KR - Mega Vibe

Torchy · Post by **Torchy** » 23 Jan 2008, 16:35

Hang on a minute ... The original RAT bypass (the first schem you posted) used a BJT for led switching and could affect the tone as explained in RGs article. It was clever but crap. RG used the FET and the reverse biassed (leakage) junction to ensure reliable led 'on' state and to reduce/eliminate pop at the output (caused by the fx reaction to the output dc pulldown). The RAT second-issue used a FET for the switching (the second schem you posted)

If you cant see the difference between the RAT second-issue schem you posted and the RG schem you posted you need to look again. Look at WHERE the pn junction is in RGs schem - it is across the G-D junction and Vcc. Look at the RAT schem - no pn junction, two series small-sig diodes (1.2V drop), and a 22M (!) resistor across G-D and Vcc. The two behave similarly but get there by different means.

The original point that the MB is true bypass is still correct.

Shitty Medeocre Tone · Post by **Shitty Medeocre Tone** » 23 Jan 2008, 17:24

Both look like FETs to me in the schematic.
I am no Rat guru by any means and do not even own one.
I am also not trying to get anybody riled up and just wanted to expand on what is and what is not "Original".

Back on track, so swapping out a diode for a resistor, at GD junctions, is different enough to claim as one's own?
The two series diodes causing the 1.2v drop is a real non-issue as you could do similar by raising the resistor value. I don't know the resistor values used in either design so I am not sure if this compensation was made or not.

My conclusion is that it's a good mod to an old design.
I also have much respect for RG and not picking on him here so please do not take it as such.

I have seen others get designs named after them that tell me it actually was not even their idea and actually borrowed from somewhere else to be used in a guitar circuit. This is no slam so please don't let any of this ruin your day.

Looking at the data sheet for the 2n5457 I see a SGD junction on the device and this too tells me it's a FET
http://www.fairchildsemi.com/ds/2N/2N5457.pdf

If I read you post correctly you are saying this schemo is from the Rat2 ?
OK, I didn't even know that there was a RAT2 ...LOL
Do you have a RAT1 schemo that we can look at?
Did the RAT2 come out after RG posted his MB schemo?
I mean it is possible, just not likely, the modding could have been the other way around.

Most curious

Another point is that I have heard, from a reliable source, that MB, while is True Bypass, can develop issues over time. It was explained to me why but I don't remember.

soulsonic · Post by **soulsonic** » 24 Jan 2008, 01:22

I'm sorry, but what is your point? I thought it was considered common knowledge that R.G.'s Millennium Bypass is an improvement of the LED switching circuit as used in the Rat2? I've known this since around '99-'00 when I first saw it on R.G.'s site. He's never tried to fool anyone about it's origins, so why are you making a big deal about it?
It really looks to me like you've got an axe to grind about something else.

Shitty Medeocre Tone · Post by **Shitty Medeocre Tone** » 24 Jan 2008, 03:42

soulsonic wrote:I'm sorry, but what is your point? I thought it was considered common knowledge that R.G.'s Millennium Bypass is an improvement of the LED switching circuit as used in the Rat2? I've known this since around '99-'00 when I first saw it on R.G.'s site. He's never tried to fool anyone about it's origins, so why are you making a big deal about it?
It really looks to me like you've got an axe to grind about something else.

No axe to grind here as this was a KR Vibe post that seems to have taken a turn into bypassing for what ever reason.

Personally how it looks to you is not important.

soulsonic · Post by **soulsonic** » 24 Jan 2008, 04:07

Shitty Medeocre Tone wrote: Personally how it looks to you is not important.

Of course not.

Shitty Medeocre Tone · Post by **Shitty Medeocre Tone** » 25 Jan 2008, 02:04

OK, so true/Millenium bypass is an improvement to the original design.
Agree or disagree?

Is smaller size an improvement or does the original layout size have something to do with the tone of the original UV?

