[stereth]

I finally got my half-naked META42 built and it's making me nervous. Listening to it from my PCDP, it sounds great, full spectrum. From my Sonica, low treble on up is rolled off and vocals are coming through a washing machine. It sounds horrible. Bass is fine.

From my PCDP, it sounds awesome into my V6s. Voltage at the TLE is split +9/-9. Voltage from ground to rails (after first BUF) is +4/-13. But it still sounds good. There's a fairly loud pop (delayed right to left) when I turn the amp off. Resistance between rails and ground is negative (caps still discharging?) then increases to 10k ohms within 30 seconds or so.

Details: I used the values in the original schematic PDF on Tangent's page. OPA2132 amp, single BUF634's on power and each output. R9 populated. R7=100 ohms.

This is too much for my brain to handle at this time of night; maybe tomorrow it will be perfectly obvious. Any suggestions would help me out a lot. Thanks.

 

[guzzler]

so you've not actually built a META42, rather a clone right??

First thing is to sort out the rail imbalance for the BUF's... did you put a 30-100uF capacitor across the rails (not to ground, straight across) before the BUF as recommended in the datasheet??

As for the different sources, that sounds like a problem with an impedance mismatch, perhaps your Sonica is having a problem driving it...

g

 

[jeffreyj]

Quote:

Originally posted by stereth I finally got my half-naked META42 built and it's making me nervous. Listening to it from my PCDP, it sounds great, full spectrum. From my Sonica, low treble on up is rolled off and vocals are coming through a washing machine. It sounds horrible. Bass is fine. ....

This is probably unlikely, but check anyway... Does the Sonica have both line and headphone outputs? Which one are you using? A headphone output might require a minimum external load to keep the output side of any DC blocking caps at 0v nominal. I would expect the volume control of the META42 to present more than enough load for that task, but check to see if there's any DC present anyway. The fact that it works fine with your pcdp but not with the Sonica makes it rather obvious that there is some sort of problem, or problematic interaction, with the Sonica.

 

[stereth]

Guzzler, what qualifies a clone? I used the exact schematic except for using BUF634s instead of EL2001s. I called it "half-naked" because the enclosure is more like a carrying case; everything is mounted to the board. I used protoboard instead of Tangent's PCB.

I looked at the data sheet for the BUF (which I should have done before) and I think I see what you're talking about. For a "pseudo-ground driver," they have a 10uF cap across the rails. I might have to add that. Also, they mention putting the BUF inside a feedback loop, and they recommend 0.1uF or 10uF tantalum decoupling caps. Do you think those are necessary?

Jeff, the Sonica only has a line out, which I've been using to power headphones for a while. It distorts at high volumes but sounds good otherwise. I would check the DC offset but I only have an analog multimeter, which has a voltage resolution down to about 0.1V. It doesn't have a problem feeding my cmoy with a 10K volume pot. My META has a 20K volume pot. Oh, and it works perfectly with both headphone out and line out of my PCDP.

Overall, do you think it's dangerous to my sources or headphones or the META itself to run the amp with these problems? That's what was making me nervous.

 

[ofb]

stereth, a meta42 is a pcb from tangent, not an amplifier topology. nobody is trying to give you the gears here, it's just helpful to establish that to be sure what you're describing.

unfortunately, 'meta42' is a common misnomer for the circuit the pcb has made popular, so there's some confusion out there and no widely accepted accurate name. but saying 'meta clone' is good enough for people to know what you mean.

i rather like the name PPL's 2001 Space Oddity, but i'm like that.

meta42 history

 

[stereth]

ofb, thanks. I stand corrected. I just did a search and found out what I was saying wrong. Not trying to detract from the team at all...you guys do wonderful stuff.

So, when building a Jung multiloop headphone amplifier with a design based on the A42, what do I have to modify to use the BUF634 instead of the EL2001? 10uF cap across the rails? 100uF? Electrolytic or tantalum? Bypass caps on the output buffers? These seem like last touches, stabilization upgrades, not something that could shift the virtual ground halfway to a rail.

And Guzzler, what do you mean "before"?

The Sonica sounds better through the particular amp in question today. It might have just needed a reboot. This computer does strange things sometimes when I leave it on for a few days in a row...

We need a name for these amps. Should I call it an A42? proto42?

