Why use millennium bypass?

[pappasmurfsharem]

I've looked to see if there was another answer apart from using a DPDT switch, but haven't really found any.

Is there any other benefit or is it just so you can get a better DPDT rather then a 3PDT that might be lower quality?

Then I guess apart from people in countries where switches are hard to come by.

Has there really been many instances of the cheap 3pdt failing?

[Lurco]

maybe some answer here? http://www.diystompboxes.com/smfforum/index.php?topic=103307.msg923370;topicseen#msg923370

[JustinFun]

These days, one good reason is to avoid a 4pdt! I recently built a SHO where I wanted 2 outputs, one of which was to be be only switched on when the effect was active (as a solo boost into a 2 input vintage amp) easiest method was 3pdt plus millennium using the pole of the 3pdt you would usually use as LED switch to switch the 2nd output in and out.

I also hear (though I haven't tried this) that it can reduce pop in vintage circuits which are not true bypass LED friendly (such as Dallas Rangemaster).

[Chugs]

I use Mil bypass all the time. I find DPDT's to be cheaper and better quality than 3DPT's.

[R.G.]

Then again, why not?   

[mistahead]

Then again, why not?   

Parts count.   

I've not seen any issues with 3PDT over DPDT foot switches RE reliability - but I do not produce hundreds of units and don't buy the cheap crap when it comes to jacks, switches, pots, the "hardware" part of the circuit. I did hear some VERY cheap and VERY nasty 3PDTs were around a while ago, but when the price was in orders of magnitude lower than the average everywhere else it should have rung alarm bells, but DIY has that branch of folks who do it because they're cheap/skint and are going to get stung and be bloody vocal about it (funnily enough).

[R.G.]

Parts count.   

That's fair. It does take a diode and either a JFET or a MOSFET to make a Millenium. The deluxe version takes two diodes and the super deluxe adds a resistor. If a person is really having a tough enough time that even two more parts is a problem, that's certainly a valid issue. 

The big issue with the Millenium was once parts cost. MOSFETs used to cost $0.50-$0.75. Today, they're $0.28 in ones, and if you think you'll ever use them for anything else, 100 of them cost $9.30 cents as of today at Mouser. A 1N194 is between $0.03 and $0.05, and as low as $0.02 in 100-baggies. Resistors are about $0.02. So the incremental cost of the Millenium has come down to about $0.20 to $0.50 above the DPDT.

I don't price DPDTs or 3PDTs these days, and the quick look had them all over the map.

I've not seen any issues with 3PDT over DPDT foot switches RE reliability - but I do not produce hundreds of units and don't buy the cheap crap when it comes to jacks, switches, pots, the "hardware" part of the circuit. I did hear some VERY cheap and VERY nasty 3PDTs were around a while ago, but when the price was in orders of magnitude lower than the average everywhere else it should have rung alarm bells, but DIY has that branch of folks who do it because they're cheap/skint and are going to get stung and be bloody vocal about it (funnily enough).

Yeah, there was a big to-do over 3PDT reliability a ways back. As I said, I pretty much don't use either of them, so I can't say.

I'll tell you my biggest beef with 3PDTs - I hate running the wires to them. Wires to controls are the enemy. 

So I'm back to - why not?   

[Arcane Analog]

Wires to controls are the enemy. 

PCB mounted pots/switches should be the real enemy. Wired leads are so much more durable. I would estimate a good 75%-80% of the repairs I do on pedals and amplifiers pertain to board-mounted pots and switches - especially footswitches. I cannot believe builders use PCB mounted 'stomp' switching of any kind. The question is then not "if it will break" but "when will it break." The answer asner for these builders is surely "after the warranty has expired."

[mistahead]

Parts count.   

That's fair. It does take a diode and either a JFET or a MOSFET to make a Millenium. The deluxe version takes two diodes and the super deluxe adds a resistor. If a person is really having a tough enough time that even two more parts is a problem, that's certainly a valid issue. 

The big issue with the Millenium was once parts cost. MOSFETs used to cost $0.50-$0.75. Today, they're $0.28 in ones, and if you think you'll ever use them for anything else, 100 of them cost $9.30 cents as of today at Mouser. A 1N194 is between $0.03 and $0.05, and as low as $0.02 in 100-baggies. Resistors are about $0.02. So the incremental cost of the Millenium has come down to about $0.20 to $0.50 above the DPDT.

I don't price DPDTs or 3PDTs these days, and the quick look had them all over the map.

I've not seen any issues with 3PDT over DPDT foot switches RE reliability - but I do not produce hundreds of units and don't buy the cheap crap when it comes to jacks, switches, pots, the "hardware" part of the circuit. I did hear some VERY cheap and VERY nasty 3PDTs were around a while ago, but when the price was in orders of magnitude lower than the average everywhere else it should have rung alarm bells, but DIY has that branch of folks who do it because they're cheap/skint and are going to get stung and be bloody vocal about it (funnily enough).

