Advice Before I Dive In

[Khron]

By C13 i meant the input cap in mobknarly's schematic, just to make it clear(er) for him

And regarding that high-pass... I'd say it was a bit of a compromise ;D Good side = you have the option available; bad side = they didn't want the extra cost of machining and mounting a switch that's accessible from the outside (possibly on its own little PCB)...

 

[ln76d]

As a reference, here's the stock schematic, from the same (quite invaluable) source

Parts names are diferent on each microphone (apex, mxl etc.)

With this mod, you actually also get to set aside a (rare-ish) 1Gohm resistor for other / later projects That's how i had spares available (for free, basically) for two AKG Perception 200's i revived (one leg of one of the 1G resistors had broken off) ;D

I found yesterday on aliexpress 1gohm resistors for cheap:
http://www.aliexpress.com/store/product/Free-shipping-20pcs-lot-1-4W-1000M-1G-precision-resistors-hundredth-accuracy-of-1-0-25W/1403812_32414713146.html
Also, even cheaper 100M (which i'll be testing) for tube mikes:
http://www.aliexpress.com/item/20pcs-lot-high-pressure-high-impedance-resistance-metal-glaze-glass-resistors-1-4W-0-25W-100M/32415962449.html

Perception series microphones can sound really good after mods, but work with it is horrible.

In fact, before starting to "dig" after a K47 capsule, it might be worth messing arround with some corrective EQ inside the mic. It's definitely cheaper than a new capsule, and even if it turns out you don't like it after all, one small cap and a trimmer's not much of an investment anyway, is it?

I proposed this before, but anyway with k47 there's no need to use any EQ, becauseh in the high frequency range there's nothing to smooth. Definately there's a little bit better low end response and peak in the hi-mid range if anyone likes that kind of sound.

http://audioimprov.com/AudioImprov/Mics/Entries/2014/2/8_EQ_Pt.3_-_Transformerless_Mics.html

I tried this type of eq for three different k67style capsules and with different adjustements -  truly i don't like it.
It can smooth high frequency, but something is wrong with it.
For me it sounds worse than extra cap on the drain, even if there is less correction.

 

[ln76d]

By C13 i meant the input cap in mobknarly's schematic, just to make it clear(er) for him

And regarding that high-pass... I'd say it was a bit of a compromise ;D Good side = you have the option available; bad side = they didn't want the extra cost of machining and mounting a switch that's accessible from the outside (possibly on its own little PCB)...

Another thing which is important (except removing C13) - in  that circuit input impedance of the preamp is 1Gohm like in original schoeps circuit - not 500M like in chinese version.
There is audible difference, i tried this circuit without C13 comparing 1G and 500M input resistors.

 

[Khron]

It's really interesting why the chinese chose to take that route - you'd expect them to want to simplify the circuits so they'd need / use the least amount of parts possible (and as a direct consequence, lower the cost) ;D

... But we digress

 

[ln76d]

It doesn't work that way, it's cheaper to use more chinese parts than hire someone to redesign or design new circuit.
(also, from mass production point of view, backplate polarisation is safer than diaphragm polarisation)
That's why for many years there were mostly two types of circuit in all of those chinese microphones.
Correct me if i'm wrong, but origin of this circuit is rode nt1/nt2 designed by Jim Williams.
Since it was made by 797 audio in china, it opened up new possibilities for chinese microphones.
Despite that NT1/NT2 was ***** schoeps circuit, glory for Rode because it opened also many options for DIY microphones!
Some time after  i started topic here about moding NT2, with all the advices from Jim Williams (collected from the net), when i understood how works schoeps circuit i realized that the original circuit and "mods" for it  are s**t as hell!
It doesn't sound good at all in both ways...

 

[mobknarly]

Thanks for the tips! I think I may be starting to figure this out. Let me outline what I have come to understand.

You change the mic from backplate polarization to diaphragm polarization. In doing so you remove the capacitor from the signal path.

You change the fet and re-bais it, using the test circuit to find the values of r7 and r10. This makes the preamp circuit input impedance 1G as opposed to 500M.  This increases the headroom of the microphone.  (Is this because of more possible dynamic range in the signal/fet?)

Referring to the GenChinaMod.pdf, I was wondering
1. How you would bypass r5 (using a jumper?)
2. What it means by (move) under that
3. How you tie the backplate back into the circuit
4. Anything I should know about removing c4
5. What is the reasoning behind changing Q1 & Q3

Thanks so much, I feel like I might be close to getting a little clearer picture of all this!!  ;D

 

[Khron]

1. You don't bypass R5, you remove it altogether And you solder the wire from the front diaphragm, to the R9/C9 joint.
2. See above
3. You solder the black wire to the stand-off where the FET gate and one end of the 1G resistor are.
4. Not really
5. The BC559C's shown in the "mod" circuit are higher-gain and lower-noise than the "stock" 2N5087's, but this is just an example You just basically want some low-noise PNP's there, not just any old run-of-the-mill ones, if possible

 

[ln76d]

5. The BC559C's shown in the "mod" circuit are higher-gain and lower-noise than the "stock" 2N5087's, but this is just an example You just basically want some low-noise PNP's there, not just any old run-of-the-mill ones, if possible

Not true
2N5087 have lower noise than BC559!
In any of these microphones i didn't found 2n5087 only 2n5551.
I will back later with updated schematic and some ideas

 

[Khron]

Nitpicker! ;D Either way, BC559's are most definitely lower noise than the more usual 5551's - happy now?

