Hi everyone , I wonder which of these two designs would be better for a high power yet simple amplifier that actualy sounds decent (i'm afraid to use the word HI-FI) even though i have used similar schematics before and with the right speakers and good signal source they do sound much better than some say.
Ok so here are the two contestants. One is a modified version of Rod Elliot's project 3a , which i have built in the oriignal form and I'm happy with it , this modified version I found on the internet.
The other one is from long time forum member Apex.
Ok so here are the two contestants. One is a modified version of Rod Elliot's project 3a , which i have built in the oriignal form and I'm happy with it , this modified version I found on the internet.
The other one is from long time forum member Apex.
Attachments
They are quite different topologies.
I would guess they both sound very similar.
The second one does have the advantage of some protection.
I have built similar amps to both and they both sounded good.
I would guess they both sound very similar.
The second one does have the advantage of some protection.
I have built similar amps to both and they both sounded good.
Why do you choose those ? because they have simple inputs stages , or because
they have lot of outputs ?
There are many , many different designs.
OS
they have lot of outputs ?
There are many , many different designs.
OS
The question will not be about the best design- there is no such thing as a generalised "best" design. Any professionally designed power amplifier built to a tested standard layout with appropriate commercial standard components and wiring will sound "decent".
The person who doodled the P3a mod should do some homework and consider how MJE340/350 are going to drive 5 pairs or even more than 2 pairs of any kind of power transistor in a CFP topology output stage. That one is a certain disaster!
I think a 1,200W PA might shred your speakers and your party friends, though. That might not sound decent at all.
The person who doodled the P3a mod should do some homework and consider how MJE340/350 are going to drive 5 pairs or even more than 2 pairs of any kind of power transistor in a CFP topology output stage. That one is a certain disaster!
I think a 1,200W PA might shred your speakers and your party friends, though. That might not sound decent at all.
Sorry guys, the second paragraph should simply read "....should consider how to drive 5 pairs or even more than 2 pairs...." etc.
Dear Ian, I think you missed counting one current gain stage. 🙂The person who doodled the P3a mod should do some homework and consider how MJE340/350 are going to drive 5 pairs or even more than 2 pairs of any kind of power transistor in a CFP topology output stage. That one is a certain disaster!
On the top side, for example, the MJE340 is not straight driving 5 x 2SC5200 😱 but through an 2SA1943 which, just out of the top of my head must have at least Hfe>20 , and probably higher.
Which will reduce its workload by the same factor.
If anything, I am somewhat worried about optimistic Apex's design.
Not sure that the meager current provided by the Vas stage is enough to drive that bucketful of power transistors boosted just by a pair of power transistors used as drivers.
If enough, it will probably be just so, with no extra margin, with the certain possibility that on some transistor batches of not being enough at all, simply because of Hfe spread.
No doubt they will be a pair of robust drivers though 😛
You were briefly right but that was the purpose of the following post. More apologies if it doesn't appear to correct the statement.
My point was not about enough drive current, but instability and current sharing in a CFP output stage where some builders find it necessary to fit individual drivers to each output device and this becomes counterproductive.
EF designs don't have this problem, as perhaps millions of working amps prove.
I think the P3A derivative is the most interesting of the two, but very much doubt it will be stable as presented. Personally I would convert the output stage to an EF3, insert base stoppers in all the drivers and outputs and liberally decouple the driver/output collectors.
The emitter resistors should be brought up to 0R22 - 0R33 to improve current sharing with that number of outputs. R6 is superfluous. If you are going to the trouble of building such a high power design, it would be worthwhile improving the input stage, starting with emitter degeneration to the LTP and then bring up the tail current to improve slew. 2mA per device and Re = 100 ohm would be a reasonable start. A degenerated current mirror would improve slew even further, which might be beneficial for an amp with high voltage swing. All standard D. Self fare.
With the radially different output stage, I doubt it will retain the P3A character anyhow.
The emitter resistors should be brought up to 0R22 - 0R33 to improve current sharing with that number of outputs. R6 is superfluous. If you are going to the trouble of building such a high power design, it would be worthwhile improving the input stage, starting with emitter degeneration to the LTP and then bring up the tail current to improve slew. 2mA per device and Re = 100 ohm would be a reasonable start. A degenerated current mirror would improve slew even further, which might be beneficial for an amp with high voltage swing. All standard D. Self fare.
With the radially different output stage, I doubt it will retain the P3A character anyhow.
