Microphone boosters?

[user 37518]

A friend of mine asked me about microphone boosters such as the cloud lifter and so on, my guess is that they are useful when recording quiet sounds with low sensitivity microphones such as ribbon mics, in some cases, your preamp may not have enough gain to fully raise the signal up to a desired level. But what if your preamp does have enough gain? these manufacturers claim that your signal will be less noisy, but I say that unless you are using a preamp with a really lousy Noise Figure or high noise floor, I don't see how it could makes things less noisy.

Don't want to sound like a smart ass but Friis equation basically says that in order to achieve the lowest noise figure, your first stage in a chain of amplifiers should have the lowest noise figure and it should also have the highest gain, because the following stages will have their noise figure divided by the gain of the previous stage, so having a first stage (mic booster) with 20dB of gain followed by a preamp with another 40dB of gain (for example) should make things worse, not better! that is unless of course your preamp is a noisy preamp.

My way of thinking is that instead of using a mic booster, you are much better off using a silent, clean, transparent , low noise pre with lots of gain in the first place.

Thoughts?

 

[Bo Deadly]

Do they really claim a general noise improvement? I don't recall seeing that exactly. They claim it provides "clean" and "transparent" gain but the only mention of noise improvement is with respect to assisting an under-performing pre or long cables.

Seems like a good item to have in your tool-box but it's way overpriced for what it is.

It's also a little annoying that they have a patent when the essence of it is pretty much the same as the old Schoeps circuit.

 

[user 37518]

Do they really claim a general noise improvement? I don't recall seeing that exactly. They claim it provides "clean" and "transparent" gain but the only mention of noise improvement is with respect to assisting an under-performing pre or long cables.

Seems like a good item to have in your tool-box but it's way overpriced for what it is.

It's also a little annoying that they have a patent when the essence of it is pretty much the same as the old Schoeps circuit.

Do you have a link for the Schoeps circuit you are refering to? A friend of mine asked me if I can make him a "booster" circuit, I was thinking on just using a 1510 with 20dB of gain. The cloudlifter mentions that they use JFETs, but to me something like the 1510 should do the trick, some 7815 and 7915  plus a 1510 with 20dB fixed gain snugly fitted in a small enclosure and powered from 48V phantom. I might aswell make him a full fledged preamp for that matter...

 

[emrr]

Every user claims a noise improvement, and no manufacturer discourages these proclamations.  Some make them themselves.  It’s impossible.  Use a line amp after a decent preamp if you need more.  Better yet, digital gain in post. 

But.  Now the kidz also believe such devices are essential for “driving the preamp”.    How do you argue with that?  They all want 5%+ distortion on everything. 

 

[abbey road d enfer]

A friend of mine asked me about microphone boosters such as the cloud lifter and so on, my guess is that they are useful when recording quiet sounds with low sensitivity microphones such as ribbon mics, in some cases, your preamp may not have enough gain to fully raise the signal up to a desired level. But what if your preamp does have enough gain? these manufacturers claim that your signal will be less noisy, but I say that unless you are using a preamp with a really lousy Noise Figure or high noise floor, I don't see how it could makes things less noisy.

Don't want to sound like a smart ass but Friis equation basically says that in order to achieve the lowest noise figure, your first stage in a chain of amplifiers should have the lowest noise figure and it should also have the highest gain, because the following stages will have their noise figure divided by the gain of the previous stage, so having a first stage (mic booster) with 20dB of gain followed by a preamp with another 40dB of gain (for example) should make things worse, not better! that is unless of course your preamp is a noisy preamp.

My way of thinking is that instead of using a mic booster, you are much better off using a silent, clean, transparent , low noise pre with lots of gain in the first place.

Thoughts?

I fully agree with your analysis when all the elements in the chain are correct.
I started an interest in mic boosters when I had a T-bone ribbon mic that performed lousily. The rest of the chain could not be seriously incriminated (Focusrite ISA430MkII, Canare cable). I found out the mic had an actual impedance of 2500 ohms!
I then purchased a FetHead, which actually improved performance, because it's input impedance of 44 kohms did not smother the mic and FET's do not have significant noise current (I found out later that it's not entirely true).
The ISA did not see a correct impedance, since the FetHead has a very high output Z, but with the elevated level it was acceptable.
There are at least three threads in this group about this very subject,
https://groupdiy.com/index.php?topic=5743.0
https://groupdiy.com/index.php?topic=72896.msg939468#msg939468
https://groupdiy.com/index.php?topic=74016.msg937308#msg937308

You may find there my own take on the subject.
It is used by several persons, so it must do something, or is it collective hallucination?
I believe in most cases it helps hiding interference noise in long mic cables.

