Active ribbon-mic

[Whoops]

Whoa! That's a Behringer product!!?!  Thanks for the tip. I was unaware of the acquisition.

That makes up my mind. I have a "No Behringer equipment" policy in my studio.   

KlarkTeknik and Midas were bought by Music Group (Behringer)

So Behringer is using the  KlarkTeknik brand name for their line of vintage clone gear (LA-2A, 1176, etc) and for some Live Sound Helper tools that in the past would be released with a Behringer stamp.
Quite clever actually, as they know some people have preconceptions on the Behringer name so they got another name to stick there that is associated with quality.

Midas name is used in a different way, the Midas products are still original products and with the Midas standards but Behringer is using the Midas name for some parts in Behringer products like:
- Midas designed Mic Preamps (in the X32 console)
- Midas Transformers (in the KT clone products), as far as I know Midas was never a transformer manufacturer

Also Midas live sound consoles and stageboxes, The Pro series, which sound outstanding had a price drop for almost half the price since the Behringer takeover.  A really good buy I have to say

 

[kato]

Quite clever actually, as they know some people have preconceptions on the Behringer name so they got another name to stick there that is associated with quality.

Very clever indeed! They got me. Even as a non-Behringer dude, I was thinking, "Hmmm, $29 isn't too much of a gamble. " LOL. I could justify it for the enclosure if I had to DIY a circuit inside and call the $29 an expensive enclosure.

Also Midas live sound consoles and stageboxes, The Pro series, which sound outstanding had a price drop for almost half the price since the Behringer takeover.  A really good buy I have to say

I've heard Glen from Spectre Media give his endorsement to the Behringer Midas 500 series gear. Sounds like a win.

Has anyone other than Lennie built this circuit?
It sounds like he had some level of success until he tried to add a polarity switch. Then it got messed up. Then the whole thread de-railed over the use of "polarity" vs. "phase" which is enjoyable for me to ponder, but we lost sight of the project.

Looks like this could be built in an hour. Finding an enclosure is probably the hardest part.

 

[Whoops]

Very clever indeed! They got me. Even as a non-Behringer dude, I was thinking, "Hmmm, $29 isn't too much of a gamble. " LOL. I could justify it for the enclosure if I had to DIY a circuit inside and call the $29 an expensive enclosure.

Well, I think it should be good, the quality is fine.

 

[kato]

I'm wondering if I could somehow wedge the circuit between two Neutrik jacks.

I'll have to ponder this a bit...

This enclosure could be highly @!#?@! as well. 

 

[kato]

Well, I think it should be good, the quality is fine.

It gets good reviews. It's out of stock everywhere at the moment.

I'll just pretend like I don't know it has Behringer inside.

I placed a pre-order. AMS says they'll have them back in stock around October 1. I can cancel my order between now and then if I change my mind.

Thanks for dropping all that knowledge.

 

[kato]

I haven't done an audio project in years, but I managed to find all the parts.

I'm thinking about wiring this in the air and stuffing it in a tube. Hoping I got Drain Gate Source right from the datasheet. (It's drawn without the orientation specified.) i.e. am I looking at it from the top or the bottom. Not clear.

Please let me know if I'm about to @!#?@! something up. Thank you brothers.

 

[Whoops]

I'm wondering if I could somehow wedge the circuit between two Neutrik jacks.

I'll have to ponder this a bit...

Any problems with this?

 

[kato]

Great minds think alike! It's gonna be a very tight fit, but I think it can be done.

Just about to order something like this (similar version with 3pin female side.)

 

[abbey road d enfer]

Looks like this could be built in an hour.

This arrangement with 4 FET's in cascode is not a good idea. It has been designed to overcome the 15V limitation of the 2SK170 FET, at the expense of increasing the already too high output impedance.

Apparently you haven't read the other threads about the same subject.
https://groupdiy.com/index.php?topic=72896.msg939468#msg939468
https://groupdiy.com/index.php?topic=74016.msg937308#msg937308
I suggest you read them from beginning to end.

 

[kato]

Apparently you haven't read the other threads about the same subject.
https://groupdiy.com/index.php?topic=72896.msg939468#msg939468
https://groupdiy.com/index.php?topic=74016.msg937308#msg937308
I suggest you read them from beginning to end.

You are correct. I just spent the evening catching up on all the threads, and the tangential ones. Thank you.

Which circuit are your circuit boards based on? Something posted here, or something of your own design.

I want to try a couple/few different variations just to see how they work out. Even the 2 transistor original Fethead is probably worth trying despite the noise complaints. Maybe it will be different with my mic + preamp combinations.

