Ribbon mic hiss (yes, it's the mic) -- ideas? Impedance issue maybe?

[midwayfair]

I have a mic built from what was originally an Austin kit. It sounds fine (I've gotten lots of good recordings out of it), but here seems to be a hiss issue.

First, I want to be clear that the mic is not CREATING hiss. It's the same hiss (same frequency) that's in the mic preamp when it's UNLOADED, but only slightly less loud. If I vary the impedance on one of my pres (from 3K down to 200R), the hiss starts to get less around, oh, 300 Ohms, but of course there's very very little signal at that point; there's not much change above 600R. I tested a dynamic (SM48 and RE20) under the exact same conditions and there is substantially less hiss. I made a recording at +66dB gain if it's helpful (ribbon, dynamic, then unloaded). (I start out at 3K, then 600R, then 200R on each: https://dl.dropboxusercontent.com/u/9878279/Ribbon%20mic%20hiss.mp3)

I'm using a 1.2um (I believe it's 2" long) ribbon and a cinemag 9887:  0.12 : 150/1 : 35. I don't know a lot about ribbon motors, but the only thing I can think of is that there is some improper loading happening with this transformer. I don't know if it's the number of windings or some other characteristic. As a last resort, I have some 1.8um foil here but I really don't want to have to cut two more ribbons and I used up a lot of the 1.2um practicing. (Plus I do like how it SOUNDS -- I just want to get rid of the hiss.)

I only have one ribbon microphone, so I can't grab another one to compare, but it seems like something must be wrong.

 

[JohnRoberts]

In mic preamps there are multiple sources of hiss..(aka noise).

#1 the thermal noise of the mic's DC source resistance.
#2 active device ein (equivalent input noise ) voltage 
#3 active device ein current times the source resistance.

If the Mic DCR is elevated, or transformer step up scales it up, there will be more noise from #3.

The input stage current density (operating current) can make a difference.. generally more operating current results in lower ein V and higher ien I.

Do you have any way to confirm mic source impedance?

JR 

JR

 

[midwayfair]

Do you have any way to confirm mic source impedance?

Can I just measure pins 2->3? Would that give me the resistance of the ribbon itself x35^2?

EDIT: Hrm. I get 25.9R when I measure pin 1 > pin 3, but nothing when I measure pin 3 to pin 2 OR when I measure pin 3 to pin 1. Not shorted, but maybe an open circuit. That's not good. Maybe I found the problem, but I don't really know the cause. There aren't any obviously broken wires and I don't see any problems with the transformer, but I guess it's possible there's something borked inside.

 

[emrr]

You likely popped/distorted the ribbon doing that, it's an invalid test AND one you shouldn't do. 

Sorry.

AC impedance and DC resistance are different things. 

A preamp noise test should be loaded 150-200 ohms resistive to simulate a typical mic.  Comparing that result to the mic in question would reveal any noise contribution from the mic, but possibly tough as you are comparing a signal to a lack or signal on top of the noise. 

 

[midwayfair]

You likely popped/distorted the ribbon doing that, it's an invalid test AND one you shouldn't do. 

Well, lucky for me the ribbon seems to be fine and doesn't sound any different now.

Correction, though: the 25.9R is between pins 2 and 3, there was no measurement when pin 1 was involved. So ... now I have even fewer ideas.

In order for this test to be any different from the one where I plugged an SM48 in (which is a known impedance), I'd have to know the output impedance of the ribbon, and I've forgotten how to calculate that based on the area, but knowing that might not help me verify for certain that there isn't something wrong with the mic.

 

[fum]

1.2 micron foil with a 1:35 ratio output transformer is going to set your mic output impedance really high.

off the top of my head, likely greater than 1K.

$.02

 

[midwayfair]

1.2 micron foil with a 1:35 ratio output transformer is going to set your mic output impedance really high.

off the top of my head, likely greater than 1K.

$.02

I was afraid this was going to be the problem. I called a Cinemag dealer to try to locate a 1:28 in case I have to go that route, or I may simply cut the 1.8um ribbon for it.

 

[fum]

Depends where you want the change.  You could put a fet buffer (fethead, cloudlifter, whatever) between mic and pre.  May or may not be an improvement, but would likely be different.

