Piezo / Mag preamp schematic, will this work ?

[NoFi]

Hello,
Hello i'm trying to build a small phantom powered/balanced onboard preamp for guitar with piezo and magnetic inputs.

Schematic :


I'd be greatful if you have some feedback/suggestions before i start building it.
I may have made some mistakes.
Are the protection diodes really necessary ?
Regards,
Greg

Edit : the R8/R9 resistors are actually 100 ohms on the balanced output.

[anotherjim]

It looks well thought out to me.
I would keep the diodes it looks to me there is quite a strong switch-on pulse thru the caps on the amp outputs until power stabilizes.
A small point, U1c is inverting, so isn't the inverting U1a really the pos/hot/XLR2 phase?

[amptramp]

The input impedance of the MAG input is essentially R3 which is a bit low at 100K and will limit the high frequency response.  Another stage like U1D may be the answer.

[NoFi]

Thank you so much for taking the time, you both made some very good points.
100k is definitely too low for the mag input, i can't beleive i didn't consider this.
I'm limited by the quad amp, and dont want to mess around with fets...

Maybe i'll try two follower input stages in parallel, and some sort of passive mixing resistor arrangement before the balanced output.
A blend knob could be useful here.
That will also get rid of the inverting stage.

But there might be some other considerations for the passive mixing relative to each branch of the balanced output... yet i think it's worth trying.

[samhay]

You do not have any voltage gain here, yet the output from a piezo and a magnetic pickup may be vastly different.
If the idea is to use this to mix the 2 signals then I would play with the gain so that they have a similar level.

[PRR]

> You do not have any voltage gain here

Moreover there is 100:1 *loss* in the output network. As Microphone inputs are more sensitive than most pickups need, some loss may be OK. But 100:1 seems a lot.

I do not understand why separate inputs for mag or piezo. Mag will drive 10Meg fine. If an ill-tuned mag pickup "needs load" to avoid screech, switch 100K across the 10Meg.

The proposed 2K output impedance is perhaps high for some Microphone inputs. 200r would be a better hiss-fit. Assuming output loss will be near 10:1, then the output networks may be 1K:100r. For the expected levels, most general-purpose opamps will drive ~~1K OK.

[samhay]

>Moreover there is 100:1 *loss* in the output network.
That threw me too, but from the OP:
>Edit : the R8/R9 resistors are actually 100 ohms on the balanced output.

[PRR]

http://oi63.tinypic.com/5wju51.jpg

[NoFi]

Samhay >  Thanks, Yep, i was planning on possibly using a pot before the summing amp to tame the piezo.
The magnetics already have the usual volume and tone.

It's 100ohms resistor on the output (my bad).

PRR > Thanks for your modifications, basically what's the advantage of having the two op amp follow each other for the balanced circuit ?
I've noticed this configuration on several schematics but i 'm not sure why it would be better since one side goes through an additionnal op amp ?
Thanks.




I'm really not sure about the passive mixing thing at the middle  .

[PRR]

> one side goes through an additionnal op amp ?

So? A TL072 inverter is practically perfect for audio use.

Also your first plan runs Piezo through three stages, for no apparent reason.

I also think a TL072 is easier to manage than a TL074.

[Phoenix]

>Moreover there is 100:1 *loss* in the output network.
That threw me too, but from the OP:
>Edit : the R8/R9 resistors are actually 100 ohms on the balanced output.

R10 and R11 should also be changed to a higher value. They should be an order of magnitude larger than the input impedance of whatever you're driving. That could be 600 ohms, or up to 10k for a balanced input. I'd recommend changing their value to 100k.
I'd also recommend changing U1A and U1B to an opamp that can drive lower loads than a TL07x. A TL07x isn't too happy about driving anything lower than about 2k, and is much happier driving 10k. A better choice would be NE5532, which will quite happily drive 600 ohms, and is the classic choice for balanced line drivers. You should still use a TL07x for U1C and U1D for their high input impedance though. This is still only two 8 pin DIP packages, so should suit you fine.

Have a read through this link.

As has already been suggested, you may need some gain, especially on the piezo input, as their output is usually significantly lower than a magnetic pickup. You will need to test the individual piezo you have though to know how much gain you might need.

Good luck with your project.

[Phoenix]

Here's some recommendations. I've changed the mixing to active and added gain controls. I've also changed the driver opamp to an NE5532 rather than TL07x for it's superior driving capabilities. If the piezo requires gain, that can be done on the input buffer. I left out your protection diodes, I think they're overkill. Hope this helps.

[NoFi]

Thank you very much for your input !
Your solution with the all inverting 5532 solves the mixing and if it's better at driving loads then i should probably use it !

I should also probably add some low pass filtering in the op amp feedback loops R5 / R7.

I'll see if i can fit everything on a small PCB.
Regards,