Hartley Oscillator on Multi-Pattern MIcs

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

deb611

Member
Joined
Jan 8, 2017
Messages
16
Location
Los Angeles CA
Hi there, long-time lurker.

I've been stalled on a project in which I created a multi-pattern stereo microphone. I've run into a significant frustration and I think I've found the issue. The schematic I used is a standard UCM multi-pattern Schoeps circuit, which uses a Hartley oscillator to ramp up the capsule polarization voltage, and also uses the oscillator to make a multi-pattern microphone. I doubled the circuit and laid it into a standard Chinese stereo mic body similar to the SP LSD-2.

The issue I'm running into is an injection of oscillation, but ONLY when both sides of the mic are on. If only one side of the mic is on, I don't get any oscillation whatsoever. Changing the pattern switch on either mic changes the pitch of the oscillation, but the oscillation pitch is shared on both microphones.

Schematic:

screen.jpg



I found this from Zapnspark in my hunt for answers:

"It's a variation on a Hartley oscillator using 2 mutually coupled inductors (Instead of the usual center tapped inductor).
Oscillates at ~ 2 mHz. You can find this type of polarizing circuit in some older Schoeps microphones. (e.g. CMC-5)
The diodes and caps provide level shifting and rectification to create a voltage multiplier.
It's a simple and effective circuit.  Very weak output current though so - it's not meant to drive ANY resistive load - just the capsule itself.  Some el-cheapo China mics violate that rule in order to get a multi-pattern mic.  Bad idea! That's another topic though."

I did as much research as I could on oscillators, but I'm very much a self-taught novice and I'm out of my depth here. I've spent a lot of time and effort getting my PCBs designed and I'd really prefer not to scrap or restart my project at this point. Any ideas on modifications I could make to prevent the oscillation injection? I've already added two 1000pF caps across R3 & R4 to roll off any potential injection above 20K, but the oscillation I'm hearing tends to be in the 3-5k range.

my next step is to try shielding in between the two boards to avoid injection, since the boards are back-to-back with less than a half-inch of separation, but I have no idea if that will work or not. In the meantime, I'm open to ideas and criticism of my complete and utter folly  ;D
 
You haven't said how the 'oscillation' manifests itself but ...

Have you got 2 oscillators in the same mike body?

This is a really bad idea.  Use one oscillator to power both capsules.  2 oscillators will beat with each other and generate stuff in the audible band.
 
Any chance you could re-wire the rectification / filtering / pattern-switching to something like this?
http://audioimprov.com/AudioImprov/Mics/Entries/2015/12/20_Modding_a_BM-800_Mic_files/Mic%208%20TLC.jpg
 
Khron's suggestion is from Henry Spragens  Schoeps like schematic, published some time ago on his audioimprov.com website. I can confirm it works, as I have built several mics using this design with no problems at all.

So, I'm with Khron's suggestion... give it a try!

EDIT: On your schematic, I see that the font side of capsule is connected direct to ground. How is it going to be polarized that way?
One  common approach for multi-pattern designs, is to have the front side of the capsule  always connected to the  polarization voltage (+50 to +60V) output from the DC-DC converter, the center plate is always to the input FET gate thru a small poly cap (1nf), and finally, a switch controlling how the rear side gets the polarization voltage:  +50 to +60V from the DC-DC converter for Omni pattern, unconnected for cardioid pattern, and -50 to -60V from inverse DC-DC converter output for 8 figure. This is clearly seen on Herny's schematic. Take a look...

Regards.

HL


 
I've used Henry's wiring / switching arrangement too, in my pair of modded Devine BM600 mics, with no issues whatsoever.
 
ricardo said:
You haven't said how the 'oscillation' manifests itself but ...

Have you got 2 oscillators in the same mike body?

This is a really bad idea.  Use one oscillator to power both capsules.  2 oscillators will beat with each other and generate stuff in the audible band.

As far as the oscillation, it manifests in a unison signal around 2k coming from both sides of the mic, but only when both sides are powered. It would definitely appear to be an over-air problem, as moving my fingers closer to and further from different components causes the pitch to change, as the oscillation is being injected into the circuit via a component lead, PCB trace, or some other antenna-acting part.

Ricardo, this is an elegantly brilliant solution. I'm going to give this a try ASAP. I can simply pull the voltage from the 1st board, and remove the oscillator circuit from the second board entirely.

Something like this?
screen2.jpg

The one problem I can see here is that the Hartley oscillator, from my understanding, is not ideal for any significantly resistive loads. I'm wondering if running two multi-pattern capsules will put cause some strange behaviors.


Khron said:
Any chance you could re-wire the rectification / filtering / pattern-switching to something like this?
http://audioimprov.com/AudioImprov/Mics/Entries/2015/12/20_Modding_a_BM-800_Mic_files/Mic%208%20TLC.jpg

I can see that being quite difficult, but maybe feasible. I'll have to look at my PCB layout. My whole goal here is to have a functioning stereo mic for concert taping (that I can run in mid-side or blumlein) without having to re-design/re-order the PCB...which is asking for a lot, I understand, but I can dream, right?  ::) I'll have to study the circuit you posted further, but it seems doable. I think I can lift the voltage multiplier section (everything above R2 dropping in for everything above R7 in my schematic) and capsule polarization wiring from that schematic without heavily modifying the signal circuit itself. But I'm not 100% here. Thanks for posting! Gives me something else to try and wrap my thick skull around  ;D

homero.leal said:
Khron's suggestion is from Henry Spragens  Schoeps like schematic, published some time ago on his audioimprov.com website. I can confirm it works, as I have built several mics using this design with no problems at all.