What you are after is a much higher input impedance on the mic pre if you are gonna try and run the mic that way.

The 300 ohm mic output impedance spec is kinda like the pirate code, in that it's more a guideline.  A mic with that spec is going to perform well working into most commercial mic pre's on the market.

If you don't care about interoperability, you are free to do whatever sounds good.  The fet buffers may have a much higher input impedance than 3K.

I've never cracked one open, so don't make a claim to know, but typically they are going to employ a matched jfet device, and jfets have really high input impedance, which this mic would like more.  There's also some details left to the implementer (they are likely going to set an input impedance using resistors, but hopefully still leave things high, like 20K or so? <-- guessing)

The 9888 with 1.2um is going to be more like 500 ohm

$.04  Carry on.

 

[PRR]

> to know the output impedance of the ribbon

Same as any other.

Feed a steady mid-band signal. In this case, a radio is convenient; commercial over-compressed music is steady enough you can eyeball "a level".

Put loading resistors on until output level drops 6dB.

Avoid distortion (non-issue in a passive mike). Listen for tonal change, indicating impedance is not uniform over the audio band.

 

[PRR]

Putting "dummy loading" across a mike (any source) is sure to drop your S/N ratio. Yes the hiss drops, but the precious signal drops more. Don't do that.

 

[abbey road d enfer]

I get 25.9R when I measure pin 1 > pin 3, but nothing when I measure pin 3 to pin 2 OR when I measure pin 3 to pin 1.

that's absolutely normal, because the xfmr's secondary is floating. You get ca. 25 ohms between the "active" legs, that's good. 25 ohms DCR for a nominal 150-200 ohms impedance is just right.

 

[abbey road d enfer]

I'm using a 1.2um (I believe it's 2" long) ribbon

That's very likely to be the culprit. Most ribbon mics use about 5 um stock, resulting in about 4 times less DCR. Indeed there are other sources of Johnson noise, in particular the primary's DCR and contact resistance. Ribbon mic designers loose a lot of hair trying to improve the contact resistance between the primary's wires and the ribbon itself.

I suggest you join the micbuilders group, where you can get a lot of hands-on info.

 

[fum]

Let's correct that a bit.

Most pro ribbon mics use a ribbon of roughly 2um.  A lot of crappily made mics use 5um, and they sound like it.

I don't have any hair, so there is that....

 

[midwayfair]

Hrm, after doing some more testing based on the suggestions today, I'm starting to think it must be some other problem than the impedance. I tried:

1) A FET preamp I made that has 25K input impedance. This is probably the highest in my house. There is more hiss with the ribbon than the dynamic, and this is certainly high enough impedance to bridge whatever's coming out of the ribbon mic. One thing that is much clearer when doing this test is that it sounds even more like the preamp is unloaded.
2) A cloudlifter, which is 3K input impedance, pretty much identical results to the 3K setting on the Pro MPA II.

Indeed there are other sources of Johnson noise, in particular the primary's DCR and contact resistance. Ribbon mic designers loose a lot of hair trying to improve the contact resistance between the primary's wires and the ribbon itself.

I'm not entirely certain what I could do about this, unfortunately.

I decided to open the mic up today and it looks like I have to re-ribbon it anyway, as I can see a couple very tiny tears about 1/3 of the way down the ribbon. I'm going to move to the 1.8um foil, shorten the xfo wires as much as physically possible, and if that doesn't help, then I'm just going to have to live with it.

EDIT: I want to specify that despite an apparent fault in the ribbon, this isn't a new issue or a S:N ratio issue. It's just that there's more hiss when this mic is plugged in than other passive mics in my collection.

Also, I shorted the Xfo wires and I don't hear a change. Re-ribbon next.

 

[JohnRoberts]

Hrm, after doing some more testing based on the suggestions today, I'm starting to think it must be some other problem than the impedance. I tried:

1) A FET preamp I made that has 25K input impedance. This is probably the highest in my house. There is more hiss with the ribbon than the dynamic, and this is certainly high enough impedance to bridge whatever's coming out of the ribbon mic. One thing that is much clearer when doing this test is that it sounds even more like the preamp is unloaded.

My comments were not about the "input impedance" of the preamp (typically 10x source impedance for bridging termination), but the "source impedance" of the mic.

