If I am not mistaken, I believe Carver used such concept of tracking PS on some of his designs.
Thanks a lot gents,
My bubble has burst...only loking. I am disappointed though in that I will have to have a 'scope though...but it is understandable. You can only go so far with mathematics.
Devastated....well a 'scope is on the shopping list then, it 'll have to be. I had thought of the figure 11.4 that has been suggested and can find the necessary circuit mod on site probably. Still though I had my heart set on Class A, first, QUAD current dump second. I have a good selection of transistors knocking about.
Tempted by the Sloane route....already designed..tried and tested...works ! ! But not my own. Something I have done. I suppose I have to learn to walk before I can run though.
So, if I try my own Class A I will definately need a 'scope ? This is the route I want to go down, for now. Once I have myself set on doimg something I have to do it. From what I have been reading I was kind of (maybe I mis-lead myself here) under the impression that with Class A, using mirror image topology, some current mirrors, good choice of output transistors (I have MJ15003/4's 2N3055h and TIP 142/147 [these are darlington's though and I thought I could use these before the 15003/4's]) and good all round application of design and principles that I would literally have to bias the ouptut devices to the midpoint of the supply rails and away to go...sort of. I am wrong here then am I ??
Thanks gents
Gareth
My bubble has burst...only loking. I am disappointed though in that I will have to have a 'scope though...but it is understandable. You can only go so far with mathematics.
Devastated....well a 'scope is on the shopping list then, it 'll have to be. I had thought of the figure 11.4 that has been suggested and can find the necessary circuit mod on site probably. Still though I had my heart set on Class A, first, QUAD current dump second. I have a good selection of transistors knocking about.
Tempted by the Sloane route....already designed..tried and tested...works ! ! But not my own. Something I have done. I suppose I have to learn to walk before I can run though.
So, if I try my own Class A I will definately need a 'scope ? This is the route I want to go down, for now. Once I have myself set on doimg something I have to do it. From what I have been reading I was kind of (maybe I mis-lead myself here) under the impression that with Class A, using mirror image topology, some current mirrors, good choice of output transistors (I have MJ15003/4's 2N3055h and TIP 142/147 [these are darlington's though and I thought I could use these before the 15003/4's]) and good all round application of design and principles that I would literally have to bias the ouptut devices to the midpoint of the supply rails and away to go...sort of. I am wrong here then am I ??
Thanks gents
Gareth
Hi Gareth
No need to be discouraged. You only really need a scope - well a spectrum analyser really - in the case of your nominated class S. Check the posts again - that is the difficult route if you wanted to build a design that has yet to be made for DIY!
Class A is really quite easy, provided you keep a couple of guidelines in mind:
1. Keep the power out down to 20W or less because the heat dissipation is much greater than the output in class A. i.e. EXPENSIVE heatsinks are needed.
2) Keep the speaker impedance up to 6 ohms or above so that a single pair of p-p outputs can be used to supply 20W /per channel amp with only 0.4 degrees/Watt sink for each channel. Picture a big 2 - 3 unit rack case with 300mm depth where each side is just one huge deep-finned sink. Lotta sink, right? Yes, you could fan-cool a little sink and let the noise wreck class A benefits or just let nice quiet big ones do the job efficiently.
Randy's designs mostly work fine but the PCBs may not be up to today's standards and if you are serious about A, you could do worse than look carefully at Nelson Pass's great designs, hot as they get.
Most of us seldom use amps at home above 10W levels and provided you don't use tiny and insensitive speakers, this is enough for anyone. Again, check the posts and also consider the popularity of JLHoods's 10W designs. There are many threads based on them here and if you would like the opportunity to try class A before committing hundreds of pounds for even a modest amplifier, try this first.
I use and really like class A myself, but IMO it is not essential or suited to all types of audio use. For instance, as the levels get wound up, the distortion that wrecks AB at low levels reduces and the apparent difference is less, so class A will be just wasted heat at parties, discos etc.
No need to be discouraged. You only really need a scope - well a spectrum analyser really - in the case of your nominated class S. Check the posts again - that is the difficult route if you wanted to build a design that has yet to be made for DIY!
Class A is really quite easy, provided you keep a couple of guidelines in mind:
1. Keep the power out down to 20W or less because the heat dissipation is much greater than the output in class A. i.e. EXPENSIVE heatsinks are needed.
2) Keep the speaker impedance up to 6 ohms or above so that a single pair of p-p outputs can be used to supply 20W /per channel amp with only 0.4 degrees/Watt sink for each channel. Picture a big 2 - 3 unit rack case with 300mm depth where each side is just one huge deep-finned sink. Lotta sink, right? Yes, you could fan-cool a little sink and let the noise wreck class A benefits or just let nice quiet big ones do the job efficiently.
Randy's designs mostly work fine but the PCBs may not be up to today's standards and if you are serious about A, you could do worse than look carefully at Nelson Pass's great designs, hot as they get.
Most of us seldom use amps at home above 10W levels and provided you don't use tiny and insensitive speakers, this is enough for anyone. Again, check the posts and also consider the popularity of JLHoods's 10W designs. There are many threads based on them here and if you would like the opportunity to try class A before committing hundreds of pounds for even a modest amplifier, try this first.
