Leach vs. LM4780 - listening test

[BobEllis]

I did a comparison between my Leach amp and an LM4780 based amp for those who are interested.

I have a 4 channel leach amp that I have been using to biamp my latest DIY speakers - roughly a JM Lab Electra 906 clone using the Focal 6W4256 and T120DX5. Looking to bring them to my bedroom, I wanted something smaller than the Leach, so I bought a couple of LM4780s to try.

I dead bugged the pair into a four channel amp, working with pretty strong magnifying glasses. I didn't expect the leads to be so close together. It is as ugly as one would expect a prototype to be, so no pictures. Everything is laid out on a spare piece of MDF. The circuit is straight out of the app note page 1, minus the capacitors in the feedback loops and I added a 10R/100nf zoebel. No magic parts were used, just digikey and mouser stuff, like my Leach. The power transformer was pulled out of a dead receiver, and produces 35 volt rails. I have 3300 uf 50V Xicon caps within an inch of the chips.

Power up went well, no oscillations visible on my scope, current draw within specs, offset <10mV. I hooked it up to my crossover and speakers and hit play. Not my typical auditioning material, but Jimmy Buffet's Bananna Wind was in the player.

Initial impressions were pretty good, then I noticed a low level click that sounded like testing a paper woofer with a 1.5 V battery. Oh, that must be the wood blocks. The other background percussion wasn't there at all, but the drums and cymbals were - to a fault. the cymbals seem harsh. I don't think I was clipping anything, since my el cheapo voltmeter says I was running 2-3 volts out.

With the Leach amp last night, I could hear a lot of background percussion - wood blocks clearly hit and resonating, tamborine(s), steel drums on these tracks. All were precicesly placed. The chip amp imaging was broader and somewhat drifty, Cybals were almost always located at one speaker or the other or both. With the Leach it sounded like a drum kit in the center.

I tried Dire Straits Brothers in Arms next. The drum intro to Money for Nothing was painfully harsh. Ok, I cranked it up a bit, but I still don't think I was clipping anything. The song sounded like it was just drums, bass guitar and lead vocals. It sounded pretty good, but the Leach showed me that there is a lot more information on the disc that just wasn't being reproduced by the chips - sort of the MP3 effect.

All in all, I think the LM4780 is midfi at best - not terrible sounding but missing a lot of detail and a bit harsh on the top end. Had I never heard anything better, I'd be able to live with it. The Leach is able to pass much more detail, with much better imaging. The Leach transmits enough detail that I can hear that the drum is being hit in different parts of the head, while with the chip it's just a drum hit. I still like my A-75 a little better than the Leach for the same reasons I prefer the Leach to the chip. I haven't tried the A-75 on these speakers, since I only have two channels.

Outcome:

I will make room for the Leach amp in my bedroom. I don't intend to mess with chip amps any further except for non-critical applications. Exotic resistors and capacitors may improve the performance a bit, as might BrianGT's (or other good) boards. It's just not my cup of tea. As always, your mileage may vary.

My son is happy that I did this experiment. The chips will be bridged/parralleled into a new bass guitar amp for him.


Other equipment: (in case it matters to you)

Pioneer DVL-919 disc player
Yamaha DSP-A1 used as a DAC/Preamp (it's DAC is noticeably cleaner than the Pioneer's)
Crossover - breadboard DIY using 5532's biased into class A with 1K to negative rail.
Interconnects - ancient Monster
speaker cable - parts express 12 ga.
2 parallel 12uf Axon caps to protect each tweeter from DC

[destroyer X]

The small distances you have inside one chip amplifier will make them oscilate very easy... to avoid, larger capacitors are used internally and some externally.

I imagine that those caps are making all the mass.

But there are some chips that can turn you confused related "who is really the better".

I could detect two Chips, only those two, the first one is the STK4142 and all cousins, for instance, STK4142II.

The other one, have not in my mind now, if you interested i can find the model number to you, the prefix is LA.

The STK4142 was used in many models from Marantz, JVC, Sony and other brands, and, not disturbed by speakers or bad pre-amplifiers with strange tonal controls, will produce a very agreable sound.

I never compare this STK with Leach amplifier, but i am sure, only using memory, that the first one has not loss in deep bass.

