Bauman Neve output in a PRR varimu..

[Svart]

I haven't done it yet but i really need balanced outputs from this and have some of Fabio's neve out boards that are half populated. So who would like to know the outcome? i need to get transformers first though. I would also like to maybe bypass the dual opamp if possible but I haven't really had the time to investigate, this is just thoughts.


any comments?

 

[JustinS]

I'd love to know the outcome, I've been looking at doing one of these for ages (along with all the other 1/4, 1/2 & 3/4 finished projects I have going), but would really need one fairly happy in +4 balanced land without workarounds. Varimu with Neve output... :grin: :grin: Mmm good!

Justin.

 

[Svart]

The PRR varimu was cheap and easy to build. The audio is clean and the unit works great as is! I would recommend building one immediately! Run to your local Radio $hack and find the oldest filters you can as there are reports of the ones you buy online being newer and spec out differently.

 

[PRR]

> Fabio's neve out boards

Post scheamtic.

 

[Svart]

oh wow, I went to find the page and Fabio's site is down.. suspended actually.. I have the hardcopy but don't have a scanner.

Let me see what I can do.

 

[PRR]

Or whichever "Neve" the Fabio plan is based on; does not have to be exactly the Fabio scheme.

 

[Svart]

http://www.jlmaudio.com/neve_ba283.htm

second schemo down. the output section.

that's the closest I could find.

I have not had time this week to do more than ponder the idea.. :?

 

[PRR]

Feed the secondary of the plate transformer to BA283 output-stage's pin L and J.

Strap 220Ω from pin K to pin J.

Connect the 5532's input to pin K. (There are other ways, but this saves thinking.)

Obviously the BA283 needs the usual external connections: power, signal ground, and pin A to power return.

 

[Svart]

thanks PRR.. I'll figure out Fabio's connections VS. the 283's connections and post the final hookup of the unit with output.

 

[Greg]

Fabio's board is the AM stage of BA283. He also have external gain control already built into his board by jumpering GAIN and RG, then selecting C4 and RGx per application.

Neve Out Schematic

Once suggestion: I'd bump up the capacitance of C3 on his schematic from 1000u to something more like 2200u-4700u. A 4700u/35V Nichicon PW fits in his board just fine. It helps LF response.

 

[Svart]

interesting..

I would likely use this similar to the original, with a set output and vary the input of the Varimu to change gain and compression so I am unsure how much gain i would need just yet. I suppose trial and error will have to do.

 

[Greg]

IIRC, the BA283AM has 15dB of gain without the extra control. Just leave the jumper out, C4 and RGx, and you'll accomplish this. :thumb:

Go to page four of the following link... good info about the BA283AM.
http://www.technicalaudio.com/pdf/neve/0283-allvers.pdf

NOTE: The part numbers on the original BA283AM and Fabio's board is differenct.

 

[Svart]

"Connect the 5532's input to pin K." I'm sort of confused by this as the NI "input" pin of the 5532 would already be connected to L correct?

I assume this is actually removing the 5532 from the circuit or am I just paralleling the two? I see now that paralleling would make much more sense.

also, 220R between K and J would now be in parallel with the 1k2 (r17 on the fabio board). should this be so or should the 1k2 be taken out?

 

[PRR]

> 220R between K and J would now be in parallel with the 1k2

Yes, I saw that. The original limiter has postamp gain of 10 and gives -10dBV, we want +4dBm so we need 4 times more gain. I figured 170Ω+3.3K and allowing for the existing 1.2K that gives around 220Ω. However I was using a wrong assumption for output iron gain: the proper value may be closer to 150Ω. Try it and see. Because it is not a brick-wall limiter, you will have to crank it up with your program material and your desired squish and see what comes out, then adjust the gain-set resistor if it is way off from what your recorder needs.

> "Connect the 5532's input to pin K." I'm sort of confused by this as the NI "input" pin of the 5532 would already be connected to L

No. You expressed an interest in getting the chip out of the audio path. I was thinking that we could take the 5532 off the transformer, and get the same signal at the 283's feedback point. Loading here upsets the 283's gain, but the impedance is ~220Ω which will not be upset by the >>100K impedance of the 5532 input.

BUT. I forgot there is DC bias at pin K. Enough to throw the 5532 to the rail. We could fix that with a acoupling cap, but the 5532 needs fairly small input bias resistors so we come near 1uFd, which is big for film-caps. Since it is only the sidechain, we could probably use 5uFd electrolytic and 10K resistor....

But I've never been real happy with that sidechain. And now that your main power is a +24V, it would be nice and may be possible to design-away the +/- power supply.

One little serendipity: with a fairly hefty 24V supply, you could wire stereo 12AU7 heaters in series and feed them from the same raw supply.

 

[Svart]

just a thought.. what IF we were to leave the 5532 in the loop but parallel the Neve output stage with it.. and *not* use the normal output but rather just the new output stage? I don't mind the IC being there...

Fabio's version does not have point K like the original circuit.


Other than that I do like the idea of merging the varimu and class A output.. especially since I have a box of 200va 30vac toroids.. :green:

 

[PRR]

> leave the 5532 in the loop but parallel the Neve output stage

That's by far the best easy answer. The plate transformer output is only a few K impedance, it can easily drive both the 5532 input pin and the 283's input, and neither input is flawed enough to upset the signal. Or go for TL072 in place of 5532: it isn't in the audio path, is good enough for the sidechain, and injects nanoAmps less noise-current. Which may not matter, but you also save a dime and one part-number to buy.

The 12AU7 grid needs just enough control voltage that it would be hard to do under a 24V supply.

> Fabio's version does not have point K like the original circuit.

??? It almost has to have a tap at the first emitter, to adjust gain. Otherwise gain is stuck at ~5.6. If there is a gain control between the Preamp Section and the Output Section, then the output really needs gain over 10 or 30. For general-purpose use, Neve set the card gain low but gave a place to crank the gain to the needs of the specific application socket and studio specification.

Huh. That sheet does say "T1... picked for gain higher than 600 to minimise input noise". I just "proved" that anything over 300 is as good as it gets. If Neve wanted better noise, improvement would come from changing R1 120K and R2 68K, as well as soaring Hfe.

 

[Svart]

it can easily drive both the 5532 input pin and the 283's input

I was hoping you would say that.. but wasn't sure myself.

What if we were to rectify the +- rails from the main trafo to get the source for our +24 rail? I don't remember the specs of the trafo i used off the top of my head but it was a rather large part.. I know you aren't a fan of regulated voltages for IC's but would you think this neve output would need much more than smoothed raw DC?

I guess i should clarify, Fabio's board has a provision for increasing gain that does tap the traces that "K" would usually be routed to.

 

[PRR]

 

[Svart]

truly a gentleman and a scholar!

 

[PRR]

You understand it? I'm not sure I do. I was hoping the sidechain geeks here would put their bootprints on it.

For stereo, the stuff lower-right of red line is duplicated, and of course there are two grid transformers to feed and two 12AU7s to hold up above ground.