Active ribbon-mic

[clintrubber]

Hi,

With all this potential phantom-damaging, why not make the ribbon-mic active ?
This of course nothing original (R-122, etc), but has the obvious advantage
that pre-amp demands (gain, inp-Z) are lessened at the same time.

Circuits could be inside the mic of course, but for a first incarnation
I was more thinking about replacing the usual wired XLR-3 with a XLR-something
(to still prevent +48 carrying XLR-3's to be used) and then adding an inline
enclosure with XLR-something_in, and the usual XLR-3_out.

Thinking about usable phantom-powered circuits for a ribbon-mic,
the ESP project#93 was the first to come to mind.

http://sound.westhost.com/project93.htm figure 6

Modification by making the input balanced (a small Beyer-TX for instance), removing the electret-feed resistor R2 & providing the required TX-termination.


Any other ideas/circuits ?

Thanks,

Peter

 

[BradAvenson]

OK, so why don't you put the electronics in something like this and just plug it in to the bottom of your ribbon mic.
http://www.switchcraft.com/products/connectors-21b.html

You shouldn't need a transformer because there is one in your microphone already. You just want to make sure you aren't loading the ribbon microphone too much. There are some preamps that have input-Z of 5K to keep from loading the ribbons too much. You might double the circuit and keep it balanced. It would be interesting to see how it works with dynamic microphones.

 

[clintrubber]

Hello, thanks for the reply.

I had a look at those cool looking Switchcraft-enclosures when I built the pre-amps of the ESP #93 project for PZMs, but as small as it already is, that ESP-PCB won't fit that enclosure well.

So I made something alike but larger, using a simple Alu-tube & 2 Switchcraft ( :wink: ) XLR-connectors (secured in the tube by their screws).

But you think that circuit could work OK for the first stage a ribbon sees ?

If I understood it correctly, you suggest to connect the balanced out (post-its-internal-TX) to the unbalanced input of the ESP#93 circuit, correct ?

It looks like it could work to connect that balanced ribbon output directly across R4 (skipping C1 & R2) and put a resistor in // to that R4 to bring down the input-impedance. - Hmm no, that might give problems when the voltage across R4 is too much, like possibly at power-on.
OK, then better to keep C1 & connect R2 to ground give it a suitable value.

(BTW, I used a film cap for C1 for the PZM-version, as large as I could fit in the tube)


Thanks,

Peter

 

[BradAvenson]

Too bad I don't have any good ribbon mics at the studio otherwise I would try it out. Maybe I'll try to dampen my octava some.

 

[gang of elk]

[quote author="BradAvenson"]There are some preamps that have input-Z of 5K to keep from loading the ribbons too much.[/quote]

i'm confused, can someone help me understand loading of ribbon mics and what causes it? some older ribbon mics i've seen are around 50 ohms, pretty low. wouldn't it be better to have a more closely matched impedance @ preamp input? or make an active circuit to serve the same purpose when a low Z preamp input is not an option? i thought this is what the active royer is all about. connecting a 50 ohm mic to a 5k ohm preamp does not make sense to me.

 

[clintrubber]

That ESP-circuit has a gain of 10dB I thought (at least could be set to do so) so that looks just the amount we could use - bringing it in line with the other kind of mic-signals. Maybe noise could be an issue ? Well, since I have that circuit already I'll try when a ribbon arrives.

Bye,

Peter

 

[clintrubber]

i'm confused, can someone help me understand loading of ribbon mics and what causes it? some older ribbon mics i've seen are around 50 ohms, pretty low.

That 50 could be the output-impedance, right ?

From the Royer-site I learned that 1500 Ohms and up (the mic-pre input-impedance, not the mic-outp-imp.) is a nice value for ribbon-mics.

 

[alk509]

[quote author="gang of elk"]connecting a 50 ohm mic to a 5k ohm preamp does not make sense to me.[/quote]

Why not? You don't want to match impedances, you want to bridge them.

Peace,
Al.

 

[gang of elk]

[quote author="alk509"][quote author="gang of elk"]connecting a 50 ohm mic to a 5k ohm preamp does not make sense to me.[/quote]

Why not? You don't want to match impedances, you want to bridge them.

Peace,
Al.[/quote]

hmm, i need to study up on this.

i've experienced the ill effects of impedance mismatching between other components within the studio, i assumed the principle of Z matching was universal, thus the benefits of bridging trafos or impedance matching trafos. i assumed the same ill effect may occur w mic/preamp connections where no physical 'matching device' is used.

i still don't understand why a low Z match may cause loading of a ribbon mic but perhaps some reading will clarify.

