Quote:



This design is already capable of more current than the typical A47, ignoring variables like the supply voltage. If you meant doubling up the opamp (voltage gain stage), it wouldn't be a benefit for a headamp, AFAIK. You just raise the gain if you want, err, higher gain, or choose a different complimentary pair of transistors for more current handling. BC327/337 are just nice for a TO-92 package but something like 2N5087/2N5088 might be nice, or step up to TO-126 sized transitors, a couple common around here are BD139/BD140.

I would suggest just building it in a basic version first, which can be done in short order and relatively inexpensively. If you had some 2N3904/2N3906 or whatever lying around it would give you an idea and you can take it from there. Just note the pinout of the transistors you choose, those Sijosae used (BC327/337) are CBE pinout (COllector, Base, Emitter), while many are instead EBC, so they'd be rotated 180' from the picture.

I have been playing with this design for the last few days.

So far I have altered the design so it looks almost like a pimeta .
I replaced the output stage with sijosaes modular diamond buffer design.
The virtual ground is now almost the same as in a pimeta, only without the jfet cascode, using a opa2132 with the diamond buffer, and a 220k voltage divider as input.
I don't know if it works though .
I'll have to acquire some new platines for edging first and BC327/BC337s. I'm holding off mail order for now because I'm still working on the schematic and don't want to order twice.
If this works, it would be a very nice portable amp.

Hey guys, just wanted to find out, which one should sound better, Class-AB I or Class-AB II.

Also, I noticed there are no input caps, would the sound quality benefit from having these or not, I've got a few Orange Drops lying around.

Thanks

Never built a Class AB II. Interesting.

I left out input caps since I have no problem with offset. Based on tangent's article, it should sound better without caps.

My Class ABs:

Thanks Eokboy, i'll be getting parts for the Class-AB II today/tomorrow and building it as soon as I get them.

I've read tangent's article, I was just wandering how much the sound would get worse by having 0.1uF or 0.47uF input caps.

Plus, in that article it has the f=1/2(pi) RC formula, but this amp doesn't have an R. To add input caps would I need to add this one too?

I'll have to measure how high the offset is on my crappy soundcard at work is and hopefully woun't need to add input caps, but just in case, I'd love to know this.

Thanks

You could try with 100K input resistors and 0.47uF caps. I just followed the original schematic.

Quote:



If you add a coupling capacitor between the volume pot and the input to the opamp, you will also need to add a resistor from the input of the opamp to ground. Without it, there won't be any bias path for the opamp and it won't function properly. The value of the resistor should be at least 10x that of the volume pot. Without the coupling cap, the volume pot itself serves as the bias path, but I prefer to add a resistor anyway in case the pot wiper contact fails (which may lead to high DC offset).

Just built one of these, getting a big problem with DC offset. Exactly 2.6V on both channels. Took me a bit to find out that the discrete rail splitter only works when you close the loop between V+ and V- before you get a good even split. After that I added all the components and I think I hooked up everything correctly.

I measured DC offset between Right/Left Out to Ground without any source connected with an AD823 as the chip. Only finished putting this thing at 1:00am before getting called into work *GRRR!*. Will play with it later today to see if something changes, but with a config like this, I'm not sure of what there is to do except scan over all the soldering to make sure nothing that shouldn't be is connected.

Quote:



if you're adding this as ground channel:

then it'll work just fine. and look something like this once finished :


or look something like this if you're running it in Jung MultiLoop configuration :


but still even after all of that, it still falls short compared to these which uses a completely different discreet buffer implementation :
, improved layout :

also running in MultiLoop, with around 27-29mA of current drain and a much better sound. all of the circuits shown above uses 2 dual channel Opamp to with one dedicated channel to split the voltage and buffer it like in the PINT/Mini3. a good sounding portable amp indeed, especially for people that has a particular fetish with through-holes and donutboards.

thedoctor, do you have teh schmatic for the best layout in the bottom right ? I might try to build one, but probably will try to etch a board for it.

Quote:



i'm not exactly educated in Eagle, but here's a sketch. trying to learn how to use it to draw schemtics, but free time is running low.

Just wanted to ask, how important is it to match the BC337/BC327 transistors? I don't exactly know how this is done so I just took the first pair of each and installed them.

Quote:



if the pair is working in an open loop (hope i got that right) like in the sijosae discreet rail splitter, then unmatched pairs could cause voltage imbalance to happen. but if the pair is working in a closed/feedback loop of an Opamp, then the Opamp will keep the offset at near ground levels.

to match the transistors, the easiest way is to use a multimeter that could measure HFE.

Uhm... and in the Sijosae Class-AB II amp, are they working in open or closed loop?

Quote:



The L/R are closed loop, while the discrete rail splitter is open loop, although in the schematic posted above by thedoctor, which he feels is the best configuration, all is closed loop.