Symasym biased for class A

[Dr.EM]

I can't seem to find any info on this so will start a new thread. I am building another Symasym amp to power planar tweeters in a 3 way active setup. My plan is to use regulated +/-30V supplies and also to increase the circuit bias to give mabye 4W of class A power. Load is 8 ohms, very much flat with the planars and active setup.

I will be using the 1943 and 5200 type output devices mounted on a 0.4c/w heatsink per amp. I am hoping to build today (am matching signal transistors right now!) but need to make it so that I can set the bias suitably high (since it is not designed for such high bias?). What will need to be changed to allow a suitable bias for 3-4W class A? Or mabye it will already do this (my other one didn't look like it though with the pot nearly all the way round for the typical 55ma bias).

Any help appreciated!

[al3xu]

I'm running my Symasym monoblocks with 350mA bias (+/-32V rails). Output devices are MJL21193/4 on fairly big heatsinks. They do turn pretty hot.

My initial target was on the 700mA side, but I think I need to add more output devices (I was thinking 3 per side) to reach it. As I don't have an appropiate way of measuring the heatsink's temperature (other than my multimeter's termocuple, and I kinnda' don't trust it) I just don't want to push it any harder given the summer temperatures these days.

I am running the monoblocks full range but as I increased (in little steps follwed by several hours/days of listening after each step) the bias current I noticed the sound getting sweeter and more detailed specially in the upper treble range.

Anyway, I am very curious about your findings. Good luck with your build!

[Dr.EM]

Your results are very promising! 350ma bias gives about 2W into 8ohm of class A? Not sure how to calculate it. Also what voltage should I look for accross the 0.22r output resistors to set this 300-350ma bias current? Can it be set this high without changes to the design or must changes be made to allow setting of this higher bias? Let me know what changes must be made if there are any! Many thanks

[al3xu]

For 350mA bias current you should measure aprox. 150mV across both 0.22R. But be carefull, my output tranzistors are rated at 200W dissipation if I remember corectly.

There are no changes for you to make to the original Symasym 5.3 by Mike Bittner.

I suggest you do the bias increasing in steps and watch over the heatsink's temperature. They will get hot, although your heatsinks seem quite big at 0.4 C/W.

One mod I made to the original design was changing the input long tailed pair tranzistors to 2sk170B. It's really a worthwhile change. (Does nothing to do with the bias current increase)

Keep up the good work and keep us posted!

[Dr.EM]

Thanks very much for your help! I shall carry on building as per the original design and will certainly keep you posted . The heatsinks are 30cm by 10cm with 3cm fins on a 1cm base, all anodised black aluminium. They form the sides of the case too so are in direct contact with the air for good convection. The transfer from transistor to heatsink will be most critical, I plan a mount with allen bolts and quality Sil pads, should do the trick

[Dr.EM]

One amp is just about ready for testing. I ideally need to get the allen type mounting bolts and change those 1.2ohm resistors for correctly rated ones (they are 1W but should be 2W, will do for testing). I'm off out soon so will get back to you tomorrow, hopefully it works! The overall mounting scheme is much tidier and simpler than my previous build

[al3xu]

Looks like a nice clean build! I noticed you didn't use the recommended wimas for by-passing.

One more advice, use extra capacity for the power supply capacitors as the increased power in class A will drain them much faster.

For example, my 2x33.000uF per monoblock only last for 3-4 seconds after I switch of supply power, vs. 10-15 seconds with the recommended bias setting (55mA).

[AndrewT]

go back to basics.
required maximum ClassA output power = 4watts.
Lets assume the driver is 8ohm impedance.

Required output voltage for 4W into 8r0 = sqrt[2*P*Rload]=8Vpk.
check:-
8Vpk across 8ohms results in 1Apk
P=Ipk*Ipk*Rload/2= 1*1*8/2=4W.

Your load only needs 8Vpk and 1Apk to meet your 4W max power requirement.

Assume that the voltage losses through the amp are 5Vpk when delivering max current.
The PSU supply rails need ~ 8Vpk+5Vpk ~+-13V. Not +-30V

For a push pull amplifier the quiescent current in the output stage ~Ipk/2 ~ 500mA. Add 50mA to avoid non linearities as ClassA limit is approached.

Heatsink dissipation ~=0.55*[13+13]~=15W shared between two transistors.

Will the symasym work on +-13Vdc, probably not, it requires a redesign. Or choose another source circuit.

[Dr.EM]

I got 100 of those little yellow box capacitors, they seemed to work in my other symasym, still polyester, 63v.

I plan a regulated supply, the design is here:

LM338 regulated snubberized PSU for audio amplifiers

This should eliminate noise problems you can get with class A. Will I still need extra capacitance before the regulators though say? Bear in mind there are no big bass peaks to handle!

I probably wasn't clear enough in my first post. I proposed 4W class A power (roughly) but by using the supply of +/-30V I would still have mabye 40-50W power on tap, albeit in class AB. It is for tweeters so the 4 watts will probably cover most circumstances but as you have often said it is still wise to have headroom with tweeters to cover possible peaks. I figure I can live with these being handled in class AB and 3-4W of class A strikes a reasonable balance.

The maths is good to see, might be able to manage the 550ma quiescent if it is only 7.5W per package. The total 15W gives only a 6c temp rise with my heatsinks (ignoring insulators etc). There will also be regulator chips on the same sinks though adding about 3W quiescent dissipation.

[al3xu]

I don't think you should worry about class A noise problems since this is not a "real" class A amplifier.

I really don't know if it will benefit much from the regulated supply proposed. I guess it wouldn't harm to try both ways and see the differences. But since I'm not much of an expert better wait for a second opinion on that, especially given the fact that the psu was designed for chipamps.