iSK Little Gem internals

[analogguru]

Now lets see:
The iSK Little Gem uses exactly the same circuit as the Samson C02 - except that Samson used a 22pF to ground at the gate of the FET.... I don't know why but maybe RuudNL has an idea. You can try to put this capacitor there and see what happens.

The nonsense I am talking about can be seen in the drawing of the output stage where two 27k resistors are mounted from emitter to collector.

In the original circuit from Schoeps these resistor are mounted from Emitter to BASE.  I think a cloner became confused with the pinout of the transistors and the rest of the cloners were following...

Sonotrad is using a very similar circuit (including the nonsense) except that they use a J305 (instead of the 2SK596) with a 1GOhm resistor from gate to ground.  Maybe this cures the noise....

 

[Khron]

No, there are no base-to-emitter resistor in the old / classic Schoeps circuit.

If anything, there's a pair of Zeners instead of those two collector-to-emitter resistors.

 

[analogguru]

No, there are no base-to-emitter resistor in the old / classic Schoeps circuit.
....

Yes, you are right, my mistake.  But in a bunch of (other) clones they are used there.

If anything, there's a pair of Zeners instead of those two collector-to-emitter resistors.

Exactly, and the zeners make sense there.  So replacing those 27k resistors by 6,8V zeners can't be a mistake.

 

[RuudNL]

I found the 22 pF between gate and ground 'strange', to put it mildly...
At first I thought it was meant as a 'pad', so I removed it, hoping that this would improve the signal to noise ratio.
But to my surprise it didn't make any audible difference.
(As far as I remember: no higher output and no improvement in noise.)

 

[Icantthinkofaname]

As expected with mics this cheap, QC is a bit lacking. I have a pair, one of them is actually relatively quiet (compared to the other one), and this one, the one I took pictures of, is noticeably louder. I just cleaned both of them and it didn't clear up any noise (not that I thought it would, I've just been opening the one quite a bit).

I've thought of using a replacement PCB as well, or rather building a new circuit, but these mics PCBs are pretty thin at something like .5 millimetres, and I don't know of any existing ones that would fit.

Edit: forgot the decimal. It's somewhere between .5mm-1mm, NOT 5mm. It's not quite 1mm, but it's a over half a mm. I'd guess it's roughly .8mmx14mmx73mm.

 

[RuudNL]

Maybe the best solution would be to throw them in the bin and buy something decent.
But of course that's not the spirit of DIY... 

 

[abbey road d enfer]

Would an EQ rolloff cap actually change the self noise though? I know a lot of the self noise is in the high frequencies, but if it's caused by the circuit and not necessarily the capsule would it help? I've been interested in adding something to roll off a bit of the high end anyway, so if it will cut down on self noise that's a bonus.

If you think there's too much of something (treble, sibilance, presence, whatever...) why don't you use an EQ, either hardware or software? In particular it will allow EQing exactly what you want, instead of altering a wide range of frequencies.
Typically, microphones have a presence hump due to diffraction that is bell-shaped and thus should be compensated with a bell-shaped cut. Applying a low-cut is gross.

 

[Icantthinkofaname]

Maybe the best solution would be to throw them in the bin and buy something decent.
But of course that's not the spirit of DIY... 

I'm probably gonna keep them around for instruments, but they're a bit too noisy for voice recording I think. I really like the bodies because they don't have that silly body vent nonsense a lot of cheap SDCs have, and the capsules sound decent enough. But it seems like only a 1mm pcb max could be held, and the body can really only accommodate a PCB 14mm (maybe 15mm) wide and around 73mm long.

I actually bought these when I was trying to figure out the issues with my MK-012s which ended up being noisy Chinese capsules.

If you think there's too much of something (treble, sibilance, presence, whatever...) why don't you use an EQ, either hardware or software? In particular it will allow EQing exactly what you want, instead of altering a wide range of frequencies.
Typically, microphones have a presence hump due to diffraction that is bell-shaped and thus should be compensated with a bell-shaped cut. Applying a low-cut is gross.

I've done a bit of EQ, and I liked it when I EQ'd it to be similar to my MK-012 (off topic but I love the iSK Pearl which is kinda like the MK-012 with the the low end rolled off). It's really just the self noise that I don't like.

