Sum bus theory and practice

[mediatechnology]

I was thinking that in the current summation approach a little bit of inductance locally right at the CFB amp's emitters could negate the VCA's 15 pF/channel output C. In a 16-channel that's 240pF or so of distributed C. The 1570 may not have the performance advantage though using "resistorless" current summation.

In the preamp tests I used 10 uH per Rgain leg. IIRC at maximum gain the -3 dB point was about 80 kHz. It sure reduced low-level rectification that I could observe in the DC error servo. In an unshielded proto without inductors I could move the servo by simply grabbing a mouse over 4 ft away. The mouse and I made enough of an antenna to radiate into the inputs and gain switch wiring which got rectified by the circuit and corrected by the servo.

[juniorhifikit]

I have used a few ferrite beads at mic inputs but never tried inductors to decouple bus capacitance. In principle it should work, not unlike the inductors inside the classic Deane Jensen opamp (i.e. tuned well above audio range).

I was pondering the same thing. I spoke with John Hardy, who makes 990's based on Dean's designs - they have 20uH inductors on the emitters. He was kind enough to send me a bag of the ferrite-bead cores (Jensen sells them too). He snakes a length of wire through the six holes in the ferrite core, and that's the inductor in the 990.

[mediatechnology]

These wind up in the emitters aka Rgain lines.



The thread is here: viewtopic.php?f=6&t=256

I'm trying to remember in the 990: Do the two input transistor's emitter inductors share the same core?

[juniorhifikit]

I'm trying to remember in the 990: Do the two input transistor's emitter inductors share the same core?

http://www.johnhardyco.com/pdf/990-2007.pdf

page three

[JR.]

I have never personally messed with a 990 but I suspect they are separate inductors. I would have to look closer to see if there are normal vs. CM considerations that argue against using a common core.

I have seen a single core variant for mic pre input filtering where the inductor presents a different impedance to CM signals than normal signals (actually a transformer of sorts IIRC).

I haven't thought about a coil in emitter of mic pre's.. I Assume the feedback r is still connected direct to the emitter? Probably no big difference either way..

While I like to keep the RFI out before the active devices, reducing the closed loop gain at HF helps the opamp keep up with AM radio slew rates.

More stuff I haven't investigated... good stuff

JR

[juniorhifikit]

I have never personally messed with a 990 but I suspect they are separate inductors. I would have to look closer to see if there are normal vs. CM considerations that argue against using a common core.

Two separate cores in the 990 (I used to assemble them for John!)

[mediatechnology]

I have seen a single core variant for mic pre input filtering where the inductor presents a different impedance to CM signals than normal signals (actually a transformer of sorts IIRC). JR

I think Benchmark used to sell one and used it in their preamps.

haven't thought about a coil in emitter of mic pre's.. I Assume the feedback r is still connected direct to the emitter? Probably no big difference either way.. JR

Yes Ra and Rb return directly to the emitter. The Ls are in series with Rgain only.

Checkout figure 6 "RL2" in the Hardy pdf: http://www.johnhardyco.com/pdf/990-2007.pdf#page=5

[JR.]

Yup.. It is a little more straightforward brute force than my MFB LPF, but I like clever too...

Worth bench testing to see how much C it can isolate.. I have made consoles with more than 100 feeds to L/R so that many VCAs could be > 1000pF on bus, admittedly an extreme case.

JR.

[juniorhifikit]

So I was looking at this SSL-type summing amp that I've been fiddling with, and I see it's very close to what's inside a 990, but I have a few questions about the differences and how it all relates to the current summing concept.



If we remove the I-V stage after the VCA (5534); remove the summing resistors (15K); remove C1 & C2; and leave L1 in place (inductor/ferrite bead), we're essentially there in theory?
And what's going on with R24 & R25 and those caps (similar to R26 & R58 + C29 & 30 back to back)?

-Scott

p.s. not sure why the image is displaying oddly - opening in a new window looks ok

[mediatechnology]

If we remove the I-V stage after the VCA (5534); remove the summing resistors (15K); remove C1 & C2; and leave L1 in place (inductor/ferrite bead), we're essentially there in theory?

Yes, I think so. You might could leave C1 and C2 if you wanted to.

I was just thinking today that the SSL 82E26 card used input inductors. http://www.ka-electronics.com/images/SSL/ssl_82E26.pdf

And what's going on with R24 & R25 and those caps (similar to R26 & R58 + C29 & 30 back to back)?

That's a biased bipolar capacitor with the middle biased up to about 1.6V. Trevor also did that in the 82E149 mic input. In the 82E149 it makes a lot of sense.

http://www.ka-electronics.com/images/SSL/ssl_82E149.pdf