ECM8000 Mods

Marik said:
Bob,

Are you talking about conventional connection? Then yes, it is self biased, and is done through gate to source forward current, which is very small. However, I am using it as 'cathode follower', and believe that source resistor (which is BTW another source of the noise--I am using here Vishay 0.1%) sets the current--

Hmmm.. all this talk of Johnson noise makes me wonder.. always thought that was the name for electrons banging around against moving atoms in the current flow, since we aren't dunking the whole thing in liquid Helium.. and the air motion noise would be called something else? ( Browninan noise?)

Anyway.. since the resistor is fairly large (10K) finding a low noise type might be an improvement.. wondering if anybody's done some experiments in usng a smaller one that would reduce noise contribution? (but probably also reduces the gain/sensitivity)

Don't know how much lower you could go, and then amp it back up downstream with the off mic preamp, before you lose whatever you gained in extra preamp noise, but it might be an approach to try, as well..

Damn.. I keep talking like this and I'll have to buy a ECM8000 and some WM61's, too!! ;-)

BTW, did a bit of research on phantom power balanced mic connections.. I see now why you have current restrictions with the fairly high value pair of resistors in series with the power.. and thereby why you can't easily draw more than a little current.

Wonder what the voltage is at the FET pair after the drop.. If it's close to 10v, perhaps putting the regulator upstream so it powers everything, and thereby has more current flowing thru it, or even having a reg for the FET's and the mic for two different voltages.. but there I go with more parts to squeeze in.. ;-)

Jon
 
<Ha! You think it's bad now, wait until you have kids! :p Not that they're not worth it, but they greatly limit your time, which is why you're doing all this and I'm not helping. :( >

Hopefully the kids are not in our plans right now--she is a dancer (to avoid any assumptions--she is a professor of modern dance at Univercity of Utah) and is pretty young, so I have some time for my projects...

<You've made a great pioneering contribution here, one that will probably stand for years to come. And I, for one, am very grateful!>

Tim, thank you for kind words. We never know what future prepares for us, but if this place becomes little bit better, if some people will start to understand how and why microphones work that or another way (here, once again, I would re-enforce Harvey's great thread), and I was just a little part of that, I will feel good.

<Hmmm.. all this talk of Johnson noise makes me wonder.. always thought that was the name for electrons banging around against moving atoms in the current flow, since we aren't dunking the whole thing in liquid Helium.. and the air motion noise would be called something else? ( Browninan noise?)>

I just checked AIP HANDBOOK OF CONDENSER MICROPHONES. Yeah, they call it Johnson noise, so I will stick with that. :)

<Anyway.. since the resistor is fairly large (10K) finding a low noise type might be an improvement.. wondering if anybody's done some experiments in usng a smaller one that would reduce noise contribution? (but probably also reduces the gain/sensitivity)Don't know how much lower you could go, and then amp it back up downstream with the off mic preamp, before you lose whatever you gained in extra preamp noise, but it might be an approach to try, as well..>

There is a noise contribution here, however, compare to other factors, it is not that crucial. The fairly large value helps to keep the FET out of saturation. The smaller value (meaning less bias) will somewhat improve sonics, however at the price of reduced overload capabilities. The gain/sensitivity will still remain the same, as source follower has unity gain=1.
Just use high quality resistor there.

<Wonder what the voltage is at the FET pair after the drop..>

it is 27V on the drain of upper shoulder. Voltage on pins 2 and 3 drops down to 35V.
 
I did the Linkwitz mod on a WM61b which seems to have lowered it's sensitivity about 12db. Here is a pic of a side by side. It was easy to cut the ground connection with the tip of a razor blade scratching around the post.
 

Attachments

  • mic (3).jpg
    mic (3).jpg
    26.9 KB · Views: 1,058
I need to redo the linkwitz mod, I have .8 oms resistance from the post to the case. I am not sure but that might explain the drastic db loss.
 
<I pulled apart my mic and here is what was inside.>

Huh baby! Swapmaster is the first person has it with transformer!
OK, I will draw schematics and post latter tonight. Pay attention at different construction of the ring around XLR connector. Mine version doesn't have this ring --it is integral part of the body. Interesting. Thank you Swapmaster.


