Drawmer Gate needs repair

[living sounds]

I've got a couple of Drawmer DS 100 SCAMP noise gate modules racked up, pretty much two mono DS201 minus the link.

Now one of them has stopped gating, which means it shows the exact same behaviour in "gate" mode as in "bypass" mode. The "range" control has no effect. Otherwise it is working as it should, key signal filter works fine, all LEDs light up properly, showing the correct states.

I've measured it through and was able to follow the DC control signal up to 3p3 cap outlined in red. The "range" pot workes fine, no short between bypass and gated signal.

Now I don't really understand what's going on in the actual gate part of the circuit, only that it's some kind of discrete VCA, and the obvious guess being one ore more of the active devices broken. Does this make sense? Is there an obvious place to measure?

Thanks!

 

[living sounds]

Here's also the complete schematics for reference.

 

[moamps]

Is there an obvious place to measure?

Yes, check the FET.

 

[living sounds]

Thanks, will do! Does the FET provide a variable resistance to ground in order to attenuate the signal, depending on SC input?

 

[living sounds]

I took out the FET and checked it with the multimeter's diode tester. With gate connected to meter ground and the other side to drain or source there is no continuity. With the connection reversed there is a resistance of ca. 700 ohms. Is that too much? Between drain and source it's at ca. 25 ohms (datasheet says 30, so that's close enough). 

With the the FET out of the circuit the gate is now permantly closed when engaged.

Can I presume from this that the FET is broken?

Also, what would be a suitable replacement for the obsolete 2N4091 (datasheet attached) in this circuit?

Thanks!

 

[living sounds]

I now did comparisons between the broken and the working gate. The bad one's FET measures -0.15V at its gate with the noise gate closed, and -0.01V with it open, whereas with the good channel it is -7.58V closed and -1.55V open.

Also, the control voltage line (marked red in the attached image) is +9.26V open and -3.33V closed on the good channel but +10V open and -0.1V closed on the bad channel.

Any hints would be great. Thanks.

 

[PRR]

"Any" N-type JFET will work.

For very lowest THD and hiss, you want one with high Vp/Vgs(off) and low Rds(on). This will typically be called a "high current switch".

 

[living sounds]

Thanks PRR! I've got a BF245A, a PN4393, J309, 2n5457. I've also compared voltages with the JFET soldered out of the broken channel, they're very close to the good channel, so I guess the FET is really to blame.

 

[JohnRoberts]

Yup, the gate of the JFET should be high impedance,  your measurements suggest it is a low impedance..

Look at at the data sheets for the old one and pick one that is similar for range of gate voltage.

The pinch off voltage varys from JFET to JFET but that 22k pot in schematic may be a trim for that.

JR

 

[living sounds]

It works with the BF245A. But now the output is 9db below the other channel even after re-biasing the FET via the pot. This also means the gate is less effective, since the remaining signal at maximum attenuation stays the same.

 

[living sounds]

With the J309 the amplitude is the same. Distortion and noise is OK. But the other channel still sounds better, it seems to react faster...

 

[living sounds]

For very lowest THD and hiss, you want one with high Vp/Vgs(off) and low Rds(on). This will typically be called a "high current switch".

Since the Vgs value is negative, does "high" in this context mean numerically high or absolutely high?

 

[PRR]

I see.

You have to be "small" compared to the 33 Ohm gain-set resistor. The circuit is designed for the <30 Ohms of the 2N4091. You want Rds(on) near 20 Ohms.

Go here:

http://uk.mouser.com/Semiconductors/Discrete-Semiconductors/Transistors/JFET/_/N-ax1rs/

Shift-click Rds(on) 12 to 25 Ohms. Your thru-hole choices are J109 and J109_D26Z (which appears to be a bent-lead variation of J109). Both are under a buck and ample stock.

That happens to be the UK site. You should take "J109" and hammer Mouser (or other vendor) into showing a country of your choice.

My BIG question is: why would this part fail?? It has "no" in-circuit stress. And JFETs are fairly proof against silk panties in dry weather static-shock.

 

[living sounds]

Thanks. I've ordered a 2N4091 from ebay now.

Why did it fail? Well, these are SCAMP modules, I have no respective "lunchbox", so I added a homemade PSU and had it all lying on a rack connected a little flimsily. There might have been a momentary short of a power line. Or the part was just an outlier. Anyway, these gates sound great, I'll put them in a proper box now.

 

[JohnRoberts]

I've seen JFETs spontaneously fail, but I've used lots them as output (turn on/off transient) clamps, so they were exposed to static and rogue externals voltages.

JR

 

[SSLtech]

Ah... This is what happens on 50% of the dbx 165 compressors that land on my bench. They use 4 FETs in parallel (to get equivalent Rds 'on' low enough, presumably) as output transient clamps, and usually one of them fails short.

 

[living sounds]

Got the new 2N4091 - it did the trick. Thanks everyone!