ECM8000 Mods

Marik

Pro Microphone Design
OK folks,

Today I went to local GC and got ECM8000--just to play with it and to see what it's capable of. Of course, before the listening, the very first thing I opened it up. I swear, I remember seeing somewhere on forums a picture of its PCB with tiny transformer. The picture on the box also shows a transformer. I got quite a shock--you know, when you expect something for sure and it's not there. It was not in mine! Anyway I don't care much for that, as you can imagine which quality it should be considering $39 for the whole thing. I opened the capsule (I had to put some strength). I heard rumors, and it indeed suspiciously looks like Panasonic WM60:

http://www.panasonic.com/industrial/components/pdf/em07_wm60_a_dne.pdf

I 'reversed' schematics. The input stage is on BC118 (silicon transistor), and output on A1349--I could not find any documentation, but it looks like two transistors in one body. The crutial signal path capasitors are surfice mounts :rolleyes:, and the board is full of other 'goodies'.
As for the mods.
The very first thing is capsule itself. The Panasonic (as many other electrets) has a built in FET. This FET is responsible for the noise and low SPL. The easiest way to go is to make a 'Linkwitz mod'. Originally, the FET is hooked up without source resistor. The mod is to hook it as a source follower, which reduces noise and increases SPL to 140Db.
There is another capsule--WM61 which has about 4Db lower noise, but also lower SPL. However with Linkwitz mod the SPL is still at very respectable 134Db. The ultimate solution would be disassemble the capsule, cut off the internal FET and instead, use 2SK170, selected for high transconductance. I will probably have to order some Panasonics and play with them.
Tomorrow I will start serious listenings, measurements, tweakings, recordings, etc. I will also try to use Oktava 012 pre (I will have to modify it a little to make it work) to see if and how much difference it makes.
So, Gentlemen,
If you are interested and don't get bored, please let me know and I will keep you posted on the progress.
 
Marik said:
The very first thing is capsule itself. The Panasonic (as many other electrets) has a built in FET. This FET is responsible for the noise and low SPL. The easiest way to go is to make a 'Linkwitz mod'. Originally, the FET is hooked up without source resistor. The mod is to hook it as a source follower, which reduces noise and increases SPL to 140Db.

Right on.

It's about time someone came up with a mod for that thing. It just seems like such an obvious candidate for it.

Keep us posted, and thanks.
 
Marik said:
...I swear, I remember seeing somewhere on forums a picture of its PCB with tiny transformer. The picture on the box also shows a transformer. I got quite a shock--you know, when you expect something for sure and it's not there. It was not in mine! ...However with Linkwitz mod the SPL is still at very respectable 134Db. The ultimate solution would be disassemble the capsule, cut off the internal FET and instead, use 2SK170, selected for high transconductance. I will probably have to order some Panasonics and play with them...
I didn't think there'd be a transformer, but you never know. That's really good 'cause had there been one, it'd be pretty cheap and would have just caused another "work-around".

As far as the capsule's FET causing the noise, I guess that kills the FET replacement idea. Anyway, the Linkwitz mod is pretty simple. It's like: Doh! Why doesn't Panasonic sell them like that in the first place??!! And do we know for sure the sensitivity isn't going to drop after changing the polarity of the FET? I wonder if we can find out what Earthworks is doing that makes theirs so much quieter.

Concerning the surface mount caps, the hard part’s not replacing them, it’s finding out what’s in there in the first place.

(edited to add more info)
 
Last edited:
This is a post from Arny Kruger on rec.audio.pro this morning:

"A well-known local choral group - the Detroit Concert Choir use a QTC1 matched pair mounted on a high stand. That plus spot mics for the soloists.

If you want to listen to some real-world evidence of how noisy QTC1s really are, here's where you can get their recordings for a nominal charge: http://www.detroitconcertchoir.org/ . You can probably email them and find out which mics were used for which recordings.

I'm using what I thought are the far noisier ECM8000s, and the noise that shows up on my recordings is basically the room, not the mics. (see additional comments below).

Here's a recent technical discussion of the self-noise of the ECM8000's from one of our sister NGs:


From: Norbert Hahn <hahn@hrz.tu-darmstadt.de>
Message-ID: <lvquuvcvafp6gv381hn4cul6hjd5tig5rk@4ax.com>

"Today I did some measurements with some of my mics, AKG Body C 480 B with CK61 ULS capsule (Equivalent noise level (IEC 651 A weighted) 13 dB, Sennheiser MKH50 (Equivalent noise level 12 dB) and the Behringer ECM-8000. Microphone preamp was a Behringer UB802."

"Hm, well the ECM-8000 has about 10 dB more noise than the Senn and the AKG, which measure the same when reduced to a single number."


