Eros intermittent volume balance issue [resolved]

[DrewTube]

All,

I have a strange gremlin haunting my turntable playback.  I have a Rega P5 (w/ Rega Exact) -> Eros -> EFPIII.  Most of the time, the image is dead-center and this thing sounds GREAT.  Very intermittently, the image will suddenly pull about 30 degrees right, which I *think* is the right channel suddenly becoming 2-3 dB louder.  Unplugging and replugging the right channel input (i.e., turntable right output, Eros right input) makes it go away ~80% of the time (the other 20%, a second iteration does the job).

Through some careful process of elimination, I'm pretty certain the problem lies either in the turntable or in the Eros (FPIII, power amp, speakers are shared with other sources and have never exhibited this problem).  I have spent hours poring over the Eros with a magnifying glass, carefully re-soldering any joints that seemed at all suspicious, and looking for any solder bridges, clipped wire ends, or "fuzzballs."  I pulled out all my NOS tubes and put back in all the stock EH tubes; I have also swapped left and right EF86 and tried a third 6922.  I had an unrelated issue with my turntable wiring, so the wiring is all redone from cartridge to output jacks (problem existed both before and after that change).  None of these changes have rid me of my gremlin.

Two key questions:

1) Is it possible this gremlin is in the cartridge/turntable?  Any way to validate that, given that I do not have another phono preamp to swap in?
2) If it is in the Eros, where do I go next?  Does this sound like a lifting ground in the signal path somewhere?


Of course, please let me know any additional needed info and I'm happy to run any tests/diags.


Thank you!

-Drew

[Grainger49]

I would reheat the right input RCA connections.  Add just a touch of solder.

You can verify that it is the table/cartridge if you swap right and left cables at the Eros.

You can verify that it is the cartridge by swapping right and left connections at the cartridge.

[DrewTube]

I would reheat the right input RCA connections.  Add just a touch of solder.

You can verify that it is the table/cartridge if you swap right and left cables at the Eros.

You can verify that it is the cartridge by swapping right and left connections at the cartridge.

Thanks, Grainger.  I've already tried re-heating that right input RCA.  In fact, it showed traces of oxidation, so I even replaced the entire right input jack and did the new soldering very carefully.  I'll try a reheat once more, to be sure.

Your other two tests are so simple!  I'm very irritated at myself for overlooking such simple tests!  I will give those a try...the biggest challenge will be that the issue is intermittent and I have to test long enough for it to happen.

I'll update as this progresses.

[Grainger49]

Drew,

Well, that clears the RCA jack from being the problem.  Don't worry, sometimes it is the simple things we overlook, thinking zebra rather than horses.

If you want to keep the right left proper while testing just swap the output cables.

[Grainger49]

Drew,

Any news?

[DrewTube]

Drew,

Any news?

Sorry to completely ghost this thread over 6 years ago!  Long story; short version is sudden, major life change and the audio equipment went into storage for years.  Boo.

Well, now it's back out and making sweet music, but my Eros gremlin persists!  In fact, it is happening more consistently now, which is annoying - but might help get it diagnosed.

Per Grainger's suggestion, I swapped the L/R of both the Eros inputs and outputs and the volume jump switched sides, so I'm now confident that it's the Eros that's the problem.  I replaced all four NOS tubes with the stock EH's and no change - so I think it's not tubes.

To reiterate, the problem is that some random time ~2 - 10 minutes after I turn on the Eros and it gets nice and hot, the right channel suddenly increases gain by ~6 dB (that's a rough guess).  Enough to pull the imaging way right and be really obvious.  The right channel audio is also then slightly distorted, sounding somewhat like an over-modulated mic.

All the LEDs are solidly lit and all my resistance and voltage measurements check out well within parameters.

I'm pretty confident the problem is in the Eros circuit somewhere and not the tubes.  I've quadruple-checked solder joints at this point.

Any thoughts on what to test and try next?!  This is driving me nutty!

[Paul Birkeland]

I would be absolutely certain that you are getting a voltage check once the image shift has happened.

I would also recommend just switching the outputs of the Eros and checking that way, as switching both would allow an issue from the TT (which I believe to be unlikely) to switch channels.

For the situation you're describing, I would be nearly certain that the voltages will change when the problem pops up.  Would you say that certain frequencies get louder than others, or does it stay pretty even?

[DrewTube]

Thanks, Paul.  I will do a bit more isolation testing and voltage checks this week when I get a chance.

The last time the image jump occurred, I did NOT see any change in voltages, which completely baffled me.

I will repeat this all with much care and will report back here in the next few days.

[DrewTube]

PB et al.,

OK, apologies for another long delay.  Combination of being busy and the issue suddenly becoming difficult to reproduce again.

