[TzeYang]

I encountered a very very weird problem with the dc offset measurement with my 3 channel diamond buffers, with AD744 as ground channel opamp. No gain stage.

Based off the diamond buffer design below:

Problem:

Situation One:
-Measuring directly from the bare output wires soldered on the board. however, the bare wires are not soldered to the output female jack yet.
-Black Test pen on output Ground, Red Test pen on output Left or output Right.
-Active Ground measured from input ground to output ground.

Results:I get decent Offset values, around +8mV per channel, 0.001mv on the active ground channel.

Situation Two:
-Measuring using a male to male interconnect connected to the output jack.
-Output jack is already soldered to the bare output wires.
-Black Test pen on output Ground, Red Test pen on output Left or output Right, all tested via the male interconnect output.

Results: DC Offset values are insane. +100mV on each channel.

Situation Three:
-Measuring using a male to Left and Right RCA interconnect connected to the output jack.
-Output jack is already soldered to the bare output wires.
-Black Test pen on output Ground, Red Test pen on output Left or output Right, all tested via the RCA Male interconnect output.

Results: DC Offset Values are normal.

Situation Four:
-Again, similar to situation one, measurements are taken from the bare wires not soldered to any jacks.
-DC Offset is measured from:
~Input Active Ground to Output Active Ground
~Input Left to Output Left
~Input Right to Output Right.

Results: DC Offset values are normal, around +8mV on L and R channels, active grounds remain around 0.001mV.


Any ideas on what happened? >.< this is insane!!

I took a risk and plugged in my KSC 75 and it sounds wonderful. No distortions, no background noise at all! What really happened?

 

[FallenAngel]

Simple, is that interconnect DIY and does it ground itself creating a ground loop and sending your amp into oscillation?

 

[TzeYang]

not DIY , it came together with gary's PA2V2. Radioshack stuffs i think.

EDIT: No opamps on Left and Right channels. No feedback. Basically something like this:

EDIT2: Edited the silly schematic i uploaded to prevent confusion.

 

[NelsonVandal]

I see that you have the same problem as I have. I just built 3 channels of Sijosaes buffers and I get 100 - 200 mV DC offset using TLE2426 rail splitter. Unlike you I dont use any opamps at all, not even in the ground channel. With "floating ground" there's an acceptable DC-offset, about 15 mV (look at my post a bit further down in this forum). There's no difference in sound with virtual ground or "floating ground". The amp sounds marvelous, but this isn't how it's supposed to be connected. I think these buffers aren't made as stand alone buffers since you can't adjust the bias. When using them in a feedback loop, there's no offset problem. But I think both you and I dislike the sound of opamps, and just want the buffers.

I think I'll try a more sofisticated buffer design next time. Maybe the LISA III buffers. Do you have any other suggestion? When you have heard the sound of a no opamp design it's impossible to go back to the colorations heard from the different opamps. I have not heard 744 yet, maybe it's more transparent than the rest. I'm planning on using them in a PPAS.

 

[TzeYang]

okay guys, i finally solved the problem.

Turns out the opamp was not stable unless it's configured at a gain of two. So i used the compensation as output.

Nice opamp, it works perfectly for my buffers.

Offset is low as well. Using a star ground scheme LOL.

 

[NelsonVandal]

TzeYang, could you please post a schematic? Especially how you have the grounds connected.

 

[TzeYang]

nelson, here you go. enjoy

. Remember to match those transistors!!

 

[NelsonVandal]

Thanks for your help. I did match the transistors for hfe from the beginning. Still 70 mV offset. I "ajusted" the 3k resistors a bit and got the offset down to 10 - 15 mV which is acceptable. This is probably not the right way to do it though. At first just two channels. Like before very good sound. Adding the ground channel introduced that karaoke thing again, but I didn't use an opamp, just the buffer. Adding an opamp for multilooping fixed that but the sound was somewhat grainy (I used one channel of an AD823, and this has never sounded good in any cofiguration). Now I'm using a LM6171 instead and the sound is better. This is a very very good sounding amp especially with my modded Philips SHP895. I'm going to try it some more with my Senn's now.

