PINT Problem

[Filburt]

Okay, tried jumpering L1 on the LM6171, and it went to 5,55 and 4,95. Tried switching 75 ohm resistor back in and it's still stuck at that.

Argh, this is just so confusing and frustrating

 

[tangent]

You're measuring OG to OR and OG to OL, right? DC volts?

 

[Filburt]

No, I was instructed to do IG to OR/OL.

I built a new one with the parts you sent me, tangent. I get about 1mv on IG to OR, and about 0,9mv on IG to OL. For OG to OR and OL, I get ~7,0v for both.

Problem, though. When I turn on the amp, with headphones on it, I get this winding up sound (e.g. starts lowish, goes high - starts at maybe 800hz and then goes to 4khz within a second or so). This is with new jacks in place. With the old jacks in place, I was getting some of the signal, but it sounded like a bunch of it was missing, and it was pulled to the centre and sounded hollow, and when I turned up the volume, it would start to crackle and have static. I have C3 installed, and the ferrites. The ground channel amp gets hot, while the other not so much. It doesn't get extremely hot, though, but it does hurt my finger to hold it there more than momentarily.

 

[tangent]

Quote:

Originally Posted by Filburt No, I was instructed to do IG to OR/OL.

That's the same thing, if IG to OG is near 0V.

Quote:

I get about 1mv on IG to OR, and about 0,9mv on IG to OL. For OG to OR and OL, I get ~7,0v for both.

Then IG to OG must give about 7V, too. In which case, the ground channel is broken.

Quote:

I was getting some of the signal, but it sounded like a bunch of it was missing, and it was pulled to the centre and sounded hollow

The amp isn't grounded correctly. Likely you've got IG or OG swapped with something, or there's a conduction path where one oughtn't be. Is this happening with the amp outside of a case, with the jacks flopping around on the end of hookup wires?

 

[Filburt]

Yes, IG to OG gives 7v. Okay, so now my ground channel is broken? *sigh*...I wonder what broke it. Do you mean the chip is broken or the connections are messed up? I have been extremely meticulous and careful in building these things...

Does the description with the whirring sound help? Is the op-amp busted? I don't have any left

Is there a way to find where the fault is?

 

[Filburt]

Okay, I put a 75 ohm resistor on L1 for the ground channel, and it got rid of the squealing sound. It occured to me to do it after I removed the cable extension for my earphones and plugged them in without it, and noticed that no squealing happened.

However, I now once again have the messed up hollow/drawn to centre sound, which sounds like once again a problem with the ground. I don't really know what to do, though, at this point...

Also note, I'm using the isolated jacks from mouser (Kubiconn).

 

[PinkFloyd]

The squealing sound and heat you mention is oscillation. From the look of the underside of the PCB the soldering is pretty iffy, go over everything with a hot tip and let it flood into the joints a bit more to form a nice mirror like finish.... feed a little dab of solder from the reel onto each joint as you melt the existing solder. It's amazing how poor soldering is attributable to 99% of the faults in an amp, double check every last pad. Haven't read the entire thread but one of those resistors doesn't look too healthy to me (the one adjacent to C1R). There's enough solder on those opamps to sink a battleship, get some desolder braid and mop it up..... you're soldering a small opamp not welding a wing onto a car. Less is more when it comes to soldering and you can actually see if the joint is ok, big wads / blobs of solder can hide a dry joint.

 

[Filburt]

Well I rebuilt it, so those pictures are the old amplifier. I checked all the solder joints and re-did them, with no effect. I did the second amp even more carefully than the first. I realise something has to be wrong, though...

 

[PinkFloyd]

Quote:

Originally Posted by Filburt Well I rebuilt it, so those pictures are the old amplifier. I checked all the solder joints and re-did them, with no effect. I did the second amp even more carefully than the first. I realise something has to be wrong, though...

It'll be something very minor, been there many times, take a rest and then go back to it refreshed and don't look for the obvious.

 

[Filburt]

Alright, sounds like good advice. Maybe I'll try remounting everything.

 

[kramer5150]

Wow this is a GREAT thread!!! Filburt has a really good sounding cmoy, and IMHO is a competent builder... Certainly higher up the ladder than my meager skills.

I think I'll hold off on a pint build... just too much black magic going on here for my meager skills.

