Transformer-coupled mic booster

[OneRoomStudio]

I was thinking about how most of the less expensive mic booster circuits use transformerless balanced circuits (either discrete or IC), and some of the more expensive ones (i.e. Soyuz) use what is likely a transformer driving a discrete boost circuit. People seem to love the sound of the Soyuz, so it got me thinking. If the goal is to inject a little "color," why not design a simple circuit using cheap transformers and a basic JFET boost? Any reason something like this wouldn't work? I realize the published bandwidth of these transformers is limited, but at low levels, it might be just fine (and add to the "color"). From my spice simulations, it looks like it would draw ~5mA, and do around 15dB of gain, which is perfect.

 

[volker]

That should work. Be prepared to adjust the bias of the JFET obviously. When you discover that these transformers are rather subpar and more than a bit of color, you can always go for OEPs or the cheap Neutriks, both good bang for the buck choices.

 

[OneRoomStudio]

I have a few OEPs and Neutriks laying around, but I was curious to see how these Xicon transformers would sound at microphone level (maybe terrible).

 

[RuudNL]

Maybe you can simplyfy the circuit even more, by DC coupling the base of the transistor to the drain of the FET.

 

[OneRoomStudio]

Maybe you can simplyfy the circuit even more, by DC coupling the base of the transistor to the drain of the FET.

Great point! Not sure why that didn't occur to me. According to my simulation, that would actually slightly reduce the current draw and slightly increase the gain by about 1dB. Plus it would be cheaper and simpler!

 

[radardoug]

And direct couple the output transistor to the fet.

 

[OneRoomStudio]

And direct couple the output transistor to the fet.

Isn't that what RuudNL already mentioned? How else would the transistor be coupled to the FET?

 

[radardoug]

True, I thought he meant the input to the fet, where you could do the same.

 

[OneRoomStudio]

I made up a little board that fits nicely in a Bud box. It works and sounds great. Does around 20dB of gain with no perceptible added noise. Total cost was around $70. Thanks to @porkyc for getting the transformers to me from the UK!

 

[emrr]

R1 600R?

 

[OneRoomStudio]

R1 600R?

Yeah, the input transformer is wired 150:600 (a little different from the original schematic).

 

[emrr]

Yeah, the input transformer is wired 150:600 (a little different from the original schematic).

That then loads the mic as a matching condition rather than the expected bridging condition of 1K5Ω+. With a dynamic mic that introduces some degree of LPF, with a condenser it lowers headroom and increases distortion. Tosses some gain out. I'd go 6K8 or higher.

 

[OneRoomStudio]

My thought was that with a goal of introducing some “color,” loading the mic a bit would make sense, but that’s probably not a great way to go about it. I’ll change it out for a 6K8. Thanks for the feedback!

 

[jokeramik]

Do have some measurements for us?

I recommend increasing C 4 by a factor of 10 to minimize distortion in the bass range.

What is the R8 resistor for? 100 ohms are far too low with windings connected in series with an impedance of 150 ohms each. The signal drops and the distortion increases.

How exactly are R9 and R10 matched?

I think R7 is too high because the load is much lower then R 7.

R 7 must be lower then the loadresistance (Tietze/Schenk: Halbleiterschaltungstechnik)

Best regards!

 

[OneRoomStudio]

Do have some measurements for us?

I recommend increasing C 4 by a factor of 10 to minimize distortion in the bass range.

What is the R8 resistor for? 100 ohms are far too low with windings connected in series with an impedance of 150 ohms each. The signal drops and the distortion increases.

How exactly are R9 and R10 matched?

I think R7 is too high because the load is much lower then R 7.

R 7 must be lower then the loadresistance (Tietze/Schenk: Halbleiterschaltungstechnik)

Best regards!

Good call-outs. The schematic at the top is actually old. The PCB is as so:

-C3 (used to be C4) is 470uF now (yes, 22uF was way too small)
-R6 (used to be R8, 150R now) terminates the OT and keeps it more linear. Not strictly needed, but helps prevent ringing.
-R7 & R8 (used to be 9 &10) are 0.1% matched
-R5 (used to be R7) is now 2K2, determined by simulation. The transistor DC current with this value is 3.18mA, so if my math is right, a 2200||150 = 140R load, which means it can swing +/- 0.45V, and since this is going into a mic pre and the input signal was in the order of 0.02V...that should be fine. It certainly works in practice.

Also, the goal of this is to have a bit of "color" anyway.

 

[jokeramik]

-R5 (used to be R7) is now 2K2, determined by simulation. The transistor DC current with this value is 3.18mA, so if my math is right, a 2200||150 = 140R load, which means it can swing +/- 0.45V, and since this is going into a mic pre and the input signal was in the order of 0.02V...that should be fine. It certainly works in practice.

Please look at Tietze/Schenk "Halbleiterschaltungstechnik" 10th Ed., Page 56: R_load > R_E.
Do you measure the signal with a oscilloscope and a distortion meter?
Best regards

 

[jokeramik]

-R6 (used to be R8, 150R now) terminates the OT and keeps it more linear. Not strictly needed, but helps prevent ringing.

Yes you made a heavy load on the output... With the input resistance of your interface/mixer you made the load mor heavier due the parallel connection of these.
So you have voltage divider. Why do you do this?
You lose a lot of level, the SNR is reduced and the impulse reproduction deteriorates.
Ringing can only be optimally prevented using a Zobel element (series connection of capacitor and resistor).
Look at the application notes made by Jensen Transformers.
Check the unit with a square wave signal and a oscilloscope.

Best regards

 

[OneRoomStudio]

Please look at Tietze/Schenk "Halbleiterschaltungstechnik" 10th Ed., Page 56: R_load > R_E.
Do you measure the signal with a oscilloscope and a distortion meter?
Best regards

I don’t speak German and don’t have access to this book. Are you able to share the contents in English?

I know that typically speaking, it’s better for the emitter resistor to be less than the load, but that’s not a hard-and-fast rule. In this case, I’m dealing with very low levels and aiming for some amount of distortion and non-linearity (as I said before, the goal of this was to be “colorful”).

 

[OneRoomStudio]

Yes you made a heavy load on the output... With the input resistance of your interface/mixer you made the load mor heavier due the parallel connection of these.
So you have voltage divider. Why do you do this?
You lose a lot of level, the SNR is reduced and the impulse reproduction deteriorates.
Ringing can only be optimally prevented using a Zobel element (series connection of capacitor and resistor).
Look at the application notes made by Jensen Transformers.
Check the unit with a square wave signal and a oscilloscope.

Best regards

This device is not intended to be plugged directly into a mixer or interface at line level. This is only a “mic booster,” intended to be put in front of a mic preamp (and powered by phantom power). It does around 20dB of gain as-is, which is plenty for this purpose.

I could see the utility of adding a switch to lift the load for preamps with very low input impedance, but for transformerless preamps with input impedances of >10K, I would want this to be in place - and in those situations, the parallel impedance is basically unchanged.

 

[jokeramik]

This book is a

I don’t speak German and don’t have access to this book. Are you able to share the contents in English?

This book is available in english.