It seems we are having a confusion of tongues.
1) In #145 you showed a different response between L & R without damping, which led me to comment that probably one channel was floating while the other was not.
2) In #159 a picture revealed a connection between TT chassis and left channel shield, confirming my suspicion in 1), so I suggested to cut that link.
3) In #171 it reads as if you did not cut that wire, but for some reason you pointed to #73 and to an attached pdf that both have nothing to do with the non floating left channel.
4) But in #172 you mention that with Cart removed there is still a 12R resistance between TT Chassis and Left channel shield.
Where and how did you measure this, because the short piece of wire in picture #159 can hardly have this high resistance.
So to conclude, IMO as long as both channels are differently connected, i.e. Right floating and Left not, without damping you can hardly expect the same response to the mute switch.
Hans
As far as I understood, the picture in post #159 doesn't show JRA's turntable, it's just a picture of an example of a shorting wire.
12 ohm could either be a weak short or an intentionally placed resistor. Either insulation got damaged somewhere or the manufacturer placed a 10 or 12 ohm resistor to ensure that the chassis is still sort of grounded when the ground wire isn't connected, but without creating a hard ground loop when the ground wire is connected.
12 ohm could either be a weak short or an intentionally placed resistor. Either insulation got damaged somewhere or the manufacturer placed a 10 or 12 ohm resistor to ensure that the chassis is still sort of grounded when the ground wire isn't connected, but without creating a hard ground loop when the ground wire is connected.
That's right Marcel and sorry Hans. I guess I didn't make it clear that the pic was from the NET, as an example of what's done by some(???) TT manufacturers. And maybe that practice is as Marcel said; provides a relatively minor 50/60Hz hum reminder that one forgot to attach the real gnd wire.
I've built my own using the OPA1656 but did not find the proto boards be a good idea for trying opamps, so I started having boards made using ExpressPCB.
And after much playing around, I found that putting the IC farther away from the actual input was an improvement with noise. To later finding that instead of one loading cap on the input, I divided the capacitance value by 3 and use three different caps on each input to the chip. In between the input on the PC board to the input of the IC for each channel, I have 47K to ground, ferrite bead, 100 ohm resistor, 33pf to ground, ferrite bead, 33pf to ground, 21 ohm resistor to IC, 33pf to ground.
The beads and resistors in series really help clean the sound up in the high frequencies. Someone in a previous post had mentioned putting a resistor in series to help lower the noise and that is what I've also done, but went a tad further.
I've integrated some filtering into the loading of the cartridge to deal with the spurious noise coming from the turntable.
And after much playing around, I found that putting the IC farther away from the actual input was an improvement with noise. To later finding that instead of one loading cap on the input, I divided the capacitance value by 3 and use three different caps on each input to the chip. In between the input on the PC board to the input of the IC for each channel, I have 47K to ground, ferrite bead, 100 ohm resistor, 33pf to ground, ferrite bead, 33pf to ground, 21 ohm resistor to IC, 33pf to ground.
The beads and resistors in series really help clean the sound up in the high frequencies. Someone in a previous post had mentioned putting a resistor in series to help lower the noise and that is what I've also done, but went a tad further.
I've integrated some filtering into the loading of the cartridge to deal with the spurious noise coming from the turntable.
Attachments
mp20748: I don't remember seeing your post. Don't know why and not sure where this Reply is going. In any case, that seems like a whole lot RCL armor you felt you needed to add to the 1656 inputs. Question I have is what was it about the sound up in the high frequencies that was not clean? Was the source actually the TT? Was it treble distortion in the music or as you say spurious noise like pops, ticks, crackles and such? And was the unwanted noise quite loud? Sorry for all the questions. Just curious.
The very fast OPA1656 has to be used with great care, meaning good power supply decoupling, preferably a ground plane and a well designed PCB around the input connections.
Despite the low source resistance driving the preamp, the interconnect’s capacitance and switchable input resistors can easily lead to stability problems causing differences in sound perceptions.
Proof of the pudding would be to use a slower opamp that’s less critical like the OPA2134 or the OPA1642 and see if the supposed sound differences are still there.
I bet they won’t.
Hans
Despite the low source resistance driving the preamp, the interconnect’s capacitance and switchable input resistors can easily lead to stability problems causing differences in sound perceptions.
Proof of the pudding would be to use a slower opamp that’s less critical like the OPA2134 or the OPA1642 and see if the supposed sound differences are still there.
I bet they won’t.
Hans
I thought your comment was on-target. Without special measures taken the 1656 RIAA datasheet circuit (a classic topology) will work using basic DIY build techniques but unfortunately not with the 1656, as shown, if/when the TT has a mechanism that shorts and opens the two pairs of cartridge wires inside the TT. And I should repeat that once playing an LP my original implementation played music and it was good. Albeit it was only one side but I heard no "audible" misbehavior - until the end of course when the Lift&Short function royally upset the 1656 causing it to convulse. Same with the 1612A. Anyway, old news. Fixed with Marcel's diazepam damper network.
And yes, other op amps were not upset by having their inputs shorted and cleared. The 626, 2107, 2134, 2604 and surprisingly the LM4562/LME49720 were well-behaved.
And yes, other op amps were not upset by having their inputs shorted and cleared. The 626, 2107, 2134, 2604 and surprisingly the LM4562/LME49720 were well-behaved.
