Modifying DI

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

warpie

Well-known member
Joined
Feb 7, 2009
Messages
1,566
Hello,

I have some Whitsle Rock Audio DI PCBs and I'd like to rack them in one enclosure but they need some modifications.

Ideally I'd like them to operate at +22V  (or + 24V ) instead of +16V and also I'd like the output impedance to be low enough so as I can plug them into my preamps' mic inputs. So maybe around 100-200Ohms.

(FWIW, almost all my preamps have transformer on the input but I don't think it really matters in this case)

Can you please help me as I'm not quite sure what resistor values need to be changed?


2mweago.jpg
 
I think this is the origin of that circuit:
Ray Marston JFET principles

See part 2, figure 8.

And it mentions  +12v to +20v. I actually did a API325 clone for eurorack (modular synths) and it has this impedance converter for guitars and such, still need to test it, but I am thinking of running it on 24 volts.
 
Thanks Jarno!

It does look very similar indeed.

Do you know if the output impedance is simply set by R9?  If so, can I change R9 to say 150 Ohm in order to be able to plug it straight into my preamps?

I believe my assumption is correct but it would be great if someone could confirm.

EDIT:
Jarno said:
I think this is the origin of that circuit:
Ray Marston JFET principles
Very nice article, thanks  :)
 
warpie said:
Thanks Jarno!

It does look very similar indeed.

Do you know if the output impedance is simply set by R9?  If so, can I change R9 to say 150 Ohm in order to be able to plug it straight into my preamps?
not even close... R9 sets output impedance only when power is off. The output impedance is dominated by the emitter follower.

More importantly the drive capability (what impedance it can drive) will depend on other factors. In the positive direction it can source a lot of current, but pulling down is limited by R8.

A good simple circuit to learn with.

JR
I believe my assumption is correct but it would be great if someone could confirm.

EDIT: Very nice article, thanks  :)
 
warpie said:
Do you know if the output impedance is simply set by R9?  If so, can I change R9 to say 150 Ohm in order to be able to plug it straight into my preamps?
No. R9 sets the load impedance. Not the source impedance. The source impedance is defined by Q3 and R8 and it's already pretty low. Probably lower than 150 ohms. It depends because the circuit is non-trivial. There's no feedback. If there were the output impedance would be very low. But as it is, it's just a simple emitter-follower buffer which means the output impedance is vaguely the source impedance of Q3 divided by the beta of Q3 although that doesn't tell the whole story because it will have limited asymmetric output swing because it can source current well but it can't sink a lot of current.

But for small signals the output Z is going to be lower than 150 ohms already so you can just use it as-is.
 
I don't think it needs the 1uf bootstrap cap for a DI input.
The CC in the source leg is nice
ALSO
Read about what happens with the Vdg goes up with certain jfets.
If you change from +16 to +24 what happens?
 
Gus said:
I don't think it needs the 1uf bootstrap cap for a DI input.
Agreed; it is needed only if the 10 Meg option is selected. I never felt the need for such a high value of input impedance. It is claimed to offer extended LF response to piezo p/u's, but at the end of the day, there is so much LF noise you need to hi-pass the signal.

Read about what happens with the Vdg goes up with certain jfets.
If you change from +16 to +24 what happens?
Not sure what you imply there... V(br) of 2N3819 is spec'd at min 25V, typ 35. I don't think there's any serious risk.
 
warpie said:
Can you please help me as I'm not quite sure what resistor values need to be changed?
The main issue there, as others have already mentioned is the output stage operating point. As it is running with about 1.5mA quiescent current, the output voltage into a typical mic input (1.5-2k actual impedance) is restricted to about 700 mV (about -1 dBu), and even that would be achieved with not unsignificant distortion. An electric guitar can be challenging, a bass even more.
I would increase the quiescent current, by decreasing R8 to 3.3k and increase R9 to 100k.
 
Thank you all.  I'll replace the resistors suggested and I'll see how it behaves.

