Oktava ML51 active ribbon mic

[mad.ax]

One of these landed on my bench. It's a small ribbon mic, with what looks to be a toroidal transformer followed by a battery powered output buffer. Output is unbalanced.
There's no switch to power it off and the battery leaked, damaging the inside.
I am aware of this thread : http://www.groupdiy.com/index.php?topic=5743.msg72562#msg72562 , where Marik posted the schematic. I traced mine and it looks the same except for R3 which is 3.3K instead of 100K.

I checked twice but the input elco has indeed its positive connected to the ribbon. Am I missing something here?
I have no idea what the transistors reference may be. I can see some Cyrillic (a kind of B followed by 'oO'), and on the second line : 8604
The mic came with a 1V5 battery in it.

I guess, it's not the original battery. Marik's schemo calls for 3V. Is anybody knows what kind of battery it is supposed to use?
Anyways, after cleaning the oxydation, I wired a 1V2 rechargeable battery that was laying around, and it works. I wouldn't call it perfect or even great, but it works.
So now I have two options:
-Wiring a 2 X AAA battery holder (there's enough room). This would leave the electronics 'stock', but I wonder how much mojo there is to expect from this gain stage...
-Ripping the electronics off, and build a brand new phantom powered balanced stage ala PRR like shown in the same thread : http://www.groupdiy.com/index.php?topic=5743.msg73762#msg73762

Anybody's gone this route before, or have some experience to share regarding the ML51?

Axel

 

[zebra50]

Thanks for the photos!

Well, I haven't played with that mic, but I have built a few versions of the buffer circuit as mentioned at the end of the thread. It works very well, but often the limiting noise factor is the (oxidised) ribbon the and transformer, and you just end up amplifying that.

Do you know what ratio the transformer is?

Another way to go would be to try one of Marik's nice ribbon transformers and skip the active bit altogether.

http://www.groupdiy.com/index.php?topic=40920.0

 

[Marik]

Hello Axel,

That circuit is noisy like hell! I would not bother with the battery, but whether replace it with some P48 (plenty to chose from), or discard it, all together.

Best, M

 

[mad.ax]

Thanks for your answers guys! Here's some more porn for you:

Even with 3V, the original circuit is indeed quite noisy... I ordered some transistor to try the PRR ckt...
I don't know what ratio the transformer may be... It's a toroid, originally asymmetrically wired, and the link to the 'hot' side of the motor was done with a screened wire, as you can see in the pic...

The ribbon itself looks fine. It is well protected by several layers of fabric

Axel

 

[PRR]

> build a brand new phantom powered balanced stage ala PRR like shown in the same thread

No. That is for a mike which is already transformed to balanced 200-300 ohms, just a little weak, and your console preamps are not lowest-hiss.

Maybe your transformer is near 300, and maybe it can be floated with good balance. But that buffer has BIG gain. That suggests impedance is very-very low or the capsule is just very weak.

In that case a single-ended input is probably the way to go.

At least I assume the original buffer has high gain. However if Q1 has hFE of 100 or more, it appears R2 will slam Q1 collector to ground, horrible gain and level. Maybe they had low-hFE transistors. Maybe they just did not care. (Maybe when assembled with the backward electrolytic, cap leakage masked the mis-design.)

> I don't know what ratio the transformer may be

Since you also don't know the ribbon resistance, we don't care what the ratio is.

The short path is to find the reflected impedance at the transformer secondary, then design a buffer/amplifier for good noise-figure from that source, and try it.

i.e. if they transformed up to 70 ohms, we'd be thinking parallel BJTs at many mA (which may not be Phantom-friendly.... cross that bridge when we get to it). If transformed up to 40K we'd start with 20uA in a BJT or maybe a clean JFET.

To find impedance: apply a voltage to unknown impedance and a reference impedance. Find what reference impedance balances the unknown impedance. In this case we hope the unknown is resistive, or very-nearly so. The reference impedance may be a box of resistors, a pot, etc.

We'll measure the transformer AND ribbon together. (We'd like to know the ribbon but its resistance is too low for simple techniques.) To be SURE of not busting the ribbon we MUST have a big drop out of the signal generator. Use 1K and 10 ohms or 100:1. Don't bring the signal generator over 1V.

We need a signal sensor to resolve 10mV and 5mV all across the audio band. Cheap DVMs may read 10mV but only up to 400Hz. Oscilloscope is linear in this range, and is handy to know if you are reading test tone or humm/buzz crap. Wire sig gen and 1K:10 divider and meter/scope, check for constant few-mV level 100Hz to 5KHz.

Put sig gen to zero volts. Connect ribbon transformer secondary and reference resistor, start near 100K. Move meter/scope to junction shown.

The ribbon will sing. It must NOT sing loud. Preferably we never hear it, though I suspect that in a quiet room a just-audible sing near 3KHz is safe (Marik may have experiences).

