P3A - 60-100W Hi-Fi PowerAmp – Dual Output

http://autoforum.com.br/index.php?showtopic=70257&st=60

http://www.diyaudio.com/forums/showthread.php?s=&threadid=24057

High Power P3A design questions. Post #1

Hi, I have been looking at building myself a version of the P3A amp from Rod's pages, but I have a
few questions about the design and it's implementation.

Firstly, Rod states that teh maximum rail voltage which should be used with this design should be
+/- 40V to get about 100W into an eight ohm load. However, as I will be building this with the
newer onsemi output devices and drivers which are rated at up to 230V (I think, that is just a figure
from memory, it could be 350 but I don't think it is any lower than 200V) as well as doubling the
number of output devices, would it be safe for me to run this design on higher rails? As I was
considering going as high as +/- 50V to get around 125W out of each channel.

The second point is concerning the aditional output devices. Now, in the article, Rod states that the
high frequency rolloff of the amp is coverned by a descreet capacitor added to aid stability. So,
would I be correct in thinking that the extra juction cappacitance seen by the drivers from the extra
output transistors will not have to great an effect on the amps frequency responce as this is not the
limiting factor governing the circuit design?

Anyway, I have included my updated scheematic below, along with a provisiona copy of the PCB
layout and would apreciate comments on both, as well as teh questions posted above.
Andrew.

Here is my provisional PCB layout to with the above design. I intend on testing this for a single
channel on a single sided PCB, but if it goes well, I intend to make the board double sided and get
few professionaly made.

Also, if anyone knows a good scource of low value, high power resistors that can be end mounted
in the Uk can you pleasse post here, as I am having trouble finding any at the moment.
andrew.
 Post #5
Here is another coppy of the board, but without the component overlay (which isn't finished yet
anyway) so you can get a better look at the track layout.

Post #6
Pay attention to your driver transistors. You may need to replace those with a higher rated type.

Post #7
Hi, the driver transistors will also be the new, very well specified, models MJE15034 and
MJE15035 from onsemi, both being rated at 350V (it was 350 for the outout devices too, 230W for
their power handling) and up to 4A each.

What about the small signal transistors that are on there? Is there anything better thant he BC546 to
use, as I always thought that these were a cheep generic component, so there must be something
that would be better out there to replace them?

Post #8
I think Rod told me to use the mpsa42 and mpsa92's for high power p3a's....someone might need to
confirm that though, I can't remember. I'm pretty sure that they'll work.

-Matthew K. Olson

Post #9
Hmm, not the fastest devices in the world (only 50MHz, compared to the 35MHz for the output
devices seems a little slow) but I guess they would more than do their job properly and the specs
for current and voltage handling seem ok. I guess I could always give a few different types a go and
see if any are noticably better in there.

Post #10
I use 2N5551 small signal transistors, in my P3A running at +/-52V.

http://www.fairchildsemi.com/ds/2N/2N5551.pdf

/Freddie

Post #13
quote:
Originally posted by bigparsnip
I found something else now, it's a toshiba part, the 2SC2240. It's rated at 120V, at
100MHz, with a typical HF in the 200-700 region, so I guess this would be my best
choice so far. Has anyone else used or seen these about before though?

2SC2240 data sheet

I use them all the time with great results. I've never had one fail even at rails up to 80V! The
2SA970 is the PNP compliment. They are very wide bandwidth and quiet too. Try 'em you'll love
it.

Re: HO P3A Post #15

Parsnip,

First, make sure and credit Rod for any work, a link to his site would be the right thing to do.

Regarding your circuit, use Toshiba 2SC1775As for the differential pair, trust me on this, the sound
improvement is not slight. They are rated >100v. The pin config is different from most, so double-
check your board. Also you MUST match the pair as the 1775's gain varies

For C4 and C5, I'd use a minimum of 100uF. The amp tends to sound a little thin otherwise. The
same goes for the input cap, a large (>=4.7uF) film cap works well. For best sound, use a non-polar
for C4, Black Gate is great, Nichicon Muse series sound nice as well.

Your chosen Onsemi drivers will work fine.