Anybody miss the expression pedal or is the big speed know good enough?

pf_fan · Post by **pf_fan** » 26 Jan 2008, 12:43

Torchy wrote:Best Vibe replacements came out of the tests as PerkinElmer VT33N2s.
http://www.alliedelec.com/Search/Produc ... 080A35E17F

Smallbear has acceptable alternatives - these are the discontinued Mouser stock that Steve bought out :-
http://www.smallbearelec.com/Detail.bok?no=343

And what about this photocells? Any experience?
http://www.smallbearelec.com/Detail.bok?no=712

Torchy · Post by **Torchy** » 26 Jan 2008, 15:42

Sorry, not used those - IIRC they have a slower rise time response but quick (dark) fall time and should work fine. Try them and let us know

R.G. · Post by **R.G.** » 26 Jan 2008, 15:51

Shitty Medeocre Tone wrote:I think RG developed MB to be another internet myth with little truth beyond coining the name for the design.
...
OK, so true/Millenium bypass is an improvement to the original design.
Agree or disagree?

You're perfectly entitled to think that. However, you're wrong in that belief. The Millenium has a highly specific - and original - improvement to the Rat bypass. I've explained this many times on line.

I agree that the original Rat bypass with the bipolar transistor was my starting point. The Millenium was an improvement to that design. I believe it was a patentable improvement, and that is an informed belief as opposed to a conceit.

The original Rat bypass used a bipolar darlington transistor and was biased with a 10M or 22M resistor. That's the one I saw.

That JFET version schematic that WAS shown in this thread was put into Rats after I published the Millenium. They changed to the JFET after seeing my work. But they missed the issue and went with a resistor for bias which works well enough in the Rat.

Jack Orman put a whole series of variants on the Rat bypass on his web page after I came up with the Millenium. They were resistor biased, and progressed in the direction of every higher-value bias resistors and more current gain. He used darlingtons, a triple-stage darlington, a CMOS gate, and finally the JFET one posted here. They all used resistor biasing.

What's novel, and frankly patentable, about the Millenium is the diode bias.
Reverse biased diode leakage is a source of nanoamperes that is almost constant over the whole reverse blocking voltage range of the diode. This characteristic is despised by semiconductor guys. They work furiously to eliminate it or make it not matter.

I have seen a lot of analog circuits, both discrete and integrated over the decades. That is the only one I ever saw which made use of the leakage current in a reverse biased diode for a constant current source.
It is possible that it existed, however I have not found any instance of it before then. It was novel at the time, it's certainly useful, and it was definitely not obvious to one skilled in the art, which is the definition of patentable. I chose to publish, not patent. Frankly, it was a gift.

The reason the Rat can use resistors is that they have a 10K volume pot which provides the DC resistance to ground that their circuit senses. The lower that value, the less pop you get when switching such a circuit. Conversely, the higher the value to ground of the output resistance of your circuit, the more pop you get from any external circuit which senses resistance to ground. That's because the effect resistance has to conduct the bias current of the external circuit to ground and has to pull that current to below a level which causes a pop when switched back in. The bias on the Milllenium is so small that you can use effects with output resistances of 1M or over, depending on the actual diode.

If you try to use the original Rat bypass with resistor bias for a circuit with a 1M output pulldown, it will pop when switched. The Millenium will not.

I frankly was flattered when Proco went to the JFET style circuit. I was dumbfounded that they didn't also use the diode bias which was the heart of the matter. But then resistors work OK for them in their one particular situation.

We can all agree that a nuclear reactor is just minor refinement to the original design of a campfire as a heat source, right?

R.G. · Post by **R.G.** » 26 Jan 2008, 16:12

Shitty Medeocre Tone wrote:If you look at the LED arrangement in this Rat schematic it is what is commonly referred to as "Millennium bypass". It seems that since the 3PDT switch is as cheap as a DPDT many prefer to not use Millennium bypass since there is less to go wrong over time.

It's important to understand the history. At the time I came up with the Millenium, 3PDT switches cost on the order of $15-$25. A couple of diodes and a FET were much cheaper. Over the next couple of years, the street prices of 3PDTs dropped to about $6. If the 3PDT was that cheap in 1999, I probably would never have had the attention span on the problem to come up with it.

And while we're at it, "less to go wrong over time" is a mistake. In general, mechanical devices, especially switches you stomp on, are about two to three orders of magnitude more likely to fail. The third section of a 3PDT switch is much more likely to fail over time than the electronics in the Millenium bypass. If you really need it, I think I can dig out the mean time to failure (MTTF) calculations; but it should be obvious.