 

[ofb]

Quote:

Originally posted by stereth what do I have to modify to use the BUF634 instead of the EL2001?

this has come up a few times but it's broken-link hell out there because both sijosae and headwize have changed servers. here's one link. i'll leave your troubleshooting to someone more knowledgeable.

sijosae shows us how (damn, he's good)

note: if you hit a broken headwize link, just change the "headwize" in the url to "headwize2". i'm afraid it's more complicated for sijosae's move.

 

[tangent]

Quote:

Not trying to detract from the team at all...you guys do wonderful stuff.

Thanks, but at the same time, we don't want the term "META42" to come to mean any Jung multiloop amp. It's just a PCB, as ofb said.

Quote:

what do I have to modify to use the BUF634 instead of the EL2001?

I hope it's obvious that the pinouts of the two chips are radically different. If you've connected all of the pins correctly, it should work. The BUF634 isn't an especially picky chip.

Quote:

Guzzler, what do you mean "before"?

Between the TLE2426 and the BUF634. He's referring to the virtual ground application example in the datasheet. But I wouldn't look to that first for a solution to this problem.

Quote:

We need a name for these amps. Should I call it an A42? proto42?

I like "Jung multiloop", personally, since that's the base topology in use. Look at the schematic in the Jung article (available through the META42 docs...history page, I think) and compare to the META42 and your amp. There will be differences, but these are tweaks, not topology changes.

 

[stereth]

Thanks Tangent. I found the link but it points to an unrelated article. I couldn't find it on Google either. (typing "jung multiloop" into google pulls up links to your site and sites mentioning Headwize or Head-Fi.)

Actually, I laid out the circuit with the BUF634 in mind, so that's not my problem. Aside from the pin layout, the BUF should be pretty much a drop-in replacement for the EL2001?

It sounds good. Sounds excellent, in fact. It really brings out the bass; it gives a nice kick to the deep bass. This is the first time I've been able to stand listening to something connected to my PCDP's line out. I can't even describe what it does to the rest of the spectrum. The music is just so much more raw. I don't have to look for details in the music; they're right there.

So, the question now: if it sounds good, should I worry about the ground being off-center? Should I add more stabilizing caps or resistors, or is it safe enough to ignore?

 

[tangent]

Quote:

I found the link but it points to an unrelated article.

Lovely -- they've recently redesigned the web site and have apparently trimmed the archives. I've put in a complaint, but probably having others complain about this will help.

Quote:

(typing "jung multiloop" into google pulls up links to your site and sites mentioning Headwize or Head-Fi.)

That's because the term is exclusive to the headphone forums. No one else calls the circuit that. Jung himself never gave it a formal name.

Quote:

hould I worry about the ground being off-center?

Yes. The buffer will have a bit of inherent DC offset, but only on the order of tens of millivolts at most. Something very wrong is going on here.

 

[guzzler]

hi stereth, not been on for a while, sorry...

i was really meaning a bypass cap across the supply rails near the BUF, just i said before because i was thinking schematic mode and i usually work left to right, so to me a bypass cap is to the left of the chip in the schematic... sorry!

g

edit: just fixing my contradiction, guzzler==donkey boy!

 

[stereth]

I wonder if I somehow fried the BUF. I took it out of the socket, jumpered from input to output, and it works fine--precision ground, 8V each way. I know the TLE can only put out 20mA, but the amp draws a pretty steady 12mA at a good volume from a line-out, so I should be safe there as long as I don't listen too loud.

Thanks Tangent for your help here and the guides and articles on your website.

Guzzler, not a problem. Sometimes the physical location of a bypass cap matters.

Some low-res pictures here. The box is from a wallet I received as a gift. I had to pull out the lining to make room for the Orange Drops. Power switch in the back left. Volume is the blue knob. Signal goes up the left side, across to the amp, down to the buffers and out the right jack. Power section is down the middle (ground is wired vertically across the entire middle section). I might eventually mount it on a thin piece of wood with battery holders on screws underneath...then I'll just have to search for a good dust-proof carrying case. Still won't be a good portable amp; just a good transportable amp.

 

[MERTON]

how hard is it to build a meta? basically i guess just how many hours did it take you to do it?

 

[stereth]

Merton, it doesn't take long at all if you have the right parts. I spent several hours laying mine out, but if you buy the PCB you won't have to worry about that. And although I used more board-mount components, I didn't save any room compared to the META42 board. Maybe a few hours to build, check, and test, then however much work you're willing to put into the case. It's a great investment of money and time compared to what you get out of it.