Simple ... I am in love with the P3A there is no other way arround that .
if the application is ok with 60-100W then so far there is nothing better in my world .....
Before many of you start to say things remeber that in my shop we repair an also listen audio machines
now days between 2500-3000 machines per year
that means that i have been listening the latest 10 years anything that is available , any topology any make, any brand , just say it and we already done it , and obvioulsy not once and not twice ....
I can remind you the reasons because some of us tent to forget why
Smaller amps requiere less safety measures ( vi limiters , miller compensations and so on )
mutliple tranistors suffer from times* the device capacitance so driver stage need to provide the current but also the ability to drive that much of capacitive load
Multiple active devices in the output there no guarantee that will "ring" at the same time .... The more the outputs then amplfier is more pcb depending
Thermals is an issue , rail distribution is an issue , decoupling is an issue , too many active devices and oscilation is waiting for you in the corner
you start with 200-300 MHZ ltp , then you might get away with 100 mhz VAS and drivers and then 30MHZ outputs
made a P3a ltp 300 mhz vas and drivers 200 and outputs 50 mhz just to examine the stability of the circuit
no problems at all and let us not forget that i have been messing around of limmits with the miller compensation
So if you sum those details and add to them minnor refinements that i ve made in the P3A this in total is what makes the P3A the amplifier that fits the picture
remember that other similar EFP amplifiers might preform alike, close or even better but the bootstrap in combination with CFP when it come to sound is unbeatable cobination any way you slice it .
there is a possibilty that a diamond input and a more in generall sophisticated input scheme or + and -VAS will provide precision but the result doesnt give the same feeling
Listeners say that the original combo playes better
But through the years and long discussions about that, obviously we have been there before and more than once .
----------------------------------------------------------------------------------------------------------------------------------
So today's question is "what if more power is needed ? )
there hass been an issue before and if i recall correct you can double the outputs as long as you add 0.47 /5W to every emmiter
i am not sure i want or like that
i could risk to skip that step and go with very close matching on the outputs fit them in a thick fin of alouminoum before the heatsink one on the lower side one on the upper side for the PNP and the NPN alike
That is to ensure that each pair shares the same temperature add that most of them i go with relatively hi bias so that prevents in away that outputs will act diferent
Any other ideas to squeeze more power ?
if the application is ok with 60-100W then so far there is nothing better in my world .....
Before many of you start to say things remeber that in my shop we repair an also listen audio machines
now days between 2500-3000 machines per year
that means that i have been listening the latest 10 years anything that is available , any topology any make, any brand , just say it and we already done it , and obvioulsy not once and not twice ....
I can remind you the reasons because some of us tent to forget why
Smaller amps requiere less safety measures ( vi limiters , miller compensations and so on )
mutliple tranistors suffer from times* the device capacitance so driver stage need to provide the current but also the ability to drive that much of capacitive load
Multiple active devices in the output there no guarantee that will "ring" at the same time .... The more the outputs then amplfier is more pcb depending
Thermals is an issue , rail distribution is an issue , decoupling is an issue , too many active devices and oscilation is waiting for you in the corner
you start with 200-300 MHZ ltp , then you might get away with 100 mhz VAS and drivers and then 30MHZ outputs
made a P3a ltp 300 mhz vas and drivers 200 and outputs 50 mhz just to examine the stability of the circuit
no problems at all and let us not forget that i have been messing around of limmits with the miller compensation
So if you sum those details and add to them minnor refinements that i ve made in the P3A this in total is what makes the P3A the amplifier that fits the picture
remember that other similar EFP amplifiers might preform alike, close or even better but the bootstrap in combination with CFP when it come to sound is unbeatable cobination any way you slice it .
there is a possibilty that a diamond input and a more in generall sophisticated input scheme or + and -VAS will provide precision but the result doesnt give the same feeling
Listeners say that the original combo playes better
But through the years and long discussions about that, obviously we have been there before and more than once .
