chipamp performance....?

Status
Not open for further replies.
Even though the 50 watts or so that the Gainclone is capable of providing, even some relatively efficient speakers have revealed it's shortcomings.
Going to 80 watts or more seems in almost every case to have more ''oomph''.
There is no comparison to massive headroom and high power in amplification.......just as long as the circuit topology retains transparency and musicality.
 
This was not one single test, but an aggregate of tests and listening experiences over the last few years...

Candidate 1: LM3886TF chipamp with 32V rails, 40,000uF caps/channel with snubbers (80,000 uF total), Blackgate NX HQ for Ci and Holcos. Roderstein resistors in feedback loop. High quality, but a little way from being the best components money can buy. Wired up P2P, and I've only let the two decoupling caps exceed the length of the chip itself, it's a compact little guy.

Candidate 2: Roksan Kandy LII

Candidate 3: Usher R1.5

Candidate 4: Cambridge Audio 840A V2

Candidate 5: Vincent SP-331

You'll notice the chipamp is the weakest of all the amps if you just look at power output. We noticed the chip was able to keep up with the amps in terms of tonal balance and sound quality, though its sound signature was a little different. I think all the subjective audiophile terms people normally use to describe this sort of thing applied well enough, I just think an amp has a job to do and it needs to do that job without getting in the way. The chipamp delivered.

Even with the very inefficient Usher Tiny Dancers (sensitivity measured @85 dB/watt) there was no shortage of impact/oomph till the time our ears were bleeding.

It was much easier to push the chipamps to the edge when we connected difficult speakers, with Dynaudio speakers the chips would clip really easy as they probably can't keep up with the current demands of the 4ohm/86dB Dynes. A parallel chip sorted that out, though the sound was no longer in the league of the single amp, even when we used exactly the same components.

One thing the bigger amps do much better than the chips is control of the speakers at high volumes. I don't know how loud you listen, but I run a pair of usher Mini Dancers (again 85 dB/watt), and most of my listening is done at about 1/2 volume, equivalent to about 12 O'clock on a regular amp. I haven't encountered any issues, but in test conditions (read insane/illegal levels) we felt that the single chip was losing impact on bass transients. This is not apparent at more normal levels, though.

In my view, the chips are severely current limited, the 10 amp rating is a red herring for sure. I don't think it can cope with low impedance speakers or other difficult loads, and it needs a load that is both reasonably sensitive and has agreeable impedance. Once those are satisfied, the amp is capable of very good performance, as long as you have the right power supply.

In short, I feel no pressing need to buy a commercial product just yet, I'm plenty happy with what I have. At listening session in my house, people are surprised that such little money can deliver such nice sound.

The only advice I can give is to make a very good power supply. As free of ripple as possible, as oversized as possible. It's pretty much make or break with these amps. The difference from a 200VA to a 400VA PSU throws up a completely different animal. Also, don't be afraid of PS caps. I was insistent on using low capacitance for a long time, till I tried high-cap supplies with a few RC snubbers. There may be cynics, but knock it till you try it. I can never go back to low-cap again, I believe in as much caps as written in the first line of this post.

Also very sensitive to layout, a poor job at grounding leads to terrible headaches later. I prefer P2P construction inspired by Peter above, and it gives me freedom to plan in three dimensions, which helps a lot in keeping ground paths separated, yet close together
 
Touche, Andrew 🙂

I'm seeing peaks of about 15V across the terminals, average about 6-8 (tough to make out with the needle all moving) right now. We aren't at ear splitting volume, but it's a little uncomfortable. That about 1-2 watt average, I think.
 
sangram said:
I'm seeing .............across the terminals, average about 6-8 That about 1-2 watt average, I think.
6 to 8volts at the speaker terminals is ~6W into 8ohms.
That is about 88dB average at the seating position if you listen to two channels @ 2.5m (8feet).

6W leaves just 10dB of overhead from a 60W amplifier. No wonder it sounds loud. It will have started severely clipping transients and gone coarse and shouty (my words for sounds horrible).
You desperately need 95dB speakers if you really like your average levels ~88dB at the listening position.
 
I don't doubt your calculations but honestly, it isn't sounding coarse or shouty, actually. It is loud, very loud, but the sound is not changed at all from lower levels, just louder. And I'm not being defensive, if you were in my country I'd have invited you to an audition 🙂 with pleasure.

And my distance is actually about 12 feet, not 8. The room is 19 feet wide, the speakers are 4 feet from the front wall and my couch is 3 feet from the back wall, so 11-12 feet is the distance depending on the position of my head (give or take). So drop another 6dB (?) off the levels.

I'm not changing my speakers, they are pretty awesome and cost a pretty awesome amount of money. They are very insensitive, but a very easy load to driver, no wild swings in the impedance, and above 6.5ohms most of the way through.
 
How good are these chipamps?. Well I do not need to be an engineer , have the most expensive audio test equipments or work for Nasa, to know that these "Little" amps can kick some a$$.Do not get wrong here. I am not saying that these fellas are the best but, They are really good.I had so many good amps and I have tested them side by side using same speakers in all and the result is astonish.What amps did I test? Portal Pananache, Bryston B60, Acurus, Adcom Integrated (rare).I did not compare the power as you can see some are more powerful than the chipamp (LM3875) I compared just the sound quality (tonally)
all these amps are good and the chipamp was next to them fighting side by side!.at the end I sold all of the above amps and now my chipamp is in the living room playing beautifully.I use good parts in my amp like:Black gate, Mundorf, caps and premium parts like from Audiosector (Peter Daniel) by the way He is a Great guy.
Well this is my impression and is just my opinion and I respect other's opinions.Maybe there are people here saying all this is @*%$#@% and these amps are crap.Again build one and give a try and if you do not like it...buy a commercial one.
:smash:
 
sangram said:

Also very sensitive to layout, a poor job at grounding leads to terrible headaches later. I prefer P2P construction inspired by Peter above, and it gives me freedom to plan in three dimensions, which helps a lot in keeping ground paths separated, yet close together

You hit the nail squarely on the head. High power and high gain can cause Excedrin headache #71.

A pair of LM3886's will do a tad better than a single LM4780, but the latter benefits from smaller footprint.

The driver chips from National do an even better job with Darlington's or MOSFETs and you can run them "hot" if you want.
 
sangram said:
M3886TF chipamp with 32V rails, 40,000uF caps/channel with snubbers (80,000 uF total..

...In my view, the chips are severely current limited, the 10 amp rating is a red herring for sure. I don't think it can cope with low impedance speakers or other difficult loads, and it needs a load that is both reasonably sensitive and has agreeable impedance. Once those are satisfied, the amp is capable of very good performance, as long as you have the right power supply...

...don't be afraid of PS caps. I was insistent on using low capacitance for a long time, till I tried high-cap supplies with a few RC snubbers. ...

..Also very sensitive to layout, a poor job at grounding leads to terrible headaches later. I prefer P2P construction inspired by Peter above, and it gives me freedom to plan in three dimensions, which helps a lot in keeping ground paths separated, yet close together

Interesting reading, especially about the caps. Is there any potential downside to 'up-capping' - sonically, or technically?
 
Well I used to have small power supply (capacitance) for a long time but, once I started trying with bigger values like 1000uf, 4700uf, 10000uf per channel and the sound became stronger and tighter(bass).since then I use hefty power supply at least 10000uf or 20000uf per channel.some members like small power supply. it is all in taste.Right now my power supply is 10000uf per channel using Mundorf caps and Black gates 1000uf next to the chip.I do not have any intentions to change the values. happy as a clam!
 
Status
Not open for further replies.