Power Cap upgrade(?) in the Corda HA-1
Jun 11, 2002 at 5:34 AM Thread Starter Post #1 of 7

Nezer

Antibacterial soap... kills bacteria... bad karma?
Joined
Dec 6, 2001
Posts
1,273
Likes
13
This weekend, after listeing to the Toshiba SD-2109 as a source I was reading another thread here and git the bright idea to replace the caps in the power supply of my Corda.

They have now burned in for over 24 hours and I'm happy to report... WOW! Maybe it's psycholoigical but the bass seems faster and the whole thing just seems more smooth.

What I did was remove the 2 2200uf caps and the 2 1000uf caps that came stock with the HA-1 and replaced all four of them with some Cerafine 1000uf 35V caps. Although techically it's a downgrade ( the 2200uf provide a better resivoir) sonically it seems to really have improved things. Oh, and now the unit doesnt stay powered on for a solid 20 seconds after the power is killed, now it's more like 3-5 seconds (not that it matters).

The only problem was making the Cerafines fit. I had to get a bit creative with how they were placed as the older caps were a LOT smaller (even the 2200uf). The front two are bent to lay on thier sides and the back two are bent back to make room for the case top.

I was never really sure that replacing the power supply caps would make that great of a difference but it really seems to!

I'm now toying with the idea of replacing the opamps (and subsequent changes to the circuit to match the current load of the newer ones to ensure proper class-a operation) to OPA627s. I might also go through and replace all the resistors with Vishay CMFs while I'm at it.

I'm also considering adding a buffer between the power rails after the voltage regulators to ensure the opamps are being fed *exactly* the same voltage on the + and -. While the voltage regulators are good they are speced to go a quarter of a volt either direction. The buffer would help balance any differences and ensure the voltage rails are always equidistant from the virtual ground. This may help the opamp perform a bit better but I doubt it would be that audible if at all.

Any thougsh or suggestions? Now that i've voided the warranty any other upgrade thoughts?

How hard would it be to replace the second stage with a buffer?

I'm afraid this simple idea has opened a whole can of worms and I'll never be able to leave well enough alone...
 
Jun 11, 2002 at 3:47 PM Post #2 of 7
You can drop in a BUF634 for the LM6171, the pinout is identical. I think you might need to disconnect the feedback setting resistor pair that connects to the inverted input of the opamp, but maybe not. You'd be running the buffer open loop, however, so you would likely have a large amount of DC offset. I have an amp configured like that and it sounds good enough, but I much prefer the sound of the LM6171 to the BUF634. I'm not sure about the pinout of other buffers. I'm pretty sure the Elantec ones have a different pinout. I think most people prefer their buffers inside a feedback loop, which is not feasible inside the Corda unless you are willing to give up the crosfeed. Why not just build a META42 with a good transformer based power supply? It looks like that is what you are trying to turn your Corda into anyway.
 
Jun 11, 2002 at 5:06 PM Post #3 of 7
Ahh, yes, the Meta42...

Believe it or not I am actually considering building one with an external power supply popped into the Meta board at the C5+ and - (and 0V too of course).

I was going to model the power supply around Jan's but give it a bit more Umph, a beefier trafo, a few EL2001s in parallel after the regulators to stabalize the voltage rails, the biggest 35V Cerafines I can find/afford.

I'm pretty serious though about swapping the opamps in it to OPA627s. It seems like going the buffer route on the second stage might be more trouble than it's worth.

I don't believe (though I'd have to check to be sure) that the two stages share a feedback loop at all so I could, in theory, mix and match op-amps between the stages (627s on the first stage, 6171s on the second or vice versa). Would doing so have any sonic benefit at all?
 
Jun 11, 2002 at 10:20 PM Post #4 of 7
I swapped the corda psu with an external lt1085/33 based psu with shottky's, beefier transformer and caps. An major improvement.

One of the best thing about the corda is it's breath-taking speed. I wonder if swapping out the lm6171 with opa627 would affect the speed of the corda? (slower). Though I was toying with idea of trying ad8610's. Will see.
 
Jun 11, 2002 at 11:20 PM Post #5 of 7
You're right there is no global feeback because then the crossfeed would not work.

KurtW replaced the input stage 6171s with OPA627s and said it improved the sound. I think he left the output stage alone. The great thing about the 6171s as the output stage is that they can provide 100mA of current, whereas most all of the other opamps you'll see used in the DIY forum provide far less current and so are not as good at driving headphones directly. Hence the big excitement about the META42 which has the buffers for the output stage, and a multi-loop topology to boot.
 
Jun 12, 2002 at 2:32 AM Post #6 of 7
Anyone know the slew rate of the Cerafines? That upgrade was amazing!

I wonder if I could just pop in 627s in the first stage without changing anything else...
 
Jun 12, 2002 at 5:00 AM Post #7 of 7
Quote:

I wonder if I could just pop in 627s in the first stage without changing anything else...


Almost. You have to lift the compensation pins of the 627 since these are no connects on the 6171 but they are connected on the pcb and go to other places. Also, the 627 runs pretty warm, much warmer than the 6171. You can solve this problem by backing off on the Class A bias current, which is way higher than it needs to be.

The 627 isn't particuarly good as a output driver, but I do like the BUF634 in the output stage. You have to lift some pins again, and if you want to use it in the wide band mode there is another connection to make. But even without the wideband mode I thought it was an improvement over the 6171 (when using the 627 in the first stage...I didn't try it with the 6171 still in the first stage). It also doesn't cost as much as the pricy 627. I measured the DC offset in my unit and it was pretty low, nothing to worry about, although this can vary from chip to chip.
 

Users who are viewing this thread

Back
Top