Good Luck!!
Garrett

 

[Filburt]

Alright, I went and tried to improve the soldering on a variety of things, to no avail.

I also tried installing 75 ohm resistors on the main audio amp's L1s to see if it'd have an effect...it didn't. If anything, it made it worse, so back in the ferrites go. I'll leave a resistor in the ground channel, though.

One thing tangent didn't make clear to me is whether IG to OR/OL is supposed to be 7v also, or if it's okay that both are about 1mv. Based on what I'm hearing, it sounds like the problem is in the grounding of the input, but I can't see anything wrong with it, unless either the op-amp or the traces are screwed up in some way. I really hope it's not the op-amp, as I have no spares. I only have one board left, too, so I'm really not wanting to have to build this up yet again, especially since I'm running out of other parts.

 

[Filburt]

Okay, here's an update and my advice to others trying to build this -

I figured out the problem with the sound at this point. Having a 75 ohm resistor on L1 for the ground channel causes it to basically just not work. I end up getting a weird hollow sound. However, it is _not_ stable, as far as I can tell, with the 50 ohm ferrite from DigiKey. It blew up on me when I tried to switch back to the ferrite. Wonderful.

The other amp appears to be stable at gain 4 with the ferrites and caps on C3.

I swapped in an LM6172 again, with ferrite, not 75 ohm resistor as, again, I end up with weird sound if I have it there. It works correctly and very stable. It sounds rather good, very similar to the Hornet, which means it's aggressive and detailed, but the soundstage is slightly collapsed and it isn't as warm as my CMoy or something like an SR-71. It's also very quiet in terms of noise floor. When I first turned it on, I checked my power to make sure it was working as I couldn't hear anything (no pop, no hiss, etc.). When I played music through, though, it became apparent that it was working

It is running very stable now, though, with all ferrites on L1, caps installed on C3, AD8397/LM6172 combo. Neither chip gets hot driving my earphones or my friend's ATH-EM7. I may try removing C3 to see what happens, if it's making much of a performance hit, but somehow I'm doubting it is.

I'm going to order another couple AD8397 from AD, as someone pointed out to me that it's cheaper there and I am running out of money, but I have enough parts for another amplifier. Based on what I know now, I have some pretty solid ideas on how to build this successfully the first time around.

After this great ordeal, and considering my limited college student budget, I am accepting donations

If anyone has any questions, let me know (or if you want to donate ) and I'll try my best to answer.

 

[amb]

Hi Filburt, as you discovered, you should not put a resistor in place of L3 (the ferrite on the ground channel output). While a resistor in place of L1 and L2 is more or less ok, it's not so with L3. This is because the ground channel is supposed to source/sink the return current from both channels, and must have very low output impedance in order to preserve stereo separation. Putting a resistor there kills the low output impedance. That's what gave you the "hollow sound".

The ground channel amp operates in unity gain, and has a direct feedback path from the output to the non-inverting input (sans any resistor). This is probably the one place where an AD8397 is going to be very unhappy if the headphone cable has significant capacitance. The L3 ferrite isolates the output of the chip, but the way it's wired up on the PINT (where the ferrite is wrapped within the feedback loop), it offers no isolation to the inverting input. See the PINT schematic.

On our Mini³-ized PINT, we did a small surgery to the PCB to put the ferrite outside of the feedback loop, giving isolation to both the output and the inverting input. This involves one small cut and one jumper to the PINT board. See pics below:

Original PINT:

Our modification:

In addition, we also put L1 and L2 outside of the feedback loop by populating R4 on the pads for C3 (leaving the original R4 pads unused), and then stacking C3 on top of R4. This is a little frustrating to do due to the small SMD part size, but it all worked out very well.

 

[tangent]

Quote:

Originally Posted by kramer5150 just too much black magic going on here for my meager skills.

In several ways, the PINT is the most difficult amp to build on my site. It's definitely not for everyone.

Quote:

Originally Posted by Filburt here's an update and my advice to others trying to build this

Thanks for the detailed report, Filburt. I'm glad you've worked out a stable configuration.

Quote:

Originally Posted by amb we did a small surgery to the PCB to put the ferrite outside of the feedback loop

It'd probably be simpler to just jumper L1G and use a ferrite bead around the OG wire at the board end.