Leach Amp

[Kilowatt]

Here's my progress

I'm building 2 Leach amps for my home stereo system. Eventually I will probably upgrade to 4. I'm done with the power supply, and I'm nearing completion of one channel. My power supply measures +/-60.5V at no load (40-0-40 1500VA toroid, 11200uF per side). This should give about 120W into 8 ohms and nearly twice that into 4 ohms according to the article about the amp. That ought to be plenty for my subs. Has anyone here built the Leach amp? How does it perform into 4 ohms? I would tend to think that it would try to put out more like 300 or more watts into 4 ohms instead of 240 because of Ohms Law, which would probably be too much for the outputs. Am I right?

I'll probably build some ESP project 03 amps for mid and trebble instead of tube amps because big tube amps are too expensive for me to build.

I am reading G. Randy Slone's High-Power Audio Amplifier Construction Manual and will attempt to build the 550W amp that's in there once I'm done with Leach amps, but before I try for my big amp.

 

[Damon Hill]

You might be pushing your luck to get 300 watts out of even two
pairs of MJ15003/4s; perhaps try three pairs and slightly higher
values of emitter resistors. You'll need a higher power supply
voltage too and that's where you'll be pushing the margins.

An alternative might be multiple sets of MJ21193/4s; anyone have
other recommendations? How about TO-3 versions of the MJE15030/31? (see next paragraph)

I've got a monster wind tunnel heat sink that's hard-wired for up
to five pairs of output devices and I'm tempted to try out an old
batch of MJ15003/4s to see if I can squeeze 500 watts continuous. I'll have to get a transformer rated for that voltage/power.

I've tested my Leach amp at full power into 3.75 ohms, but never
for more than a few minutes--the heat sinks are too small and
poorly ventilated for that kind of power. I'm rather pleased with
the sound of the ver 4.5 boards and can't wait to get a set of
matched MJ21193/4s installed, to see that improves the sound.

Probably can't go wrong with Sloan's designs, either. Just be sure
you've got lots of heat sink area and air flow.

 

[Kilowatt]

I'm not out to get much more than 200W from each Leach amp because that's what my speakers are rated. I've got 4 Sony Xplod 12s. They're 200W RMS, 4 ohms. I'll have two of them in series for each of my 2 Leach amps (120W/channel), and when I make my other 2, there will be a Leach amp for each sub (at least 200W/sub). So the 2 x MJ15003/MJ15004 with my supply should be fine, but I might add more outputs in parallel, so I can get more power out reliably, maybe even into 2 ohms.

I don't know well 1500VA will supply 800W worth of Leach amps, and I might want to add more filter caps too.

I wasn't thinking when I made my ver. 4.5 boards, and I ended up with my foil pattern a mirror image of the original artwork, because I just printed it out (foil side), layed it ink-side-down on my board and developed it like that. Because of this, I have had to turn all my transistors around, so the drivers (TO-220 case) do not lay flat on the board as they are supposed to, and I had to bend the base lead on the metal can transistors the other way, but it shouldn't cause any real problems with the function of the amp. At least I hope not.

By the way, Damon Hill, when you tested you're Leach amp into 3.75 ohms, what was the supply voltage, and what was the power output?

I wonder if I could use 2N5684/2N5686 transistors for outputs especially for 2 ohm loads. Those things are beasts! 300W, 50A, 80Vceo. They're expensive, over $5 each, and I have 2 of each in the junk box, so I might as well try to use them for something.

 

[djk]

I have several factory built Leach amps in addition to home brew. The LSR&D 101 has a 90VCT 1KVA transformer with a pair of 20,000µF cans for +/-63VDC and does 300W+300W@4R. A couple of these have been driving 2R for the last 10 years. LSR&D used 3 pair of MJ15011/15012 with one pair of MJE15030/15031 as drivers. The Super Leach from Heathkit, AA1800, used four pair series connected MJ15024/15025 and did 400W+400W@8R with +/-95VDC. The AB International 1100A uses the Leach topology with a tiered power supply and does 1000W+1000W@2R with only three pair of MJ15024/15025 with rails at +/- 0, 48, and 95VDC.

 

[Damon Hill]

For whatever it's worth, I've just replaced the MJ15003/4s in my
Leach amplifier with MJ21193/4s. Initial tests and listening
suggest, unsurprisingly, no difference in stability or sound
quality.

Both batches of transistors tested very close together for beta
at 1 amp, so matching seems to be a given. They replaced
a motley collection of unmatched devices with four different date codes.

The newer devices have a greater SOA, somewhat better
linearity, and a Ft of 4 Mhz vs. 2 Mhz; I suppose it may be possible
to modify Leach's design to take better advantage of the new
devices, but I'm not sure where to begin.

I recently upgraded my amplifier to the ver. 4.5 circuit and am
very pleased with the sound quality. Imaging and sound field
are much better, and a slightly strained and edgy quality from
the older 2.something circuit seems to have vanished; the sound
is almost effortless now.

Needs a reality check against some known high-quality amplifiers.
I may audition it on a friend's Martin-Logan electrostatics.