 

[ln76d]

Nitpicker! ;D Either way, BC559's are most definitely lower noise than the more usual 5551's - happy now?

It's not about that, don't take it personally
Please remember that "a lie repeated a hundred times becomes the truth" - i see that often in the net!
Am using 2n5087 very often in these circuits also BC416 and some others.
2N5087 are easy to find and do excellent job here!

So here's a schematic for Ian, there's nothing new but the devil is in the details!
I hope that it wouldn't be problem for Zapnspark, that i used his schematic
I have no time right now to draw new one!
Zapnspark - if this would be a problem i will remove it
Description i will make later

I wonder what would happen if R9 would be 1Gohm?!?

You change the mic from backplate polarization to diaphragm polarization. In doing so you remove the capacitor from the signal path.
You change the fet and re-bais it, using the test circuit to find the values of r7 and r10. This makes the preamp circuit input impedance 1G as opposed to 500M.  This increases the headroom of the microphone.  (Is this because of more possible dynamic range in the signal/fet?)

Not exactly! Compare both schematics of original circuit and modified circuit!
Changing fet and set bias doesn't change impedance to 1Gohm.
If there sits 2SK170GR or 2SK170BL (better), you don't have to change FET - at least for start.
If you will look on original circuit, both 1G R5 and R6 are connected in parallel - here's 500Mohm as a result.
Removing R5 you have 1Gohm impedance from R6.
When you changing polaristion - C4 is useless.
In "original" circuit C4 is for blocking DC voltage (distributed by R5) in the way to the FET.

For Q1 and Q3  there's also important to match them!
You can use for start multimeter with HFE measurement option.
Even if some guys here would say that multimeter matching is not worth anything  i would say it is always better than no matching at all.
You can use both models of transistors mentioned above. Look in datasheets of original transistors and a replacement, to check is the pinout the same, because it don't need to be

More tomorrow

 

[Khron]

I know, that's what the smileys were supposed to point out - that i'm aware it's all in the name of good fun

Nitpicker! ;D Either way, BC559's are most definitely lower noise than the more usual 5551's - happy now?

It's not about that, don't take it personally

mobknarly, the thing about changing this polarisation around, is that the capsule itself IS a capacitor

In the stock chinese circuit, the backplate is grounded, and the diaphragm is biased to (something around) +60v - you can't feed that straight into the FET's gate, or you'll instantly release the magic smoke Hence the need for a DC blocking cap - in this case, C4 (on the Zapnspark schematic).

By applying the bias voltage to the backplate, we can use the capsule itself "as" a/the DC blocking cap - killing two birds with one stone, as it were

(Re)biasing the FET has the purpose of maximising the undistorted / unclipped headroom. If the gate (via the 1G resistor) is biased too far off the optimal point, if we assume a constant sinewave input, one half of the signal would reach clipping before the other half. The idea is for both halves (positive and negative) of the signal to reach clipping at the same time. Is this making any sense?

 

[mobknarly]

Awesome! Thanks for the help. @Khron & @ln76d  you have been awesome... great info!  I have read through everything. It helped me under stand a lot of stuff.

@ln76d I looked up resistors in parallel and better understand the reason why it changed from 500Mohm(<-- is this correct?) to 1Gohm.  I think that if R9 is 1Gohm then the circuit would switch back to 500Mohm because then it would be 2 resistors of equal value in parallel again. (Let me know if I'm right  ;D )
I have looked over the schematic you posted, thanks for putting that up.  I think I have some of it figured out but I'm excited to hear you explanation! 

@Khron Thanks for the info about the capsule being a capacitor, and the capacitor being used to block DC to the FET. I didn't put all the pieces together until you explained it.  I have now read more info about capacitors.  Thanks for pointing me in the right direction, I'm getting a lot more figured out! Also, your other answers helped me better understand the physical changes I would have to actually make to the mic ( I've been looking at pictures of the actual boards)

Thanks for taking the time to share this knowledge!

