No, it's the main line-level preamp (lacks a phono stage) that, in this case, provides system Gnd reference via interconnect cable shields to the upstream RIAA metal box then upstream again to the TT via its cable shields. Pretty universal. Now, the line-level control preamp is third-prong grounded to the wall.
Jean-Paul: In many cases the Grn wire Earth is connected to DC power supply minus. That minus is audio shield gnd so yes, audio gnd for single-ended I/O is referenced to Earth ("typically") - when there is a 3-prong AC plug. Now some gear will have a "ground lift" switch which disconnects the audio from Earth and allows referencing from another piece of gear. It's a hum fixit switch for ground loops. And some gear can earn "double or reinforced insulation" status from UL or CE and are not required to supply a grounded AC cord. Those units float - have no direct relationship with house Earth.
Unlike Jean-Paul, I don't think it is related to RF in the air. I think it's an oscillation that depends on the source impedance - a cable resonance getting undamped by the amplifier. I only mentioned susceptibility to RF signals because one of the possible fixes has reduced sensitivity to RF as a benificial side effect.
All known here
but what about your setup? It is harder to solve trouble when PE is lacking I can tell. In my trouble shooting the multi-approach seems to work. PE connections, shielding connected to PE, decoupling, a metal box with PE connection over the electronics and lifting it up/putting it down, checking the mains side of things, ferrite rings over the cables, the order how equipment is placed on eachother (!), LED lighting on/off, wireless internet on/off, smart phone test, other RCA cables etc. Whenever I can trigger or break the phenomenon I know I am in the right direction. Had it several times that one of the RCA cables had GND disconnected for instance. I don't care too much what is behind the issue as long as I can get it solved.
It can be a combination of both oscillation and RF pickup.
Last edited:
I agree with Marcel that it must have to do something with LF problems.
Since we don’t know exactly what the switch is shortening, is it really only shortening the cart wires or is connecting both cart wires to the TT chassis ?
And since you mentioned that one channel had a low resistance between cold and TT chassis, my first question is: what if you change channels in your test, problem still the same or not ?
Second question: what happens when you disconnect the TT from your Riaa preamp, insert just a 1.5 meter RCA interconnect to it and shorten and open the RCA plug at the other end ?
To simulate a cart at LF you could connect a 1K resistor to the far end or whatever the DC resistance of your cart is.
Hans
Long shot for Johnc124: (See post #581) Of the three big players - 1656, 49990 and 1612 - do all three use input biasing compensation? Datasheets for the former and latter indicate yes with the +/- sign. Sometimes it goes unsaid so do you know if the LME49990 is bias compensated? Lastly, of the list of "good guys" the 4562/49702 didn't fit the GBWP profile (too high). Datasheet indicates no input bias compensation used. True?
Now I have more questions to test and answer. Thank you all for hanging in with me. I'll be back.
The 49990, 1612, and 4562/49720 all have input bias cancellation. The OPA1656 does not because it's a CMOS amplifier, and the low input bias current comes from the fact that the inputs are the gates of a MOSFET rather than the base of a BJT.
I'm not sure how you came to the conclusion that 4562 doesn't have IB cancellation. Just from the voltage noise I would guess it's running ~500uA of tail current in the input pair. To hit the 10nA of typical input bias current, the transistors would need betas of 25000 if there was no bias current cancellation. Furthermore, the input current noise would follow the shot noise equation (sqrt(2*q*IB)) and would be 57 fA/rtHz if there were no IB cancellation (in reality it is 1.6 pA/rtHz). Definitely IB cancellation on all of those bipolar op amps.
OK. For all the PE and grounding concerns I have a verbal explanation and a hopefully helpful illustration. Open attached PDF.
>>>>>>>>>>>>>>>>
Main line-level preamp accessible metal = 3rd prong PE.
Main line-level preamp chassis-mount RCA connector shields = PE
Now:
The connection from the isolated outboard ±15Vdc linear supply → RIAA metallic box is a 2-cond balanced mic cable with braid shield. The shield is the ±15Vdc center mid-point common.
This mid-point “0.0Vgnd” is bonded internally to the metallic RIAA box and also serves as the circuit/signal Gnd for the Circuit Card.
All four RCA female chassis-mount I/O jack shields are isolated from the metallic RIAA box.
All female I/O jack shields become “0.0Vgnd” / metallic box via individual “signal + ground” connections from the Circuit Card to each of the isolated RCA female signal and shield tabs.
Thus via the main preamp → RIAA preamp interconnect shields, the metallic RIAA box becomes PE as does the bonded ±15Vdc mid-point “0.0VGnd” and the op amp Circuit Card signal Gnd.
Thus the RIAA preamp → TT cable shields are also then = PE
Bottom line is that ALL PEs originate at the main system’s preamp Earthing.
>>>>>>>>>>>>>>>>
Main line-level preamp accessible metal = 3rd prong PE.