Set sig gen to 1KHz, bring it slowly up from zero V until you get a clear reading (or hear singing!). With 100K it should not sing and the reference resistor voltage is probably very near the voltage at the 10 ohms (means transformer impedance is much lower than 100K).

Reduce reference resistance until voltage drops to about half. Now ref-R is equal to transformer impedance.

No real precision is needed. If 470 ohms gives 0.6 of full voltage, it's 300 ohms for all practical purpose.

Try several frequencies from 100Hz to 5KHz. If all within 2:1 it is resistive-enough for design. Mild resonant bumps to 4X impedance are no concern. If impedance rises or falls about as fast as frequency, then reactance dominates impedance. Since the naked ribbon can't be very reactive, that would suggest the transformer is cheap crap (what a shock!) and you should consult with Marik about determining actual ribbon resistance so a better transformer can be found.

 

[mad.ax]

Thank you for your advice Paul!
This afternoon I tested the transformer as per your schematic. Fed exactly 10mV at the center of the voltage divider, and lowered the pot until I reached a stable 5mV between the pot and transformer. Level didn't vary much with frequency. After desoldering the pot I found 1K.

I was curious so I wired an XLR to the transformer and listened through my cheap bench mixer (Tapco MIX60). Despite using the full 50dB gain of the mixer, it sounded much better and less noisy than with the buffer!
Looks like the Oktava toroid is less crappy than its electronics after all...

I'll bring it to the studio this week to listen with real preamps, and compare with other ribbons, as it seems that passive may be an option...

Axel

 

[zebra50]

There is no reason why the tranny shouldn't be OK. Some of the old Oktava ribbon mics are very nice indeed. I have ML11 and ML17, both with toroidal transformers, and they are both great mics in my opinion.

Stew

 

[PRR]

> I found 1K.

1K is good.

I'm under a stopped-up toilet right now but will ponder a preamp.

> Looks like the Oktava toroid is less crappy than its electronics after all...

At low impedance and low level, transformer design is "easy". At least capacitance does not trouble us. And the toroid will have very low leakage inductance.

In this case the VERY low impedance on the primary does leave us open for trouble. The primary wire must be very fat and very short or we have excess noise resistance. This winding is a long way away from the ribbon, and fat wire is hard to wind. But let's assume it is pretty-good, and think about that truly ugly booster/buffer.

> using the full 50dB gain of the mixer, it sounded much better

How much more gain would you like? It may be hard to judge on workbench. Can you plug a SM57 panned left, this ribbon+toroid panned right, both a few feet from your mouth, and find gains that put your voice "center"?

 

[mad.ax]

I'm under a stopped-up toilet right now but will ponder a preamp.

Sometimes I wonder if there's anything you don't DIY Grin Thanks for your help and interest!

In this case the VERY low impedance on the primary does leave us open for trouble. The primary wire must be very fat and very short or we have excess noise resistance. This winding is a long way away from the ribbon, and fat wire is hard to wind. But let's assume it is pretty-good, and think about that truly ugly booster/buffer.

I wonder why they did run the hot wire to the top of the motor assembly. Using the bottom would have shortened the path by a good inch...

How much more gain would you like?

Not more! I was just pointing out that my crappy bench mixer gets noisy when used at max gain, but still, total noise was less than with the Oktava buffer and 30dB gain.

It may be hard to judge on workbench. True!
Can you plug a SM57 panned left, this ribbon+toroid panned right, both a few feet from your mouth, and find gains that put your voice "center"?

I haven't a 57 handy here. That's why I'm planning to test it at the studio, so I could see if the passive route is OK for real world use...

Axel

 

[Marik]

Thank you for your advice Paul!
This afternoon I tested the transformer as per your schematic. Fed exactly 10mV at the center of the voltage divider, and lowered the pot until I reached a stable 5mV between the pot and transformer. Level didn't vary much with frequency. After desoldering the pot I found 1K.

I was curious so I wired an XLR to the transformer and listened through my cheap bench mixer (Tapco MIX60). Despite using the full 50dB gain of the mixer, it sounded much better and less noisy than with the buffer!
Looks like the Oktava toroid is less crappy than its electronics after all...

I'll bring it to the studio this week to listen with real preamps, and compare with other ribbons, as it seems that passive may be an option...

Axel

Hello Axel,

The ribbon itself on that microphone should be 0.26 Ohm and output impedance of the mic should be in 300-400 Ohm range. The 1K you got suggests 1) Contaminated contacts, or 2) Higher ratio transformer. From the picture of the transformer you posted the Pri looks like any other Oktava ribbon transformer. Those are wound on M6 stack--probably not the best solution for that application, but nevertheless, they are very fine transformers.

Could you disconnect the transformer and measure its ratio? Feed the 1KHz signal from the secondary and measure on the Pri. Usually, for ribbons I adjust "in" to get 100mv "out". Nice and round, so I don't need to reach for calculator  .

