A from the designer you know exactly where to go
Hi michaelkiwanuka, let's go back to topic
don't critics other design if you can not make a better one,
if you wish just make a new thread okay.
So here's my PCB design that I keep
& now I need some critics So anyone feel free to critics me
there is two red & orange and I don't sure which one is better.
Thanks to Rod Elliot who give us free schematic
of course if you guys want the real P3A from the designer you know exactly where to go
Attachments
Last edited:
Yes, you did make these comments, as well as telling everyone that they should avoid all my designs because they are "sub standard". The exact quote is ...
That is a highly inflammatory (and defamatory) statement for anyone to make. This is particularly true because there is no evidence that you've actually built or tested any of my circuits. You've just looked at them, and decided that they are no good.
So, the use of a 100pF compensation cap with no degeneration is likely to cause slew induced distortion? Have you built one to test it? Done any measurements perhaps so you can prove your point?
Many hundreds of people have built the amp, and are very happy with the results. It's not a Douglas Self 'blameless' amplifier, and makes no pretense to be anything that it's not. The amp works very well and countless people have told me that it sounds very good indeed.
Despite your obvious (and pointless) criticism, there are vast numbers of amplifiers (commercial and DIY) that use very similar topology, many of which have achieved critical acclaim. While triple Darlingtons remove loading from the VAS stage, they are also prone to thermal runaway unless extreme care is taken with the bias servo.
Not everyone has the skill and patience to build very complex amplifiers that offer vanishingly low distortion and DC to daylight frequency response. Most home builders want an amplifier that they can build and actually complete and get working - preferably without destroying expensive parts.
P3A is exactly what I claim it to be - a simple amplifier that is easy to build, allows transistor substitution without breaking into oscillation, and is ideal for general hi-fi use. It's not SOTA and makes no claim to be.
I find your general tone and attitude to be perplexing - you make many assertions but they appear to be from a purely theoretical perspective. It is necessary to actually build and test things for yourself before launching into a debate on the 'failings' (real or imaginary) of other designs.
Perhaps you could dazzle the DIY Audio readership with a built and tested amplifier of your own design, rather than defaming people with whom you have never held a discussion or even made contact.
What you're pointing out is sources of distortion that are also present in commercial and consumer designs. In other words, most of us have been listening to it our whole lives. While concepts like this may be easy for some of us, they are beyond a DIYer that is still struggling with the basics.
I had already used a lot of the circuits on Mr. Elliot's site well before I was even aware of it. But I've been at this for 40 years too. Mr. Elliot provides classic circuits that are optimised, and they are practical for DIY use. Mr. Elliot also points out caveats in detail and provides tips that will be useful when the DIYer steps it up to the next level.
I first looked at Douglas Self's "Sources of distortion in audio amplifiers" about 6 weeks ago for the first time. While I had never examined some of the topics he addressed, I had used a couple of the solutions he implemented. I have never built an amplifier with all the bells and whistles of the Blamess amplifier, but the design is so simple and elegant that it's a no-brainer for me.
Many of Mr. Elliot's circuits are equal to or better than industry standard circuits. Is there a better solution for tone controls (other than tweaking turnover frequency and boost) than the classic tone control circuit in Mr. Elliot's "hi-fi" preamp? If there is, I would surely like to know it.
Back to circuit board layout...
@John Bali: 1) It would help if you posted a layout with the component reference (in any form) superimposed on the layout, or component values, or something like that.
2) My personal preference would be a layout with electrolytic capacitors for rail decoupling on the board. RodE may know this is not absolutely required, but then again, he knows exactly what to expect. Making a new layout from his schematic, I would follow common practice and use at least one per rail (like the yellow layout). Sakis has posted about larger decoupling caps here, and I would suggest reading his posts about component selection for other tips. I have found his comments on the P3A very informative.
3) At a glance, I can't find the trim pot on the yellow layout (see my first point).
4) Ground trace on the first layout (red) should probably be a bit bigger. Personally, I would err on the side of safety, so to speak, and just make power and ground traces bigger than necessary. Since they are to the outside of the board, one can safely make a circuit board bigger in outside dimensions, while leaving the signal circuits as tight as desired in the middle.
5) I see the transistors lined up for a common heatsink, but I would also say make sure your intended heatsink will fit, space on either side seems a bit tight.
@John Bali: 1) It would help if you posted a layout with the component reference (in any form) superimposed on the layout, or component values, or something like that.
2) My personal preference would be a layout with electrolytic capacitors for rail decoupling on the board. RodE may know this is not absolutely required, but then again, he knows exactly what to expect. Making a new layout from his schematic, I would follow common practice and use at least one per rail (like the yellow layout). Sakis has posted about larger decoupling caps here, and I would suggest reading his posts about component selection for other tips. I have found his comments on the P3A very informative.
3) At a glance, I can't find the trim pot on the yellow layout (see my first point).
4) Ground trace on the first layout (red) should probably be a bit bigger. Personally, I would err on the side of safety, so to speak, and just make power and ground traces bigger than necessary. Since they are to the outside of the board, one can safely make a circuit board bigger in outside dimensions, while leaving the signal circuits as tight as desired in the middle.
5) I see the transistors lined up for a common heatsink, but I would also say make sure your intended heatsink will fit, space on either side seems a bit tight.
Last edited: