Alchemy vs Millennium

[JimRayden]

I just found this thing from the Alchemy website: http://www.diystompboxes.com/analogalchemy/pedals/ledtrick.html

I was wondering, why is Millennium so much more complex? For less leak? How much is the "LED Trick" vs the latest "Millennium" leak relationship?

----------
Jimbo

[cmat]

I havent seen this.  I wonder if it "pops" like the mill. 1?

[stm]

I havent seen this.  I wonder if it "pops" like the mill. 1?

Upon initial inspection of the circuit it doesn't seem to have an element that may cause popping. Should this happen a small capacitor (1nF) between the transistor's base and ground should avoid it.  Larger capacitors might produce slow turn on/turn off delays--might be cool!

With the values shown, one would expect a current around 3mA on a standar RED LED, assuming an hfe of 150 for the transitor. You could use a high gain transistor like a 2N5088 if you increase the base resistor to 1 Meg (thus keeping approximately the same current). Don't be afraid of experimenting.

If using a True Green or Cyan or Blue or Violet or White LED you might need to reduce the base resistor or choose a transistor with more gain (hfe).

In fact the base resistor can be calculated as:  Rb = hfe * (Vcc - Vbe - Vled) / Iled

Where:

'hfe' is the measured gain of the transistor you are using
'Vcc' is the battery or supply voltage (typically 9V)
'Vbe' is the base-to-emitter voltage of the transistor (typically 0.6V)
'Vled' is the forward voltage of the particular LED you are using (can be obtained from datasheet or your memory)
'Iled' is the desired ON current for the LED

Best regards.

[JimRayden]

Thanks for the instructions, I'll make sure to play around with it.

------------
Jimbo

[R.G.]

Maybe I can help.

The LED trick, the Millenium and the Rat bypass all rely on the output resistance of the effect pulling something to ground. They differ in how much current the output of the effect has to pull down.

The LED trick uses a 2N3904 as an emitter follower, the Rat bypass used a darlington transistor, and the Millenium 1 used a JFET. The forward bias on the base/gate of the active element has to be sufficient to allow the active device to turn on when the control line is open, setting a minimum current that must be "eaten" by the output of the effect. Remember the output of the effect also has to pull the output capacitor substantially to ground to keep from popping.

In my ruminations designing the first Millenium, I arbitrarily set a limit of a few millivolts of allowable DC put on the output capacitor. That means that the bias element and the output resistance have to divide down 9V to a few milivolts or you will get pop. So, just in ballpark numbers, you need to use a bias element that is about 1000 times the output resistance of the effect.

If the effect pulldown resistor is 10K, no problem, a 100K resistor to switch the active device works fine. The original Rat had a 10K output pot, and that made the LED switching work great. However, some effects don't like a 10K pulldown resistor. The LED trick is fine for output resistances that are low as well. If you use a darlington device, you can up the output resistor to the 100K - 500K region and use a several-megohm resistor for the bias resistor.

I wanted the Millenium to work with any effect, and even with a 1M pulldown resistor. The bias device for the Millenium is not a resistor, it's the reverse leakage of a semiconductor diode.  Some of these have higher leakage than others, and I picked diodes that have higher leakage to get the bias UP to the point where the LEd would switch quickly enough. The leakage of a high leakage silicon diode is something like 25 nanoamperes maximum, typically 10nanoamperes.To pull 10 nanoamperes to 10millivolts, you only need R= 10E-3/10E-9 = 1E6 ohms, or a megohm. That's the value of output DC resistance that the effect must have with a Millenium bypass.

Then you must be able to make your switching device switch with that little current. To switch a 5ma LED current with 10nA,  You need a current gain of Gc=5E-3/10E-9 = 5E5 or 500K. Practically speaking, you can't do that with bipolars. If you use enough bipolars in series (i.e. darlington) to get that, you can't live with the multiplied leakage that the devices have themselves. FETs, both JFET and MOSFET have unmeasurably high current gains because they eat substantially no current, and can't.

So...

I was wondering, why is Millennium so much more complex? For less leak?

The Millenium 2 is only one part more complex, and that part is optional - it's the protection diode I threw in for careless users, the low leakage diode. The Millenium 1 works with the same setup, one JFET, one diode, and one LED, but other parts were added to cure the faults that some JFETs have, the high Vgsoff causing a dim glow when the LED should be off.

All of the Millenium bypasses, 1, 2  and C, have exactly the same leakage current, from that one high leakage diode, and it's there deliberately, being what controls the operation. The LED trick and the Rat bypass have equivalent "leakages" determined by their pull up resistors.

How much is the "LED Trick" vs the latest "Millennium" leak relationship?

As I noted at first, the Millenium bypasses all have leakages to the output resistor of under 25nA (if built correctly; that's an insanely low current). The LED trick and Rat bypass have leagakes of 9V/their pullup bias resistor, and are orders of magnitude larger.

I havent seen this.  I wonder if it "pops" like the mill. 1?

Built properly, the Mill 1 does not pop audibly at all; if you instrument it, you can find a transient in all of them. The Millenium has transients orders of magnitude LESS than the other two under identical conditions, by design.

Upon initial inspection of the circuit it doesn't seem to have an element that may cause popping.

That's because you aren't seeing the whole circuit. The pull down resistor of the effect is actually part of the circuit. The pull UP resistor in the LED trick and the pull down resistance of the effect form a voltage divider and put that voltage on the output cap of the effect to make a pop when the effect is switched back in.

Should this happen a small capacitor (1nF) between the transistor's base and ground should avoid it.  Larger capacitors might produce slow turn on/turn off delays--might be cool!

Indeed it might - the Mill 1 does this. You're actually watching the rate at which a 10nA current source (the leakage diode) charges the gate-source capacitance of the JFET.

[Marcos - Munky]

I use Joe Davisson's Led Trick in my pedals. Never tried the Millenium, they're both simple to build, I just use JD way because everything is readly available in the stores where I buy stuff, and darlingtons and FETs are a little more difficult to find.

[R.G.]

The bottom line is always use what works.

[stm]

I use Joe Davisson's Led Trick in my pedals. Never tried the Millenium, they're both simple to build, I just use JD way because everything is readly available in the stores where I buy stuff, and darlingtons and FETs are a little more difficult to find.

Hi, do you use JD's Led Trick stock?
How bad is the transient upon switching?

After reading R.G.'s explanation I stand corrected: the circuit will generate a small pop due to charging of the output capacitor. In this respect placing a capacitor from the base of the transistor to ground won't help.  JD's circuit is attractive due to its low parts count and easy to get by components, while the Millenium 2 is the "Pro" version. Bottom line is to experiment and see if you can live with the glitch that might produce the former circuit.

[RDV]

When I use this one and these exact parts it works flawlessly. Substitute and you'll be disappointed.


RDV