[thoppa]

Hi,

I just thought I'd start a thread to see if anyone else has built one of these and what they think of it.

I've seen amps like the Keces use these chips and that sells for something like US$450.

I got the pcb on Ebay for US$9 and got some sample chips from National to see how it sounded. I'm just using a simple LM317/337 +/-15V psu. It's dc coupled. I added a deep bass boost into the feedback loop, using a 1uF Soniccap to equalise HD650 (double-helix cable). All the electrolytics are SilmicII and they are bypassed with Vishay MKP1837. Resistors are Dale RN55. IIRC I set the gain to 4.

http://cgi.ebay.co.uk/Low-distortion...item4a9f775327

The PCB doesn't have a ground plane and the decoupling caps are quite a way from the 49710, but there is no oscillation (tested on a scope) or audible noise (depending on the source of course). I used a star earth and isolated the ground of the inputs from the case I used.

The initial dc offset was huge (2 volts or so) but instantly dropped very fast and then started dropping slowly from about 200mV. After getting fully warmed up and burned in, it's settled to about 8mV.

I'm amazed by the definition and clarity. The soundstage is great and the tonal balance is excellent, dare I say perfect. It has plenty of punch and the chips have a huge bandwidth. I'm serioulsy thinking about building a four-channel fully balanced amp powered by a sigma22. I reckon that would be a giant killer.

Anyone else tried this as an amp ?

cheers,

Tom

 

[ROBSCIX]

..interesting and a good little deal if you have the components which many of use do.
Are you asking for other peoples opinion of it for a comparison to yours or you don't own one and want to know what people think of it?

I might have to grab one or a few!

 

[FallenAngel]

So what did you do with the 1uF cap?

 

[Nisbeth]

Got the PCBs but haven't done anything with them, so thanks for the impressions.


/U.

 

[FallenAngel]

Damn, just chips alone are pushing $65 + shipping. Guess the whole amp is around $100 without PSU.

Wondering why they chose to use 2x LME49600, the integrated diamond buffer already can output 250mA of current.

The DC offset is a little worrisome.

Have you taken a look at TwistedPearAudio's Ventus? Uses the same buffer, but only a single one and has a discrete input stage instead of opamp.

 

[fierce_freak]

Something like this or the Ventus without any gain stage would be a good project. I've found most headphones don't need any gain (for my ears).

 

[thoppa]

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif Damn, just chips alone are pushing $65 + shipping. Guess the whole amp is around $100 without PSU. Wondering why they chose to use 2x LME49600, the integrated diamond buffer already can output 250mA of current. The DC offset is a little worrisome. Have you taken a look at TwistedPearAudio's Ventus? Uses the same buffer, but only a single one and has a discrete input stage instead of opamp.

Yes, the chips' prices are the downer. I don't know why this design chose a dual buffer but I imagine it is to keep the power/heat dissipation down for low impedance cans. I think for high impedance, it would be quite easy to modify the board and use one buffer.

I hadn't seen the Ventus - very nice ! I see it uses an output inductor which would seem to indicate that it could be prone to oscillation/instability.

This design keeps the op amp and the buffer in the same feedback loop so stability isn't a problem and the full bandwidth can be maintained. It is pretty much the same design as given on the LME49600TS data sheet just with a dual buffer.

I used the Soniccap in parallel with a 5.6K resistor and connected these in series with a 3K feedback resistor as the feedback loop. There is a 1k resistor connecting the feedback to earth. This gives about a 2.5 (8dB) bass boost at ~30Hz with a corner frequency of about ~60Hz,and an overall gain of 4 (12dB) for the amp.

Nisbeth - get to it ! It's a sweet amp imo.

Rob - just want to get it 'out there' really. It's a great amp, easy to build and even a wally like me hasn't managed to stuff it up. So I think anyone can build one of these and have a 'state of the art' solid state amp for less than US$200.

These chips are stunning performers - as good as it gets imo.

EDIT: hmmm, just started to wonder what it might be like with an ad797brz.....

 

[linuxworks]

does that seller have all his own designs? I wonder. something looks familiar in some of his offerings.

can't say I'm at all impressed with the look of this pcb. probably worth totally redoing it and adding a GP (at least?)

 

[CodeToad]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif does that seller have all his own designs? I wonder. something looks familiar in some of his offerings.

It's a free-for-all in the wild, wild East

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif can't say I'm at all impressed with the look of this pcb. probably worth totally redoing it and adding a GP (at least?)

Sounds like a new open source headphone amp project may have just been born. Excellent.

 

[chesterqw]

the LME49600 documents have a simple headphone amplifier evaluation board which seems very cool to me.

 

[thoppa]

Yeah - bring it on ! Integrate it with a y2 and make it a fully balanced dac and amp.

 

[glingrem]

Hi!
I think, that CFA in output stage of composite scheme will be better.
LT products 1206 and 1210 CFAs which are good for that.
Better than LME49600 for low resistance loads in first.
And no need in doubling.

 

[thoppa]

Hi,

A current feedback amp won't work in this design. Check the circuit in the datasheet.

Tom

 

[fierce_freak]

Quote:

Originally Posted by thoppa /img/forum/go_quote.gif The initial dc offset was huge (2 volts or so) but instantly dropped very fast and then started dropping slowly from about 200mV. After getting fully warmed up and burned in, it's settled to about 8mV.

I ordered a couple boards. I'm not sure if or when I'll get to it, but I have a quick question for you, regardless. Is the DC offset still huge on every startup?

 

[thoppa]

Hi,

It was in constant use for quite a while, and during that time, turning off and on made no difference.

However, if it hasn't been in use for a long time then the dc offset comes back, but drops away even faster. It'll drop to 100mV in less than a couple of minutes with no load.

I'll need to do some long-term testing to see if this settles down with time. It's too early to say but it is obviously changing.

The buffer chips come in a dry pack with instructions to bake them before soldering if the humidity paper is pink. So I wonder if this has anything to do with the chips needing to get warm before their dc offset drops away.

Anyway, for the time being, I keep my cans disconnected whenever I switch on unless she's already fully warmed up.