 

[abbey road d enfer]

The nonsense I am talking about can be seen in the drawing of the output stage where two 27k resistors are mounted from emitter to collector.

In the original circuit from Schoeps these resistor are mounted from Emitter to BASE.  I think a cloner became confused with the pinout of the transistors and the rest of the cloners were following...

Actually these 27k resistors do not make much difference, whether they are E-C or B-E.
There could be a difference in operating current with resistors connected B-E, but their value would need to be much lower (<10k) to be significant. Then there may be other issues with the loading of the 1st stage.
In any case they are nonsense but a benign one.

 

[Icantthinkofaname]

Has anybody in the thread built a quiet mic using electret capsules? I have a couple iSK Pearls (also electret) that advertise a self noise of 16 dBA. They certainly don't sound noisy enough for me to doubt that.

Maybe I'll have to open them up and see what looms different from the Little Gems.

 

[abbey road d enfer]

Has anybody in the thread built a quiet mic using electret capsules? I have a couple iSK Pearls (also electret) that advertise a self noise of 16 dBA. They certainly don't sound noisy enough for me to doubt that.

Maybe I'll have to open them up and see what looms different from the Little Gems.

16dBA translates to 78dB S/N ratio.
I you look at mfgrs specs, not many are capable of that. At JLI Electronics, there's only one, the JLI-2590A (just says >70dB), at Primo, they have 4. Transound have a couple of capsules that look suitable, though only the TSB2555 - a 25mm element - specs 80dB S/N.
AudioTechnica announces 16dBa for their 4051b, the most expensive of the line.
In short, I have some doubts about most of the claims of electret pencil mics with less than 20dBA self noise.

 

[Icantthinkofaname]

16dBA translates to 78dB S/N ratio.
I you look at mfgrs specs, not many are capable of that. At JLI Electronics, there's only one, the JLI-2590A (just says >70dB), at Primo, they have 4. Transound have a couple of capsules that look suitable, though only the TSB2555 - a 25mm element - specs 80dB S/N.
AudioTechnica announces 16dBa for their 4051b, the most expensive of the line.
In short, I have some doubts about most of the claims of electret pencil mics with less than 20dBA self noise.

I believe the Pearl uses the TSB2555 or a similar capsule. It's 25mm in  diameter, and it's noticeably bigger than the Oktava MK-012 capsules. The iSK Pearl/BM-88 actually sounds fairly good too. Kind of like my MK-012s with less bass, which is why I have nearly no desire to mod those ones, unless it's a mod to make the capsule omni or hyper cardioid, because it's not an interchangeable capsules mic.

The Little Gem uses 16mm electret capsules, so maybe the noise is coming from the capsules, but I don't know which models they use.

 

[analogguru]

....
I have a couple iSK Pearls (also electret) that advertise a self noise of 16 dBA. They certainly don't sound noisy enough for me to doubt that.

Maybe I'll have to open them up and see what looms different from the Little Gems.

There is not much difference:
1.) a 56pF capacitor is switched from gate to ground for the -10dB pad.
2.) the two 100nF capacitors are switched in parallel to two 10nF capacitors for the low-cut filter.

(BTW, are the capsules interchangeable from size ?)

More and more I suspect that the 2SK596 are selected from the zillions produced (and used) for el cheapo electred capsules.  I have to try if there is so much difference between them.

The funny part is that you can get complete capsules which use this FET inside (e.g. those marked with "SG") cheaper than the FET alone.

 

[Icantthinkofaname]

There is not much difference:
1.) a 56pF capacitor is switched from gate to ground for the -10dB pad.
2.) the two 100nF capacitors are switched in parallel to two 10nF capacitors for the low-cut filter.

More and more I suspect that the 2SK596 are selected from the zillions produced (and used) for el cheapo electred capsules.  I have to try if there is so much difference between them.

The funny part is that you can get complete capsules which use this FET inside (e.g. those marked with "SG") cheaper than the FET alone.

Would replacing these help at all? Or is the JFET not an issue?