 

[TzeYang]

btw NelsonVandal,

How did you measure the Offset? And when you said multilooping did you mean adding a negative feedback and closing the loop with an opamp?

Frankly, i felt kinda happy to again meet another person who doesnt want a gain stage.(it's kinda rare you know) Buffers and Unity gain for the win!!!

 

[NelsonVandal]

Quote:

Originally Posted by TzeYang /img/forum/go_quote.gif How did you measure the Offset?

From output ground to output left/right
Quote:

And when you said multilooping did you mean adding a negative feedback and closing the loop with an opamp?

Yes
Quote:

Frankly, i felt kinda happy to again meet another person who doesnt want a gain stage.(it's kinda rare you know) Buffers and Unity gain for the win!!!

The same to you. I can't understand that people want to use amps with gain stages. I have 1V output from my portable and 2V from my CD player. Even though I use my 300 Ohm HD650's most of the time, I don't need more output than unity gain.

I rebuilt this little amp in a cheap plastic box almost exactly the size of my iRiver IHP120. I'll post some pictures later. I think it sounds very good, but I'm going to give it some time. The more you listen, the more you learn about colorations. It sounds brighter than I'm used to. Probably more neutral because the Xtra X1's buffers are a bit dark and muddy.

My next project is PPAS with LMH6321 as buffers. I'm going to try it without L/R opamps at first. From what I've read, and heard from AD746 in my stationary Sijosae amp, I think AD744 is the only opamp I could live with.

 

[NelsonVandal]

Here are the pics. Everything but the box and battery is surplus. Not the prettiest.

 

[NelsonVandal]

After some listening I started to hear that also this amp colors the sound. Voices and instruments that are supposed to be centered were vague and to wide. There was a bit of lack of slam, a bit of softness. This is how my CMOY clone with LM6171's sounded like. I tried it with LMH6654, and there was more slam and better soundstage, but the offset went up to 50 - 60 mV. Now I have OPA2132 (just using one channel) inside. Voices are where they're supposed to be, and the offset is lower than with the other chips.

I never thought the ground channel amp would affect the sound signature like this. Maybe I should go back to two channels, or get a grip on myself and order some of AD744's.

 

[TzeYang]

NelsonVandal, do you have an MSN messenger contact or something? it would be a pleasure to talk to you about amps.



This is the 3 channel buffer i built.


I'm actually having plans of making a 627AP ground channel opamp based 3 channel diamond buffers. Havent decided on the power section yet. However this too requires alot of planning, especially the part where stabilizing the opamp is concerned.

 

[NelsonVandal]

NelsonVandal, do you have an MSN messenger contact or something? it would be a pleasure to talk to you about amps.QUOTE]

OK, but I'm not very clever when it comes to electronics, as you can see in my posts, but I know what I like and what I don't like in sound. My quest is to find the most musical, non-colored, likeable sound possible without spending too much money. Reading this kind of forums is a great help to me. I'm not used to MSN messenger but I have it installed in my computer. Send me a PM and I'll try to contact you.

 

[NelsonVandal]

After further listening I'm starting to notice a lack of detail and aggressivenes. This is not placebo because I've A-B:d lots of times between the Senn's to line out from iRiver IHP120 and via the amp.

I wired it back to just left and right channels with passive ground. Now there's more bite to it, more focus, and it's harder to tell the difference between amp and line out. Maybe, just maybe, I lost some "air". More listening time will tell. I'm going to try 3 channels again when I get some AD744's. What I've gained except sonic improvement is twice the battery time. It consumes about 9 mA now. With those two 3300 uF capacitors inside it plays for quite a while when disconnected from the battery.

Are 3 channel designs relly better than 2? I'm going to start a thread to find out.