 

[rafafredd]

I don´t think it would be a problem to load most ribbon mics with impedances down to 500Z. But it´ps said that generally ribbon mics sound best with higher impedances, like 5K-10KZ...

 

[pstamler]

[quote author="gang of elk"][quote author="alk509"][quote author="gang of elk"]connecting a 50 ohm mic to a 5k ohm preamp does not make sense to me.[/quote]

Why not? You don't want to match impedances, you want to bridge them.

Peace,
Al.[/quote]

hmm, i need to study up on this.

i've experienced the ill effects of impedance mismatching between other components within the studio, i assumed the principle of Z matching was universal, thus the benefits of bridging trafos or impedance matching trafos. .[/quote]

Depends on the age of your gear. Impedance matching was the standard through the 1960s; beginning in the 1970s, bridging impedances (input impedance significantly greater than source impedance) began to take over, and today that's the almost-universal standard.

Matched impedances are better if the piece of equipment was designed for them. Most ribbon mics weren't; they were designed to be loaded by higher impedances, and will have optimum high-frequency behavior when so loaded.

Peace,
Paul

 

[Marik]

All the ribbon microphones "feel" impedances very differently. It depends on a transformer used and how it is matched with ribbon (it does not necessaraly says about transformer's quality). Some work best when unloaded, some would "ring" when unloaded, some would need load for best HF response. The general rule of thumb is when it is loaded too much, the LF will be chocked. You need to make experiments to find out what works the best with your particular mic.

 

[clintrubber]

So what do you all think, would the ESP-circuit above be suited ? (using a proper termination determined experimentally)

Any other suggestions for such a first circuit that the ribbon-mic sees ?
(preferrably phantom powered)

Thanks,

Peter

 

[Rossi]

I've read somewhere on the Royer site, that their active ribbons use higher than usual transformer ratios for extra voltage gain. The active component (FETs) is only for current gain and impedance conversion. The rationale being that the active component should add as little noise as possible.

I wonder what a simple impedance converter could do for a budget ribbon mic. There wouldn't be any voltage gain, of course, but you could present a pretty high impedance to the transormer (maybe 10 K) so there would at least be no gain loss. Also, lowering the mic's output impedance might help to improve the performance with not-so-spectacular mic pres. I think I've seen two different specs for the Nady/Thoman ribbon. In one case output impedance was rated at 200 ohms, in the other it was 600 ohms. And who knows, it could be even higher than 600 ohms. So an impedance converter might do some good.

 

[clintrubber]

Interesting. So they use more 'TX-gain' but then arrive at a higher output-impedance and then bring that down again by the active buffer, right ?

Also, lowering the mic's output impedance might help to improve the performance with not-so-spectacular mic pres

I thought the main issue was that the higher gain-requirements for ribbons would bring some preamps into trouble but yes, the impedance might be an issue too.

I'm a bit confused by the impedance-thing, since ribbons would be lower-imped. than average. But then, with those high-TX ratios it jumps to the other side of the usual impedances.

The active component (FETs) ...
... a pretty high impedance to the transormer (maybe 10 K)

Mmm should brush up my theory, but with such high source impedances the use of FETs sounds logical yes.

Bye,

Peter

 

[Rossi]

Yes, of course it's mainly the higher gain requirements that causes trouble. But since we don't know the real impedance of the budget ribbons, it might be worth a try to interface it with an impedance converter. If the output impedance was say 600 ohms, there would be a gain loss of about 3.5 dB if you used it with a preamp of relatively low impedance of say 1200 ohms. Shure has a calculator on their site: here
I hope the link works like that.

If you interfaced the microphone output it with an impedance converter of say 10 K input impedance (I don't know what Royer uses) the gain loss would only be 0.5 dB. So the output would be 3 dB hotter.

Given the high gain requirements, perhaps noise performance could be improved, too, if you brought the mic's output impedance as low as would be reasonable to do. At least below 200 ohms. Perhaps even as low as 50 ohms. Of course the impedance converter would have to work with what phantom power can easily deliver. I'm sure PRR would know better than me what's reasonable and/or doable.

 

[David Curtis]

I'm new here, so please be gentle...

I've been experimenting with adding an active output stage to my Oktava ML52. I replaced the original transformer with a Jensen JT-34K, which was a major improvement all by itself. I used an Analog Devices AD797 for the active stage. The AD797 is (in Analog's words) an ultra-low-noise, ultra-low-distortion op amp. In general, I've been pleased with the results, though I'm still in experimentation mode.