I was triggered by a post that mentioned adding an 82ohm resistor that made me think about the preamp I built, that has an 100 ohm resistor in between 47k/100pf loading and the IC's input. I added that resistor as well, a ferrite bead idea from Yamaha that is common on most of their earlier magnetic preamp designs. And in my research about why did Yamaha add the bead, I was told its most likely to cut down on some of the noise that would be at the phono input. It was explained to me that the high gain of the preamp would attract a lot of common high frequency noise that is everywhere these days. These noises may not be in the audio range, but getting into the opamp could cloud its performance, and that turned out to be true. So I then implemented something I remembered on a different circuit where the input was spaced from the source using series components. So I then spread the 100pf capacitance by using three 33pf caps going to ground and in between the beads and 100ohn resistor. On that particular board I've added another 33pf right on the chips input that I change out depending on the cartridge I'm using. I built three identical boards to play with this.
And after noticing a difference I took my preamp to a guy that repairs audio stuff, and he also explained that it made sense because of the noise from the high gain input. He then pointed out that the 4 cartridge wires are not shielded for at least 12 inches from below the tone arm to right near the cartridge. So to me that confirmed why Yamaha would add the beads right at the input of their preamp. The audio repair tech provided a set of speakers and an amplifier and album. I took my Denon DP35 turntable with me that was using an Ortofon red cartridge that I paid $99 for. I told the audio guy to see if he could notice a difference and he said yes immediately and that I was using a $99 cartridge. But the bigger thing on my board is that I use regulators for both +/- supply and a lot of decoupling between the chip's supply to the regulators.
I may have used the wrong word to describe the noise, when HASH noise would have been the better word to use. One thing for sure, these chips perform better when better decoupled with lower noise on the input. The board design is also critical to minimize any parasited oscillation, which will also cloud up the signal.
Hans / Marcel: I found the 12Ω weirdness in the Left channel. I had the TT upside-down to check/replace electrolytic caps and saw it again. Left channel shield was not isolated from the TT Gnd wire. I had long ago installed a grounded RCA jack bracket on the rear of the plinth and there was this tiny whisker of braided shield wire touching the backside of the bracket and that there was the ~12 ohm mystery.
MP20748: I’ve been using the 1656 in my MM pre for over a month now with no issue. Since the Marcel Diazepam fix went in I can plug ‘n play whatever I have here. My DIY board is not pretty; there’s no gnd plane, no flooding – just semi-random point-to-point wiring (like tube p-t-p wiring). It is (happily) dead-snake quiet. So I have to wonder what else is driving your need to multi-stage-filter the 1656’s input in order to get rid of HASH noise. Obviously it shouldn’t know how to do that in the audible range. Is the 1656 getting hot? Do you have a scope? Oscillating..?
MP20748: I’ve been using the 1656 in my MM pre for over a month now with no issue. Since the Marcel Diazepam fix went in I can plug ‘n play whatever I have here. My DIY board is not pretty; there’s no gnd plane, no flooding – just semi-random point-to-point wiring (like tube p-t-p wiring). It is (happily) dead-snake quiet. So I have to wonder what else is driving your need to multi-stage-filter the 1656’s input in order to get rid of HASH noise. Obviously it shouldn’t know how to do that in the audible range. Is the 1656 getting hot? Do you have a scope? Oscillating..?
Could it be external RF noise getting in the line? It would depend on how close MP20748 is to a city centre or other source of RF: I know coaxial speaker cables made a noticeable improvement in downtown Montreal [1], whereas a listening group was straining to hear any difference in a sparsely populated Denver suburb [2].
Similarly, Wendy Carlos discovered that Faraday shielding her entire New York City studio reduced the noise floor on some gear to the point where previously unusable modules became quite musical [3].
From this I figured RF interference (RFI) could hide as internally-generated noise in a system, and since then my gear's had plenty of shielding, balanced I/O, and input RF suppression to avoid the problem.
This article has useful information on fighting RFI:
UNDERSTANDING and controlling RF interference
[1] Quad 405 amplifier, KEF 103/3
[2] Some beefy solid-state amp driving large Magneplanars
[3] http://www.wendycarlos.com/other/PDF-Files/KbdInterview-82*.pdf
Last edited:
No I never did those experiments again. Should have just to confirm but got lazy and just moved on.
One change I made which I believe makes technical sense re RFI. All of the RCA I/O jacks mounted to the RIAA metal box were of the isolated type where the RCA shield does not contact the grounded metal shell. I changed that to non-isolated so all shields bond directly to the box's outside "skin". You might remember that the box gets its Earth Gnd via the downstream main preamp RCA cable shield. When isolated that Gnd/shield is discontinuous in that the shield is taken from the jacks inside tab and wired to the board's signal Gnd, which is the midpoint of the +/- power supply and then from there just a wire to the box interior.
Now there's a continuous gnd/shield starting at the preamp, includes the box's shell at the outside and continues on to the TT. I remember hearing that likened to a gourd if you can imagine a gourd (the enclosure) with two stems and a continuous skin. No audible result, did no harm, there was nothing needing fixing but I feel better.
One change I made which I believe makes technical sense re RFI. [...]
Now there's a continuous gnd/shield starting at the preamp, includes the box's shell at the outside and continues on to the TT. I remember hearing that likened to a gourd if you can imagine a gourd (the enclosure) with two stems and a continuous skin. No audible result, did no harm, there was nothing needing fixing but I feel better.
I like XLR connectors because the shield is distinct from signal lines, whether hot & ground for unbalanced, or positive & negative for balanced connections. If you don't want the bother of driving the negative line then you could always run quasi-balanced, with that line attached to ground through the same impedance as the output driver would exhibit. A differential receiving end will cancel interference in the same way as if the connection were fully balanced, albeit with 6 db less signal-to-noise ratio since only one line of the pair is driven, after all.