As I read this, these measurements put ECM8000 self-noise in the 22-23 dB range. But there are some other relevant issues that are discussed in more detail about 2/3 through this post.

Note that the mic preamp used was hardly SOTA, but the mics in question are pretty sensitive so the mic pre wasn't seeing terribly low-level signals.

Ironically 22-23 dB noise puts ECB-8000s in the same range of self-noise as the spec on the QTC1 per http://www.earthwks.com/ns/qtc1.html . Yes, this is pretty stinky by the standards of some large-diaphragm mics.

Note that the ECM8000 is flat up to about 25 KHz, while the QTC1 is flat up to about 40 KHz. It's probable that the QTC1 has a far smaller diaphragm, which makes the modest self-noise spec more impressive.

Returning to the comments of the person (Norbert Hahn) who tested various mics for self-noise:


"However, the frequency curve of the unfiltered self noise is quite different: The ECM 8000 has almost white noise between 300 Hz and 15 kHz, dropping then with 6 dB per octave. Between 300 Hz and 30 Hz the self noise rises by 18 dB at lower frequencies. Thus, the self noise of the ECM 8000 sounds like a mixture of pink and white noise."

"The frequency curve of the self noise of the AKG looks like pink noise between 30 Hz and 3 kHz and like white noise between 3 kHz and 45 kHz."

"The frequency curve of the self noise of the Sennheiser MKH 50 is quite different. Between 30 Hz and 4 kHz is looks pretty much like pink noise (a little more rise at lower frequencies though), but above 6 kHz the noise rises by 8 dB within one octave, peaking at 18 kHz. Above that the noise drops by 18 dB/octave. So that mic has the least amount of noise in the same octaves where the ear is most sensitive."


This is one of those things that doesn't seem to show up on many spec sheets - the spectral content of noise very much affects how we perceive it. Spectral content is much of the difference between the sound of wind blowing, the sound of sea crashing on rocks, and a 150 psi air hose hissing its heart out given that the 20-20 KHz or A-weighted SPLs were the same.

My own take on this issue is that small-diaphragm omnis can get a bad rap for noise because they are so truly omni. If there's a noise in the room they'll probably pick it up. For example, many of my recordings include trivial but personal comments made by the musicians and people in the audience, made in low voices and whispers while they are waiting for things to start. Not so with cardioids in the same general vicinity.

I still lust after a pair of QTC1s, but its probable that I am getting a fair simulation of them with my ECM8000s. Probably good enough of a simulation to judge a number of relevant issues, like "Are small omnis for me and my application?"
 
I'm actually not bothered as much by the self-noise thing as I am with the SPL handling. It's one of those things where it would be nice to be able to take advantage of an omni for the sake of close-mic'ing without as much proximity effect. There might also be accoustic considerations, which can only be exaggerated by the characteristics an omni mic brings to the table (let's face it -- a lot of our basements aren't Abbey Road), furthering the utility of being able to close mic.

Problem is on most loud sources, it just plain clips, so really close-micing isn't an option in a lot of situations. And it is an issue with this particular mic.

Marik, I'm not sure if this helps at all, but I have a couple of Audix TR-40's here. I might be willing to lend one of them to you for a few days (for scientific purposes) if I have your word that it will be returned quickly and in one piece. :D
 
I only use my 8000s for drum overheads and they do a great job. No noise issue at all and no clipping trouble. They are a bit noisy if you try to use them for quiet acoustic recording though, so I just don't do it. For 40 bucks they're hard to beat.
 
Last edited:
Harvey, Thank you for very interesting and as always, very informative post. I will go to RAP later to check it.

<Note that the ECM8000 is flat up to about 25 KHz, while the QTC1 is flat up to about 40 KHz. It's probable that the QTC1 has a far smaller diaphragm, which makes the modest self-noise spec more impressive.>

The diaphragm in ECM8000 is actually very small--6mm diameter, the rest is just a housing. If you look at the mic from the capsule, the black thing is the actual size of the diaphragm.

Flatpicker,

<As far as the capsule's FET causing the noise, I guess that kills the FET replacement idea. >

I don't know what kind of FET is used in the capsule and its specs. It might be a very good one, but it might be not so. Only experiments can show it. The 2SK170 is a very nice sounding and very low noise device, so IMO it would be worth at least to give it a try. If it doesn't work I just loose $5 and my time, but what I learn is worth more than that. Oh yeah, and if I don't try I just won't sleep at nights :).

<Why doesn't Panasonic sell them like that in the first place??!! And do we know for sure the sensitivity isn't going to drop after changing the polarity of the FET? >

I wish they were selling them without FET at all. The sensitivity is not the biggest issue--it can be easily taken care of later, in the preamp circuit. The simple changing of bias of the first stage will do the job.