Yeah, there was a big to-do over 3PDT reliability a ways back. As I said, I pretty much don't use either of them, so I can't say.

I'll tell you my biggest beef with 3PDTs - I hate running the wires to them. Wires to controls are the enemy. 

So I'm back to - why not?   

I prefer to solder wire to nine lugs (eight, nine, whatever the specific thing I'm doing is) after doing a mechanical "twist" and "pinch" on each one, on a device that is more tolerant to my imperfect soldering skills... the M's all have more heat/fail-due-to-me components. So there's two schools of though there.

And parts count for me also reduces the possible points of failure (rather than specifically cost), the M's don't offer functional redundancy by adding parts - just more things that can pop/wear/etc. The reduced points of failure makes for faster identification of issues and repair there of.

PS: I'm having an intelligent conversation (I think) with R.G. - does this forum do achievements?   

[R.G.]

Well, OK, **my** enemy. 

PCB mounted pots/switches should be the real enemy. Wired leads are so much more durable. I would estimate a good 75%-80% of the repairs I do on pedals and amplifiers pertain to board-mounted pots and switches - especially footswitches. I cannot believe builders use PCB mounted 'stomp' switching of any kind. The question is then not "if it will break" but "when will it break." The answer asner for these builders is surely "after the warranty has expired."

There's good ways and bad ways. At my day job, we use exclusively PCB mounted controls, and the failure rate (which I track) is only barely above the intrinsic failure rate of the parts, near as I can tell.

I think that this good fortune relates to a lesson I was taught by a grizzled mechanical engineer back when my pointy engineers' ears were still wet. The way he explained it is that solder has mechanical limits, although they're laughably small compared to other engineering materials. What does it in is any kind of tensile load or any repeated bending load, as its fatique cracking limits are - well, laughable.  His advice was to never attach a PCB with controls in more than one plane, and never go more than 6 inches without a mechanical support other than the solder/controls, as well as never building in a static load.

Especially a flexing load. It is substantially impossible to fix a PCB in more than one plane without introducing a bending load or a shear load, or both. Well, maybe if you're holding it in place by casting material, but that's not too practical for our purposes.

So all of our controls mount to the PCB in such a way that the support distances from the panel are in a single plane within about 0.005", down at the mechanical tolerances of the parts, and that any secondary planes with controls on them have secondary PCBs, too. Other than that, it's wires.

Single PCBs with controls and jacks at right angles give me a case of the shakes.

Worse yet, I can't use that "after the warranty has expired" thing. We just upped our warranty to lifetime.    Gulp.

It's possible to do it well - but you have to go learn the lessons, do the homework, practice, design hard, and keep on working at it. One lasting lesson I had is that product design is not a rifle shot - it's a public appearance that you can't get out of.

[R.G.]

I prefer to solder wire to nine lugs (eight, nine, whatever the specific thing I'm doing is) after doing a mechanical "twist" and "pinch" on each one, on a device that is more tolerant to my imperfect soldering skills... the M's all have more heat/fail-due-to-me components. So there's two schools of though there.

And parts count for me also reduces the possible points of failure (rather than specifically cost), the M's don't offer functional redundancy by adding parts - just more things that can pop/wear/etc. The reduced points of failure makes for faster identification of issues and repair there of.

PS: I'm having an intelligent conversation (I think) with R.G. - does this forum do achievements?   

We are having an intelligent conversation, although I think it means you've stooped to my level. 

The sad thing is that I'm more likely to hose up the wire cutting/stripping/soldering than I am the board mount parts, so our schools of thought are actually the same: don't do the stuff you're likely to mess up. We just mess up different stuff. 

[R O Tiree]

I *almost* always mount the 3PDT switch and LED to the board. Because I have a small desk-top mill I can cut slots, rather than holes, for the lugs, so they are a snug fit. Not a force-fit... just snug. I hate a rat's nest of wires crammed into a pedal enclosure. Gently push the switch home so it's properly seated perpendicular to the board, nine dabs with the soldering iron and it's done - neat, tidy and very quick to do.

I have only had one failure which I very strongly suspect was due to the client trying to reverse-engineer the pedal... It wasn't just the switch that was bent/trashed. The LED leads were twisted up and the body of the LED was cracked, components had been "pushed over" (fingers gripping the board) and pot lugs were bent out of shape. Clearly, quite some force had been applied to the board in an attempt to remove it and all the damage occurred when the LED popped out of its bezel followed by the switch body separating and thus the board assembly came free very suddenly, pulling on the wires to the pot lugs. The solder joints to the base of the switch were absolutely solid, however

[pinkjimiphoton]

Parts count.   

That's fair. It does take a diode and either a JFET or a MOSFET to make a Millenium. The deluxe version takes two diodes and the super deluxe adds a resistor. If a person is really having a tough enough time that even two more parts is a problem, that's certainly a valid issue. 