----------------------------------------------------------------------------------------------------------------------------------
So today's question is "what if more power is needed ? )
there hass been an issue before and if i recall correct you can double the outputs as long as you add 0.47 /5W to every emmiter
i am not sure i want or like that
i could risk to skip that step and go with very close matching on the outputs fit them in a thick fin of alouminoum before the heatsink one on the lower side one on the upper side for the PNP and the NPN alike
That is to ensure that each pair shares the same temperature add that most of them i go with relatively hi bias so that prevents in away that outputs will act diferent
Any other ideas to squeeze more power ?
With the CFP outputs, put a .1 ohm (or even .068 or .047) resistor in each emitter when paralleling extra devices. The “emitter” resistor for the whole CFP would stay the same. The better match the parallel transistors are, the smaller those extra resistors can be made.
It never behaves quite as cleanly with paralleled outputs. Unfortunate, but nothing you can really do about it. And the higher your supply voltage and the bigger VA your trafo the higher the risk of destruction from accidents if there is no SOA protection. It will run just fine till you short the speaker wires.
It never behaves quite as cleanly with paralleled outputs. Unfortunate, but nothing you can really do about it. And the higher your supply voltage and the bigger VA your trafo the higher the risk of destruction from accidents if there is no SOA protection. It will run just fine till you short the speaker wires.
Doesnt bridging change the SQ of the amp?
Since I'm up to bat, might as well take a swing. What if you stacked everything vertically downstream of R6 (560), sharing R5, R4 and C3 into Q2? Could the output Z of Q1 still drive two "Q4"s? Perhaps a second C4 wouldnt be needed. Would be fun to set the bias with two layers connected to the output terminal?
Since I'm up to bat, might as well take a swing. What if you stacked everything vertically downstream of R6 (560), sharing R5, R4 and C3 into Q2? Could the output Z of Q1 still drive two "Q4"s? Perhaps a second C4 wouldnt be needed. Would be fun to set the bias with two layers connected to the output terminal?
Once you get above 100 watts I tend to add some sort of protection.
As a minimum DC protection and if possible over current detection.
What price do you put on your speakers failing due to blown output transistor(s)?
I have designed and built loads of amps with 4 output transistors and they all sounded fine to me.
As a minimum DC protection and if possible over current detection.
What price do you put on your speakers failing due to blown output transistor(s)?
I have designed and built loads of amps with 4 output transistors and they all sounded fine to me.
That is part of the fun with higher power amps.
Massive amounts of output devices.
Bias stability and current sharing is important.
Maybe.. eventually, the extra resistors wont bother you.
Even slight amount of degen in driver transistors.
Massive amounts of output devices.
Bias stability and current sharing is important.
Maybe.. eventually, the extra resistors wont bother you.
Even slight amount of degen in driver transistors.
That is something well known for a long time, you're right.
But in the world of home entertainment, I'd think that a single pair amp would be plenty, unless you're already deaf or trying to impress/annoy your neighbors.
A pair of MJL2114/MJL21193's would be robust enough for that.
You just need better speakers for more sound with less power. You are in Greece which had a Peavey factory. Search out a pair of used SP2-XT speakers. 101 db 1w 1m. In my living room I used mine at 1/8 watt (1 Vav) to 50 W peak (rare). Very low distortion, IMHO. Successor SP2 (2004) had second Harmonic Distortion measured at 20 db below 5 W output. 98 db 1w1m. If you find some with blown drivers, the originals are still in production.
I used an Apex AX6 as amp. One pair output, built in $3 speaker protector against DC. One channel popped a wire loose & developed a hum problem, so I'm now using a Peavey M-2600 amp which has 2 pairs outputs for 70 w/ch. I really don't hear any difference. My hearing stops at 14 khz. M-2600 has no DC protection, though.
Last edited:
For day to day use, you just don’t need very much power. 10 or 20 watts is enough, 60 more than adequate. Just use the amp the way it was intended and enjoy.
