LA-2A DC issue??

Just finished my LA-2A clone, and have a problem I can't seem to find discussion on.  I am experiencing significant DC offset (or rather extreme 2nd order distortion?).  Here's what my humming looks like:



As the signal gets louder it seems to cross a rather dramatic threshold into obvious clipping -- very easy to see that threshold in the picture.

I've been over the build several times now, (used Kent's layout, but with values of original drawings instead of his Jensen-type I/O mods.... used original UTC iron in my build, so stuck with the values on the original drawing).  Everything seems to check out.  I've swapped tubes with entirely different set, but still have same result.

Oh, and the result is the same with the T4B opto removed from the unit or with it in the unit, but the 'reduction' knob full CCW (no limiting).  When it's compressing this distortion is quite reduced, and with lots of reduction can be made to not cross into the ugly clipping at all, though it still sounds a bit odd... I think half the waveform is still a bit trippy.

Any ideas what might be causing this?  I'd be tremendously grateful for help.  It's unusable, and I'm currently at a loss.  Will hit the sack, and perhaps a fresh brain might make sense of it tomorrow....

Thanks so much!!!

rascalseven said:

When it's compressing this distortion is quite reduced, and with lots of reduction can be made to not cross into the ugly clipping at all, though it still sounds a bit odd... I think half the waveform is still a bit trippy.

Click to expand...

If the distortion all but disappears under extreme compression then it would point to a fault in the output stage?

Have you checked for DC at the primary of the output transformer?

I replaced the 10u/450v electro at the output transformer's primary, but it didn''t changed anything.

Still looking for some wiring mistake, but haven't found anything after several times of searching.

when I scope the output with my oscilloscope, I don't see the DC weirdness (bottom of the waveform clipping), but it clips, evenly, with only half the throw of the gain pot, which is much less output than I get with other things.  This clipping appears symmetrical, though.  The black lead of my scope is connected straight to ground, so when I connect it to pin 3 of the LA-2A output and the red lead to pin 2, the output transformer's secondary is essentially grounded (running the scope off the same circuit as the LA-2A on my bench).  Would this contribute to the symmetrical clipping? 

I haven't run up to my studio yet to have this play in my balanced system, but I'm wondering if the connection is the problem, or a major factor.  Anyone know??

Thanks,

JC

JC,

I have attached my schematic with the correct voltages.  Don't worry about the GR side which is custom, the output stage is stock.

From the description you gave, it sounds like the sudden onset of oscillation overloading the stage.  The DC feedback path must be DC, not come from the ac side of the output cap.  The DC stabilizes the output WCF and boosts V1 cathode's voltage up to reasonable working levels, otherwise it would be too low, a nice touch.

Check around the circuit for high frequency oscillation, might be ~100kHz.  If it is you might need grid stoppers on the 12BH7 or another tube.  Check for mis-wires when you are fresh, a good nights sleep is brain processing time!
best

DaveP

Thank you for your help, Dave! 

My feedback is from the DC side of the output electro, so I'm good there. 

FWIW my B+ is 250v, not 265 or 275 as I've seen on other drawings.  I'm using Allied's 6K88VG power transformer with the traditional power supply design.  After the 4k7 resistor (R29) I'm getting 250v.  Voltages read in the ballpark of ones I've seen published throughout the unit, so I'm not too worried about it.... doesn't seem anything major wrong, but welcome your input.  Perhaps I should change R29 value?

Also, I see a couple of 1k resistors on your output stage drawing that aren't on the LA-2A schem -- at pins 2 and 7.  Are these your grid stoppers?  Values seem rather low for that, from what I'm reading.  If I can't find the problem elsewhere do you recommend I add these resistors?

Lastly, the output trafo I'm using is old as dirt (an ancient UTC HA-113), and I'm not certain about how reliable it is, though it measures no open windings.  Will swap with a new Edcor 15k:600 for testing, until I get the kinks worked out.  One less variable to be concerned about.

I'm so grateful for your help.  Thank you!

JC

if feedback through wiring, does moving low level signal leads have effect; are shields grounded...

Rascal,

The grid stoppers work with the capacitance in the circuit/tube and as such they only operate at frequencies beyond hearing to cancel parasitic oscillation.  If they were as high as 68k they would affect down to 15k or thereabouts.

Check this out: http://www.aikenamps.com/InputRes.htm

Also I would reduce R29 to get the correct HT voltage, there is no point or virtue in wasting power in that resistor, just cos some other guy used that value with a different transformer.

Good luck
DaveP

I put in the 1k grid stoppers, and things seem to have improved.  I'm taking it up to my studio to play properly with all the real gear (instead of my Mbox 2 here at the house) to give it a more thorough workout.  Keeping my fingers crossed!

Thanks!

JC

Excellent news!  It works perfectly (and sounds lovely!).  It probably worked well all along, as it turns out.  I discovered the line input (padded mic input) of my Mbox2 is the culprit of the crazy distortion and DC offset.  Don't know why, but when I undid the pad and turned down the gain of the LA-2A, it got clean.  Once I put the pad back in it got spitty and funky.  I've since tried the unit with legit gear and it works perfectly.

In the process I altered the B+ to a more proper +272vdc and redid the heater wiring to lay flat against the panel holding the tube sockets (which I should have done in the first place), so all the messing was probably worthwhile anyway.

Thanks so much for your help, gentlemen!  Hope I haven't troubled you.

JC