 

[Bo Deadly]

Do you have a link for the Schoeps circuit you are refering to? A friend of mine asked me if I can make him a "booster" circuit, I was thinking on just using a 1510 with 20dB of gain. The cloudlifter mentions that they use JFETs, but to me something like the 1510 should do the trick, some 7815 and 7915  plus a 1510 with 20dB fixed gain snugly fitted in a small enclosure and powered from 48V phantom. I might aswell make him a full fledged preamp for that matter...

After two clicks on abbey's links, everything you need is here:

  https://groupdiy.com/index.php?topic=5743.msg959936#msg959936

This is the Schoeps circuit:

  https://billthompson.us/bt/sch/cmc5-schematic.jpg

It's a mic circuit so it's got a lot of extra bits that you wouldn't need. Using the phamtom supply impedance as the load of some transistors is essentially the same conceptually but I must admit the whole JFET cascode thing is simpler and more elegant. I would just copy that CL1 circuit. You might need to sort transistors though. You probably won't find 2SK170 but there are active parts that come close enough (actually the fake 2SK170 on Ebay are actually not bad but you would definitely need to test and sort them).

 

[abbey road d enfer]

Do you have a link for the Schoeps circuit you are refering to? A friend of mine asked me if I can make him a "booster" circuit, I was thinking on just using a 1510 with 20dB of gain. The cloudlifter mentions that they use JFETs, but to me something like the 1510 should do the trick, some 7815 and 7915  plus a 1510 with 20dB fixed gain snugly fitted in a small enclosure and powered from 48V phantom. I might aswell make him a full fledged preamp for that matter...

You can't use a 7915 with phantom power, which is positive only. The main challenge with all these in-line boosters is the limited current available.
At 20dB gain, the 1510's EIN is 4.6 nV/sqrtHz, which is about 10dB higher than a discrete FET differential stage.

 

[JohnRoberts]

Seems like pretty good marketing and overly generous claims.

In their marketing they claim patented technology but did not provide patent number.

A search at USPTO revealed Pat # 8,433,090  Cloud ,  et al. April 30, 2013

The patent describes a rounded magnet structure and "novel" (cough) phantom powered JFET preamps. Curiously there is no mention of the preamp in the claims, so it is not patented, or novel apparently.

Doing a search of JFET cascode circuits I found almost the exact circuit

The only difference is that their circuit uses a resistor in place of the J511 current source, and does not use drain resistors on the cascode JFETs just connecting them directly to the mic inputs (using the phantom power resistors for drain loads.

JR

 

[user 37518]

You can't use a 7915 with phantom power, which is positive only. The main challenge with all these in-line boosters is the limited current available.

Yeap that's what I just remembered that the 7915 is for negative voltages only, I'm just so used to it that I forgot.

As for the noise case, I just looked at the 1510 datasheet, indeed the noise spec at 20dB of gain is 4.5nV/rtHz, those pesky resistors in the gain select circuit are the culprit,  I know that theres an incentive to use an all-discrete stage due to the high voltage, but I might just go with a discrete BJT input stage followed by an opamp and a voltage regulator, why use a FET?

 

[user 37518]

This is the Schoeps circuit:

  https://billthompson.us/bt/sch/cmc5-schematic.jpg

It's a mic circuit so it's got a lot of extra bits that you wouldn't need. Using the phamtom supply impedance as the load of some transistors is essentially the same conceptually but I must admit the whole JFET cascode thing is simpler and more elegant. I would just copy that CL1 circuit. You might need to sort transistors though. You probably won't find 2SK170 but there are active parts that come close enough (actually the fake 2SK170 on Ebay are actually not bad but you would definitely need to test and sort them).

squarewave, thanks for the links, that circuit you posted is for an internal mic amp coming straight out of the capsule. The output of a mic will be balanced and low Z, so that 1Gig resistor is not needed and the circuit should be different. 

 

[JohnRoberts]

Yeap that's what I just remembered that the 7915 is for negative voltages only, I'm just so used to it that I forgot.

As for the noise case, I just looked at the 1510 datasheet, indeed the noise spec at 20dB of gain is 4.5nV/rtHz, those pesky resistors in the gain select circuit are the culprit,  I know that theres an incentive to use an all-discrete stage due to the high voltage, but I might just go with a discrete BJT input stage followed by an opamp and a voltage regulator, why use a FET?

JFETs have lower input noise current, but higher noise voltage. They are generally optimal for higher source impedances. Lower source impedances do better with bipolar (lower noise voltage but higher noise current).  Evaluating noise involves an equation with multiple variables.

JR 

 

[user 37518]

JFETs have lower input noise current, but higher noise voltage. They are generally optimal for higher source impedances. Lower source impedances do better with bipolar (lower noise voltage but higher noise current).  Evaluating noise involves an equation with multiple variables.

JR

Thanks John, I am aware of that, my question is why did they use a JFET with this low source resistance were a Bipolar is better?

