Help me change this DI input to 10Mega impedance

[Whoops]

Hi,
this is an interesting DI project from member Kevin.

As stock the input impedance is 1 Mega, could you please help me out in modifying this circuit for 10 Mega input impedance?
What components should be changed?

The IC is a TL072

Thank you so much

 

[Bo Deadly]

The impedance is not 1M. It looks like it's only a little more than 4k7+510. But there are other issues. I would choose another circuit entirely IMO.

 

[Whoops]

But there are other issues. I would choose another circuit entirely IMO.

Could you please explain what are the issues you see? and why would you choose another circuit?

If you know an active DI circuit that you would recommend please let me knot

thanks

 

[abbey road d enfer]

There's nothing very wrong about the active part of the circuit (except R10 saying it's 1 milliohm (m is for milli, M is for mega).
It's the rest, all the attenuators that load the input.

 

[Bo Deadly]

Could you please explain what are the issues you see? and why would you choose another circuit?

There are so many problems I'm not going to break it all down but the origin of most of the problems is that the schematic is just horribly incorrect. Where did you get this schematic? There are dots on nets to indicated that they are connected where they are not supposed to be. And all over. Not just in the pad network.

But the problems are not just with the schematic. The feedback of the first op amp yields a gain of 3 at DC but is unity at AC. The LED network looks like it's rectifying the signal but in a way that could distort the output and then maybe change the bias on the second amp. The battery doesn't make any sense. Presumably there's a virtual ground in there somewhere.

This circuit is actually kind of interesting because it's so bad that it's good. It could be used in some kind of class about how to not design a circuit. It's like a funny puzzle circuit where you're challenged to find all of the things wrong with it.

 

[JohnRoberts]

I suspect the schematics were published somewhere but back in the 1980s I designed a direct box based on TL07x for Peavey (AMR EDB-1).

Basically pretty simple non inverting amplifier using high impedance bifet op amp. Some extra complexity to grab power from mic phantom power. Output resistively padded down to mic level.

The input impedance is basically the resistive termination. IIRC I probably used 3M. I don't recall any complaints about that being too low.

JR

 

[Bo Deadly]

I found the source of the schematic:

P96DI + A72DI Direct Injection Boxes

and it really is that horribly bad.

I have to guess that maybe they added a bunch of nonsensical connections and ground symbols to obscure the design. But I have no idea why they would do that for what should be a very simple circuit. Why put the schematic in the build guide at all if you're going to deliberately make it completely nonsensical?

 

[abbey road d enfer]

The feedback of the first op amp yields a gain of 3 at DC but is unity at AC.

Think it over. Tip: C2.

The LED network looks like it's rectifying the signal

How?

but in a way that could distort the output and then maybe change the bias on the second amp.

?

The battery doesn't make any sense.

Yes it does. Tip: R and S of input jacks when a mono plug is inserted.

Presumably there's a virtual ground in there somewhere.

Yes. R11, R13, C6.

This circuit is actually kind of interesting because it's so bad that it's good.

Actually it looks suspiciously similar to a DI box I designed in 1976, and sold about 40 000 in all its different iterations. People must be silly, buying things that are so badly designed.

It could be used in some kind of class about how to not design a circuit. It's like a funny puzzle circuit where you're challenged to find all of the things wrong with it.

I agree that the input attenuator and the output to the XLR are wrong. I can assure you my version didn't look like this.

 

[abbey road d enfer]

and it really is that horribly bad.

This company IS bad. they select good products, probably think they understand what they do, and butcher the design.

I have to guess that maybe they added a bunch of nonsensical connections and ground symbols to obscure the design.

I can assure you that it makes some sense. It's working. Not as it should, but still not a complete nonsense.

 

[JohnRoberts]

If you are going to copy somebody's DI design, find a better one.

You probably won't find many with 10M input impedance bacause that is excessive for audio applications.

JR

 

[gyraf]

Discussion thread - P96DI & A72DI - Passive and Active Di Boxes in Business Card Format

 

[abbey road d enfer]

If you are going to copy somebody's DI design, find a better one.

The basic circuit is good, it's what they have grafted that's wrong.

You probably won't find many with 10M input impedance bacause that is excessive for audio applications.

That's pretty much my opinion, but some members insist that I'm wrong...
They have their ears, I have mine...

 

[Whoops]

I have to guess that maybe they added a bunch of nonsensical connections and ground symbols to obscure the design.

I don’t really think that “they” applies here, it was designed and being sold by member “Kevin” in this forum.
I also don’t think that Kevin wanted to obscure anything at all he just provided the schematic to help me out on modifying it for an higher input impedance as it’s clear in the thread link you posted

 

[Whoops]

You probably won't find many with 10M input impedance bacause that is excessive for audio applications.

