Help running THAT1200 on single-sided supply

[chris p]

Help!!

I'm mapping out a design for a super JLM BA - based around the BA Dual, but with vari-Z and HPF, LED peak metering and a balanced insert point.

The BA's run of a 48V DC single sided supply. I'm planning on a darlington - zener combo to drop this to around 36V DC for powering the insert point circuit, which will utilise THAT Corp's 1200 and 1646 ICs, but these are based on a bipolar supply (max +20 to -20V).

Now Prodigy has a schematic for a single-sided supply for the 1646, but not for the 1200. The 1646 circuit creates a 1/2V voltage buffered by an opamp to provide a low impedance ground reference - see http://www.prodigy-pro.com/diy/index.php?topic=31692.0 - and the DC is then removed from the outputs with an electro cap.

I'd like to adapt this to the 1200. In looking at the equivalence circuit on page 1 of the 1200's datasheet http://www.thatcorp.com/datashts/1200data.pdf, my very lowbrow interpretation is that OA1, 2 and 4 are all about the CMRR bootstrapping, and the actual differential engine that converts balanced to unbalanced is OA3. Providing a low impedance 1/2V reference at the Ref pin (pin 1) biases the inputs and maintains a low Z ground point for the bypass resistors, and the added DC can be stripped by a 220uf or so cap on the output. In other words, use the same 1/2v reference circuit from the prodigy circuit to feed the 1200's pin 1 reference, and strap an cap on the pin 6 output - voila, single sided power operation.



Before I experience the delights of 36V DC zapping through my body and the delightful pop of exploding capacitors, can anyone confirm, deny or comment?

[rob]

yep that's pretty much it, Chris

the key is to take Vref to 1/2 Vcc. This can be as simple as say two 10K R in series from Vcc to 0V and the Vref taken off the junction of the two R. Then a bypass cap from this point also to ground, say 220uF. Or do as is done in the example for the 1646, whereby the Vref is buffered by a opamp. I would also tend to decouple the opamp output with a small R and a large C as any noise on the opamp output will be injected at unity gain through to the 1200s output.

Both the inputs and outputs will need to be coupled through caps, with the +ve side facing the 1200 and the -ve side pulled to ground ( say 47K ).
All other implimentational aspects re the CM 220uF cap and the RF protection ( if you want ) remain as per the datasheet
Do not use fig 15 design for the input coupling caps. This won't work

btw all this is from a quick scan of the datasheet and a bit of scrambled thinking....no promises :)

[chris p]

Cheers Rob - I appreciate your input

So the circuit is now looking more like this



What's the thinking on the 1u cap in the opamp output decouple - maybe a 10u like the decouple on the initial voltage divider?

Yeah, no promises, clearly understood. I'll breadboard this up and switch on from a safe distance, like all those nuke tests.

Thanks again.

[rob]

that's a 1nF not uF, purely there as a HF stability / noise filter thingy, should be fine.
Increase the cap on the output of the 1200 to 100uF, keep that -3dB down into the sub Hz region. The value of this coupling cap is one of the things that seems to separate the men from the boys. Whilst the maths tells you that using even a 10uF when a typical circuit load might be around 10K gives you a -3dB point of 1.6Hz and you'd think that would be fine. When examining what real consoles use, you'll find that 10uF or sometime 22uF in many consoles, however in SSL and Neve consoles you'll find 100uF! They were thinking more about phase shift at 20Hz not freq response.