 

[Bo Deadly]

Thanks John, I am aware of that, my question is why did they use a JFET with this low source resistance were a Bipolar is better?

Probably because it would require a lot more parts. You would need bias resistors and coupling caps. And the noise improvement would be marginal because there's not much gain.

 

[user 37518]

Probably because it would require a lot more parts. You would need bias resistors and coupling caps. And the noise improvement would be marginal because there's not much gain.

That makes sense..

 

[Vac11]

You may find there my own take on the subject.
It is used by several persons, so it must do something, or is it collective hallucination?
I believe in most cases it helps hiding interference noise in long mic cables.

Not hallucinations, your design works great!

 

[john12ax7]

Don't want to sound like a smart ass but Friis equation basically says that in order to achieve the lowest noise figure, your first stage in a chain of amplifiers should have the lowest noise figure and it should also have the highest gain, because the following stages will have their noise figure divided by the gain of the previous stage, so having a first stage (mic booster) with 20dB of gain followed by a preamp with another 40dB of gain (for example) should make things worse, not better! that is unless of course your preamp is a noisy preamp.

My way of thinking is that instead of using a mic booster, you are much better off using a silent, clean, transparent , low noise pre with lots of gain in the first place.

Thoughts?

You don't necessarily need the first stage to be the largest gain,  it just needs to be high gain with low noise,  then the downstream stuff becomes negligible. You also want to have this high gain with low noise before any  significant system losses.

I agree with you that you are generally better off just just using a high quality preamp to begin with.  My guess is many went out and bought SM7B mics and then were disappointed using them with budget on board pres,  which created the need and demand.  Plug the SM7B into a good pre to begin with and it's unnecessary.

The other case would be a really long cable , so it makes sense to amplify before this loss.  You would better off just moving your mic pre closer,  but if that's not practical it could be a decent solution.

 

[user 37518]

You don't necessarily need the first stage to be the largest gain,  it just needs to be high gain with low noise,  then the downstream stuff becomes negligible. You also want to have this high gain with low noise before any  significant system losses.

Yes, thats what the Friis equation says.

In the case of my buddy, he just bought a ribbon mic and wants to record acoustic guitar with it, he says that his API pre is almost at 3/4 of the max gain and that he is hearing considerable hiss. I think the noise is mainly coming from the mic rather than the pre, but it wouldn't hurt miking one of these boosters. The API 3124 spec sheet says that the EIN is -129, but doesn't specify if its dBu or dBV, either way it ought to be lower than the mic noise.

 

[JohnRoberts]

Thanks John, I am aware of that, my question is why did they use a JFET with this low source resistance were a Bipolar is better?

Sorry if you mentioned what the actual source impedance is, I didn't see it.  I am speaking in broad strokes and most of my low noise design experience was last century, back then bipolar was the only game in town for low Z (150-200 ohm) mics. Just taking a quick search for low noise JFETs I found one claiming 1.8nV/rt Hz and that was not a rigorous search so there are probably better parts now.  FWIW the low noise bipolars I've been using since the 70s were well less than 1 nV/rt HZ.

IIRC Brad was working on a crazy low noise JFET mic preamp before his death... (no I don't have any details, but I think his design involved a trick custom PS, because of poor PSRR.)

That said a JFET with low enough input noise voltage is a win-win, but they also have to deal with input capacitance. Perhaps why that Cloud design used a cascode topology, to reduce input capacitance feedback from the drain voltage swinging with output signal.

I don't recall these parts from back in the day when I was designing but the 2sk170 claims 0.95 nV/rt HZ... not too shabby.  I used a 2sk117 for a phono preamp front end back in the 80s and it didn't suck. There may be better parts than that around today.

JR

 

[Bo Deadly]

Note that JFET and bipolar noise is very comparable above 100Hz or so. It's just low frequencies that really ruin the numbers. So if it were possible to add a low-cut option to the CL1 circuit, that would be a triple plus because it would 1) increase dynamic range 2) reduce hum common in many environments and 3) make the noise performance of the JFETs comparable to bipolar.

 

[abbey road d enfer]

why use a FET?

Because BJT's have more THD, so require NFB.
The way BJT-based in-line boosters are arranged, they use voltage-to-current NFB, simply by installing resistors between collector and base, which reduces significantly the input impedance. The BJT boosters proposed by PRR and Ricardo have an input Z of only a couple hundred ohms, which is not desirable. Dynamic and ribbon mics want to see at least 2k. (Ricardo's circuit is for very low Z mics, typically 30 ohms).
Due to current constraints, it is not possible to use voltage-to-voltage NFB, because in order to maintain proper noise performance, it would involve very low resistors in the NFB loop, which the circuit would not be able to drive.
Operated with almost no NFB, FET's offer acceptable THD performance, as well as adequate noise level.