[lostspyder]

Ive been studying the MINT and PINT schematics. Why did they get rid of the "railsplitter" and the buffers? Is it so they can save board space by eliminating the need for two big power caps and replace them with a single one? Other than that it seems like it requires more board space than a "railsplitter" - the "OpAmped ground" needs two smaller caps and resistors. I thought that the big revelation of the PIMETA was the addition of the railsplitter and buffers - now they get rid of them?

From my meager breadboarding of amps in the past few days, the presence of buffers has made more of a diffrence than any other single thing I've done.

The real reason I'm asking is because I'm working on a 'tiny pimeta'. A railsplitter, two AD744s and two HA3-5002's on the output without an opamp in the ground (basically a MINT).

 

[pinkfloyd4ever]

try to get your hands on a PPAS board..that's a surface mount PPA. The boards were made awhile ago but you should be able to find an unpopulated one around here of all places that someone's willing to sell

http://www.head-fi.org/forums/showthread.php?t=186961
http://apuresound.com/ATAT/PPAS/PPAS.htm

 

[lostspyder]

That thing looks awsome, but I don't have the money to dump into is. All my parts are DIP and I was planning on building it on perfboard. I will defiantly borrow some of the design tho.


Quote:

Originally Posted by pinkfloyd4ever /img/forum/go_quote.gif try to get your hands on a PPAS board..that's a surface mount PPA. The boards were made awhile ago but you should be able to find an unpopulated one around here of all places that someone's willing to sell http://www.head-fi.org/forums/showthread.php?t=186961 http://apuresound.com/ATAT/PPAS/PPAS.htm

 

[TzeYang]

the pint itself uses the ZOMG ARGHGHGHGHGHGH AD8397, which outputs plentiful of current by itself without relying on external buffers. Oh refer to Mini Cube for more info.

That is why a BUF634 isnt neccessary for this application.

 

[NelsonVandal]

If HA3-5002 are OK, you're about to bild a great sounding amp. Use AD744 comp pin out. AD744's sound better at 18 than 9 V. If you want take it a step further, make room for two DIP sockets so you can add an active ground channel when you feel like it. If you want to push it further towards "state-of-the-art" make room for isolating JFET's or any other CCS to the 744's like in PPA, PPAS or LISAIII.

PINT sounds good, but it isn't the answer to all prayers. AD8397 has a unique sound signature. AD744 is more neutral and detailed.

 

[lostspyder]

Mabie I will have to hunt down two AD8397s

Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif If HA3-5002 are OK, you're about to bild a great sounding amp. Use AD744 comp pin out. AD744's sound better at 18 than 9 V. If you want take it a step further, make room for two DIP sockets so you can add an active ground channel when you feel like it. If you want to push it further towards "state-of-the-art" make room for isolating JFET's or any other CCS to the 744's like in PPA, PPAS or LISAIII. PINT sounds good, but it isn't the answer to all prayers. AD8397 has a unique sound signature. AD744 is more neutral and detailed.

I'm going to breadboard and possibly construct a PIMETA with separate railsplitters for each channel with HA3-5002s and AD744s (Is it still a PIMETA after all that? ). I was building this mini amp out of leftovers

The LISAIII is rediciliously complex looking

 

[ericj]

Quote:

Originally Posted by lostspyder /img/forum/go_quote.gif Mabie I will have to hunt down two AD8397s

The downside of the AD8397 is it's propensity for oscillating itself into a small fire. And the fact that it's very easy to kill with a slipped probe.

It's very, very sensitive to layout issues.

 

[lostspyder]

Mabie I wont mess with AD8397 then I have a tendency to make mistakes



I'm reading something about bending pin 5 into pin 6 and etc on the AD744. I don't really understand it because it's using an adapter and stuff.

But I gather that pin 5 of the AD744 has the opamp running in class A and therefore does not require Q1 and Q2 in the Pimeta design.

IE: I should leave Q1, Q2 out and use pin 5 as the output instead of pin 6?

 

[tangent]

Quote:

Originally Posted by lostspyder /img/forum/go_quote.gif Why did they get rid of the "railsplitter" and the buffers?

1. It's "he", not "they". I'm just zis guy, you know?

2. There is a rail splitter in the PINT. It just happens to be made from two resistors and a buffering op-amp, instead of being a monolithic device. As to why I did it that way, a dual op-amp takes the same PCB real estate as a single, so the second op-amp in the ground channel was "free". The two additional resistors to create the vground divider are smaller than any TLE2426 package.

3. Buffers are popular because most op-amps can't put out anything like the amount of current that a buffer can. But when the design goal is to use AD8397s, that reason goes away.

Quote:

Is it so they can save board space by eliminating the need for two big power caps and replace them with a single one?

That's an entirely different question. You can use just one cap with a bare TLE2426 if you want. Not advisable, but you could.

Quote:

I thought that the big revelation of the PIMETA was the addition of the railsplitter and buffers - now they get rid of them?

Do you see now that the PINT and PIMETA really are fundamentally the same topology?

Quote:

From my meager breadboarding of amps in the past few days, the presence of buffers has made more of a diffrence than any other single thing I've done.

Yes. But using high-current op-amps would do the same thing.

Quote:

Mabie I wont mess with AD8397 then

I definitely wouldn't recommend it for your breadboard project. This is easily the crankiest chip I've used. Ever.

Quote:

I gather that pin 5 of the AD744 has the opamp running in class A

A typical op-amp has two or three separate gain stages. It's very very common for the first one or two to be running in class A, because the next stage only draws tiny currents. Only the last stage has to supply the output current, which is why they never run in class A, which is why we talk about op-amp biasing in the first place.

The AD744 is one of a very few chips out there that gives external access to the outputs from these intermediate gain stages. Pin 5 is simply the point right before the final gain stage, so it can't put out any current at all to speak of. What it lets you do is replace the 744's own output stage with your own -- the buffer.

 

[majkel]

Could anybody say what is the difference between sound signature of the AD744 buffered from pin 5 and 6?

 

[NelsonVandal]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif Could anybody say what is the difference between sound signature of the AD744 buffered from pin 5 and 6?

I think the sound signature is about the same, but with greater clarity/details. I have more experience with the dual AD746 than with AD744 normal output, but they should be the same. I always found AD746 a bit dull, and I find AD744 comp pin out more exciting/livelier.

 

[lostspyder]

Ok, I put my MINT together (only one channel so far) and there seems to be a negative gain. I bypassed my POT (because volume clearly is'nt a problem) and still.I built it to the stock specs, except I'm using 744 and HA3-5002's without D2 and R11 (those shouldn't be the problem, or could they?).

I have to crank my source (knockoff shuffle) to hear anything.

edit: I suppose it's probably more of a factional gain (ie 1/2) than it is negative hu?

 

[majkel]

NelsonVandal, it's not the reason you took the signal from the comp pin. I guess its 746 not being quite the same chip. It's like 2*OPA134 is better than OPA2134 though the same overall sound signature. AD744 from the output is not dull for me. There is everything OK for me, especially the reality and precision of placing the instruments, but overall this chip is on the cold side which I don't like. I will try someday the comp pin signal as it will be buffered so no high current draw will occur.