Page 24 of 28

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Wed Jan 17, 2018 7:35 pm
by mediatechnology
You're welcome.

How big are the step sizes? Were they 0.5 dB or 0.25?
I did a spreadsheet for the MTC.
There's an incredibly long formula in the xls that converts exact values to nearest E96.
I can send the xls if you have a large number of steps to calculate.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Wed Jan 17, 2018 11:21 pm
by Gold
They are 12 pos MBB switches. The steps in dB are */- 0.5, 1.0, 1.5, 2.0, 3.0.

I can always use a pot and measure. I usually do that because I never trust my math.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Thu Jan 18, 2018 4:52 am
by mediatechnology
Paul it just occurred to me (as I woke up this morning) that the input Rin needs to be doubled (approx 28K) to keep the -6dB loss built in.

The actual adjustment range needed is -9 to -3 dB.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Thu Jan 18, 2018 11:39 am
by Gold
Can the resistor and trimmer that feed the signal to the rotary switch be mounted on or near the rotary switch or do they have to be mounted near the op amp? That would be the equivalent of R5 and VR1 on the MTC schematic. I don't see a good way to mount the resistor and trimmer on the PCB and not have it flap in the wind. I could deal with the resistor but trimmers are so delicate I'd want some physical strength in the way it's mounted.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Thu Jan 18, 2018 1:18 pm
by mediatechnology
Rin should be local to the op amp's inverting input.
The trimmer in series with Rin could be off-board.
Rather than an actual trim use a smaller value resistor you pick on final test. That will ease mounting.

If you wire the cable so that the shield is driven by the output, then you should be able to locate Rstop at the switch.
The cable capacitance at the inverting input from the center conductor to shield will still be across the feedback network.
If the capacitance is too much >100pF then you might see some HF rolloff.
That however would be about 3' of cable so no worries.

Capacitance from the inverting input to ground is de-stablizing and should be avoided.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Wed Jan 24, 2018 2:03 pm
by Gold
Would there be a benefit in slipping a braided shield over the driven shield cable and connect it to chassis? The Neumann PUE74 phono preamp uses a driven shield. They specify a custom cable using coax runs for the stereo driven shield and an overall braided shield. I've been told that the downside to using a driven shield is that the shield becomes less effective as a shield.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Wed Jan 24, 2018 4:21 pm
by mediatechnology
I don't think there would be any benefit in that short of a run inside a box.
The output impedance of the op amp is "almost" 0Ω so at hum and in-band audio frequencies its effectively at AC ground.

If the cable were long and extended outside the shielded enclosure there would be benefit to having an overall shield at RF frequencies. In that case a build-out resistor would be needed to isolate the op amp output (the inner shield) to the outer one which is grounded.

IIRC Neuman's driven shield was for a low-level phono cart run and the objective was to reduce inner conductor to shield capacitance.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Thu Jan 25, 2018 11:11 am
by Gold
Thanks Wayne. That makes sense.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Thu Jan 25, 2018 11:13 am
by mediatechnology
the objective was to reduce inner conductor to shield capacitance
In this case we take advantage of that capacitance.

Re: Mid Side M-S Matrix Uses No Precision Resistors

Posted: Fri Nov 16, 2018 1:53 pm
by Gold
I'm running into a problem using the M/S board as described in the previous posts. To recap, I want variable gain for both the M channel and the S channel. Since U9 is available for gain I tried using the Decode section as the Encode section and the Encode section as the Decode section. There is no variable gain happening yet. I have the PCB stuffed exactly as in the schematic. I'm just routing the Encoder Output to Decoder Input and testing the L/R Out after going through the Encoder and Decoder. I'm getting terrible crosstalk specs this way. Below is a picture of the test setup for this.
MS_Xtalk-1a.jpg
MS-Xtalk-1.jpg
When I test the M/S board set up using the Encoder as the Encoder and the Decoder as the Decoder I get much more respectable numbers. Am I doing something wrong or do I have to go to plan B?
MS_Xtalk-2.jpg
MS_Xtalk-2a.jpg