The more I think about it the more I'm beginning to think the capsules are just using cheap ones from JLI or similar. I can't imagine the capsules are of very high quality (though they would be fine if not for the self noise), and most of the smaller electret capsules I find online are supposed to be quite noisy. And if the weak part of the mics is the capsules I don't think there's much I can do, since I can't open up the capsule heads to swap out the actual electret element. I guess if there are true condensers that would fit on these bodies I could build a proper Schoeps style circuit, but I'd need a PCB that would fit, or I'd have to use perf board.

What I find interesting is that the CM-20 line, which is similar but apparently employs a transformer and doesn't have swappable capsules, is advertised as having 20 dB self noise and having the same capsules. Of course I have no idea what the actual self noise of the CM-20C I used to have was, and I can't remember if it was quieter or not, and can't test it since I gave it away last winter.

 

[RuudNL]

To test this, I would do the following:

1 - Unsolder the wire to the capsule and put a 47 pF capacitor between the 'gate'  and ground.
Most probably, you will get the same amount of noise.

2- Unsolder the FET and put a resistor (3.9K or 4.7K) between the points where the drain and source were soldered.
If the noise level is now much lower, the FET is causing the problem.
If the noise level is still high, the noise may be on the supply voltage of the FET stage.

 

[snakethumper]

Hi

I've spent ages drawing  the Samson C02 circuit directly from the pcb. Here it is for any who wish to use it. The only things i werent able to identify were the 2 SMD caps. This diagram is hand drawn for now, if i get time later i'll do it properly in CAD software.

I have a few of these mics, and one of them is noisy, which im trying to fix. So far, i've only replaced Q2 and Q3, which hasnt fixed it. Currently sourcing 9.1V Zener and 2SK660 etc..

Enjoy!

[http://www.everysolution.com.au/Files/Samson_C02_Schematic.pdf]

(Please let me know if the link doesnt work)

David

 

[Icantthinkofaname]

Hi

I've spent ages drawing  the Samson C02 circuit directly from the pcb. Here it is for any who wish to use it. The only things i werent able to identify were the 2 SMD caps. This diagram is hand drawn for now, if i get time later i'll do it properly in CAD software.

I have a few of these mics, and one of them is noisy, which im trying to fix. So far, i've only replaced Q2 and Q3, which hasnt fixed it. Currently sourcing 9.1V Zener and 2SK660 etc..

Enjoy!

[http://www.everysolution.com.au/Files/Samson_C02_Schematic.pdf]

(Please let me know if the link doesnt work)

David

It might be because of the capsules. I gave up on the Little Gems because it seemed like the capsules were the biggest bottleneck.

 

[Khron]

If we're talking noise, don't discount ceramic caps. I've "revived" two Shure condensers (a KSM27 and a KSM141) which were noisy. Replacing all the tiny anti-RF caps cured the issue (and there sure were a LOT of them!).

https://groupdiy.com/index.php?topic=61079.msg773749#msg773749

Hi

I've spent ages drawing  the Samson C02 circuit directly from the pcb. Here it is for any who wish to use it. The only things i werent able to identify were the 2 SMD caps. This diagram is hand drawn for now, if i get time later i'll do it properly in CAD software.

I have a few of these mics, and one of them is noisy, which im trying to fix. So far, i've only replaced Q2 and Q3, which hasnt fixed it. Currently sourcing 9.1V Zener and 2SK660 etc..

Enjoy!

[http://www.everysolution.com.au/Files/Samson_C02_Schematic.pdf]

(Please let me know if the link doesnt work)

David

 

[xeawr]

I fully agree with that! Although it is stated that this modification would not increase the self noise of the microphone, my experience is different. Adding series resistors to the base connections of the output transistors will increase the noise level!
With 10 K there is an audible difference in self noise. Although this modification is sold (for too much money!) by a well known seller of 'microphone parts', I would never use it. If you need HF rolloff (for a K67 style capsule), there are better ways.

Hi RuudNL, you got me curious here. What would a good way to tame a chinese K67 style capsule (=add lpf) in a schoeps-style circuit?

 

[Khron]

A nF-range capacitor (or one in series with another resistor), in parallel with the JFET's drain resistor.

Hi RuudNL, you got me curious here. What would a good way to tame a chinese K67 style capsule (=add lpf) in a schoeps-style circuit?