Questions about zeners and voltage in Hansen DI

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

hg_man

Well-known member
Joined
Jan 21, 2005
Messages
56
Location
Indianola WA
I've been working on putting together a circuit to buffer a piezo pickup, powered by phantom power. Several people suggested looking at Bo Hansen's DI:

http://web.telia.com/~u31617586/

For my own understanding and for making the PCB, I redrew it with the changes Bo suggests:

Hansen_DI2.gif


Now I have two questions, possibly related:

What are ZD1 and ZD2 doing? Why are the pointing away from each other? (Or did I misinterpret Bo's schematic?)

Why does Bo show +24V in the upper center? I would think that the voltage at this point would be +48V because it's being fed from the phantom power on pins 2 and 3, both of which are at +48V.

Your help in understanding will be MOST appreciated.

Alden
 
Looks ok to me. There will be 24V at the junction of R3 & R4 to provide Q1's
Bias. Zd1 & Zd2 provide a bi-polar 12V clamp on the input for protection.
Does this help?

RonL
 
Bo's label indicates +24V at the junction of R8 and R9, not R3 and R4. That's what's got me confused. I kept trying to figure how ZD1 and ZD2 gave a 24V supply, but they are in the wrong place and facing the wrong way.

In thinking about ZD1 and ZD2, I think I get it: if the input goes above +12V, then ZD1 lets through everything above +12. It gets to ground through ZD2 because ZD2 conducts anything greater than the usual diode forward threshold. If the input goes below -12V, ZD1 is already conducting, and ZD2 now conducts also because its anode is more than 12V below its cathode.
 
You have Q2 upside down. It is a PNP and should be connected as an emitter follower.

The (roughly) 24V as Ron points out and as you've reconciled indeed has nothing to do with the zeners, which are merely there as input overvoltage clamps. The current drain of the circuit pulls down the 48V upside of the 6.8k's to about 24V, hence the notation.
 
You might want to take the - of C2 and connect it to the emitter of Q1
this will bootstrap the input so the input R is much higher to work with piezos

You could even make C2 a film.
 
Gus: Make C2 a film as in polypropylene or polyethylene? Same value?

Is the connection between C2 and the Q1 emitter INSTEAD of to ground, or in addition to it? No, wait, thinking while typing... That would be instead of going to ground. If the lower side of C2 was connected to Q1 emitter AND to ground, then R5 would no longer be part of the circuit, and the output from Q1 would be all bleed into the ground as well instead of going to Q2. I think I'm starting to understand some of this. :grin: Still a long way to go...

bcarso: Thanks for pointing out my Q2 orientation error. Oops. (This is why we show these things to other people before starting to build, right?) :wink:

This is one circuit I'm starting with. Another uses a FET for the first stage, which (as I understand it) provides a higher input impedance. The product may be a hybrid of the two, or more.
 
Check the transformer wiring - there would be no output as shown - as the windings appear to be connected in anti phase!

On the secondary side Pins 1 and 4 are +ve - so should be connected to XLR pin 2, with pins 2 and 3 being connected to XLR pin 3.

The 24v shown on the original diagram is the dropped voltage from +48 volts when the current of 3.5mA is drawn by the circuitry.

The back to back zener diodes are there to give over voltage protection.

Although the circuit (when corrected) - will probably work OK - it is possible to simplify it considerably by using a suitable op amp (TL072 would be fine) - and doing away with the very expensive transformer - which is not actually giving much isolation.

Go back to the active circuit and improve on this - and save a small fortune for no loss of quality - and possibly actually increase performance, as you will be able to 'tailor' the input impedance to whatever you think best - or to suit the transducer to be used.....
 
[quote author="AudioJohn"]
Although the circuit (when corrected) - will probably work OK - it is possible to simplify it considerably by using a suitable op amp (TL072 would be fine) - and doing away with the very expensive transformer - which is not actually giving much isolation.

Go back to the active circuit and improve on this - and save a small fortune for no loss of quality - and possibly actually increase performance, as you will be able to 'tailor' the input impedance to whatever you think best - or to suit the transducer to be used.....[/quote]

I'll second the suggestion.. while I've used discrete JFETs on piezo pickups a BIFET opamp will be adequately high input Z to not load them down. I designed a phantom powered direct box at my old day job that is pretty much what you describe, TL072 with passive pad/no output transformer. It worked nicely and they're still selling them some 15 years later.

JR
 
Back
Top