M&M PA-250 amplifier investigated

[tsmith1315]

Or, a quick lesson in don’ts.

This amp was built by a company with a good reputation, and was part of their introductory line of amps released in Aug 1990. Price was likely around $550. Their 50W x 2 was $350, and 125W x 4 (just two of this model in one case) was $926.

As amc32 already pointed out in this post, I’ll reinstate:

- don’t mount your vibrating automotive PCB’s through your BCE’s.

While tracing the circuit to make the schematic, I had a couple of “do WHAT?” moments.

Anyone else bored enough to nitpick this one?

[djQUAN]

no emitter resistors for the PS switching transistors?

[tsmith1315]

Well, that's typical. Maybe I mislead by representing the jumpers in the schematic as R's.

But those little 0 ohm jumpers must pass the entire primary winding current, which is probably 99% of the current drawn by this amp at full output. According to the manual, that's 45A !!!


Apparently that was a compromise for the long skinny PCB shape.

>>Lesson 2 – don’t get so stuck on a PCB geometry that you make major compromises to high current or sensitive traces in your PCB layout.

[ppia600]

-don't hang the transformer in mid-air... from the pcb's which are hanging from the switching transistors
-don't parallel tip35's or 36's with no emitter resistors (already stated)
-don't use long slim caps and not at least glue them securely

Can't see the schematic for some reasons

[tsmith1315]

Originally Posted by ppia600

Can't see the schematic for some reasons

To get a readable full size image (IE), I have to click the "expand" button on the bottom left of the schematic -after it has already expanded from thumbnail size. Same with the specs page. Original file size was about 800k, but it looked good resized to 200k on this end. I can post it as a link if need be.

Or is it not showing up at all?

It's about as simple as a supply gets and may be helpful for someone looking into how things work.

[Eva]

Class AB amplifier designers trying to do SMPS and car audio. The result looks like a joke, particularly the toroid hanging and the twisted magnet wires.

btw: There was a better way to do a SMPS with TIP35 and TIP36 and the parts available by that time, it was just a matter of copying the proportional drive topology from the old PC power supplies (still good enough to be in use today).

[tsmith1315]

What, better than a NAND multivibrator driving all 8 transistors in unison? Nah.

I haven't counted, but I swear I don't see enough turns to get rails suitable for 125W x 2.

Thanks for chimimg in Eva, care to elaborate on the proportional drive a little?

[Eva]

Proportional drive uses a current transformer to sense collector current, and applies the output to the bases, so instantaneous base current is always proportional to collector current for best switching efficiency and lowest losses when the amplifier is idle. The optimum proportion is typically 1:5 (1:5:5 or 2:10:10 turns). This also allows to get the lowest losses at high collector currents.

The circuit uses some gadgets (with resistors, capacitors and diodes) to provide a negative base current spike at the end of the cycle for quick charge removal and fast turn off, and provides negative bias for the bases of the transistors that are off, something that allows to use them close to their ratings safely.

A TL494 (available since early 1980s) usually assists the circuit to keep switching frequency stable, and to provide duty cycle control and regulation if desired, although it can be made self oscillating (no active components) by using a saturable core for the current transformer (a kind of ferrite whose saturation is very abrupt and predictable). So yes, regulated efficient SMPS in car amps were already possible by that time.

This topology is good enough to be used today in most low cost electronic halogen lamp transformers and fluorescent ballasts. I tried it in a 230V to 14V 125A adjustable power supply prototype a few years ago and I managed to get nearly 1800W reliably with just four MJE13009 TO-220 bipolar transistors on the primary side and efficiency comparable to current MOSFET technology. The design of the current transformer is critical, though, coupling has to be very good, leakage inductance degrades switching speed a lot.

[tsmith1315]

Originally Posted by Eva

I tried it in a 230V to 14V 125A adjustable power supply prototype a few years ago and I managed to get nearly 1800W reliably with just four MJE13009 TO-220 bipolar transistors on the primary side and efficiency comparable to current MOSFET technology.

Very impressive.

And thanks for the explanation!

[gisewhcs]

Just for your interest: I've sent my MM250 (most probably the next generation of M&M amps) to somebody who benchmarked it for me.
This one has a better concept all over...
It's a Mantz!

And here's the link:

M&M Electronics MM250

It's an interesting site at all and if anybody needs german words to be translated, just ask!