I'll also try the PRR BJT one from 2005. That one seems to have the most amount of thought / math / knowledge poured into it. It looks like PRR's circuit only boosts by 10db, while most of the commercial alternatives claim 20dB.

Still 10dB at lower noise is a better outcome than 20dB with a boosted noisefloor.

How have yours turned out?

 

[Whoops]

I want to try a couple/few different variations just to see how they work out. Even the 2 transistor original Fethead is probably worth trying despite the noise complaints.

No problems with any noise when I used my Fethead.

I like it , it gets the work done

 

[abbey road d enfer]

Which circuit are your circuit boards based on? Something posted here, or something of your own design.

My own design, although based on textbook circuits.

I want to try a couple/few different variations just to see how they work out. Even the 2 transistor original Fethead is probably worth trying despite the noise complaints. Maybe it will be different with my mic + preamp combinations.

While I was experimenting with my design, I compared with the FetHead circuit that I built with the same LSK389 fet than my design (which laso can accomodate the cheaper LSK170). I still maintain that the very high output impedance of the Fethead is an issue, particularly with transformer-based mic pres. The potential for exceeding the breakdown voltage of the LSK170 is also an issue.

I'll also try the PRR BJT one from 2005. That one seems to have the most amount of thought / math / knowledge poured into it. It looks like PRR's circuit only boosts by 10db, while most of the commercial alternatives claim 20dB.

It should not be too difficult to increase gain, if only by increasing the collector resistors (which would require adjusting bias), at the expense of increasing the output impedance.
Basically, there is a sheer impossibility in having a relatively high input Z, a relatively low output Z, and low distortion with a single stage.
That's what I, SE and maybe others have addressed. I'm not sure the Bumplebee uses the same approach; the absence of pertinent data leaves us in guesswork.

Still 10dB at lower noise is a better outcome than 20dB with a boosted noisefloor.

I believe PRR did not build his circuit, he designed it and optimized it tehoretically. Correct me if I'm wrong. Theoretically, based on semi-conductors data, PRR's circuit should not be quieter than one with LSK170/389's.

How have yours turned out?

I built a couple, that I don't really use, and I sold about 50 PCB's to various individuals. I'm proud to have Ken Newman (Barry Manilow's SE) among them.
I'm in discussion about having an official kit put on the market anytime.

 

[kato]

Thank you Abbey and Whoops for helping. I appreciate all the thinking that's gone into this.

No problems with any noise when I used my Fethead.

I like it , it gets the work done

Good to know. My most used mics are my MD421 and MD441 so the Fethead piqued my interest years ago. I have quality, 65dB preamps so I don't necessarily *need* a FetHead but the flexibility to use them them with my field recorder and other lesser gear is intriguing. So FetHead has been on my watch list for many years, since back when it was still a $69 product.

Just as I was ready to buy one, I noticed they came out with a germanium model which raised suspicions about the authenticity of their motives. Putting a Gearslutz style marketing buzzword on a utilitarian product made me question its legitimacy. Then I found this this thread. and discovered they blasted back at Abbey for asking a question. Uh oh.

Engineers with legit credentials — guys like Dan Kennedy and John Hardy, and others — can happily explain any aspect of their design choices without defensiveness.

Do you suspect yours is an older 2 fet model, or one of the newer 4 fet jobs?

 

[kato]

My own design, although based on textbook circuits.
...cut...
I'm in discussion about having an official kit put on the market anytime.

It looked original. I noticed you have 6 transistors inside. Congrats on a working original design.
I might be interested in your kit if I can find the LSK389's. Or I suppose I could just use the K170's. The ones I have on hand are B-grade (not "A") and might draw too much current, I dunno.

...
I still maintain that the very high output impedance of the Fethead is an issue, particularly with transformer-based mic pres. The potential for exceeding the breakdown voltage of the LSK170 is also an issue.
...
Basically, there is a sheer impossibility in having a relatively high input Z, a relatively low output Z, and low distortion with a single stage.
...
I built a couple, that I don't really use, and I sold about 50 PCB's to various individuals. I'm proud to have Ken Newman (Barry Manilow's SE) among them.

Hey, congrats on having an original design in Ken Newman's studio.

Thanks for lending your thoughts to the discussion. (i.e "PRR's circuit should not be quieter than one with LSK170/389's.") PRR seems to agree that it's asking the impossible to get optimal performance out of 48v and a transistor based circuit.

Most of my preamps have transformers on the input. I could try it on the SSL channel.
I also have very low-gain preamps in my Metric Halo 2882 (only 42dB gain) which would benefit greatly from a front-end boost.

It's interesting that Sanken's HAD-58 can go up to 40dB.  (!)  I can't justify that expense at the moment, but good to know it's out there.