Well , thanks for the opinions , since english is not my first language maybe if someone is willing could you please make some quick corrections to the p3a project schematic or just explain a little more naming where the change needs to be made.
like what is an ef3 or base stoppers etc ?
thank you.
like what is an ef3 or base stoppers etc ?
thank you.
Either amp should be converted to EF3, if you want to build either of them and have the minimum fuss getting it working. There's no way you can use EF2 on +/-110V - there just isn't enough gain. The thermal stability problems with EF3 (bias wandering around) can be dealt with by splitting up the bias stack and compensating driver and output devices individually. You can even get away with splitting up the + and - output banks if they are compensatied individually. To do that you use Leach's bias stack and sprinkle the diodes around. One on an NPN output, one on a PNP, one on each driver. And then dare the bias to move.
Hi Guys
The output stage in the first circuit is much like the old Crown DC-300, where the output devices run cold up to a certain load power level at which point there is enough voltage across the 1R base-emitter resistors to turn them on. Rod uses this circuit as a sub amp and makes no claims for it being high fidelity full bandwidth. For Rod, that speaks volumes, since he is generally not too fussy about high performance.
As others have pointed out, there are many "usual" improvements that can be made without adding significant cost. Adding the emitter follower to drive the VAS would make the single biggest improvement to the first circuit, requiring a change to the collector load of the input stage. The Lin-Thompson topology is very amenable to improvement since C-dom keeps everything stable. Adding all the improvements possible turns the first circuit into the Blameless; doing so for the second circuit turns it into Bob Cordell's mosfet amp but with a BJT output stage.
The CFP of the first circuit is not problematic. If one encounters oscillation, a small cap across the base-emitter of the driver fixes it. Rod already has this in place as C6 (100pF).
Self points out that multiple parallel output devices can lead to oscillation issues - essentially saying it is impossible to not have oscillation, in his normal style. This is more likely to be a problem with the second circuit than the first where all the outputs are on at idle and low signal levels.
Either way, "simple" is not necessarily a good way to go for high power. There is simple, then there is too simple.
Have fun
The output stage in the first circuit is much like the old Crown DC-300, where the output devices run cold up to a certain load power level at which point there is enough voltage across the 1R base-emitter resistors to turn them on. Rod uses this circuit as a sub amp and makes no claims for it being high fidelity full bandwidth. For Rod, that speaks volumes, since he is generally not too fussy about high performance.
As others have pointed out, there are many "usual" improvements that can be made without adding significant cost. Adding the emitter follower to drive the VAS would make the single biggest improvement to the first circuit, requiring a change to the collector load of the input stage. The Lin-Thompson topology is very amenable to improvement since C-dom keeps everything stable. Adding all the improvements possible turns the first circuit into the Blameless; doing so for the second circuit turns it into Bob Cordell's mosfet amp but with a BJT output stage.
The CFP of the first circuit is not problematic. If one encounters oscillation, a small cap across the base-emitter of the driver fixes it. Rod already has this in place as C6 (100pF).
Self points out that multiple parallel output devices can lead to oscillation issues - essentially saying it is impossible to not have oscillation, in his normal style. This is more likely to be a problem with the second circuit than the first where all the outputs are on at idle and low signal levels.
Either way, "simple" is not necessarily a good way to go for high power. There is simple, then there is too simple.
Have fun
It would be much better to use a standard design for which a PCB is available and no mods are required. First though, what speakers and impedance are you attempting to power, how many and how loud do you want it? i.e. Size of room, outdoors, background music, concert volume, incredibly loud etc.
It's ok Ian , I like some moddying as that is a step forward and makes you think instead of just assembling a kit.
So the story is simple , I have a few luxman amps that I got from the scrappyjard did some repairs and their working great , I use them for mids and highs.they reproduce with no problems. then I have a graphic equalizer after the luxman which is set for lower mids and low frequency aka bass , I have seperate amplifiers just for lower mids and bass alone as that is where msot of the energy goes especially listening some of todays heavy bassline music etc.
I guess it's called biamping and I think its quite a nice way to split the right frequencies for the right speakers and also amplifiers.
So I myself kinda like the Rod's topology more with some changes I guess it could sound nicely.
So I will attach a picture of what I understood from you guys so far what should be done.I edited the two transistor driver stage to a emitter follower one.
could someone explain what you mean by the EF2 OR EF3?
please be so kind as to say what more changes should be made to this schematic and why emitter follower is better in this case? Does it provide more current for the VAS part aka output transistors?
