CAD GXL 3000

[AUDIO FREQ]

Anyone have any experience with this microphone? There is one locally available for sale for 60 bucks, used in good condition. Seems to have a transformer in it. I wander what kind. Dual diaphragm, electronics look like cheapo chinese parts, but maybe I can swap or mod circuit. What do you guys think?

 

[Khron]

Depends what you're hoping to achieve..?

 

[AUDIO FREQ]

Hey Khron! To be honest, just want another project. Bored during the holiday season, and saw this online for cheap ($60 bucks only!). It has a dual diaphragm, I only own cardioid pattern mics. Maybe I could split the diaphragms up into 2, and make 2 mics out of it? Or maybe it would be better to buy it, and just mod this and have my first omni pattern mic ever! .

 

[Icantthinkofaname]

Hey Khron! To be honest, just want another project. Bored during the holiday season, and saw this online for cheap ($60 bucks only!). It has a dual diaphragm, I only own cardioid pattern mics. Maybe I could split the diaphragms up into 2, and make 2 mics out of it? Or maybe it would be better to buy it, and just mod this and have my first omni pattern mic ever! .

It's a flat circuit so you'll either want to switch it with a good K47 style capsule, CK12, or edge terminated capsule like the RK12. Or you could always just add the components to roll off the high end.

 

[AUDIO FREQ]

do you happen to know what transformer ratio it is?

 

[Voyager10]

CAD Audio GXL3000 | RecordingHacks.com says 'in the neighborhood of 2:1'.

 

[LevinGuitar]

I do find quite hard to mod that kind of circuits in a economy way, saving the capsule and transformer. As soon I use a deemphasis or try to "flatter" the response, the transformer saturates, so there have to be gain stage adjustment and low cuts, but after it, when the mic start sounding decent, the noise level becomes a small problem for a pro microphone.

Stock circuit is not a top one for an expensive capsule, and stock capsule not worth expensive transformer and circuit. Neither the body is something super for a expensive mic. For a cheap diy projects and experiments, yes, it's great.

 

[Khron]

As soon I use a deemphasis or try to "flatter" the response, the transformer saturates

So when you *reduce* the amount of (high frequency) signal going into the transformer, it saturates? I'm having trouble with the logic of that; how did you determine that's the case, though?

 

[mjmajilton]

I use the 3000 as the donor for a great multi- pattern FET mic. The 35mm K67 capsule has a lower 10k lift than the usual k67. It’s only a little bright and not harsh ( at least with my circuit changes). Major circuit changes and a 8:1 3U transformer. A Flat 47 is a good option.
Some work for a good result.
IMHO

 

[Gus]

That microphone can be a good learning experience
To adjust/build a microphone you need at a minimum a DMM, scope(or another way to view the signals without loading the circuit down) and signal generator to start to do things correctly.

If the JFET is a 170 like in the link in the recordinghacks link(Voyager10 posted)you can change it to a type that is better for this style input you will need to adjust R9 for biasing.
I like to direct couple the PNP base to the JFET drain removing C6, C5 and R7 and changing R8 and R10 resistor values and adding a part
Match up the output parts 2.2ks and other parts between the output winding and the XLR maybe remove and jumper the 47ohms or install a different matched part(helps with CMRR)
Leave the stock capsule for now
After all that then try different C4 coupling caps types and values. I often leave a 4.7uf tant in if the microphone had one stock

I inject a signal and adjust the bias looking at input and output traces on an oscilloscope

If you are clever you can adjust the drain resistor and use a cap in place of R7 to form a LPF however, you should check the the PNP is still stable you should add a part or two.

Someone I know that bought tube PA, home mono and other amps has a really good ear and would let the "sound" of the circuit set what type of guitar amp circuit would be installed or adjust the stock circuit(this was 20 or more years ago). I think the "sound" was more of the output transformer, B+ voltage and rectifier used, tubes used and some other parts.
My point is sometimes it is more fun to adjust the stock circuit changing as little as possible.