Some comments based on my experience with this project:

Noise is an overriding concern. Each element in the signal chain affects the noise characteristics of the system, often by interacting with other elements in non-obvious ways. For example, noise optimization for an amplifier like the AN797 consists of balancing voltage noise, current noise, and thermal noise of resistances at the input. The thermal noise is determined (mostly) by resistors in the feedback loop, which also determine gain and input impedance. The gain and input impedance requirements are, in turn, strongly influenced by the characteristics of chosen transformer. All of these things are interdependent, and the relationships can be convoluted. Finding a configuration for optimal (or approximately optimal) performance is a complex problem, enough so that simple rules of thumb about higher turns ratios or higher impedance loads being better can be misleading.

Regarding transformer choice--as mentioned above, the choice of transformer is highly interrelated with other design choices. An earlier post mentioned wisdom from the Royer web site that higher turns ratios are perferred for active ribbon mics. Although I'm reluctant to take issue with the guys at Royer, my own experience is somewhat to the contrary. High-performance audio transformers with high turns ratios are harder to design and build (and thus, more expensive or lower performance) than those with lower turns ratios. One of the advantages of an active ribbon design is that you can use a transformer with a lower turns ratio and either save money or get higher performance, or possibly both. You can make up for the loss in gain with the amplifier, as long as you have noise under control. The JT-34K has a turns ratio of 1:37, which is relatively high. Assuming a ribbon impedance of 1 ohm, the output impedance is about 1.4K ohms. The AD797 has optimum noise performance with input impedances under 1K, best around 500 to 600 ohms. When I can afford it, I want to get a JT-347, with a turns ratio of 1:24. With a 1 ohm ribbon, its output impedance is just less than 600 ohms. Also, its overall performance is significantly better than the JT-34K--lower distortion, higher levels, and flatter frequency response. [Note: the JT-347 is considerably more expensive that the JT-34K, which would seem to contradict my earlier statement. The JT-347 is a much larger, much higher performance transformer than the 34K. I haven't found a transformer with a higher turns ratio that comes close to its specified performance.]

Regarding amplifier--I know that many people prefer discrete designs to integrated op-amps, but I don't share this bias. For this application in particular, even an excellent circuit designer would be hard-pressed to design an amp with better performance than the AN797. Its input noise characteristics are better than most discrete low-noise devices, and it's overall performance as an amp (e.g., THD, gain bandwith product, slew rate) is stellar. The Linear Technologies LT1028 looks like a good alternative candidate. There are, no doubt, others. One drawback: the AD797 requires more current than standard P48 phantom power can supply. A production microphone would require special cabling and an outboard power supply, like older tube mics.

It's also probably worth mentioning that, while the JT-34K can fit (just barely) into the bottom can of the Oktava, there isn't any room left for an active stage, so I've had to use a makeshift enclosure for prototyping. The JT-347 will definitely NOT fit into the Oktava can.

Regarding transformer ringing (mentioned in an earlier post)--with high turns ratio transformers, the inter-winding capacitance in the secondary can be sufficiently high that it forms a self-resonant LC circuit. This resonance can be damped with an RC network across the outputs of the secondary. The Jensen datasheets include information on how to do this, what values to use, etc. This is another case where variations in loading the output of the ribbon/transformer can have a big effect. All the more reason to use a built-in active circuit to optimize the loading and isolate it from the down-stream mic pre.

--Da5id

 

[Rossi]

So far I haven't even thought much about using opamps, but if an opamp serves the purpose, why not?

Using the stock transformer (or a higher quality replacement) and getting extra gain from a more elaborate active circuit would be a more cost effective alternative to the Royer design. Good quality high high turns ratio transformers aren't cheap and usually hard to get. I would prefer using just phantom power, though.

 

[David Curtis]

I would prefer using just phantom power, though.

Me too, me too. I've looked at lower-power options, and I'm thinking about trying the Analog AD8011. It draws less than 1mA quiescent current, and has decent noise characteristics and good overall performance. It's designed for video, so it's a current-feedback amp, and it has huge bandwidth (400MHz). I'd want to do some BW limiting, to avoid the possiblity of HF oscillation.

 

[Rossi]

When I'm done with my impedance converter experiments, I wanna try opamps, too. But I was thinking of something more modest such as NE5532 or MC33078. Noise performance would probably be the main concern; I don't think we'll run into much trouble regarding bandwidth, slew rate etc. If I remember correctly, optimal noise performance for NE5532 is at about 8 K input impedance; MC33078 at about 9 K. I might be wrong, though; I have to take a look at the figures again. Current draw may also be an issue with phantom power.