Transistor choice for P3A

[ChrisJennings]

I am currently gathering parts to build a four channel amplifier utilising the P3A design. I am considering changing the output transistors to 2SC3264 and 2SC3264 any comments would be welcome.

Regards,

Christian

[XEAGLEKEEPER]

Chris
How about MJL4281A and MJL4302A these are reccomended by Rod Elliot and you can get samples( 5 each) from Onsemiconductor for shipping cost only.
Dave

[AndrewT]

Hi Chris,
stick with recomendations, it's safer, if this is your first.

Why 4 channel?

Where are your speakers?
All next to each other or spread around the room?
Then put the power amps beside the speakers and make the system perform better with SHORT speaker cables and use long interconnects instead.

[ChrisJennings]

Thank you for your response, I think I have decided to use two pairs of 2SC5200/2SC1943 in parallel although I am slightly concerned about matching. I will be running 42 volt rails and driving an 8 ohm load. I hope this will give a noticeable improvement over the MJL4281A and MJL4302A.

AndrewT, is there really an advantage to using long interconnects and locating the amplifiers near to the speakers? I don’t think this would help in my case as my listening room is 5m x 5m. I was under the impression that if needed, longer runs of speaker cable are preferred than long runs of interconnects. Is this wrong?

Regards

Christian

[Greg Erskine]

Hi Chris,

Just in case you didn't know, Rod Elliott (ESP) has his own forum were topics like this get discussed quite regularly.

http://sound.westhost.com/phpBB2/viewtopic.php?t=1320

I think single 2SC5200/2SA1943 will be OK. My P3A uses single 2SC5200/2SA1943 with 35v rails into 6 ohms. When I built mine Rod was recommending single Toshiba 2SA1302/2SC3281 at 42v into 8 ohm.

regards

[vynuhl.addict]

MJL3281/1302 are also a good choice, I recall hearing that their real world specs are still better than the 4281/1302. The On Semi devices are more robust than the Toshiba counterparts in an output section.. I used 3281's in my original p3a, but I didnt order the board from rod but instead used perf board and point to point wiring..As far as matching goes, i found beta matching to be sufficient for transistors from the same lot, i also didnt find any discernable difference in close output matching as opposed to matching the differential and driver transistors as close as possible in this design..


Colin

[AndrewT]

Hi Vyn,

matching the......... driver transistors

what parameters did you match and to what accuracy/tolerance?

[AndrewT]

Originally posted by ChrisJennings
is there really an advantage to using long interconnects and locating the amplifiers near to the speakers? I don’t think this would help in my case as my listening room is 5m x 5m. I was under the impression that if needed, longer runs of speaker cable are preferred than long runs of interconnects. Is this wrong?

you'll gather from my post that I believe the opposite holds.

The speaker cable is a load that affects the amplifier and particularly it's stability margin.
I further believe that much of the cable sound that is often described is NOT due to the sound of the cable but more due to the effect the cable characteristics have on the stability margin and the resulting change in behaviour of the amplifier that then sounds different. It can move from quite a damped output signal (inductive load) to severe overshoot (on capacitive load) and these WILL SOUND DIFFERENT. These output characteristics can change with frequency as the load changes.

Now back to long interconnects. They in general have less influence on the technical spec/output of the receiver and if connected/designed correctly should have no influence on the final sound quality.

Finally long interconnects (even balanced) are cheaper than long speaker cables.

[ChrisJennings]

Thank you all for your response, lots more to think about and read. I’m off for Christmas dinner then to play with my new UV exposure box that Santa brought!

Merry Christmas

[vynuhl.addict]

Hi Andrew,


The parameter I matched was strictly beta for the differential and drivers I got really lucky with the mje 15032/33 and got exact matched beta just using a multimeter, simple and sort of crude but the dividends were in the sound quality. I have long since scrapped the p3a project as i went with a design with a ccs vas rather than bootstrapped and a current mirror in the differential and emitter follower outputs similar to the blameless amplifier by doug self.. I prefer source colorations (vinyl)rather than amplifier colorations which I found the p3a not to lend itself well to all types of music..



Colin