Topic: lowering noise on a DIY mic

[Al Taylor]

Hi all,
I just built a DIY mic with a transound capsule, lifted the schematic from the mp3forkidz website.  I'm very happy with the sound that the mic is capturing, but the noise floor is way too high.  I notice that when I cut phantom power to the mic it becomes dead quiet and very sensitive until the 220uf filter capacitors drain, so I assume that I would be capable of lowering the S/N ratio of the mic during normal operation if I made the right adjustments. 
My experience up until now has been mostly with tube amplifiers, and I don't really know the source of the noise issues, whether it's coming from the source resistors or electrolytic caps, or if there's a property of the transistors that I'm not taking into account.  Any suggestions would be welcome!
I was thinking of putting a couple of low value inductors, maybe 10-100uH in series with the mic output, but I thought I'd ask for advice before I went and wasted too much time on anything.  I know a good number of people have built this mic, so maybe this is a common problem?
Thank much,
Albert

[Al Taylor]

Thanks for a) all the suggestions and b) the link.  That's a great site and helps to explain a lot of the questions I had about the components I was using.
I was mistaken, I did actually use the panasonic capsule that the website has up there, and yes it's rewired for source follower mode.
Your thoughts on the noise sources sound very reasonable, so I'll focus on those for now.  I have electrolytic 1u coupling caps between the capsule and output FETs, but I replaced them with some film caps I had and didn't notice a difference, so I assume the major sources of hiss are somewhere else.  The 100n caps in parallel to the elytics seems like a relatively painless solution and I have some handy so I'll try them out.  Likewise, I'll play around with biasing the FET in the capsule.  This is my first experience messing around with biasing solid state components, so I don't expect to find the answer two quickly, but I did buy four capsules, so I'm not afraid of making a few mistakes!
Thanks again, I'll keep you updated and post clips once I get signed up on a web hosting server.

[Al Taylor]

Your intuition about the feedback at the FETs appears well founded.  I put 10K resistors between the emitter and base of each FET, lowered the value of that 6k2 resistor in the PSU to 1K (1W), and  raised the bias resistors of the FET in the capsule to 4.7K and the noise floor dropped down to nothin'... Well not really nothing, I have the revised circuit on a breadboard at the moment, so there's a lot of EM noise that better go away once I clip leads and pack the thing in it's housing (energy saver light bulbs in the shop...)
Thanks so much for your help, I really appreciate it!
-A

[Al Taylor]

oh good, thanks for letting me know.  I thought the BJTs were FETs and the FET was a JFET.
Ah, learning curves...

[Al Taylor]

I made a short recording with the two mics next to one another.  I kept the gain the same on both mics, it's very apparent how much more sensitive the modded one is.  I also removed the windscreen so that I could get a recording from the exposed capsules.  The two mics are in different housings, so I wanted to eliminate that variable. Here's a link to the mp3s. 

https://webshare.uchicago.edu/xythoswfs/webui/_xy-2485172_1-t_17PVm2HU

I'm happy with the noise floor on the second one, though I haven't really tested it out thouroughly yet.  You can hear a little whine in the background, I believe that's actually the fan from my laptop and not hiss from the mic, but I could be wrong.
My only question now is whether the SPL tolerance in the new circuit has been lowered.  I thought that maybe increasing the value of the bias resistors of the FET would bring the headroom down a little.  Any idea if that would be a concern if I wanted to record a drum kit?
Thanks again for your help!