The chips can be well matched inside, this is some advantage, but the construction can be noisy and hard to trust, because of small metal surface to transfer heat to the sinks.

I thank you very much for your informs, this is a precious test, and valuable hearing analises, so serious to my person, that i will never even think in assemble that chip.

Loss of details, erasure of instruments, those informs are enougth to send the chip to junk!

thank you

Carlos

[BobEllis]

Carlos,

I actually expected it to sound pretty good not be a throwaway. This chip is pretty much a pair of LM3886s in a single package and that chip has had such good press. I wasn't expecting nirvana, but I had at least hoped for something listenable.

I didn't compare deep bass capabilities, since I have a separate sub. - forgot to mention that initially. It's a JBL2245H powered by a Hafler 500 and a JBL XO.

[destroyer X]

I am doing that for long years, and always learning new things and seeing clearly some foolishes i used to make.

There are some type of matching, some "conversation", some harmony or not harmony between the amplifier output and some speakers.

This makes me mad, as some amplifier can sound wonderfull with speaker "A" and sound very bad with speaker "B".....other amplifier can invert that result.

This way we elected Altec Lansing as our reference speakers... something must be fixed, if not you turn crazy.

Some JBL speaker use LM3886 inside, some amplified speaker, and sounds very nice....and i could heard LM3886 sounding bad with another speaker.

This "hearing evaluation" is so difficult, that many people goes to specs and try to forget those subjective evaluations, as have too much variable things to control....one of that is feelings....the listener mood, this is diminished in importance when you call a lot of people to evaluate together, and elect one speaker.

But it is not impossible that Altec lansing turns a bad load for someone.

Those tests must be repeated several times, and comparison with many speakers and many amplifiers, and i do not think can reach 100 percent precision, as humans cannot agree this way, as people is different in many aspects, including sound preferences.

My people, Brazilian people, use too much percursion, this way, if that part of the spectrum was not evaluated, we loose one of the main aspect of our musical culture, the rithm and percursion.

In the reality, talking really the deep true, this chip cannot be so bad, the ones that developed are not stupid...i have some suspections that with other speakers it can change some.

regards,

Carlos

[percy]

Carlos, you are right that there are many variables involved in an sonic analysis of an amp and opinions can vary widely.
BobEllis, thanks for starting this thread. If you hadn't I would have anyway

Now, for a person in a situation like me who is trying to decide whether to spend his limited budget (and plan possible future extension) on a Leach amp or a Gainclone it helps to get favourable or unfavourable feedback for the amp(s) in question, from the forum in any shape or form.
Atleast it helps to determine whether its worth even trying or not.
Just get the "feel" of it if you know what I mean.

For example, I am not totally discouraged from BobEllis's experience. Looks like GCs are sensitive to implementation. Had BobEllis implemented it using BrianGT's boards and method (ps, layout, zobel and all..) and still got the same results I would have been skeptical of the gc, but not just yet.

[Luke]

The LM3875 is by far the best chip I have tried out of about 6 of the most common flavors on this forum. Dont give up on chip amps till you try it

[PMA]

Originally posted by Luke
Dont give up on chip amps till you try it

... and then you'll easily find their limitations. Beginners luck (it is easy to built and works quite well), but far from the results achievable with very good discrete design - I am saying that after having compared.

[Luke]

... and then you'll easily find their limitations.

apart from a few chip amps, Ive built the jlh and aleph 2, I prefer them both but the chip amp is really quite good

[BobEllis]

If the LM4780 amp was the first I'd built I'd be happy with it. It sounds petter than whatever is in my DSP-A1. My complaint is that the low level detail heard with the Leach amp is missing.

Carlos - evertything except the amp was held constant. I put the chip amp on top of my Leach amp and swapped connections.

[percy]

..and then you'll easily find their limitations.

could you help me by mentioning what are they ?

but far from the results achievable with very good discrete design - I am saying that after having compared.

Does "very good discrete design" = $$$$ ? Which ones are you referring to ?


Two channels can be built with a GC (incl. psu) for ~$100 - maybe even less. While the most cost effective discrete designs I could find so far were the Leach and P3A, and those were coming to about twice that cost.

Keep in mind though that I am not challenging your thoughts or saying one is better than the other, just trying to find more information...