-Ian

**I wrote this reply earlier but the site was down for an hour or 2, was it that way for anyone else? **

 

[Khron]

No worries, glad i (or "we"? ) can help

I think i speak for all contributors when i say that the knowledge we've shared, as been collected (if not "cherry-picked") from a whole bunch of sources along the years. And even *IF* all this information is available in one place, chances are it might not be all that easy to "digest" for a new-comer

I just hope you won't hate us a couple years down the line, for "infecting" you with the "bug" (and/or "enabling" you/it) ;D

 

[ln76d]

Ian,

i strongly reccomend to look for older posts written by Ricardo and RuddNL.
In case of microphone topics - these Guys are my hereos on this forums
You will find a good piece of knwoledge.

Before you start collecting parts (especially capacitors) first buy microphone
Check what you can fit etc.
For unsoldering/soldering use PB tin - not PB free.
First put new tin on soldering point - then unsolder it.

Part of circuit description!

XLR - you need to reverse wiring between pin 2 and pin 3
Unsolder both wires put in their places one side of inductors 47uH (from pcb side make shortest legs as possible), put heatshrink tubes on the wires, solder wires to the second side (hanging side) of indcturs, isolate with heathsrinks.

C2 and C28 are probably monolythic X7R ceramics - you can leave them without problem.
R1 and R2 - put jumpers in place.
D1 and D2 - check if those are zener 6.2V, if not replace it. Any 6.2V zener at least 0.5W
C3 and C7 - remove
R3 and R8 - these need to be matched - leave it to the end of job - if you will make rest of the circuit, then back to it.
You will find so many stupid descriptions over the net about these resistors - like "use 100k Dale rxx" and best part "it sound the best"...
You can use stereo potentiometer to set the values of it. Let say 200k linear pot or two 50k (47k) resitors and 100k pot.
Look on original schematic (attached simplified version - looks familiar?  ), there you have voltage values at some points of the circuit. You can determine current consumption from ohm law. If you don't want to do this go in the middle and match two 120k resistors.
C5/C6/C14 - nothing fancy schmancy here - go for some low esr Panasonic, Philips, Nichicon, Nippon, Elna, Roederstein - something good and cheap.
R4 - 6.8k  - probably sits on PCB (check) if not - put 6.8k


Rest later

 

[ln76d]

C1/C10 - should be matched, of course proper matching of capacitors isn't anything simple but for start mutimeter option isn't bad - it's always better than no matching at all
1uF is the limit value, higher value have no sense.
I didn'tt notice the negative impact of capacity in this circuit - i made AB comparision switching three values of capacitors here.
I noticed that in some circuits with transformer output, for some transformers higher capacitance gave worse results than lower.
Best option are polypropylene capacitors here, but for apex style body (and pcb) is hard to fit anything. For MXL 2006 you can forget about polypropylenes. MXL 990 without any problem to fit.
With apex is possible to fit axial polypropylene capacitors, when switch is removed  and R3/R8 with changed orientation. It's a gymnastic but possible
For axial polypropylene JB Capacitors JFGC 1uF/250V are really good and really cheap
JB makes really good capacitors!
For MKT - WIMA MKS and Evox Rifa MMK are my favourite capacitors. Roederstein also are really great.
R2/R11 can be also matched.
22nF - extra cap for K67 - for this WIMA MKP2 22nF/100V is great choice - polypro which fit without any problem on any PCB.
Q2 - 2SK170 (especially BL) are really great here. There's several other FET which are fine here - 2N5457, 2SK30 for example.
I had always problem to find J305, i tried Siliconix and Calogic in few circuits and somehow it's not my favorite FET for microphones. Jim Williams glorify it. Me not Calogic somehow was better for me than Siliconix version.
Potentiometer for fet adjustement - for 2sk170 500k trimmer is enough. You can leave it in pcb, don't have to change for resistors.
R12 - as in schoeps schematic here can be also trimmer - 2k. You can adjust plarisation voltage.
D5 - zener 7.5-10V - this is also for polarisation voltage change. If memory serves me well with 10V you will get 70-72V.
you can try first for example 9V and make little adjustement (in both ways) with R12 trimmer for comparision.
C13 and C15 are mostly NP0 ceramics here - no need to change.
C12 isn't crictical, mostly X7R are there, you can put NP0 or MKT (wima mks02) or leave it.
Q4, L1, L2, R14, D3, D4, R9 can be stock without problem if the parts are the same as in schematic.
C9, C11 can be mentioned earlier Wima MKP2 22nF - if you will buy it more, you can put them also for C2/C8 - it's not necessary but also not hurting They can be also stock caps but i recommend to go for some 100V rated capacitors.  Value can be even  47nF for example.

For parts that's it.
Matching and adjustements are important in this circuit - this will give you better results than fancy schmancy components.

 

[ln76d]

Errata for schematic which i posted.

22nF capacitor "only for K67" schould be placed on the drain R11, not R2 source resistor.
Also i recommend to reverse the capsule connection and put polarisation to the backplate and the diaphragm to the fet.
Still XLR out wires need to be reversed, and 22nF for k67 on R11.