<I wonder if we can find out what Earthworks is doing that makes theirs so much quieter.>

I heard many times that Earthworks use Panasonics which are selected for low noise and frequency response, but again, it's just rumors.

Chessrock,

<I'm actually not bothered as much by the self-noise thing as I am with the SPL handling.>

Same here, my main intention is to see what this small diaphragm is capable of sonically and how far it can go. If along the way I will be able to drop the noise by some dbs it would be nice...
Of course, I will completely rebuilt the preamp section as well.

<Marik, I'm not sure if this helps at all, but I have a couple of Audix TR-40's here. I might be willing to lend one of them to you for a few days (for scientific purposes) if I have your word that it will be returned quickly and in one piece.>

You've got my word! :) It would be very nice of you, and of course, I'd be very interested to compare to mine, when I have something done with it.

Finally, I found data for A1349, so later tonight I will redraw the schematics nicely and neatly (as much as I can), and post it along with some pics, if you are interested.
 
Marik said:
You've got my word! :) It would be very nice of you, and of course, I'd be very interested to compare to mine, when I have something done with it.

Feel free to PM me with your contact info if you'd like, and I'll try to get a hold of you in the next week or so. It seems to me like the Audix does slightly better with the noise floor, and has a lot better SPL handling, so it would be interesting to know what they do to achieve that. Who knows -- you might be on to something. I'd hate to see you put Stapes out of business, though. :D
 
Marik said:
I don't know what kind of FET is used in the capsule and its specs. It might be a very good one, but it might be not so. Only experiments can show it. The 2SK170 is a very nice sounding and very low noise device, so IMO it would be worth at least to give it a try. If it doesn't work I just loose $5 and my time, but what I learn is worth more than that. Oh yeah, and if I don't try I just won't sleep at nights :).
So are you going to take the existing FET out of the capsule and swap it with a 2SK170?
 
MCM
Part Number Mfr. Name Description Unit List Price Stock On Hand Buy
Part No.
1 P-2SK170GR MCM TRANSISTOR SUBBING WITH 2SK170BL Each $2.61 0 (Lead Time 31 days)
P-2SK170GR
2 P-2SK170BLV PANASONIC TRANSISTOR Each $2.61 0 (Lead Time 17 days)
2SK170BLV
3 P-2SK170 PANASONIC TRANSISTOR Each $3.71 0 (Lead Time 31 days)
2SK170
4 2SK170 MCM TOSHIBA FET TO-92 -40V .01A .4W DGS *SS Each $0.96 420
2SK170
5 P-2SK170V PANASONIC TRANSISTOR Each $2.79 0 (Lead Time 31 days)
2SK170V
6 P-2SK170BL PANASONIC SUB ONLY - USE ITEM# P-B1CACB000002 Each $2.04 0 (Lead Time 0 days)
2SK170BL

From http://mcm.newark.com/NewarkWebComm...archResults.jsp?action=0&First=0&QText=2sk170

Anyone know the differences? - seems to be a big price difference on the highlighted one. In particular what is "DGS *SS"? Would this one be adequate for the mic mods?

Thanks
still crazy after all these years
www.crazy_as_a_loon.com
 
Scott Dorsey informed me that the Behringer ECM-8000 uses a capsule made in the Phillipines. He says it's very similar to the Panasonic capsule and quite good.

Eric Blackmer told me some years ago (at an AES show) that he uses a special Panasonic capsule, made to his specifications.
 
This is an interesting thread. In fact, I've been planning to do very similar stuff. I've actually ordered a 2SK170, because I've seen it used in many microphones. I haven't gotten one yet, though. They seem to be hard to get in Europe or at least Germany. I've also bought a number of Panasonic capsules. Unfortunately you can't get them directly from Panasonic here but there is a distributor that sells them as MCE2000 capsules. I've bought those from two different dealers and got slightly different capsules. The ones I got first had small SMD transistors, the others had larger ones. It's pretty difficult to disassemble the capsule without destroying it.

Another difficulty is the Linkwitz mod. It's virtually impossible to solder a wire to the housing of the capsule as it is made of aluminum. I'd be happy for suggestions how to solve that problem.

Anyway, thanks for the helpful info to Marik and Harvey in particular. I'd be very interested in the schematics. Keep up the good work!

BTW: a chinese guy who once posted here said the ECM8000 is made by Tenlux. And I'm pretty certain he's right, as I saw a very similar measurement on their website when it was still accessible.
 
<So are you going to take the existing FET out of the capsule and swap it with a 2SK170?>

Flatpicker,

Do I hear sceptical notes in your voice? Tell me what's wrong with that?

Rossi,

<It's pretty difficult to disassemble the capsule without destroying it.>

A friend of mine is a jewelry maker. After I bust a few with my shaky hands, I'll bring them to her. Her tools make me trembling :) Besides, I have another idea, which I have to check first. Anybody wants to guess what's that?

