G7 pattern minor issue

[zebra50]

Hi!

This week I have put together a couple of G7-ish tube mics, with the aim of doing Blumlein pair recording. I used the power supply and pattern control scheme from the G7, and a slightly altered amp circuit that is closer to the 'Royer' (i.e. 5840 tube)

The mics work nicely and sound good, but when checking them out in the studio I noticed an obvious difference between the sensitivity of the front and back capsules. For this application I want two well matched mics with accurate patterns (within the limitations of dual capsules!)

Looking back over the schematic, I noticed that there is a subtle difference in the way the polarisation voltages are handled.

Here's the G7 schematic: http://www.gyraf.dk/gy_pd/g7/gic_s.gif

For the backplates, the polarisation is set by a voltage divider (2 x 470k) which comes after a 10K dropping resistor. For the back diaphragm, the divider (2 x 100K) comes before a 10K resistor.

My measurements are as follows:

Omni mode gives 0-63-0, so no problem there.

I'm measuring 5V difference on cardioid - 63V vs 68V. 5V difference. Theoretically gives a slightly altered shape to the pattern, but inaudible to my ears.

However in Fig8 it's Front=0V - Backplate=63V - Rear=136V. So the difference between backplate and rear diaphragm is 73V, 10V more than the difference between the front and rear, which is certainly enough to hear an obvious difference in sensitivity between the two sides.

The fix is simple - either change one of the 470Ks, or put a dropping resistor before the 2 x 100K. I want well matched mics so I may try a pot in the PSU so I can set it accurately. I will report back when I've got good tested resistor values.

Cheers!
z50

PS I really don't want to seem to be criticising the G7 design - it has been absolutely inspirational to me and many others here. :thumb:

 

[Jonkan]

That was exactly the same reason to why i redid the voltage dividers in my mic, so that both sides would match.

Im not sure how much difference it makes in the end though.
/J

 

[gyraf]

Where exact pickup pattern is of serious importance - like in a MS- or a Blumelin setup - I'd suggest you mount trimming pots for the polarisation-voltage resistor divider in the PSU.

Also, this gives you the opportunity to tilt the cardioid pattern a bit towards hypercardioid, which can often sound very good at close miking because of the extended proximity effect.

Jakob E.

 

[zebra50]

Thanks Jakob. I will try this.

My other idea for a quick fix is to bypass the 10K resistor at the top right of the schematic.

Cheers!

Stewart

 

[rodabod]

[quote author="zebra50"]
My other idea for a quick fix is to bypass the 10K resistor at the top right of the schematic.
[/quote]

I thought about this too, but wondered why it had been placed there. I'm sure it would be fine to omit, but I don't like to make assumptions in case I'm missing something.

 

[zebra50]

I always assumed it was there, in combination with the 2u2 cap, as an extra filter stage.

(Haven't tested this yet - I took the day off and planted vegetables at my allotment instead, in the rain. :grin: )

 

[rodabod]

[quote author="zebra50"]I always assumed it was there, in combination with the 2u2 cap, as an extra filter stage.[/quote]

Ah, of course. Like you see repeatedly in guitar amp PSUs. Might be worth a try seeing how it is without the 10K as you say.

(Haven't tested this yet - I took the day off and planted vegetables at my allotment instead, in the rain. :grin: )

DIY to the max.

 

[zebra50]

I needed the mics for the next session so I took the easy route. I simply connected the voltage divider to the other end of the 10K. This sorts out the polarisation without altering the amplifier circuit:

The polarisation section still sees a stabilising cap, so this should be trouble free.

In my two mics, the capsule voltages after this mod were 72V on the backplate, and 0 - 72 - 144V on the rear capsule, to within a couple of tenths. They sound very good indeed.

Thanks to everyone for their ideas and suggestions.

z50

PS If you If you haven't tried Blumlein pairs then do - it's a wonderful technique.

:thumb:

 

[vertiges]

zebra50,

I'm confused... I thought It wasn't possible to make the measurements with a traditional multimeter because of the high impedance of that part of the circuit. (Actually, maybe you used an electrostatic one )

Besides, my understanding is there is no DC voltages drops accross the R10s so the voltage potentials at the junction of the two 470k and the two 100 k wouldn't be the same (I mean in cardioid)

If these 10k are influencing the polarizations voltages I want to know the explanation !!!

EDIT : as rodabod mentioned in this thread, the 10k is in series with the 470k. So I was wrong...

Thanks,

eD

 

[zebra50]

Hi Ed,

I think you answered your own question!

The 1 gig resistors don't drop any voltage, so we can measure with a good standard meter before the 1 gig, and assume the capsule has the same potential.

 

[vertiges]

[quote author="zebra50"]Hi Ed,

I think you answered your own question!
[/quote]

:green:

eD