Millenium 1 led lights up over time... when it is supposed to be off

[rikkards]

I recently built a opamp buffer (which works great) with a MosFet Booster using a DPDT switch. This works fine. However I wanted to use a Millenium Bypass for an LED indicator which also works fine (layout identical to what is in geofex).  I used an 1n9414a for the diode and a 2n5457 for the JFET.

What appears to be happening is when only the buffer is being used (booster is activated by the switch), the LED will turn on slowly over the next 5 minutes. The measurements for the JFET are below:

D: 8.4V (battery is a little old)
G: 64mV
S: 1.6V

What I think is happening is the current is trickling higher at the Gate over time which is activating the LED slowly. But I can't figure how to stop this from happening. Any ideas? Not a big deal as it is noticable if the buffer is on or off but it is nagging me to get this to work.

[R.G.]

There was, after all, a reason for the Millenium 2. The Mill 1 is picky. The Mill 2 and Millenium C are much easier. Look at the Millenium C.

[rikkards]

There was, after all, a reason for the Millenium 2. The Mill 1 is picky. The Mill 2 and Millenium C are much easier. Look at the Millenium C.

Thanks for the reply
Interestingly, I took your advice and went to this link:
http://www.geofex.com/Article_Folders/Millenium/The%20Next%20Millenium.pdf
and looking at the Millenium 1 Plus which is what I built, R2 should be 10k-100K so that it stays dark when off. I used a 4.7K as per the schematic on http://www.geofex.com/Article_Folders/Millenium/millen.htm. I may see if I substitute a 100K pot and see if that dims it down.

If not, then I may do the Millenium C. What is a good CMOS chip to use in the Millenium C?

[rikkards]

There was, after all, a reason for the Millenium 2. The Mill 1 is picky. The Mill 2 and Millenium C are much easier. Look at the Millenium C.

Thanks for the reply
Interestingly, I took your advice and went to this link:
http://www.geofex.com/Article_Folders/Millenium/The%20Next%20Millenium.pdf
and looking at the Millenium 1 Plus which is what I built, R2 should be 10k-100K so that it stays dark when off. I used a 4.7K as per the schematic on http://www.geofex.com/Article_Folders/Millenium/millen.htm. I may see if I substitute a 100K pot and see if that dims it down.

If not, then I may do the Millenium C. What is a good CMOS chip to use in the Millenium C?

Answered my own question. I think this may do the job:
74HC05 http://www.datasheetarchive.com/pdf/2611821.pdf

[rogeryu_ph]

That' what happened when I used 2n5457. Try to use 2n5458 this will solve your problem if you have.

Roger

[miqbal]

That' what happened when I used 2n5457. Try to use 2n5458 this will solve your problem if you have.

Roger

Yeah, I never have a fail Mill1, if an 2N5458 is in place.

[rikkards]

That' what happened when I used 2n5457. Try to use 2n5458 this will solve your problem if you have.

Roger

Yeah, I never have a fail Mill1, if an 2N5458 is in place.

I'll try seeing if a higher value resistor works and if not, then my next order to Small Bear (BSIAB2 and Tychobrahe Octavia) will include the 5458s

Thanks all!!

[amz-fx]

Be sure to use the two 1N914 diodes connected to the source as shown on the Rat bypass:

http://www.muzique.com/lab/true_bypass.htm

This helps maintain the source at a voltage which the gate can be pulled below to switch it off.

regards, Jack

[rikkards]

Be sure to use the two 1N914 diodes connected to the source as shown on the Rat bypass:

http://www.muzique.com/lab/true_bypass.htm

This helps maintain the source at a voltage which the gate can be pulled below to switch it off.

regards, Jack

Damn too many choices on how skin this cat! Wish I had come across that one before posting I would have tried that before.

[R.G.]

As I say, go to the Millenium 2 or C.

[rikkards]

I would except the pedal has already been built and the board is too small for anything like the C. I used a Vero board with about several columns of two strips of 6 holes separated. When I built it on a proto board everything worked and if I had probably waited about 5 minutes I would have probably seen the light kick in.

Ah well, next time I will probably look into doing the Millenium C.

As I say, go to the Millenium 2 or C.

[R.G.]

Try this: clip the control line for the Mill 1 loose and connect a 1M resistor to the control line of the Mill1 temporarily. The light should go off and stay off, not increase over a few minutes. Let me know what it actually does.

[rikkards]

Didn't get an opportunity to try that as I changed the circuit to the one here: http://www.muzique.com/lab/true_bypass.htm before reading your response.
But it was still lighting up. I then took a 1M resistor and threw it between the output after the coupling cap  to ground and it has been about 10 minutes and no glow. Doesn't look to be affecting the sound either but then again I figure I have a bit of a tin ear. 
If this still doesn't work, I can always flip it back to what it was and try your idea. The advantage of small part count.
Thanks for the help!!!

Try this: clip the control line for the Mill 1 loose and connect a 1M resistor to the control line of the Mill1 temporarily. The light should go off and stay off, not increase over a few minutes. Let me know what it actually does.

[R.G.]

um.... there was no DC resistance to ground after the output coupling cap?

If so, that was your problem, not any funny business with diodes and JFETs.

The Millenium in all it's forms is an excruciatingly sensitive detector of resistance to ground. When there is a resistance to ground of 1M or less on the control line, the LED goes off. When there is an open circuit to ground, the LED goes on. When there is a *capacitor* there, no DC resistance to ground, you see the LED get brighter over a period of minutes as the leakage from the high leakage diode charges the cap. You're actually watching the cap ramp up over that time, being charged with 10-25 NANO-amperes leaking through the diode.

The 1M to ground on your coupling cap output ought to be there for reducing switch pop anyway. It will not affect your sound, and is standard practice among those with experience.

So no, voodoo diodes won't help, but the DC resistance to ground is critical to any version of the Millenium working.

[rikkards]

I feel so stupid.  (we need a slap on the forehead emoticon)

After reading that it makes complete sense. Doh!! ~( : o (|)

You da man and thanks for making it so clear.

um.... there was no DC resistance to ground after the output coupling cap?

If so, that was your problem, not any funny business with diodes and JFETs.

The Millenium in all it's forms is an excruciatingly sensitive detector of resistance to ground. When there is a resistance to ground of 1M or less on the control line, the LED goes off. When there is an open circuit to ground, the LED goes on. When there is a *capacitor* there, no DC resistance to ground, you see the LED get brighter over a period of minutes as the leakage from the high leakage diode charges the cap. You're actually watching the cap ramp up over that time, being charged with 10-25 NANO-amperes leaking through the diode.

The 1M to ground on your coupling cap output ought to be there for reducing switch pop anyway. It will not affect your sound, and is standard practice among those with experience.

So no, voodoo diodes won't help, but the DC resistance to ground is critical to any version of the Millenium working.

[amz-fx]

Didn't get an opportunity to try that as I changed the circuit to the one here: http://www.muzique.com/lab/true_bypass.htm before reading your response.
But it was still lighting up. I then took a 1M resistor and threw it between the output after the coupling cap  to ground and it has been about 10 minutes and no glow. Doesn't look to be affecting the sound either but then again I figure I have a bit of a tin ear.

Glad you got it working.

regards, Jack

[R.G.]

I feel so stupid.  (we need a slap on the forehead emoticon)

I have this flat spot on my forehead from slapping it.  Things like that are the HARDEST things to find because your mind has already checked them off as OK. It's not being stupid - it's being human.