Main line-level preamp chassis-mount RCA connector shields = PE
Now:
The connection from the isolated outboard ±15Vdc linear supply → RIAA metallic box is a 2-cond balanced mic cable with braid shield. The shield is the ±15Vdc center mid-point common.
This mid-point “0.0Vgnd” is bonded internally to the metallic RIAA box and also serves as the circuit/signal Gnd for the Circuit Card.
All four RCA female chassis-mount I/O jack shields are isolated from the metallic RIAA box.
All female I/O jack shields become “0.0Vgnd” / metallic box via individual “signal + ground” connections from the Circuit Card to each of the isolated RCA female signal and shield tabs.
Thus via the main preamp → RIAA preamp interconnect shields, the metallic RIAA box becomes PE as does the bonded ±15Vdc mid-point “0.0VGnd” and the op amp Circuit Card signal Gnd.
Thus the RIAA preamp → TT cable shields are also then = PE
Bottom line is that ALL PEs originate at the main system’s preamp Earthing.
The no cart, cable-only short and un-short testing was done with 4ft of low-cap, 17pF/ft RCA cable with one end plugged into a RIAA phono input and the other end terminated with a 1KΩ R. See pic.
Shorting was done with the shaft of a small screwdriver placed or touched across the leads of the R (=SHORT). Or, previously placed across the leads, the scope set to trigger and then quickly removed (=OPEN). The 1656 sees 1K//47K then 0Ω or sees 0Ω then 1K//47K through a 4ft cable, resp. I practiced this maneuver until I got it right. The real mute slide switch is not that clean anyway. The scope pics are of that channel’s output. (Please see the peak-to-peak voltage in the lower right corner labeled "dV" b/c I forgot to include that value in the captions). They look somewhat familiar to the real 1656 Lift and Lower shots in as far as gross amplitude and the very low frequency content if nothing else. Oh, and the 3rd scope shot was when I bobbled my hold on the screwdriver shaft and resistor leads so this is a series of interesting Short-Open, make-breaks, repeat, etc.
Shorting was done with the shaft of a small screwdriver placed or touched across the leads of the R (=SHORT). Or, previously placed across the leads, the scope set to trigger and then quickly removed (=OPEN). The 1656 sees 1K//47K then 0Ω or sees 0Ω then 1K//47K through a 4ft cable, resp. I practiced this maneuver until I got it right. The real mute slide switch is not that clean anyway. The scope pics are of that channel’s output. (Please see the peak-to-peak voltage in the lower right corner labeled "dV" b/c I forgot to include that value in the captions). They look somewhat familiar to the real 1656 Lift and Lower shots in as far as gross amplitude and the very low frequency content if nothing else. Oh, and the 3rd scope shot was when I bobbled my hold on the screwdriver shaft and resistor leads so this is a series of interesting Short-Open, make-breaks, repeat, etc.
Attachments
My purpose was to look for gross common denominators relating to the players on the two #581 lists. I got over my head re input bias compensation (or not) here:
https://www.analog.com/media/en/training-seminars/tutorials/MT-038.pdf
Is there a +/- Ib or Ios there or not..? Yes = Compensated. If not it might have not made it to the datasheet... The 4562/49702 has no +/- figures for Ib and Ios so I thought hmmm, it doesn't. Maybe that's a clue. That's all. Thanks for your answer.
O.k, your experiment with the separate interconnect shows that your TT + Cart can be excluded from the equation.
That narrows the search for a cause.
My next suggestion would now be to shorten C4, the 100uF cap.
This will cause a DC offset at the 1656 output, but below 1Volt so still acceptable for this test.
Your main amp probable has a dc blocking cap at it’s input.
Hans
That narrows the search for a cause.
My next suggestion would now be to shorten C4, the 100uF cap.
This will cause a DC offset at the 1656 output, but below 1Volt so still acceptable for this test.
Your main amp probable has a dc blocking cap at it’s input.
Hans
Do you call 50 MHz low frequency?
The thing with oscillations is that their amplitudes tend to keep growing exponentially until some non-linear effect, like some stage that clips, stops the growth. If that clipping happens asymmetrically, as is almost always the case, it causes a DC shift.
To me, the logical next step would be to try RC damping networks or RC filters, like 10 pF-330 ohm, 15 pF-220 ohm, 33 pF-100 ohm, 47 pF-68 ohm or 82 ohm. See if it helps and if so, try to stay away from the edge of the cliff.
“Stay away from the edge of the cliff”
Good advice.
But I assume this is just a prototype for a decent PCB.
So investing much time in getting this prototype working, could become useless in the final version.
I think we all agree that this opamp needs a proper layout including a decent groundplane.
Hans
Good advice.
But I assume this is just a prototype for a decent PCB.
So investing much time in getting this prototype working, could become useless in the final version.
I think we all agree that this opamp needs a proper layout including a decent groundplane.
Hans