 

[kato]

This arrangement with 4 FET's in cascode is not a good idea. It has been designed to overcome the 15V limitation of the 2SK170 FET, at the expense of increasing the already too high output impedance.

What are the practical downsides of 4 fets in cascade / too high output impedance? More noise? Worse sound? The potential to shut down at the gates?

I'm intrigued by the simplicity of Lennie's design, as well as the Alctron MA-1 schematic that was posted by rogs. But Lennie showed up, posted and schem saying it was working, but asking If other people would let him know if they had problems. So it's relatively untested. Then he disappeared when he got called out for using "phase" when he meant polarity.

I dunno, maybe he'll come back when he realizes it's not personal. Engineers like to ask questions dispassionately.

Winston mentioned that  cascading 2SK170's into another pair of 2SK170's was a bad idea as well. He suggested more general purpose FETS for the cascode. But I'm not sure which pair should remain 2SK170's. I suppose I could socket the transistors and experiment.

Do you know which pair should remain 2SK170's?

Thanks for all the help. I'm ready to start experimenting.

 

[abbey road d enfer]

It looked original. I noticed you have 6 transistors inside. Congrats on a working original design.
I might be interested in your kit if I can find the LSK389's. Or I suppose I could just use the K170's. The ones I have on hand are B-grade (not "A") and might draw too much current, I dunno.

My experience is that LSK170B in parallels draw too much current. 2SK170GR is supposed to be equivalent to LSK170A. I have ordered a bunch of them for experimentation. That could be a cheaper alternative to LSK389A.

It's interesting that Sanken's HAD-58 can go up to 40dB.

If you look at mic pre specs, you will see that their EIN inevitably degrades when reducing gain. Typically, better than -128dBu is for gain >60dB, 40db gain results in -124/-126, and 20dB gain yields -120/-122.
That is why none of the xxlifters, xxhead et al can improve the noise performance of a well designed preamp in conjunction with a suitable mic.
These apparatus offer valuable performance only when one of the following problems occur:

• mic impedance too high; I bought a Fethead just because I had a crappy Thomann ribbon mic that had an impedance of 2.5kohms!
• Noisy microphone lines, where the increased level actually helps getting over noise.
• mic preamp with not enough gain

 

[abbey road d enfer]

Winston mentioned that  cascading 2SK170's into another pair of 2SK170's was a bad idea as well. He suggested more general purpose FETS for the cascode. But I'm not sure which pair should remain 2SK170's. I suppose I could socket the transistors and experiment.

Do you know which pair should remain 2SK170's?

I think he xas concerned with the 15V limit of the LSK. Being in series, the max possible voltage would be 30V. 
Since the noise performance is in the bottom FET's, it's those that need to be low-noise. The noise gain of the upper transistor in a cscode is roughly unity. The noise gain of the bottom transistor is the nominal gain.

 

[kato]

I think he xas concerned with the 15V limit of the LSK. Being in series, the max possible voltage would be 30V. 
Since the noise performance is in the bottom FET's, it's those that need to be low-noise. The noise gain of the upper transistor in a cscode is roughly unity. The noise gain of the bottom transistor is the nominal gain.

Thanks Abbey.

RE: "bottom" and "upper" pairs. Is this what you and Winston mean by upper and bottom?

 

[abbey road d enfer]

"bottom" and "upper" pairs. Is this what you and Winston mean by upper and bottom?

Yes. Cascode circuits are usually drawn with the input at the bottom, at least for tubes, NPN transistors and N-type FET's.

 

[kato]

Yes. Cascode circuits are usually drawn with the input at the bottom, at least for tubes, NPN transistors and N-type FET's.

Groovy, thanks for clarifying Abbey.  It all starts to become intuitive the more I think about it.

Thank you for indulging me. Now, I will stop posting, and start soldering.


For future travelers lost in the woods, here is the Winston quote.
And yes, he is talking about a different schematic, but with a similar topology...  [pasted below]

The schematic Rogs posted shows the 2nd pair of J-Fets connected as common gate to the first pair,  in other words, a cascode connection.
Cascode can be useful here (as I also wrote in the other thread)  but I'm not sure that another pair of the same J-Fet  type are the best choice  as you generally want a bit more voltage across  the bottom Fets than will be provided with a K170.

Back when Toshibas were easily available without selling your kids, the usual choice for the cascode transistor was the 2SK246.  But you don't necessarily need anything so special here, just something which self biases with more gate to source voltage.  I think the old J110 would work and there'd be no noise penalty for using it if the lower parts were still  K170's or a single K389.

The pair labeled "bottom" is also referred to as "2nd pair" and "lower parts" if I'm understanding him correctly.