Best, M

 

[mad.ax]

Hi Marik,

2.4V on the secondary gave me 40mV on the primary. 1:60 ratio

Is that different from the ML51 you dissected?
Mine seems to be an early one as it is engraved 958 on the headbasket...

Axel

 

[Marik]

Hi Marik,

2.4V on the secondary gave me 40mV on the primary. 1:60 ratio

Is that different from the ML51 you dissected?
Mine seems to be an early one as it is engraved 958 on the headbasket...

Axel

So 0.26(ribbon DCR)x60^=936 Ohm--close enough to the impedance you measured earlier. Little high for passive. You could add a simple 2x2SC3329 buffer (take a Schoeps circuit output), or if you'd like to go passive you could just re-wind the Pri. I don't remember, but think the transformer has 14 or 16 Pri turns. Unwind the polyester tape, then unwind the Pri (note the direction of the wire). Take some #19 or #18 wire of appropriate length and scale up the number of turns to get some 1:32 (for 300 Ohm output). The whole surgery won't take much time to accomplish. With more turns on Pri because of higher inductance you will also get better bass response. Just try to keep the wire as thick as you can to keep noise down. Doubling thinner wire like #21-22 would make it even more practical.

Best, M    

 

[mad.ax]

Humm... Makes me wish I had a pair to compare the active vs passive option...
I'll have to look if I have the appropriate wire. If targeting for 300 ohms, shouldn't I rather get 1:34?
(just checking if I got the math right! I guess 266 ohms would be quite good...)

Axel

 

[Marik]

I'll have to look if I have the appropriate wire. If targeting for 300 ohms, shouldn't I rather get 1:34?
(just checking if I got the math right! I guess 266 ohms would be quite good...)

Yes, of course! That what happens when I make calculations in my head, being lazy enough to reach for calculator... :

With an older motor (which is not nearly as efficient as modern Neodymium magnets) for low noise you'd  actually, prefer lower ratios even for an active. If you use low noise low Rbb BJTs you will be much better in the noise department, as the resistance noise with higher ratios goes up faster than the stepped up signal.

Best, M

 

[mad.ax]

So, I opened the tranny. Primary was 12 turns of 22 gauge...

I hadn't any spool of wire handy, but an old computer ATX PSU volunteered itself as a donor for the cause.... Plenty of 21 gauge on the big tranny....
I carefully winded a pair as tight as I could (very first time I'm doing this!) and ended up with 20 turns. Couldn't have made one more turn, not enough space inside the tore, but bingo! 1:34

Found a Shure Unidyne 588SB (kinda Beta58 grand-dad) and used it for comparison purpose with my Tapco bench mixer.
Exactly the same level, but the Oktava has less noise than the Shure, and seems to sound better!
I will report more Saturday, after testing in the studio, but for now, I'm an happy winder  ;D
Thanks again for your advises!

Axel

 

[zebra50]

That's pretty cool stuff Axel - you've just opened up a whole new world of possibilities to play with!

 

[mad.ax]

Thank you Stewart, but in all honesty, credit should go to Marik, as he is the one who walked me through rewinding the primary!

Axel

 

[zebra50]

I recall Mark and others here helping me with a similar problem five years ago or probably more, rewinding a broken ribbon transformer. I learned a lot on that journey, and spent lots of evenings winding bobbins and toroids by hand. Fun times!

 

[Marik]

Hey Axel,

Glad it worked and worked well!

Keep us posted.

I recall Mark and others here helping me with a similar problem five years ago or probably more, rewinding a broken ribbon transformer. I learned a lot on that journey, and spent lots of evenings winding bobbins and toroids by hand. Fun times!

Jeez!!! Don't remind me that  !

In fact, the Oktava transformer was the one I got hooked to toroids. I still remember having 8 different types of toroid cores and winding them by hand. Takes hours and hours... and hours.

Then I made a winder and shuttle myself--greatly facilitated the process. Later that year I got an old and broken Arnold toroid winder, but it had set of different shuttles, so after fixing it I could start doing some serious work.

Now I have automatic and CNC industrial winders. I am glad I started doing toroids by hand--without that I wouldn't have nearly as much feel for them...

Best, M

 

[mad.ax]

In fact, the Oktava transformer was the one I got hooked to toroids.

Now I think I understand what you mean... Amazing how a few turns of copper wires can drastically change the sound!

I'm back from the studio. It was  far from an extensive test... Just some quick vocals checks with the transformer equipped Studer console, and comparison with  Melodium 42B, Melodium RM6, and Royer stereo ribbon... But it was obvious from the start, this Oktava is now playing in the BIG league! Quiet, with Lots of everything you wish in a good ribbon mic, very nice highs, big bottom, living presence... You name it!
I'm just impressed, and Teo, who owns quite a nice collection of wonderful mics, definitely wants one of those ML51!

I think I'm in love :

Axel