The fact is, many prefer the 3PDT to the Millenium because (a) the 3PDTs have gotten so cheap and (b) wiring a switch is mentally simpler than understanding and wiring up a circuit if you can barely solder. I think that the fact that most PCBs for clones do not have the Millenium circuit on board accounts for that. If the thing were on the PCB, hobbyists would simply put in the parts and cope with it. When presented with a choice of wiring a switch, which they already have to do, or making a new circuit board which they aren't skilled at, they go for the switch wiring. But it is very much divided and each person is their own special case.

Shitty Medeocre Tone wrote:I wonder if anybody will make a Uni-Vibe like the original or if there is a demand for such a thing.

I posted a PCB layout for making a clone of the Univibe to the usenet news groups back before there was a world wide web. It contained the "cancel" option of the original. That has persisted throughout all of my updates to that.

You can buy a licensed version of my Neovibe layout today from General Guitar Gadgets which contains the same cancel option.

dai h. · Post by **dai h.** » 27 Jan 2008, 02:29

That JFET version schematic that WAS shown in this thread was put into Rats after I published the Millenium. They changed to the JFET after seeing my work. But they missed the issue and went with a resistor for bias which works well enough in the Rat.

when did the Rat II bypass come out? I have a hand drawn Rat II (I drew myself--unit borrowed from a relative) from before I got on the internet ( before around '97?--to be clear here I mean I recall "getting on" the internet around late '97--I drew the schematic before then when I also lived somewhere different) which shows a 2N5458 (and 1M R, 2 1N4148s, one 2.2k R, and one LED). Since I had my hand drawn schematic, at the time I saw the millenium bypass it didn't seem like a big deal (doesn't mean there wasn't a big deal since I didn't/don't have too much technical understanding) because it appeared the Rat II bypass was the same thing (again, I would not have been able to discern any fine technical differences--this was just my sentiment at the time).

R.G. · Post by **R.G.** » 27 Jan 2008, 04:30

dai h. wrote:when did the Rat II bypass come out? I have a hand drawn Rat II (I drew myself--unit borrowed from a relative) from before I got on the internet ( before around '97?--to be clear here I mean I recall "getting on" the internet around late '97--I drew the schematic before then when I also lived somewhere different) which shows a 2N5458 (and 1M R, 2 1N4148s, one 2.2k R, and one LED). Since I had my hand drawn schematic, at the time I saw the millenium bypass it didn't seem like a big deal (doesn't mean there wasn't a big deal since I didn't/don't have too much technical understanding) because it appeared the Rat II bypass was the same thing (again, I would not have been able to discern any fine technical differences--this was just my sentiment at the time).

I don't know.

If the JFET version predated the Millenium, then I have inserted my big foot into my mouth; it's happened before. However, I saw the bipolar darlington one before the Millenium was released, the JFET version afterwards, and it was my assumption that they picked the JFET up from me. If you have good evidence that the JFET version of the Rat bypass predated the Millenium, I have no reason to doubt your veracity, and I'll go research the history to be sure of my facts.

Which is kind of a moot point. Neither version uses the diode leakage bias which is the essence of the Millenium, and what lets it be so sensitive to DC resistance to ground. Correct me if I'm wrong, but don't all versions of the Rat use a 10K or so volume pot that's connected to their bypass to switch the LED? That low value lets them use resistive pullups without popping.

The key technical issue in the Millenium - and it's not obvious unless you're looking for it - is that the tiny bias current lets you just paste this thing onto effects which have only a capacitor pulldown resistor of upwards of 1M or more without it popping. It can be retrofitted onto a much wider range of effects.

dai h. · Post by **dai h.** » 27 Jan 2008, 05:53

I can't be absolutely positively 100% sure since I don't have the unit and don't have a picture. On the paper I have that circuit section (with the FET, R, diodes) drawn out separately with "On when no signal" and the FET drawn as bipolar but no arrow (probably an attempt at trying to understand how it worked) and "low freq./low noise N-Channel FET, and MOTOROLA 2N5458 1-99 .50, 100-499 .41, p.106" and it's on the back of a Digi-Key "thank you for your interest" type of letter, so I'm thinking I didn't know what what a "2n5458" was, until I was able to look it up in a Digi-Key catalog (pre-internet) and might have re-drawn it as an FET (a pencil was used and not ink) after I was able to get that info (I vaguely recall doing that drawing out the circuit from my smallbox Rat--no LED and has a 2N5458 at the output).