I use and really like class A myself, but IMO it is not essential or suited to all types of audio use. For instance, as the levels get wound up, the distortion that wrecks AB at low levels reduces and the apparent difference is less, so class A will be just wasted heat at parties, discos etc.
Thanks for that. Maybe I am being nostalgic in my choice of designs. I think I will take your advice and take a good look at Nelson's and JLH's designs before I go any further. The excitement of building my first proper amp is running away with me maybe.
Thanks again
Gareth
Mmmm.............. about 50watt of Class A would be nice though.
You don't need to get a spectrum analyser... Nice but not required.
A scope of 20MHz should be sufficient, preferably dual beam.
Add to this a decent digital multimeter and a basic signal generator/function generator and you have the 3 items of test gear that I would regard as the bare minimum.
A scope of 20MHz should be sufficient, preferably dual beam.
Add to this a decent digital multimeter and a basic signal generator/function generator and you have the 3 items of test gear that I would regard as the bare minimum.
You don't need to get a spectrum analyser... Nice but not required.
A scope of 20MHz should be sufficient, preferably dual beam.
Add to this a decent digital multimeter and a basic signal generator/function generator and you have the 3 items of test gear that I would regard as the bare minimum.
Well I have a Fluke multimeter that I use at work and I also have a function generator that operates well up into the MHz region.
Funds needed for new 'scope !!!! Great, some illegal driver has recently wiped my car out too, not funny (sorry to go off topic a bit !). I need to start looking around then
Gareth
50 Watts in class A is really pushing it! Remember that very few commercial amps that claim to be class A genuinely are once you get beyond 20W or so...
I can think of an amp that claimed 185W but actually only delivered 18 in class A!
Certainly you will need fan cooling if going beyond 20W or so.... or humongous very expensive heatsinks if not fan cooled. Fan cooling makes a HUGE difference! Think in terms of reducing overall heatsink size/volume by maybe 5 times if fan cooled.
I can think of an amp that claimed 185W but actually only delivered 18 in class A!
Certainly you will need fan cooling if going beyond 20W or so.... or humongous very expensive heatsinks if not fan cooled. Fan cooling makes a HUGE difference! Think in terms of reducing overall heatsink size/volume by maybe 5 times if fan cooled.
Seriously ? I never knew that. I know of some of the crazy power measurement methods manufacturers used in the past - the numbers game I suppose - and I am surprised at what you have said.
20 watt's it shall be then. Why have 10 when you can have 20 ?
Off to the drawing board !!!
Thanks a lot
Get a second hand Tektronix 465B. All the scope any one will ever need and cheap these days (£100 or so). Cost as much as a car when new!! Make sure you can test it yourself or get a guarantee though. There's loads of them about but they are old now and very complicated. Trust me, you don't want to be repairing one....
gareth,
well, no reason for been discouraged, but more so been a little bit realistic. By what I read between the lines of your posts, I am getting the sense that you are not very familiar with this amplifier matter. May I suggest before you dive in and go on a scope shopping spree, just pick up a few good books. Pretty much all aspects of audio amplifiers have been covered to most degree in a few good books. You may find there are various very specialized test setups and rather exotic test equipment mentioned and needed for performance verification. What is not covered in books you shall find here, there are many people here that share 'tribal' knowledge when one is in need. Either way you have got a LOT of reading to do.
As far as Test equipment goes, I consider a scope as a nicer DMM and bare minimum to have. Depending on how serious you are with this Class (A) Amplifier, you may very well need additional equipment such as a Distortion Analyzer or at least a Spectrum Analyzer. Besides decent soldering tools, you may also find a Curve Tracer is handy to have.
As far as I can tell, designing, building an amplifier and making it work is only half of the job, tweaking it to perform is the other half of this job.
If you consider forced air cooling, keep in mind the fan in use will create some audible noise. (assuming the amplifier will be in proximity) This audible noise may proof to be rather unwelcome when trying to enjoy Class A performance.
Anyway, just my 2 cent...
Cheers
Kay
well, no reason for been discouraged, but more so been a little bit realistic. By what I read between the lines of your posts, I am getting the sense that you are not very familiar with this amplifier matter. May I suggest before you dive in and go on a scope shopping spree, just pick up a few good books. Pretty much all aspects of audio amplifiers have been covered to most degree in a few good books. You may find there are various very specialized test setups and rather exotic test equipment mentioned and needed for performance verification. What is not covered in books you shall find here, there are many people here that share 'tribal' knowledge when one is in need. Either way you have got a LOT of reading to do.
As far as Test equipment goes, I consider a scope as a nicer DMM and bare minimum to have. Depending on how serious you are with this Class (A) Amplifier, you may very well need additional equipment such as a Distortion Analyzer or at least a Spectrum Analyzer. Besides decent soldering tools, you may also find a Curve Tracer is handy to have.
As far as I can tell, designing, building an amplifier and making it work is only half of the job, tweaking it to perform is the other half of this job.