JR

There’s quite a few DI options in the market with input impedance higher than 1Mega. Normally to use with Piezo pickups, there’s the Radial SB4 piezo DI with 5Mega, there’s also another Radial where you can select 1M or 10M, there’s also the countryman DI with 10Mega.
But it’s nothing new here, the input impedance of Piezo preamps are normally higher than 1M, the lower I’ve seen used by a manufacturer was 2,2M, but normally is 3M, 5M or 6M.
So it makes totally sense if you connecting a Piezo directly without a preamp/buffer to provide at least 5M in the input impedance of the DI

in Piezo duty it makes a big improvement in my opinion and many others, besides that 1M is fine for all other applications

 

[Bo Deadly]

Think it over. Tip: C2.

How?

?

Yes it does. Tip: R and S of input jacks when a mono plug is inserted.

Yes. R11, R13, C6.

Actually it looks suspiciously similar to a DI box I designed in 1976, and sold about 40 000 in all its different iterations. People must be silly, buying things that are so badly designed.

I agree that the input attenuator and the output to the XLR are wrong. I can assure you my version didn't look like this.

Ok, so I understand this circuit well enough now. The bridge at the bottom is just for protecting the amps. And the other diodes are for switching between battery and phantom. So that's all fine I suppose.

But I still don't understand the two feedback paths of the first amp. And the pad is clearly not correct. I'm not going to try and decipher what that's supposed to be doing. It looks like maybe someone tried to retrofit a pad in there without knowing what they were doing.

As for what kind of DI would be good, if want 10M then I have to assume that it's for tapping into a guitar amp input that already has a defined impedance. For that, you're pretty much committing to a specialized circuit so making the input Z adjustable to say 1M for just straight guitar or 10K for a usual line in is problematic. So for really high Z, then I would just make a specialized device just for that like a single amp with a bootstrapped input like this page from ESP:

but with an impedance balanced output instead of a fully balanced output so that it doesn't have the "tascam problem". And with the right amp the battery could last a long time depending on use. KISS.

 

[JohnRoberts]

There’s quite a few DI options in the market with input impedance higher than 1Mega.

Specsmanship... more is better so the higher the input impedance the better right? There is lots of hearsay evidence (I hear 10M is better ). This is generally harmless in use as the source should dominate that input node impedance. The largest inexpensive resistor we had in the system at Peavey was something like 3M. It might be worth spending more for a larger value resistor to stand out at point of sale from a sales/marketing POV.

Normally to use with Piezo pickups, there’s the Radial SB4 piezo DI with 5Mega, there’s also another Radial where you can select 1M or 10M, there’s also the countryman DI with 10Mega.
But it’s nothing new here, the input impedance of Piezo preamps are normally higher than 1M, the lower I’ve seen used by a manufacturer was 2,2M, but normally is 3M, 5M or 6M.
So it makes totally sense if you connecting a Piezo directly without a preamp/buffer to provide at least 5M in the input impedance of the DI

K&K Sound a piezo pickup manufacturer says

"With piezo-ceramic pickups, you will need to use the high-impedance (high-ohmic) input in your amplifier. The ideal level is about 1 Mega Ohm, if the value is less, this will result in less volume and a certain loss of high frequencies/brilliance."

in Piezo duty it makes a big improvement in my opinion and many others, besides that 1M is fine for all other applications

I am not advocating using 1M input impedance since it is effectively the same cost to use 3M or 4.7M standard parts. In fact I just looked at digikey and 10M SMD resistors are like $0.10 in onesy twosey... So do whatever floats your boat.

If you start getting crazy high impedance the the op amps input capacitance could matter, but lets not open that can of worms.

JR

 

[L´Andratté]

If you want to make a di just for piezos (like for an upright bass), I have found charge amp designs giving the best sound, there are a lot of examples on the net, best explanation I found was
here.

 

[abbey road d enfer]

But I still don't understand the two feedback paths of the first amp.

Designed by someone who does not know how to do it right, or simple mistake. Who knows?

 

[abbey road d enfer]

If you want to make a di just for piezos (like for an upright bass), I have found charge amp designs giving the best sound, there are a lot of examples on the net, best explanation I found was
here.

The main advantage of a charge amp is that it minimizes the influence of the cable capacitance, which is mainly to reduce the level, since the source is itself a capacitance.

 

[Rossi]

IMO it's a good idea to make the input impedance switchable 1M /10M

Obviously, guitar/bass pickups are designed for guitar/bass amps whose input impedance is typically around 1M. So if you want to simulate this interaction with your DI, 1M is a good value.

However, if you use your DI in combination with a guitar/bass amplifier, your DI input is in parallel with the amp input. If both are 1M, your guitar pickup now faces 500k. Which does change the sound a bit - many players won't care, but some do. To minimize this sound change, your DI should have a much higher input impedance. 10 M would be a good value as it's the largest resistor value you can get cheap.