There are times when you will really want to crank it. The solution is a separate system for that. The LOUD one can have whatever compromises are needed. SOA protection, DC relays, 10 output transistors all biased at 3mA. High order crossover distortion be damned. You wont hear it at club levels, or really care. My bigger rig amps have very acceptable SQ even at 50 milliwatts, but I do admit my little unfinished 20 watt Nelson Pass SE-A sounds better. I’ve even swapped out my Yamaha P3500 for something smaller and more optimally biased (100 watts, class H, very low heat, built in non defeatable limiter) because I just don’t need the headroom and got sick of changing ribbons every time there was an accident. My big stuff was designed and built for PA use, and the best thing I ever did was to separate normal and high level usage. Chasing it all with one system was futile. Especially when it comes to speakers. Kappa 15 LFs, 8” mids and a compression driver can take a hell of a lot more beating than Scan-Speaks and ribbons.
About 2 years ago, the last of my home brew PA amps was retired to the PA racks permanently and replaced with two 200 watt tube monoblocks. The system lives in the shop - the only place where higher levels are acceptable. And if I just want background levels there, I just use a 40 watt per channel SS solution.
There are times when you will really want to crank it. The solution is a separate system for that. The LOUD one can have whatever compromises are needed. SOA protection, DC relays, 10 output transistors all biased at 3mA. High order crossover distortion be damned. You wont hear it at club levels, or really care. My bigger rig amps have very acceptable SQ even at 50 milliwatts, but I do admit my little unfinished 20 watt Nelson Pass SE-A sounds better. I’ve even swapped out my Yamaha P3500 for something smaller and more optimally biased (100 watts, class H, very low heat, built in non defeatable limiter) because I just don’t need the headroom and got sick of changing ribbons every time there was an accident. My big stuff was designed and built for PA use, and the best thing I ever did was to separate normal and high level usage. Chasing it all with one system was futile. Especially when it comes to speakers. Kappa 15 LFs, 8” mids and a compression driver can take a hell of a lot more beating than Scan-Speaks and ribbons.
About 2 years ago, the last of my home brew PA amps was retired to the PA racks permanently and replaced with two 200 watt tube monoblocks. The system lives in the shop - the only place where higher levels are acceptable. And if I just want background levels there, I just use a 40 watt per channel SS solution.
I use 2 pairs of output devices in my 70W amplifier. It sounds better to me, comparing to the single pair version.
2 (or more) pairs is a double edge sword. I think it sounds better too - except when playing at milliwatt levels. Overall harmonic and IM distortion is lower at all levels. But the high order crossover distortion products can be worse. The overall transfer characteristic is more linear, but with a larger number of wiggles in it. There is a bigger STEP in input capacitance as one set of devices turns off entirely. Not to mention that the paralleled devices dont necessarily turn off together, leading to more than one step to deal with. Which sounds better? More distortion of lower order or less of it that spreads higher in frequency? It will depend on the program material, listening habits, how high your ears go anymore, personal experience, and your own bias.
I think the input capacitance issue is why I’ve gravitated toward tubes recently. The crossover distortion happens at a higher power level making it less audible to begin with, and the tube input characteristic is free of the abrupt solid state diode type CV curve at turn off. IV is just as nonlinear when forward biasing g1, but the CV is much cleaner. Less energy storage, less high frequency distortion for the NFB to correct for.
I think the input capacitance issue is why I’ve gravitated toward tubes recently. The crossover distortion happens at a higher power level making it less audible to begin with, and the tube input characteristic is free of the abrupt solid state diode type CV curve at turn off. IV is just as nonlinear when forward biasing g1, but the CV is much cleaner. Less energy storage, less high frequency distortion for the NFB to correct for.
Indeed, it's a sensible statement, for people understanding the meaning of being sensible.
So many blogs around the internet promoting "high end" and "monster wattages" like it's some kind of prize or some "awsome" thing to reach for.
Well, look at the automotive competitive audio "scene", where one cannot possibly be in the car whilst the performance measurements are taking place! Hint: it's not because the measurements would be interfered with...