So the story is simple , I have a few luxman amps that I got from the scrappyjard did some repairs and their working great , I use them for mids and highs.they reproduce with no problems. then I have a graphic equalizer after the luxman which is set for lower mids and low frequency aka bass , I have seperate amplifiers just for lower mids and bass alone as that is where msot of the energy goes especially listening some of todays heavy bassline music etc.
I guess it's called biamping and I think its quite a nice way to split the right frequencies for the right speakers and also amplifiers.
So I myself kinda like the Rod's topology more with some changes I guess it could sound nicely.
So I will attach a picture of what I understood from you guys so far what should be done.I edited the two transistor driver stage to a emitter follower one.
could someone explain what you mean by the EF2 OR EF3?
please be so kind as to say what more changes should be made to this schematic and why emitter follower is better in this case? Does it provide more current for the VAS part aka output transistors?
Attachments
Oh , it's a SUB you want to run.... with a EF3.
(below 1) is a EF3.
Your schema is right (in concept) , but you need to decouple the driver/prdriver
and "stopper" (4.7R) each output device (below 2). plus , shunt the predrivers to the rails (33pf's).
The EF3 will oscillate otherwise.
EF3 is the only way for a sub.
OS
(below 1) is a EF3.
Your schema is right (in concept) , but you need to decouple the driver/prdriver
and "stopper" (4.7R) each output device (below 2). plus , shunt the predrivers to the rails (33pf's).
The EF3 will oscillate otherwise.
EF3 is the only way for a sub.
OS
Attachments
Last edited:
well it's not only sub mayeb one day io decide to use a full range box with tis amp but mostly it will be midlow to low.
Ok I did some editing , tell me folks how does it look , I have a feeling something else should be changed , probably a transistor polarity somewhere reversed , its late can't see all yet.
BTW , in your green background color schematic , why does the +- supply rails have a transistor on each one of them in series with the input stage ? are they for voltage stability or anti oscillation or for what ?
Ok I did some editing , tell me folks how does it look , I have a feeling something else should be changed , probably a transistor polarity somewhere reversed , its late can't see all yet.
BTW , in your green background color schematic , why does the +- supply rails have a transistor on each one of them in series with the input stage ? are they for voltage stability or anti oscillation or for what ?
Attachments
Last edited:
Those are cap multipliers(voltage stability) , some of the inputs stages really benefit from it.
PS - that amp's artwork is free - and more have been built than ESP's "monster".
Also , APEX has moved on to class H and better designs for high power , as well.
Edit - you have it ... on your schema , also - the drivers can be fast outputs as well (2sa1294/2sc3263).
For you to safely drive 5-10 pairs (or more).
OS
Last edited:
Input stage is important , too.
Your input stage is not ideal for a high power output stage.
I built one like your schema - BIG thump with turn-on (bootstrapped amp)
It takes too long to settle - but has a nice sound with a smaller amp.
A current sourced voltage stage and input stage settles much more quickly
and remains super stable. very important factors with such high power.
example below -
OS
Your input stage is not ideal for a high power output stage.
I built one like your schema - BIG thump with turn-on (bootstrapped amp)
It takes too long to settle - but has a nice sound with a smaller amp.
A current sourced voltage stage and input stage settles much more quickly
and remains super stable. very important factors with such high power.
example below -
OS
Attachments
my last schematic was kinda wrong , put the driver base resistors in the wrong place, now changed that so I'm welcomed to your opinions.
BTW isn't the amp in your pictures overly complicated for it's purpose? I mean the transistor count is enormous, especially at the input.
I will go check out some input topologies.
BTW isn't the amp in your pictures overly complicated for it's purpose? I mean the transistor count is enormous, especially at the input.
I will go check out some input topologies.
Attachments
Last edited:
Thats funny , There are some "guru's" that think mine are simple and have 40 device
input stages.
But , yes ...you could condense the 2 separate current sources (below).
....get rid of 5 transistors and still have 90% performance. It would still be
safer than either of the first schema's at the beginning of the thread.
For a first time input stage , nearly foolproof. AND it is very safe with a large
EF output stage.
PS - this is why the forum's amp is based on this - a great beginning. 😉
OS
Attachments
Last edited:
Redesigning (or full designing) a full, working, high power amplifier may be a little too much for a few posts in a thread 😉
I think our friend crazymechanic would be better served by straight suggesting him a fully developed , working , well proven circuit. 🙂
Just sayin' 🙂
I think our friend crazymechanic would be better served by straight suggesting him a fully developed , working , well proven circuit. 🙂
Just sayin' 🙂
Last edited:
- Status
- Not open for further replies.