<It's virtually impossible to solder a wire to the housing of the capsule as it is made of aluminum.>

Once I tried to solder aluminum using Aspirin as a flux--it worked. However, you don't need to connect pin 1 to the ground if you use 3 wire connection. The biggest problem is not to overheat the capsule, as you can easily destroy electret. A good and healthy heatsink will help. For those who doesn't know about Linkwitz mod, you can find it on:

http://www.linkwitzlab.com/sys_test2.htm
 
Last edited:
Marik said:
Flatpicker,

Do I hear sceptical notes in your voice? Tell me what's wrong with that?
No skepticism here. I just wasn’t clear whether you were going to do the Linkwitz mod or try and swap the FET in the capsule. If you do swap the FET, take pictures! :D

Also, seems like someone said that DigiKey was discontinuing these WM-60AYs. I can’t remember the details, though.

BTW, that's way cool using asprin for flux.

Another note: I’ll try and get you’re ML-52 pics up by either Friday or Saturday.
 
OK folks,

First report.... but before, let me tell you a story. You remember my line:
"Besides, I have another idea, which I have to check first. Anybody wants to guess what's that?"
So, my idea was to destroy the FET in the capsule, hoping that the gate will become connected to source or drain (kinda what happens to bipolars), and thus, not to disassemble the capsule in order to put in the 2Sk170. I had a few old electrets, so I disassembled them, found 1A 50V PSU, and connected it between source and drain--nothing. The FET measures perfect. I opened my NAD amp--4amp 70V (35+35) PSU, I connected it to the FET, measured the FET--nothing. Measures perfect. At this point I got so mad that I just connected the FET to the ~120V line for a moment. Encountered some nice sparks, measured the FET--nothing--measures perfect. It drove me completely nuts. I just stuck the wires from the line to the poor FET and held it there. All my fuses went off. Since I did not have any spares, I went to Fred Meyers, got the fuses and 12 pack of Full Suspension (pale ale). Here....
at the register I realized that I forgot my card at home, because earlier I made an order from Digikey for some WM60 and WM61s, and took my wallet out the pocket. When finally I got the fuses, put them in, measured the FET in the capsule, the resistance between gate and other ends was 750Ohm WTF is that? Quality control???
Sorry for telling you all this crap--I've had already half of 12 pack, and earlier drove my fiancée to the airport, so I had to express myself....

Now seriously.
I recorded some tracks through my Jordan mini monitors (sorry, my Anton Richter baby grand is terribly out of tune, so I cannot record it) with the stock mic. I placed it at 2 feet from the speakers. The tracks were excerpts form: Sibelius, Finlandia, with Malcolm Surgent, 1st movement; Prokofiev violin sonatas with Shlomo Mintz and Efim Bronfman; and Jaco Pastorius's Donna Lee. I was quite impressed with recording. The mic got very nicely the tonal balance and also, all room noises, Harvey was talking about, which usually I don't hear.
Next, I made a Linkwitz mod. Recorded the same tracks. While recording, I had to adjust the gain control on the pre about 8-10db up--of course, it is no wonder, as a source follower has a unity gain. I was very impatient to see results, as even monitoring through the headphones I immediately could hear the change in the clarity. OK, time to playback and compare.
Gentlemen, have you ever experienced the difference between, let's say, MXL 2001 and Ela251? To my ears the difference was even bigger! While original, stock mic reproduced the right tonal balance and was quite accurate, the "Linkwitz mod" had opened the sound tremendously. The main thing was that with the same tonal balance the notes became much more separated from each other, and each instrument got its own place in the sound stage. At this point the stock mic sounded flat and dull.
However, I got a problem. The circuit started oscillating. For half a minute it is OK, then little by little boiling noise starts to appear. After I turn off and then on a phantom, it disappears, but then again all the same. I will check it tomorrow.

The next steps--I will substitute the source resistor with FET constant current source and recording it. And the next, connect the capsule's FET as the first stage of the cascode amp (as always, Kevin Carter's ideas are invaluable).

So, stay tuned Gentlemen, to be continued...

(edited for spelling)
 
Last edited:
Marik said:
So, Gentlemen,
If you are interested and don't get bored, please let me know and I will keep you posted on the progress.
Yes, please keep us posted. Thanks.

PS... if you want to destroy the FET, zap it with a stun gun. ;)
 
Woah. I'm sorry if I sound stupid.. but are you making a $40 mic sounding a million times better for very little extra?

If so.. I might just have to get into this.
 
Hey Marik, I love your way of engineering! We must be identical twins separted by birth. :D

Thanks for the tip with the aspirin, I'll try that on the weekend.

Keep us posted, this is interesting!
 
Back
Top