Is there any good way to take some measurements (i.e SNR, freq response, etc..) of  the input without using an actual guitar or synth? I have both a sig generator and a soundcard.
 
warpie said:
Is there any good way to take some measurements (i.e SNR, freq response, etc..) of  the input without using an actual guitar or synth? I have both a sig generator and a soundcard.
The simplest most direct way is to use a soundcard (or preferrably a fancy USB audio interface) and some software. There's lots of different software out there. I used to use something called DSSF3 from ymec.com but a lot of people recommend Room Eq Wizard.

To use the software, there are usually different measurements that you can "run" that put out a tone or noise or an impulse response or some kind of stimulus. You pipe that in the Device Under Test (DUT) and then pipe the output back into the sound card and then play with levels until you get the right behavior.

It can be tricky getting the levels and impedances right though. You have to think about it and probably make some special cables. For example, for a mic pre, you need to attenuate probably -40dB or more and make the source impedance ~150 ohms. Doing that correctly is not trivial.

But your DI is pretty much unity gain and high impedance in and pretty low impedance out so you can probably just connect it to the sound card / audio interface as-is.
 
warpie said:
Thank you all.  I'll replace the resistors suggested and I'll see how it behaves.

Is there any good way to take some measurements (i.e SNR, freq response, etc..) of  the input without using an actual guitar or synth? I have both a sig generator and a soundcard.
To make meaningful measurements you need to terminate the input with a passive load similar to the expected source impedance you will be buffering.

For a circuit this simple I would confirm frequency response and signal headroom. Noise from a unity gain buffer should be quite low.

JR
 
Sorry I wasn't very clear. I do have all the necessary hardware and software to do the measurements. My question was more related to the termination resistor(s).

So perhaps I should terminate the interface's line output with a 10K resistor and plug it into the DI's input in order to mimic a passive pickup? 

Once again, many thanks for all your help  :)
 
warpie said:
Sorry I wasn't very clear. I do have all the necessary hardware and software to do the measurements. My question was more related to the termination resistor(s).

So perhaps I should terminate the interface's line output with a 10K resistor and plug it into the DI's input in order to mimic a passive pickup? 

Once again, many thanks for all your help  :)
A passive electric guitar is more like 50k (Fender)-100k (Gibson), but providing an adequate source impedance is relevant only to noise measurements. For frequency response and headroom, you can just connect directly the soundcard output to teh DI input.
 
warpie said:
Can you please help me as I'm not quite sure what resistor values need to be changed?

I would do next changes:
- decreasing or leave out C2,
- decreasing  R6 to 2k2 or 3k3,
- increasing R7 to 15k,
- replacing R8 with current source set to 5mA (additional BC550  +  150 ohm emitter resistor)
- increasing C3 to 100uF,
- increasing R9 to 47k,
- and maybe omit C2.

Also  an attenuator of 20dB or so at the output may be needed.



 
moamps said:
I would do next changes:
- decreasing or leave out C2,
- decreasing  R6 to 2k2 or 3k3,
- increasing R7 to 15k,
- replacing R8 with current source set to 5mA (additional BC550  +  150 ohm emitter resistor)
- increasing C3 to 100uF,
- increasing R9 to 47k,
- and maybe omit C2.

Also  an attenuator of 20dB or so at the output may be needed.

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

Also, I redrew the circuit. Do you think it's correct (especially the current source you suggested)?
 

Attachments

  • dι.jpg
    dι.jpg
    100.7 KB · Views: 16
warpie said:
Also, I redrew the circuit. Do you think it's correct (especially the current source you suggested)?
Definitely not corrrect; you've drawn a Darlington, not a CCS.
Please note that the input impedance without C2 is about 2Megohms. With R1 switched on, it results in about 600kohms, which is correct for most applications. With R2 switched on, it becomes about 1.6 Megohm.
My suggestion was to get rid of C2 AND R1, R2 & SW2 altogether. Having a 10 Meg input impedance looks nice on the spec sheet, but it's useless.
 

Attachments

  • Output CCS.jpg
    Output CCS.jpg
    103.7 KB · Views: 5
Right, so hopefully now it looks OK?

"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?

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

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

FWIW, I'd like to use this DI for both passive/ active pickups as well as for synths.

thanks for all your help. It's much appreciated!
 

Attachments

  • di2.png
    di2.png
    78.7 KB · Views: 14
Back
Top