Mid Side M-S Matrix Uses No Precision Resistors

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by radiance »

Thanks for the thorrough reply...
..no, this +6dB is not bothering me. There was a thread on the Lab forum about one of the ms boards and it made me think. Someone questioned the amount of IC 's in the signal path and I was just thinking about another way to achieve -6dB attenuation somewhere along the line...

Here a link to that thread> http://www.prodigy-pro.com/diy/index.php?topic=37821.20

someone called Chriss posted a schematic of his take on the ms encoder/decoder, very cool...
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

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

Post by mediatechnology »

Someone questioned the amount of IC 's in the signal path and I was just thinking about another way to achieve -6dB attenuation somewhere along the line
Well keep thinking and let me know if you figure it out.

There aren't really a lot of ICs in this one if you strip out all the circuitry required to make it balanced - 4X 1246 and 4X 1646 - then it's just a pair of THAT1240s (or a single 1280) for either encode or decode and a lot more precise. That reduces the path length considerably. (The gain error of the balancing/debalancing stages inside the matrix is what reduces the precision. The 1646 isn't as gain accurate as the 1240/1246.)

Unbalanced Matrix
Image
There are not a lot of ICs in this picture

The key to matrix accuracy is not only the resistor precision but also the cross-connected inputs to produce sum and difference. Any -6dB active stage added to M but not to S is going to increase crosstalk because the phase delays won't be balanced. With 1240/1246 parts we have to deal with the values and ratios they're made with and unfortunately don't have a part that allows us to sum through the internal resistors equally and give us -6 dB. But, they're precise, inexpensive, track over temp and measure clean.

The best way to get the M gain to -6dB is to use ultra-precise (and expensive) fixed resistors in the ratios required to reduce M gain. I don't think you'll get there with trimmers due to tempco issues. The whole point of my gizmo is to avoid very expensive resistors. In checking cross-talk in the precision bypass mode even an Ohm, a single Ohm, of connection resistance inside the M and S inserts will reduce the performance. It makes me think about how difficult it would be to trim down to -90 dB. Anyone got a price and delivery on 10K 0.005% <50 ppm tempco resistors?

I'll happily plead ignorance on what goes on at Prodigy other than what Roger passed on to me about HF crosstalk. Most of the Soundcards I've used - and some of the data that's been sent to me - show rising HF crosstalk in loopback making HF measurement of the matrix appear to be the culprit. The data in the upper octave is usually about the same as the Soundcard itself. I think I've said that in earlier posts but it may have gotten missed or just ignored. viewtopic.php?f=6&t=112&start=169
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by radiance »

Sorry for stirring things up here with prodigy leftovers;-)
Also, I'm not complaining here, just want to share my thoughts and maybe hear your opinions..I'm perfectly happy the way my ms encoder /decoder (with THAT ic's) is working.

Reading my initial post I think I misunderstood something
radiance wrote: How about using a THAT1246 for the encoding part but with precision 6K resistors in series with the sense and ref (i.e. R1 & R2) pins? ..
The fact that the 1246 is -6dB is BECAUSE r1 & r2 are 6K right? Somehow I thought the -6dB was caused by something else. (not the resistors). :oops: :oops: :oops: :oops:
With this knowledge I think it's not possible the way I thought it. Yes, we're stuck with the +6dB and yes, even if there was some kind of workaround, precision & temperature drift would have been the culprits...

Best way to use ms in a mastering chain would probably be in combination with a buffered input and output control (for both left and right) with inserts.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

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

Post by mediatechnology »

Sorry for stirring things up here with prodigy leftovers;-)
Hey it's cool. I couldn't resist. They've been a little hard on Roger recently.
Also, I'm not complaining here, just want to share my thoughts and maybe hear your opinions..I'm perfectly happy the way my ms encoder /decoder (with THAT ic's) is working.
Glad you like it. I didn't really see it as a complaint but I thought it was a good idea to revisit the +6 dB Mid discussion since it was last discussed "way early" in this 21 page thread. We may have some new viewers here linking into the middle of this thread.

One thing we do have going for us is that the balanced inputs are -6 dB attenuating. So our internal level feeding the matrix (for +4 dBu levels) is -2 dBu. So what L+R giveth the THAT1246 input taketh away, then the 1646 giveth it back again. What would work well here is a THAT1640 unity gain part. I'm not the first guy that's asked for one. ;)

Another factor affecting precision in any matrix I should mention is termination and impedance differences in the insert sends and returns. The THAT1646 has 25R build-out resistors and most any op amp output, IC or DOA, is going to have to have some build-out resistance to drive capacitive loads. If a 3K input terminated compressor is in the mid channel and a 10K input terminated EQ is used in the side channel they won't have any where near the crosstalk performance when bypassed due to gain errors in the MS domain just due to loading. This makes the case for the precision bypass mode shown earlier. Points C to F here: viewtopic.php?f=6&t=112&start=18
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by radiance »

raf wrote:Radiance, no worries here!

It's all good ...
emrr
Posts: 570
Joined: Sun Mar 02, 2008 10:27 pm
Location: NC, USA
Contact:

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by emrr »

Your patience wore thin faster than mine Roger. :D I've been stopping myself on that one for several days.
Best,

Doug Williams
Electromagnetic Radiation Recorders
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by mediatechnology »

Oh, you like my caption?

And where are the resistors? They're not Dales or Roedersteins. :roll:
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by mediatechnology »

Hey, for get the Dales, go for Roderstein! Or better yet Holsworthy H8 resistors....
And how much do they cost in 0.005% tolerance? Are they in stock? What's the lead time on those? :mrgreen:
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by mediatechnology »

OK, I figure the build requires at least 16 of them.
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Mid Side M-S Matrix Uses No Precision Resistors - Boards 5/1

Post by radiance »

oh..oh what have I done ;)
Post Reply