But all that being said, I would stick to Rod's recommended max rail voltages. The slight increase
in output is simply not worth it. As advertised on Rod's site, limiting power to 100 watts ensures an
amp that will be operational for years (probably decades). If 100 watts isn't enough for your
application, simply use his bridged design, or purchase/build more efficient speakers!

250w Post #31

Parsnip,

The maximum safe output for your Onsemi output transistors is:

@120 deg farenheit - 186 watts

@130 deg 176
@140 166

At 50v I think you will have difficulty keeping the tranny temperatures below 120 unless you use
fans. You can see that every 10 degrees in excess of 68 deg farenheit reduces the maximum power
output by roughly 10 watts. Also you'll find the quiescent current hard to stabilize without some
cooling for the drivers.

So at 130 degrees (55c) and 4 ohms, you'd be exceeding Onsemi's recommended operating area at
anything greater than roughly 42v.

Better safe than deaf,

Post #56
Hi, after a bit of a hectic weekend moving from one end of the country to the other, I have finaly
finished teh second revision of my high power P3A design, so I thought I would post teh revised
scematics and board layouts so people can look for more mistake or give me some more pointers.

First up is the ne schematic, which has teh correctly paralled output devices, allong with the correct
small signal transistors that I intend to use. I have also included a resistor at the base of T7 to try
and keep things a little more stable as the gain devices will be running under a little more stress
than normal.
 Post #57
Here is the new board layout, it does contain one more wire link, but I have manged to get the
output devices close enough to the edge of the board to be directly bolted to a heatsink.

Post #58
Here is the board without the component overlay to give you a better view of the layout of the traks
on the board.
drivers thermal runway Post #62
Plz Pay attention to Drivers of output transistors
They must be thermally coupled very well unless the thermal runaway might occur due to
multiplication of positive thermal conduction coefficient feedback due to SIZIKLAI PAIR
CONFIGURATION. As a care there temperature must not exceed safe limits of about 64C'

Post #63
I just replaced bd139/bd140 and mjl193/194s with mje15034/35 and mjl4302/4281 in 2 channels of
my 4ch p3a amplifier. My bias current went up from 95 to 240 after changing transistors. Should I
run higher bias with different transistors or turn it down ?

This weekend I plan on comparing my p3a channels to see if I can hear any difference between
different transistors . Also, I am adding heatsinks to the Q5 and Q6 pair just in case.

Post #64
quote:
Originally posted by Jean
I just replaced bd139/bd140 and mjl193/194s with mje15034/35 and mjl4302/4281 in 2
channels of my 4ch p3a amplifier. My bias current went up from 95 to 240 after
changing transistors. Should I run higher bias with different transistors or turn it
down ?

This weekend I plan on comparing my p3a channels to see if I can hear any difference
between different transistors . Also, I am adding heatsinks to the Q5 and Q6 pair just in
case.

Do as you like, I run my P3A at 200mA bias current. I think it sounds better at 200mA than at
100mA.

/Freddie

Comparing transistors Post #66

I just finished a second P3A built with the the new OnSemi parts.
My initial impression is that I dont like it, the higher frequencies are "harsh or edgy" and the vocals
sound "nasal" if that makes any sense. My wifes ears also hear it as different " the singer has a
cold".
I only played it about an hour so Im hoping for some "run in" either of the parts or my head or
better yet an assembly problem.
My recolection of the 1st P3A was that it sounded great imediately and much better than the
LM3886 that I assembled on an ESP board. Unfortunately this P3A sounds like the LM3886, and
not at all layed back on the upper end or vocals.

The amp was built with some higher quality parts polystyrene caps at strategic positions as
described in the assembly link at ESP. Power supply is identical for both amps. Ive got parts for
another 8 channels so Im really hoping there is a distinct problem I can solve.

Any ideas would be greatly appreciated

Re: Comparing transistors Post #67

quote:
Originally posted by Fritz
I just finished a second P3A built with the the new OnSemi parts.
 My initial impression is that I dont like it, the higher frequencies are "harsh or edgy"
and the vocals sound "nasal" if that makes any sense. My wifes ears also hear it as
different " the singer has a cold".
I only played it about an hour so Im hoping for some "run in" either of the parts or my
head or better yet an assembly problem.