I've made some progress.  I'm more certain that the issue is the Eros.  I tested the following changes to the interconnects and then noted which channel the issue presents in:

• All connections normal: Right
• Eros inputs swapped:  Right
• Eros outputs swapped:  Left
• Eros inputs & outputs swapped:  Left

I'm pretty confident that rules out the turntable as the source of the issue and confirms it's in the right channel of the Eros.

I also very carefully and methodically repeated all my resistance measurements and voltages measurements while the issue is NOT occurring, so I know that everything is within specs and I have a recent baseline.

Now I'm just waiting to reproduce the issue again so that I can take careful voltage measurements while it IS occurring.  Of course, as gremlins do, now that I'm all prepared for this, the issue has suddenly become difficult to reproduce.  I'll keep plugging away and will update you.

[DrewTube]

OK, this gets weirder.  It appears that I was having trouble reproducing the issue because it's somehow related to the stuff on the underside of the chassis getting hot.  With the chassis flipped over to expose stuff so that I can do voltage measurements, the cooling was too good.  I eventually placed a cover over it to keep some of the heat in and had some luck reproducing the volume increase in the right channel.

I was able to get almost a complete set of measurements while the issue was occurring.  Basically, they are not materially different.  The one very weird thing is that attempting to measure IA while the issue was occurring would immediately (but temporarily) fix the issue.

What	Normal Value	Issue Value
IA	223.8	223.8(?)
OA	165.6	163.5
breg	94.4	95.8
kreg	0.728	0.775
OB	92.7	92.3
bA	0	0

To answer PB's earlier question, I don't hear any really obvious change in the frequency response when the image shift happens.  The right channel gets louder, gets a bit distorted (as I've said before), but I don't hear any obvious frequency response changes.

[Paul Birkeland]

The only part that ought to get warm is the 1085 regulator mounted to the heatsink, but it is common to both channels.  It is possible that maybe you have loose heater wiring (pins 4/5) on one of the EF86s, but if a channel gets weak from that you will see the voltages move around a lot.

[DrewTube]

PB et al.,

I don't think that the issue reproducing when things get hot is related to anything that is supposed to get hot.  I think I have a flaky component, wire, or solder joint somewhere that starts to fritz when it gets warm.  I just need to narrow down exactly which wire/joint/component that is so that I can replace it!

I made another interesting discovery.  I noticed that the screw that attaches the first terminal strip to the chassis was ever-so-slightly loose.  This is the terminal strip hosting terminals 1-5, with the screw in question under terminal 3, right above A5.  When I fully loosened this screw, it reproduced the right channel volume jump issue 100% of the time.  When I fully tightened the screw, it resolved the issue 100% of the time.  With it slightly loose, it would start off OK, then when things got warm...bleh.  I can't believe it's a coincidence that this terminal strip is in the right channel signal path where I have been having the problem.

However, this screw being loose would seem to only directly affect connectivity of terminal 3 to the chassis/ground.  The only thing connected to terminal 3 is the drain wire of the heater wiring for the socket A tube (right channel EF86).  Could the drain wire of the heater wiring lifting actually cause such a dramatic issue?  Or is it more likely that I have a flaky solder joint or component attached to terminal 1,2, 4, or 5 and the slight shifting of the terminal strip when the screw is loosened is triggering that?

[Paul Birkeland]

The drain wire won't cause that kind of issue, but I think you've narrowed down where your problematic connection is.

[DrewTube]

Alright, courtesy of the long holiday weekend, I got to spend some time on this.  Upon very minute inspection of terminals 1-5, I still didn't see any solder joints or anything that looked the least bit suspicious.  All the resistance, capacitance, and continuity checks I did all check out fine (the two resistors and two capacitors appear to measure normally when removed, as well).

I went ahead and sucked the solder out of all five terminals (top and bottom, where used) and re-flowed them all with fresh solder, making sure to get them hot for proper flow and that they cooled slowly to a nice shiny patina.

All the resistance and voltage measurements still checked out within specs and effectively the same (a few changes after the decimal point here and there), so I put the stock EH tubes back in and fired it up.  I am now not able to reproduce the volume/balance issue by loosening or tightening the terminal mounting screw - or via (very careful) physical manipulation of the terminal strip itself.  After leaving the Eros on all day to ensure it was good and hot, and after ~5-6 hours of listening without any issue or change in sound, this appears resolved!

Of course, given this was a very intermittent issue, at least to begin with, I won't be completely confident this is truly fixed until I have several dozen hours of listening under my belt without issue.  But, it looks good so far!

I really wish I knew exactly what was wrong.  Given the components involved in terminal 1-5, I'm at a loss for even theoretically how a flaky joint would cause this specific issue, but I'm far from solid enough on the ciruit theory to really know.

Thanks to PB and everybody else for their time, help, and attention.

[Paul Birkeland]

A flaky solder joint in the EQ could lead to a volume imbalance and no shift in DC operating points.