Page 8 of 13

Re: Mid Side M-S Matrix Construction Information

Posted: Sun Aug 11, 2013 9:23 am
by emrr
It should drive an output transformer fine in most cases. Do you know the ratio of the output transformer?

Re: Mid Side M-S Matrix Construction Information

Posted: Sun Aug 11, 2013 10:37 am
by Nescafe
Hi Doug,

right now I use 150:600 (+6dB) and like the sound, but it can be configured to 600:600 (unity gain).

Thank You.

Re: Mid Side M-S Matrix Construction Information

Posted: Sun Aug 11, 2013 1:06 pm
by mediatechnology
The THAT1646 can drive the transformer primary tapped to 600R to +27 dBu without any problems.

I've never tried a single 1646 driving 150 Ohms at full level but have paralleled them to get the additional current.
This thread discusses transformer driving using the THAT1646: viewtopic.php?f=6&t=140

Ultimately it resulted in the Dual Class-A line driver which will drive 150 Ohms (or less) quite comfortably: viewtopic.php?f=1&t=523

Re: Mid Side M-S Matrix Construction Information

Posted: Sun Aug 11, 2013 2:13 pm
by Nescafe
Hi Wayne,

Thank You for the useful link info, I will try to drive 150 with MS board, if I can't found the sound I will try 600 or try using the Dual Class A board :)
More question, is that correct that use MS as balancing and debalancing card, the MS mode and stereo mode will switch use point:

Stereo Mode: B to comp & G from comp?

M/S Mode: D to comp & E from comp?

and for point D & E just connect permanently the cold with ground in the board for unbalanced use? Any modification parts value if using unbalance?

I think I ever see the post about that, but cannot find again.......

Thank You.

Re: Mid Side M-S Matrix Construction Information

Posted: Mon Aug 12, 2013 11:45 am
by mediatechnology
Stereo Mode: B to comp & G from comp?

M/S Mode: D to comp & E from comp?
Yes and yes.
and for point D & E just connect permanently the cold with ground in the board for unbalanced use? Any modification parts value if using unbalance?
Correct. No mods just make sure the "cold" ("-") output is grounded at the 1646 output (D).
At the input (E) whatever provides you the lowest noise.

Re: Mid Side M-S Matrix Construction Information

Posted: Mon Aug 12, 2013 2:55 pm
by Nescafe
Thank You Sir Wayne.....
mediatechnology wrote:At the input (E) whatever provides you the lowest noise.
it his mean that I can leave the cold floating with lower noise.....? :shock:

I think the lowest noise will be connecting D to E without the tube compressor.......... :mrgreen:

Re: Mid Side M-S Matrix Construction Information

Posted: Mon Aug 12, 2013 9:32 pm
by mediatechnology
it his mean that I can leave the cold floating with lower noise.....?
No, by grounding the cold you reduce the noise.

Floating the cold to provide an unbalanced output will increase the wideband noise floor from -100 dBu typical to -60 dbu.
Grounding the cold will also restore the gain of the 1646 to +6 dB due to OutSmarts (tm) and drop the noise by 40 dB.

If you're driving a transformer then connect the primary to "+" and "-" and have ground on the MS PCB remain unconnected.

Re: Mid Side M-S Matrix Construction Information

Posted: Thu Aug 15, 2013 9:11 am
by Nescafe
Thank You Mr. Wayne......I will report the result soon after I try it.

Thank You.

Re: Mid Side M-S Matrix Construction Information

Posted: Mon Jan 13, 2014 1:49 pm
by fuccimain
Hi.

I'm looking to add a hi pass filter to the side signal, at point "C" . I'm curious if this has been discussed before. I was talking to Wayne about this briefly via email and we talked about using the decode half of the board as the encode, so the HPF side could have a variable make up, but I'd like a variable makeup post hpf pre main insert, then again at the output for the LR. So I was thinking about modifying the encode half of the board to duplicate the decode with adding the section of the decode circuit that has the 5532 to the encode half...btwn the 1240 and the 1646, basically making 2 encode halfs. Any reason this wouldn't work? The was thinking of a basic RC hpf pre the 5532. Unless using an active HPf would be better..

Sorry if anymore this has been discussed before,

Re: Mid Side M-S Matrix Construction Information

Posted: Mon Jan 13, 2014 2:22 pm
by mediatechnology
I think duplicating the inverting 5532 stage would be fine as a means to provide variable gain and buffer the RC filter at the input.
Doing it that way allows the 1646 to see a low source impedance.

One "trick" in the decode path that's not obvious is that to correct for the 5532 polarity inversion, the 1646 output PCB silkscreen polarity is reversed.
If you invert at point C in the encode path the "+" output silkscreen will be "-" so make a mental note.

If a single-pole HPF response is steep enough in the side channel I see no reason why a simple RC at the 5532 input wouldn't work just fine.
If it's also a variable gain stage you would only want to only vary Rfb to maintain a constant input impedance.