Now there's a whole lotta making sense!
Why is no one in the car when it’s being measured? Because the occupant would be DEAD - with a “D”.
And the test signal doesn’t sound very good anyway.
And the test signal doesn’t sound very good anyway.
I find that those car blaster systems are utterly stupid, and once the fools eardrumbs are shattered and they're deaf, they'll understand.
By then its too late.
By then its too late.
Sure. For more pairs you need more powerful driver to handle the increased capacity. My experience that increased driver bias, and triple darlington is mandatory. So drivers are running with 25mA instead of the common 5-10mA. Check out D'Agostino amplifiers: drivers are same as the output devices (NJW3281/1302), and they use 2 pairs of them to drive 6-8 pairs of outputs. Relentless monoblocks are absolutely overkill with 48 output pairs...
My request has gone very wild, members are placing answers without reading the OP
I AM well familiar with all the things written above espacially the part with the power and how much is actually needed
For my personal use if the P3A was rated 15 W continous 40 peak will be more than enough for every day use
I will try to explain more :
The P3a i make as an integrated amp matched with preamp+ dedicated PSU ,implemented in specific pcb's, dual mono psu,1302-3281 engine , relay input selector , refinments and so on and on, has been outperforming a number of consumer amplifiers ( some of them far more expensive some others far more power some other with big names on them ) and it is doing that for a very long period of more than 10 years .
For one time only i am called to outperform one amplifier that is 6 times the power of a well made P3A , overdesign supplied with 600W trafos/ch , driving a 4R and 90 db given speaker .
-----------------------------------------------------
the amplifier that i am willing to outperform ( at the time i am not willing to say model and brand ) in the first place is BTL topology that to begin with means in a few words that you have 2 amplifiers 150w each summed on the speaker and with 180 degrees diference they produce 300W
Given fact when you have 2 amplifiers summed like that BTL, you have double the LTP, double the suport current sources current mirrors, double the VAS double the drivers , double the safety measures double everything
Only thing that is half is the capacitance of the output stage which is now shared between the 2 amplifiers .
So the P3a A will outperfrom the big amplifier in terms of sound any way you slice it due to far lower part count
i take this as a fact
let us say that 150 W of power will be more than enough to outperform the beast with the given speakers .
so the original question is how you produce 150 W of power but without loosing all the sweet features of a P3A
there is a chance that the answer is: " you dont"
this is what i am willing to learn
I AM well familiar with all the things written above espacially the part with the power and how much is actually needed
For my personal use if the P3A was rated 15 W continous 40 peak will be more than enough for every day use
I will try to explain more :
The P3a i make as an integrated amp matched with preamp+ dedicated PSU ,implemented in specific pcb's, dual mono psu,1302-3281 engine , relay input selector , refinments and so on and on, has been outperforming a number of consumer amplifiers ( some of them far more expensive some others far more power some other with big names on them ) and it is doing that for a very long period of more than 10 years .
For one time only i am called to outperform one amplifier that is 6 times the power of a well made P3A , overdesign supplied with 600W trafos/ch , driving a 4R and 90 db given speaker .
-----------------------------------------------------
the amplifier that i am willing to outperform ( at the time i am not willing to say model and brand ) in the first place is BTL topology that to begin with means in a few words that you have 2 amplifiers 150w each summed on the speaker and with 180 degrees diference they produce 300W
Given fact when you have 2 amplifiers summed like that BTL, you have double the LTP, double the suport current sources current mirrors, double the VAS double the drivers , double the safety measures double everything
Only thing that is half is the capacitance of the output stage which is now shared between the 2 amplifiers .
So the P3a A will outperfrom the big amplifier in terms of sound any way you slice it due to far lower part count
i take this as a fact
let us say that 150 W of power will be more than enough to outperform the beast with the given speakers .
so the original question is how you produce 150 W of power but without loosing all the sweet features of a P3A
there is a chance that the answer is: " you dont"
this is what i am willing to learn