Modifying DI

[moamps]

Abbey, thanks for assistance. You started to drawing blueprints, that's something new 

Could you please elaborate a  bit? Why these changes are needed?

R6 change -  increasing the current of the left CCS, to get slightly better working point of FET when VDS increases. 
R7 change - keeping the current thru diodes at same level when PS voltage is higher
R8 replacing to CCS - CCS has for AC very high impedance, it helps to get higher headroom
C3 change - lowering LF cutoff  point
R9 change - it loads the output so it should be high, just enough to keep C3 - pol DC wise on ground

 

[abbey road d enfer]

"Having a 10 Meg input impedance looks nice on the spec sheet, but it's useless."
Useless in the sense that it's unnecessary or that it'll affect the circuit in a negative way?

I already said it doesn't give any operational benefit, with the disadvantage of making the input oversensitive to electrostatic interference when unloaded.

Also, what about the changes in the values that moamps recommended (underlined in red)?

I agree globally with them. R7 may need (or not) to be adjusted considering it biases two CCS instead of one. R9 could be almost anything between 10k and 220k. It is just there to dump to ground the DC leakage of C3.

BTW, I simmed  this circuit and it seems to have very high headroom (around 5Vrms or so).

Have you done the sim with a relevant load (1.5-2k)?

As long my results are correct, I don't think it needs an input pad but does it really need a pad at the output?

It depends on the mic pre that you connect the DI to. Can it tolerate a +4dBu level at the input?

 

[warpie]

OK so I ordered the parts and since the circuit has changed a bit I might as well make a couple of new PCBs or maybe even veroboard it.

However, it occurred to me that if this circuit is plugged into a pre where the phantom is (accidentally) ON, it'll get damaged since it's single ended? Is this case and if so, is there any solution?

 

[moamps]

Just use bipolar type for C3 rated for 50 or 63V and everything will be fine.

 

[warpie]

Simple as that.. thank you moamps 

One last question if you don't mind. I noticed that by decreasing R10 (150R) to something like 100R, I get a 3-4dB better headroom. Is it alright to do so?  FWIW, I'll be mostly using it with a 1073 so the pre's input impedance would be around 1.2k

As far as the pad is concerned, I think I can skip it from the DI and  use my pre's pad instead?

 

[moamps]

You can do experiments with R10. If you use  100 ohms, the CCS current will be about 7 mA.
And the pad at the output of DI isn't needed if your preamp can handle high signal at its input.