Correct me if I'm wrong, but don't all versions of the Rat use a 10K or so volume pot that's connected to their bypass to switch the LED? That low value lets them use resistive pullups without popping.

on my hand drawn schematic (of the Rat II), I have a 10k fixed R bypassing the output pot (value unmarked), so that seems to fit.

soulsonic · Post by **soulsonic** » 27 Jan 2008, 06:20

I recently traced my early/mid 90's Rat2 and it does indeed have a JFET in the arrangement Dai has described with the 10K pulldown resistor before the 100K Volume pot.
Knowing that the whole time hasn't affected my opinion that R.G.'s Millennium is an improvement, though I had always assumed he was improving that earlier JFET design specifically and not the earlier Darlington one. If anything, that makes R.G.'s idea seem that much more ingenious for not having been aware of the JFET version, and he is correct in saying that the JFET version uses a large resistor to bias with, just as the BJT one did, and it required the 10K resistor in parallel with the pot to make sure it would function correctly. Clearly, the Millennium is a substantial improvement.

The workings of this era of Rat are fully documented with gigantic pics in my build blog at: http://www.illuminist.tk

dai h. · Post by **dai h.** » 27 Jan 2008, 07:29

The key technical issue in the Millenium - and it's not obvious unless you're looking for it - is that the tiny bias current lets you just paste this thing onto effects which have only a capacitor pulldown resistor of upwards of 1M or more without it popping. It can be retrofitted onto a much wider range of effects.

ahhh (

ding ding), yeah that would make the millenium more useful (would not need that 10k for it to work if I understand correctly).

R.G. · Post by **R.G.** » 27 Jan 2008, 14:31

soulsonic wrote:I recently traced my early/mid 90's Rat2 and it does indeed have a JFET in the arrangement Dai has described with the 10K pulldown resistor before the 100K Volume pot.

The workings of this era of Rat are fully documented with gigantic pics in my build blog at: http://www.illuminist.tk

OK. I stand corrected on the history. Doooh! What a dummy!

The diode bias can be used in other ways - as in the Millenium 2, where it turns on a MOSFET as an inverter, not a follower. That gets rid of the annoying tendency of the JFET in the Mill 1 to leak a bit into the LED. It can also turn on a CMOS gate, as in the Mill C which lets you do other stuff besides just an LED, such as muting an input or some internal point in a pedal's circuit.

I have since then used the diode-leakage bias for biasing CMOS opamps. It will work for some JFET input opamps as well and some comparators. Some of them need too much input bias current, but there are some that work.

omarvolta · Post by **omarvolta** » 16 Mar 2008, 22:51

It would be great if you would post a pic of whats under the shroud in the mega vibe.

and two more...

dai h. · Post by **dai h.** » 18 Mar 2008, 16:14

hey thanks.

What's up with the fifth(?) photocell?

MKB · Post by **MKB** » 18 Mar 2008, 17:49

FWIW, and this has probably been mentioned elsewhere, with some modification of the cap values you can use quite a range of photocells as long as their response time is sufficient. I had a bad problem finding photocells with both the response speed and that would have a low enough Ron to give a good sweep. I finally found some at Mouser that were fast enough, but didn't sound very good as they wouldn't go very low in resistance. After reading RG Keen's excellent Univibe article and seeing how the photocells react with the caps in each stage, I tried scaling the caps in each stage to give the same response with the higher resistance photocells. The new cap values work extremely well; this copy (built on RG's old layout with an added FET buffer input) has plenty of wobble and sounds great. If anyone is interested I'll try to dig up the spreadsheet where I figured out the values.

omarvolta · Post by **omarvolta** » 18 Mar 2008, 18:16

dai h. wrote:hey thanks. What's up with the fifth(?) photocell?

that photocell goes in between the lamp and the second LED which indicates the rate of the vibe.