The above are starting points I adjust some other things for that type of microphone circuit

 

[micolas]

Hi, AUDIO FREQ

The gxl3000 is a reasonable microphone for its price, but it has a rather bad signal-to-noise ratio due to the jFET transistor used, which has a high input capacitance, about 30pF. The bipolar transistor does not have an important contribution to the noise increase.
In addition the jFET is not properly biased. R10 should be removed and R9 replaced with an adjustable resistor (eg 25kohm), and a manual BIAS done. I would replace the jFET with something else (eg 2sk30a).
All this has been analyzed in detail several times by the group members, the scheme is quite generic.
However, without changing the capsule you don't get a radically different character.

 

[LevinGuitar]

So when you *reduce* the amount of (high frequency) signal going into the transformer, it saturates? I'm having trouble with the logic of that; how did you determine that's the case, though?

It sound muddy and with lot less detail applying the native values for U87 kind of deemphasis and without modding the rest values of amplifier circuit. So I had to lower the low cut capacitor in the deemphasis, source fet capacitor and the output capacitor to achieve lower saturation/coloration

 

[Khron]

It sound muddy and with lot less detail applying the native values for U87 kind of deemphasis

As in, modding the entire JFET stage to be connected like the U87? How certain are you that it had the intended effect (and not cutting waaaay more high frequencies than you were hoping)? Did you measure anything, or just swapped out components semi-randomly?

"Muddy and a lot less detail" is pretty much what one would expect when low-passing a signal

 

[LevinGuitar]

I don't change whole biasing, that low ratio transformer can't drive the U87 gain level bias, I rather lower the gain stage there even more.
No, for now I just try/change (having an idea of what I will get) and listen the result, learning that way. I know ears can play tricks, but have fun that way and achieve what I want soon or late

 

[Khron]

Well, if it was "too muddy", why mess with all the other components not related to that, instead of reducing the value of that HF-attenuating capacitor? Y'know, the possibly obvious(?), simplest thing first?

 

[LevinGuitar]

It's not the HF response the problem after the mod, but the midrange and lows. The thing is to achieve a similar curve to a standart single fet, high ratio transformer and the k67 deemphasis implementation with this kind of circuit without a way bigger coloration and noise, that's somehow tricky, at least for me.

 

[Khron]

Ok, do you have a measured / "objective" reference curve you're trying to achieve? Like minus-some-amount-of-dB-of-attenuation at 5kHz, 8kHz, 10kHz, 15kHz?

I hope you're aware i'm not trying to be facetious or mean, but do you get what i'm asking?

Ok, in the U87 circuit, let's say they get -5dB of attenuation @ 10kHz. If your implementation of that (got any schematic, by the way?) somehow happened to get -10dB attenuation @ 10kHz, and/or attenuation starts lower in the frequency range... Lacking highs (and possibly mids) does lean towards what i understand by "muddy". And automatically attributing that to transformer saturation seems a bit arbitrary. But maybe that's just me...

 

[AUDIO FREQ]

Hi, AUDIO FREQ

The gxl3000 is a reasonable microphone for its price, but it has a rather bad signal-to-noise ratio due to the jFET transistor used, which has a high input capacitance, about 30pF. The bipolar transistor does not have an important contribution to the noise increase.
In addition the jFET is not properly biased. R10 should be removed and R9 replaced with an adjustable resistor (eg 25kohm), and a manual BIAS done. I would replace the jFET with something else (eg 2sk30a).
All this has been analyzed in detail several times by the group members, the scheme is quite generic.
However, without changing the capsule you don't get a radically different character.

I have a LSK489 sitting here, can I use one side of that instead?

Also, does anyone know who made the capsule? Is it a 797 capsule? God I hope so!

 

[micolas]

Hi AudioFreq!
Happy New Year!

You can only use half, the other half's pins can be connected to ground, you'll probably want to select one with a low Idss.

 

[kingkorg]

I have a LSK489 sitting here, can I use one side of that instead?

Also, does anyone know who made the capsule? Is it a 797 capsule? God I hope so!

Not 797, but a larger sibling of the infamous 32mm capsule. The one i had wasn't bad at all, it also came from CAD mic. However if you still have that modified Perception mic i don't know if you want to take the same route towards u87. Headbasket is not appropriate either. Maybe just swap the capsule for k47? Or some edge terminated?