If you consider forced air cooling, keep in mind the fan in use will create some audible noise. (assuming the amplifier will be in proximity) This audible noise may proof to be rather unwelcome when trying to enjoy Class A performance.
Anyway, just my 2 cent...
Cheers
Kay
Right, so a Tektronix 456B on the shopping list (with calibration certificate) and forget about THD meters and Curve Tracers for now. I feel that I understand the principle sof what is happening inside an amplifier although not in a complete way. Having read a number of books, perhaps too many..I don't know. I also understand that to get good stability, no self-oscillating, generally a good working amp is no easy task, otherwise we would all be building them. Common-emitter, common-base, common-collector and their associated impedance and phase characteristics, complimentary, quasi-complimentary, cascode, output triples.... I know...there is a lot to learn, but I am willing and determined. Start small and work up is the plan and although I have an idea in my head of what I want to, eventually, achieve I will keep chipping away. Thanks again.
Gareth
Gareth
There are a number of free simulators out there, you would do well to get one and learn how to use it, because it can give you some insight as to hwat to look for in the real world. And, you can reasonably debug your own ideas without making too much magic smoke out of your expensive parts
Most simple designs actually simulate quite close to real life in most respects.
Regarding your original question, have a look at current dumping, which is actually a gross misnomer of a very elegant feedforward technique (many followed which were less elegant/optimized due to patent rights) but they work.
A concept similar to Sandman in that it uses a low power amp to 'correct' a high power amp was also used by Sansui in ther Super Feedforward amplifiers (AU-D9 and AU-D11). Schematics for these are readily available on the net, it's worth a look. However, I would still advise looking up Pete Walker's original patent for Quad, re Current Dumping. It has recently run out so in principle everyone can use it now. The beauty of it is that it can be used on top of many classic topologies and offer real improvement.
Most simple designs actually simulate quite close to real life in most respects.
Regarding your original question, have a look at current dumping, which is actually a gross misnomer of a very elegant feedforward technique (many followed which were less elegant/optimized due to patent rights) but they work.
A concept similar to Sandman in that it uses a low power amp to 'correct' a high power amp was also used by Sansui in ther Super Feedforward amplifiers (AU-D9 and AU-D11). Schematics for these are readily available on the net, it's worth a look. However, I would still advise looking up Pete Walker's original patent for Quad, re Current Dumping. It has recently run out so in principle everyone can use it now. The beauty of it is that it can be used on top of many classic topologies and offer real improvement.
Thanks Ian, that is sort of what I was thinking in my mind. Use two amplifier's. The first a low-current high quality with the error signal negated by it feeding into a high power but lower quality (if that is the correct way to term it) amplifier thus producing a good quality and high powered output.
The sound of this design, to me, is great. The cancellation of any 'bad' signal and the high power output following the low. Sounds good. Would it be correct in that any signal changes (thus erratic) due to the load - the loudspeaker - would also be compensated for by the low-power first stage of the amp ?
The sound of this design, to me, is great. The cancellation of any 'bad' signal and the high power output following the low. Sounds good. Would it be correct in that any signal changes (thus erratic) due to the load - the loudspeaker - would also be compensated for by the low-power first stage of the amp ?
I simmed some Black's FeedForward amps - it is pretty complicted to get the matching, rolloff right
http://www.diyaudio.com/forums/soli...in-composite-op-amp-circuits.html#post2415805 post #10
given the issues seen playing in sim I would highly recommend just building a decent high loop gain amp out of Self, Cordell's books
http://www.diyaudio.com/forums/soli...in-composite-op-amp-circuits.html#post2415805 post #10
given the issues seen playing in sim I would highly recommend just building a decent high loop gain amp out of Self, Cordell's books
Don't try to walk before you can crawl
Theory from books is one thing, real life has trial by earthing protocol, layout issues, mystery parasitic oscillations that didn't appear in the book or the simulation.... then there is the magic smoke sealed in all components during manufacture and which can be provoked to escape all to readily
Don't try to walk before you can crawl
Theory from books is one thing, real life has trial by earthing protocol, layout issues, mystery parasitic oscillations that didn't appear in the book or the simulation.... then there is the magic smoke sealed in all components during manufacture and which can be provoked to escape all to readily
Cheers Jez mate, your filling me with confidence there !!!!! We have to start somewhere don't we ?
I have returned to the books, as previously suggested, and I have decided to go for a Class A for my first build. I can't really class a headphone amplifier in the same category really, can I ?
I am going to use an exixting NAD pre-amp I have and so only want to build a dual-mono power amp. Having read Slones' book on amplifier construction there is a good Class A design in there, which when driven hard goes into Class AB operation. What I am wondering is because I plan to use my NAD as a pre-amp and it has an output of 12v at 200ohm and the Class A design offered by Slone is designed to accept line level, 2v to 2.5v max I suppose, then can I simply reduce the value of the input resistor in the first stage of the amp (and also a resistor in the feedback loop which has to match), his is around 20kilohm, or can I go straight to the V.A. stage of the amp and omit the first stage ? Not sure what to do here.
I would have thought that I would still have to build the first stage and not omit it ? This is where the crawling begins....
Thanks
Gareth
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.