Active DI box running off phantom power

Go To Last Post
8 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've got a mic point, XLR, at some venue. I want to run my notebook's audio out into it, but connecting ground and audio into the point directly results in the audio being minimally above the noise floor. This, I suspect, is because the notebook's audio cannot source enough current to bring the output signal levels to above the noise floor.

I've attempted to use a pair of opamps, one in inverting and the other in a non-inverting configuration, to basically build a DI box. The quick veroboard prototype with LM741s work, but that was with a separate power source. I want to run the box off phantom power, saving me the need to carry along replacement batteries.

Attached is my initial design, which I suspect may have some kind of problem, as potentially 24V may be feeding back into the opamps when the cap is discharging the other way. I'm asking for another set of eyes to scan that and find any further problems, before I go etch a board.

And before anyone mentions it, yes, I'm perfectly happy with the performance of the 741 in this application, and I do not happen to have alternate parts sitting around.

Attachment(s): 

Last Edited: Sun. Feb 22, 2009 - 10:06 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For those unaware, the XLR connector carries floating antiphase audio on pins 2 and 3; pin 1 is a ground. In a phantom power mode, the idea is to apply positive power to both pins 2 and 3 and apply audio to that signal

I think on this design, Tim, you've failed to provide a midpoint 'earth' for the opamps. That means your opamp outputs will be around 0v rather than around 12v where you need them; you have no negative extension on the signal.

Have a look at the designs here: http://sound.westhost.com/projec... - particularly the last one. Similar, but with the extra reference.

Neil

(ePanorama is great for pro video/audio stuff, I find)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Oops, forgot to connect a resistor somewhere.

Hmm, I thought I provided that with splitting the +24V to GND reference with R10 and R11. AGND is connected to the audio ground, and should place the zero input at the +12V level. Or so I assume. C1 should remove the DC from that, and send both opamps to GND level when the input is idle, no?

The last design there appears to have both opamps with differing gain, but with the same phase. I don't quite like that idea, as this mixer doesn't quite like me when there's DC between the balanced input during idle.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I failed to notice the agnd points - but even so, a single-ended output will be relative to the ground even though it may be above or below that ground. And the references to your opamps are still referenced to the ground from the phantom power; they need to be referenced to the mid-rail.

Elliot's design does have the opamps antiphase; the input is provided to the unity gain non-inverting U1A and the output of that goes to the unity gain inverting UA2 - R2 is providing a mid-rail reference to bias the input of U1A, not a signal path to U1B, I think.

I note he uses a very low current through the divider, too; that's to keep the load on the phantom power down, I suspect; you could probably get away with a similar splitter. The capacitors across those dividers are a good idea, too.

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Now that you mention it, yes. I have completely failed to notice how U1B is wired up. I'd like to blame Sunday afternoon syndrome though. :D

I will take a closer look at Elliot's design, having had a nice night of sleep. Thanks for the advice, Neil.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It looks like you are using ring for ground rather than sleeve in the tip-ring-sleeve j1. This works with a mono plug, but not with a stereo plug. If this isnt part of the problem, sorry. Just thought I'd point it out.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Strange that you should mention that. I got very confused with kicad when I first saw that symbol, but it's connected to sleeve and tip. The third pin is for jack detect.

I'm probably going to borrow the ESP design, as soon as I figure out why the inverting opamp is connected through R4 to the output of U1A, instead of the net at the positive side of C1.

I refuse to build something I fail to understand completely. :D

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm not sure there *is* an absolute definite reason for doing it that way; or at least, that there is any great advantage in so doing it.

My thought is that by taking that point, it's a guaranteed low output impedance point; if he took it from the input - either side of R3 he could affect the half-rail biasing and perhaps the input impedance of the circuit. But I haven't analysed it in detail.