This source impedance will exhibit thermal noise by itself, and additional noise from the active preamps input stage current noise (times that source impedance). .

2) A cloudlifter, which is 3K input impedance, pretty much identical results to the 3K setting on the Pro MPA II.

Indeed there are other sources of Johnson noise, in particular the primary's DCR and contact resistance. Ribbon mic designers loose a lot of hair trying to improve the contact resistance between the primary's wires and the ribbon itself.

I'm not entirely certain what I could do about this, unfortunately.

I decided to open the mic up today and it looks like I have to re-ribbon it anyway, as I can see a couple very tiny tears about 1/3 of the way down the ribbon. I'm going to move to the 1.8um foil, shorten the xfo wires as much as physically possible, and if that doesn't help, then I'm just going to have to live with it.

EDIT: I want to specify that despite an apparent fault in the ribbon, this isn't a new issue or a S:N ratio issue. It's just that there's more hiss when this mic is plugged in than other passive mics in my collection.

Also, I shorted the Xfo wires and I don't hear a change. Re-ribbon next.

Does it make more signal along with the more hiss? If yes, don't worry about it. If more hiss for only the  same signal as from other mics  you can worry.

JR

 

[midwayfair]

Okay, re-ribboned with 1.8um. No audible change. Unfortunately I only got one ribbon out of the whole sheet this time.

>Does it make more signal along with the more hiss? If yes, don't worry about it. If more hiss for only the  same signal as from other mics  you can worry.

Well, it's less signal strength than the 48 or the RE20. But both of those are expected to be higher sensitivity. I'm going to try to borrow another ribbon microphone to run through the same preamps. At this point I'm not sure what else can be tested.

 

[fum]

How about a couple pictures of what you are up to.

We're almost at 1000 words, so it will be worth it.

 

[midwayfair]

Sure thing.

The wires are physically as short as I can make them and still get the motor in the body. To be honest, they're at the length where it's a major pain to get the assembly in and out. The only thing I could think of that would make them shorter is to not twist the ones coming off the motor.

The ribbon, when lying flat, is about 1/2 way between the middle point and the surface of the table. I left about 3 corrogations of slack. It seems to accordion fine when it's upright. (Again, though, the actual sound seems to be fine.)

 

[fum]

Electrically, that looks OK.  Don't sweat the length of those primary wires, they are fine.  Longer would be fine too.

One thing, and it may just be the angle of the picture.  I see solder on the mechanical/ground lug of the the Switchcraft connector.  Just want to make sure that in no way it is making connection to pin 2 (the orange wire).  The body ground lug is tied to Pin 1 in most cases.

Secondly, as long as those copper bits are making good contact with the ribbon (on the clamp side of things), and the clamps are tight(any unevenness in the the copper bit, where it's making contact would be bad, needs to be flat), then you should be OK on the wiring side.

Whether the ribbon is made well or not, the output should be OK.  You've a bit of a sag and a twist in your ribbon, but it should pass audio.  If you want an example of what you are shooting for, it should be more like the bottom picture in this link:

https://repair.ulrigg.com/cascade-fathead-ii/

With that combination of foil, and output transformer, the sensitivity should be decent.

The other possibility is there is nothing wrong with the ribbon mic, and that's just the noise level.  Are you using the microphone to record sources, or just listening to the noise?

 

[midwayfair]

Whether the ribbon is made well or not, the output should be OK.  You've a bit of a sag and a twist in your ribbon, but it should pass audio.  If you want an example of what you are shooting for, it should be more like the bottom picture in this link:

https://repair.ulrigg.com/cascade-fathead-ii/

Wish I had a way to corrugate a ribbon that well

The other possibility is there is nothing wrong with the ribbon mic, and that's just the noise level.  Are you using the microphone to record sources, or just listening to the noise?

I've used it to record a ton of stuff ... most of the electric guitar and bass tracks I've done for the last year since I built it.

My main reason for starting the thread is that I had recently done a comparison clip using it and the RE20 and the noise (hiss) level was noticeably higher. So I plugged in a few of the other dynamic mics I had around the house and they were all basically the same as the RE20. Since a dynamic mic doesn't generate hiss on its own, I was afraid something might have been wrong with it.

I really appreciate all the help with troubleshooting even if it turns out to be nothing, seriously.