So, I'm with Khron's suggestion... give it a try!

EDIT: On your schematic, I see that the font side of capsule is connected direct to ground. How is it going to be polarized that way?
One  common approach for multi-pattern designs, is to have the front side of the capsule  always connected to the  polarization voltage (+50 to +60V) output from the DC-DC converter, the center plate is always to the input FET gate thru a small poly cap (1nf), and finally, a switch controlling how the rear side gets the polarization voltage:  +50 to +60V from the DC-DC converter for Omni pattern, unconnected for cardioid pattern, and -50 to -60V from inverse DC-DC converter output for 8 figure. This is clearly seen on Herny's schematic. Take a look...

Regards.

HL

Thanks for the input. On this design, the backplate holds a steady voltage (lets say 48V, because I don't remember what it actually is), and the front of the capsule is grounded, thus a potential difference of 48V. But the backside is varied between 0V and 96V, thus giving a potential difference between 48V and -48V, thus the multipattern. At least, this is my admittedly troglodytic interpretation of it.

Thanks all! I'm going to try the simplest solution first, Ricardo's, and I'll post results.
 
That sort of resistive divider being used to derive the capsule bias works only in tube mics, where you have a +100-120v with a good few mA available, so putting that through those 51meg resistors is no issue (an insignificant additional load for the "main" power-supply).

Unfortunately, this sort of oscillator used in phantom-powered solid-state circuits is far less potent, current-wise, so that's most likely where your problem lies.
 
Khron said:
That sort of resistive divider being used to derive the capsule bias works only in tube mics, where you have a +100-120v with a good few mA available, so putting that through those 51meg resistors is no issue (an insignificant additional load for the "main" power-supply).

Unfortunately, this sort of oscillator used in phantom-powered solid-state circuits is far less potent, current-wise, so that's most likely where your problem lies.

This completely makes sense.

Well, I have good and bad news. The good news is that with only one oscillator, I get good signal from both capsules. The bad news is that my S/N ratio is now terrible. There's pretty significant background noise, far too much to be usable in a studio environment.

Is that a consequence of using a single oscillator to power both sides of the microphone? To test, I just lifted the front of D8 and C13, as well as the junction of R8 to R7, on one board. Then I wired the D8/C13/R10/R11 junction of the board with an oscillator to the same point on the board without an oscillator, sharing power. Do I need to lift the ground points of the "unplugged" oscillator as well? (L2, C7, and D4).

I suspect that I can't just sub out the 51M resistors for smaller ones.
 
If anything, you'd want LARGER value resistors (which would pass / draw much less current than the 51M ones).
 
Khron said:
If anything, you'd want LARGER value resistors (which would pass / draw much less current than the 51M ones).

derp derp derp  :eek: (that's my impression of me).

But I'd like to know to see if I'm crazy or not - is my suspicion of higher noise resulting from an over-draw by the oscillator section accurate?
 
deb611 said:
with only one oscillator, I get good signal from both capsules.
This is exactly as expected.

The bad news is that my S/N ratio is now terrible. There's pretty significant background noise, far too much to be usable in a studio environment.

Is that a consequence of using a single oscillator to power both sides of the microphone?
No.  As various people have said, this oscillator only delivers very small current.  But it has been used successfully in multi-pattern commercial mikes.  You just have to ensure you take the least amount of current possible.

Your 51M resistors should be at least 100M or even bigger if you can find convenient stuff that fits.

Arrange your switching so a single resistor chain is used for both mikes.
_______________________

But there are a couple of other issues in your design which might lead to poor noise.

C15/16 need to be 100n.  Change these first (with the 100Ms) and tell us how you go.

You may also need to add another stage of decoupling between your Zener D3 and the drain resistor(s) R7.

You ARE using a SINGLE Zener for both mikes aren't you.  Zeners are noisy.

Have you checked the voltages at the Source & Drain of your FET Q3?

You could remove the Zener and if the voltages are still sensible, the noise will be better.

It would be nice to have a schematic with clearer text.  Stuff that's difficult to read puts people off giving advice.
 
FWIW, I had a problem with a Studio Projects LSD2 which manifested itself as noise showing up occasionally.  I cured this by supplying the polarising voltage to both capsules from one board only. A short wire between the two boards is all that is necessary.

Although one might expect to, I never saw spurious signals at the output that were the difference of the two oscillator frequencies.  Having built several polarising oscillators for other microphones using this circuit, I am also not convinced that the layout of the two inductors is critical.  But I do put the oscillator in a small metal box, to ensure total absence of 2 MHz on the polarizing voltage.

David
 
Last edited:
Hi Khron, how did you manage the 2 coils of 100µH and 150µH? Are they separated components or are they belonging to a only one component with ferrite or air core?

The website is not mine, but the two inductors are indeed two separate components. You'll want to locate them right next to each other, since magnetic coupling between them is important.
 
The website is not mine, but the two inductors are indeed two separate components. You'll want to locate them right next to each other, since magnetic coupling between them is important.
Indeed the coupling was the concerning thing. Is there any particular type of inductor would you suggest to use?
 
Back
Top