Need help debugging Millenium bypass 2 [SOLVED]

[arjespen]

Hey.
I built a millenium bypass version 2 into my xotic ac booster clone.
I used a bs170, 2n3904, 1n914, 1k5 resistor, and a purple ultrabright led.
The led lights up fine when the effect is on, and turns almost off when the effect is switched off.
The key problem being "almost" - as I can still see a faint light from the led when it should be off....
I think I read that this behaviour could occur with the first version of the millenium, but this is the second, using a mosfet.

Anyone got a suggestion to what might be wrong?

Regards.
Anders

[R.G.]

The Millenium Bypass, all versions, is a sensitive detector of the difference between an open circuit (as in, a couple of inches of air) and a resistance to ground. The LED turns on when the control point is open and turns off when the control point is connected to ground by any resistance lower than a few megohms.

You can test the Millenium Bypass by connecting the circuit to 9V (including the LED) and manually touching the free end of a 1M resistor the other end of which is connected to ground.

The Millenium II is much sharper in discrimination than the first, JFET Millenium. It is generally fully off or fully on; in fact I've never seen one with a "soft on" like you're describing. I can hypothesize some possibilities, though.

First, test it in isolation. Open the control wire from the Millenium control node and see if the LED goes off when it's open, fully on when it's shorted. Then try the 1M ohm resistor in the control line. If this works, the fault is not in the Millenium.

If this does not work, there are really very, very few things that can cause it to not work. First, take out the 2N3904 "low leakage diode". That's a protection part, and isn't needed for most situations. Next, try a different "high leakage diode". The original 1N914s were gold doped for quicker shutoff speed, and as a result had higher leakage than ordinary silicon diodes. Modern diodes can sometimes meet all the datasheet specifications without gold doping, and so have much lower leakage, which is good for the diode manufacturer, bad for the Millenium builder.

The only thing that comes to mind other than those issues is a damaged MOSFET or oscillation in the circuit when it's bypassed. MOSFETs usually fail solidly when they fail, but I suppose it's possible for one to just get leaky and still kind of work. If your circuit is oscillating when it's bypassed, it will be driving the MOSFET on slightly with signal. Some effects don't like to have their inputs open.

Then there is the usual set of suspects: bad soldering, wrong component values, wrong pinout, wiring errors, etc. These happen vastly more often than bad parts.

[arjespen]

Thanks for your reply. :-)

I just breadboarded it with all new components of the same type.
It helped a lot, but funny enough, I can see the the led is still glowing very faintly when I connect the control to ground through a 100k resistor. (I tried a 1m too, same thing).
It's better than before, but I'm still curious as to why the led actually has that very faint glow. As mentioned this is with all new components... when the control is open: nice bright light. when using a resistor to ground: very faint glow.
As mentioned this was a purple ultrabright 5mm led. I tried with a blue led of the same type, and it had a brighter glow in both positions.
Also, I tried taking the 2n3904 away - no change.

[arjespen]

Well... turned out to be bad bs170's
tried out a few of them, and it seems the supplier had sent me a few bad ones 

[R.G.]

Ack!
It does happen. Good work tracking it down.