<I did the Linkwitz mod on a WM61b which seems to have lowered it's sensitivity about 12db. Here is a pic of a side by side. It was easy to cut the ground connection with the tip of a razor blade scratching around the post.>

Yes, it should lower the gain, but probably only by ~5-6db. The source should be well isolated from the body of the capsule.
What is a tyope of transistor Q1 on the PCB?
 
Here is the reverse engineered schematics. It is hard to see component values, so if you can, please add it to the picture.
 

Attachments

  • behrschema.jpg
    behrschema.jpg
    17.3 KB · Views: 1,029
A simple mod.

While shooting the pictures of my ECM8000, I noticed that the writing on the body was flaking off in some places, so I decided to clean it with a little Chemtool B-12 from Berryman (it's engine cleaner).

I put a little Chemtool on a paper towel, and the writing came off completely. No risidual stuff, and now nothing to identify it as anything but an unmarked omni measurement mic.
 
To do the Linkwitz mod on a WM61b. Scratch around the grounded post with the tip of a razor blade to cut the ground. I have done a few of them now and it only takes 2 or 3 minutes. The other post easily solders to the silver area with the green dot in it (see picture in earlier post). That's it, you are done, less than 5 minutes.
 
Here is the schematics with values.
Yes Swapmaster, it seems that it is easier to cut and make connection with the WM61B vs. WM61A. Thanks for the tip.
Have you connected it to the mic yet? Since this capsule has a much higher output, I would expect the output remain the same with WM61 and Linkwitz mod, compared to stock connected WM60.
 

Attachments

  • behrschemafinal.jpg
    behrschemafinal.jpg
    34.9 KB · Views: 1,132
The .8 ohm resistance dropped db's by12 over an unmodified one. Got a good connection now and it is about 5 db down from the stock mic and I am very happy with it. I don't have a mic preamp yet so I removed the preamp from the mic and made a steel adapter to mount an RCA jack in the bottom of the Behringer. I feed it into a modified Radioshack SPL meter which has a reasonably clean and very flat preamp then from the line out of the meter to the laptop.
 

Attachments

  • mvc-001s.jpg
    mvc-001s.jpg
    33.1 KB · Views: 1,069
Marik, here is a closeup of the unknown cap. How can you take closeups of these things easily? I am using a digital camera looking through the little high mag bubble on a magnifying glass.
 

Attachments

  • mvc-004s.jpg
    mvc-004s.jpg
    14.1 KB · Views: 1,061
Marik, how will the preamp in my mic with the transformer compare to the newer ones without the transformer? And can it be modded to the same characteristics you are getting with yours?
 
Last edited:
Marik said:

<Anyway.. since the resistor is fairly large (10K) finding a low noise type might be an improvement.. wondering if anybody's done some experiments in usng a smaller one that would reduce noise contribution? (but probably also reduces the gain/sensitivity)Don't know how much lower you could go, and then amp it back up downstream with the off mic preamp, before you lose whatever you gained in extra preamp noise, but it might be an approach to try, as well..>

There is a noise contribution here, however, compare to other factors, it is not that crucial. The fairly large value helps to keep the FET out of saturation. The smaller value (meaning less bias) will somewhat improve sonics, however at the price of reduced overload capabilities. The gain/sensitivity will still remain the same, as source follower has unity gain=1.
Just use high quality resistor there.

Aha.. I'm in the group that Bob mentioned upthread a bit, and a pair of very interesting links that were posted show the various noise sources in electret mics, as well as a NASA study of using WM60's as cheap mics for a large array system that also goes into response and internal/external noise contribution.

Short answer.. yep, the resistor noise for a follower isn't a large part of the picture. Neat stuff though..

http://groups.yahoo.com/group/micbuilders/message/469

<Wonder what the voltage is at the FET pair after the drop..>

it is 27V on the drain of upper shoulder. Voltage on pins 2 and 3 drops down to 35V.

Aha.. There is a problem in that many regulators have too much of a input/output voltage limit to drop the full 48v - 10v required, I imagined with the resistors dropping it, it would be less than that. Makes it eaiser..

Did some searching around on Digikey last night.. there are a few Linear low noise regulators that would be interesting to try, (only 20uV of noise!) and in smal SMT packages as well.. but only no good values fixed, have to use a variable.

And 10v isn't that rare of a value as I thought, for the more everyday 78Lxx series of TO-92 parts..
 
Last edited:
Back
Top