Transformer DI for modulars

Don Tillman don at till.com
Tue Dec 26 01:48:33 CET 2000


   From: DCMagnuson at aol.com
   Date: Sat, 23 Dec 2000 13:38:33 EST

   I'm considering making a DI module for the output of my modular
   synth.  Basically, I'd like to have a small mixer feed into a
   transformer so I can provide a balanced output to my mixer.

The ARP/Rhodes Chroma does this -- stereo balanced outputs with XLR
connectors in addition to the main outputs.  On the Chroma each of
these outputs is a floating transformer driven by an opamp follower
with extra output transistors.

   1) Can anyone recommend a suitable transformer?  I imagine I'd want
   a high input impedance, and around 600 ohm balanced output.  I hear
   good things about Jensen trannys... they are supposedly quite
   clean.

I don't have an answer here, but do check out the Jensen Transformer
web site for a huge collection of application notes and schematics of
balanced input and output stages.   http://www.jensentransformers.com

   2) I figure I'll need some attenuation to drop the signal levels
   from 10V p-p to standard line level.  Should this pad be put before
   or after the transformer?

No, you want the transformer winding ratio to do this for you.

   3) What precautionary measures should be taken to avoid damage from
   phantom power?  My mixer has globally selectable phantom power, so
   I need to protect the DI from damage.  Will the current limiting
   resistors in the mixer provide enough protection to avoid smoking
   something?

You're talking damage from accidently connecting to a phantom power
input, right?  As opposed to running into a phantom-powered input as a
regular activity?  I figure the latter won't be happening as folks are
eager to keep extra mic preamps out of the signal path.

The typical phantom power input is a 6.8k ohm pullup resistor to +48.0
volts, for each of the inputs of the pair.  I would diode-protect the
output signal for excursions above the supply voltage.  The output
devices should have no major issue with the 7mA.

   From: "Tony Allgood" <oakley at techrepairs.freeserve.co.uk>
   Date: Sun, 24 Dec 2000 09:56:02 -0000

   I have just done this too. I used a EBOS, 'electronically balanced 
   output stage' instead of a transformer. This is basically three
   op-amps. 

It's a three-op-amp circuit that emulates a floating transformer
(which I think is a fascinating engineering problem!).  The use of
positive feedback in the classic three-opamp approach makes me a
little skeptical -- I wonder what situations would compromise the
stability. 

Check out the Analog Devices SSM-2142; it's a balanced driver
output stage in an 8-pin chip.
http://products.analog.com/products/info.asp?product=SSM2142

This is used in a lot of products.  For example, the Rane AP13
Acoustic Preamp, schematic at http://www.rane.com .

These balanced drivers are a little weird.  They try to impress the
input signal across the two outputs while letting both outputs float
up and down together.  So you can't just use, say, the positive output
of the chip and leave the negative output floating, you also need to
connect the negative output to ground.

I'm pretty sure Mackie mixers use these (or something similiar) on
their outputs because of this characteristic -- the output is a
1/4-inch stereo jack, balanced or unbalanced depending on whether you
use a stereo or mono plug.  Use a stereo TRS plug and you've got
tip=positive, ring=negative, sleeve=chassis.  Use a mono TS plug and
you've got tip=positive, sleeve=chassis, and the negative output is
tied to chassis ground through the mono plug.  (You couldn't do this
with simple differentially driven opamp outputs; you'd end up running
an opamp into a short.)

Also, I figure a balanced driver stage would have seriously degraded
performance driving a phantom power input; both outputs would slide up
toward the positive rail.

  -- Don

-- 
Don Tillman
Palo Alto, California, USA
don at till.com
http://www.till.com




More information about the Synth-diy mailing list