My recolection of the 1st P3A was that it sounded great imediately and much better than
the LM3886 that I assembled on an ESP board. Unfortunately this P3A sounds like the
LM3886, and not at all layed back on the upper end or vocals.

The amp was built with some higher quality parts polystyrene caps at strategic positions
as described in the assembly link at ESP. Power supply is identical for both amps. Ive got
parts for another 8 channels so Im really hoping there is a distinct problem I can solve.

Any ideas would be greatly appreciated

Try higher bias. It sounds like you are hearing some switching distortion.

Post #69
quote:
Originally posted by Fritz
OK Ill try more bias, it now occurs to me that I assumed the bias was the same as for
the old transistors, about 100mA.

Any suggestions or should I just increment up and see if the sound improves?

Thanks

Fritz
Use an o'scope and use a sine wave at a low power level and look for 'bumps' in the curve that
indicate transistor turn on/off. Adjust to get rid of these. Be careful to not be fooled by the problem
being before the outputs and measure the signal at several points along the path from input to
output to see if it is clean.

Post #73
Yes I do, but I wouldn't recomend using the free version of eagle to try and produce a board with
eight output devices, as the one I have posted here is as physicaly large a board that the free version
will allow. So I'm not too sure you would be able to get all of the output devices on there easily
without using two or more boards (although, perhaps if you ran them down both top and bottom
sides of the board...).

Anyway, if you want the schematic, I have atached it below so you should be able to work ok from
there (Just be careful with the power supply rails I have a feeling I may have acidently joined them
at one point when I was doing my PCB layout.
Andrew.

Attachment: p3a-extended rev2.zip

Post #75
Nice information about modifications possible on the P3a. 5 months later is there any progress in
this design? I'm pretty curious about the actual performance!
Post #76
I abandoned it when I found out that adding more output transistors into a CFP creates all kinds of
issues. One of them was ringing on square waves at high frequencies. As Rod once stated, the P3A
is a simple amplifier for about 70W of power, going higher requires a more advanced design.

I'm currently working on a scaled down version of the Leach amp with 2 pairs of output transistors
to run at about 100W into 8 ohms from 45V rails. For those wanting more power, use the Leach as
is. It's an excellent amplifier.

eed sugestions for my pcb project Post #77

hello.

some days ago i found this upgrade for ESP3a and i want to build one. I made pcb project and i
need opinions before i start with pcb-making.

note about heatsinks:

- all power transistors on big heatsink;

- drivers(T5, T6) on small 2 hestsinks.
- T7 on big heatsink

any comments are welcome.

regards

This is last version of my PCB for HO ESP_P3A.

In this version power transistors and drivers are on one heatsink.

Schematic, PCB and some infos are here:

http://powa.w.interia.pl/ESP3A_ext-by_powa_rev2.zip

powa'h
 Post #79
For best thermal stability the drivers should be mounted on a separate heatsink. The design of the
output stage of the P3a is such that the driver temperature is what should be sensed by the Vbe
multiplier, not the output stage. Forcing the driver transistors to be as hot as the output stage can
result in less than optimum quiescent current.

__________________
I love it when a plan comes together

Post #80
that's sounds not so good for me. previous version was with 2 heatsinks - separately for power
transistors and drivers. so back to the roots

ok. tommorow(now is midnight in my city ) i'll try to make new version. with separate heatsinks.
i have no idea about value of R25 - emmiter resistor of T7. any suggestions?

here is schematic: http://www.diyaudio.com/forums/atta...tamp=1071610659

BTW: thanks for reply.

I think you don't need a resistor there.

But if you want to put one, try 1 Ohm.

xt version(better?) Post #82
Hello.

here is next version of my pcb, without emmiter resistor and with separate heatsinks for drivers.

BTW: one more question: is heatsink needed for drivers in normal(one pair of power transistors)
configuration in ESP3A? how big must be for HO configuration? on my board is not so much
place...

all sugestions are welcome.

Post #83
Your board looks nice. What software do you use?

I think you will be OK without a heatsink on the drivers, but you could get a small clip on one
maybe?

__________________

Post #84
It would be nice to have enough area around the drivers to mount heatsinks on them . Nice pcb
layout !