The big issue with the Millenium was once parts cost. MOSFETs used to cost $0.50-$0.75. Today, they're $0.28 in ones, and if you think you'll ever use them for anything else, 100 of them cost $9.30 cents as of today at Mouser. A 1N194 is between $0.03 and $0.05, and as low as $0.02 in 100-baggies. Resistors are about $0.02. So the incremental cost of the Millenium has come down to about $0.20 to $0.50 above the DPDT.

I don't price DPDTs or 3PDTs these days, and the quick look had them all over the map.

I've not seen any issues with 3PDT over DPDT foot switches RE reliability - but I do not produce hundreds of units and don't buy the cheap crap when it comes to jacks, switches, pots, the "hardware" part of the circuit. I did hear some VERY cheap and VERY nasty 3PDTs were around a while ago, but when the price was in orders of magnitude lower than the average everywhere else it should have rung alarm bells, but DIY has that branch of folks who do it because they're cheap/skint and are going to get stung and be bloody vocal about it (funnily enough).

Yeah, there was a big to-do over 3PDT reliability a ways back. As I said, I pretty much don't use either of them, so I can't say.

I'll tell you my biggest beef with 3PDTs - I hate running the wires to them. Wires to controls are the enemy. 

So I'm back to - why not?   

I prefer to solder wire to nine lugs (eight, nine, whatever the specific thing I'm doing is) after doing a mechanical "twist" and "pinch" on each one, on a device that is more tolerant to my imperfect soldering skills... the M's all have more heat/fail-due-to-me components. So there's two schools of though there.

And parts count for me also reduces the possible points of failure (rather than specifically cost), the M's don't offer functional redundancy by adding parts - just more things that can pop/wear/etc. The reduced points of failure makes for faster identification of issues and repair there of.

PS: I'm having an intelligent conversation (I think) with R.G. - does this forum do achievements?   

i remember that happening... but i blew it.

here ya go dude... welcome to the club...

[merlinb]

And parts count for me also reduces the possible points of failure (rather than specifically cost), the M's don't offer functional redundancy by adding parts - just more things that can pop/wear/etc.

A 3PDT contains three switches, so the chances of one wearing out are higher than for two switches (DPDT) and the M bypass.  Putting the three switches into one plastc package just creates the illusion of fewer things to fail. Solid-state solutions are always more reliable long term.

[Arcane Analog]

No matter how you spin it PCB mounted components have nothing to do with designing a better product - you can design a great layout with wire leads in mind. PCB mounted components has everything to do with cutting costs. The best practice here is wiring leads to the PCB and keeping it isolated from as much stress as possible. A 3PDT mounted to a PCB sends stress through the entire PCB each and everytime the switch is used. Likewise, a stray foot to a pot or another external control sends stress throughout the PCB.

[bluebunny]

Mmmm...  Rob will be salivating!   

[R.G.]

No matter how you spin it PCB mounted components have nothing to do with designing a better product -

Once again, "better" needs definition about how it's measured to have any meaning; even more important, in the many measurements of "better" some of the directions of measurement are either orthogonal or opposing.

A lot of consumer-people think that lower price to the end consumer *is* better. A lot of very, very fussy people to whom cost is not much of an object would say that consistency, every one coming out the same, is better. Many people think smaller is better. Some people even think that "smells like hot tube equipment" is better. Beauty - and "better" is in the eye of the beholder.

PCB construction, just like tagboard, point-to-point, and flying wires, is one tool in the tool kit of building things. I put every single control and switch in the Workhorse amps on flying wires; I was solving a different problem than is found in most pedals.

So I think I'd say that PCB mounted components may or may not be involved in designing a better product. I would also say that flying leads may or may not be involved in designing a better product. You have to say what "better" means before you can go any further than that.

you can design a great layout with wire leads in mind. PCB mounted components has everything to do with cutting costs.

IMHO, that does not follow. Yes, you can design a great layout with wire leads in mind; Workhorse amps again. PCB mounted components MAY be PARTIALLY about cutting costs, but it's as likely to be about smaller, more economical to manufacture (which is NOT the same thing as "cutting costs"), more reliable, and easier to service. PCB mounted components may have everthing to do with being able to make the thing at all, as some components are simply unavailable in other than PCB-mounting packages.

Again, the choice of wires, PCB, SMD, custom ASIC, FPLA, whatever, is to make a "better" product. But you have to state your objectives before you know if you're getting there. As Lewis Carroll said, if you don't know where you're going, any road will get you there.

The best practice here is wiring leads to the PCB and keeping it isolated from as much stress as possible.

"Best" - how's that measured? What units?

A 3PDT mounted to a PCB sends stress through the entire PCB each and everytime the switch is used

... unless the 3PDT's bushing is firmly bolted to a panel, in which case the stress is between the button and the bushing. I'd probably say the same about DPDTs as well. 

Likewise, a stray foot to a pot or another external control sends stress throughout the PCB.

More possible than with the stomp switch. However, in my day job, we would occasionally get an effect which had been "flat topped". A guitarist would use his size-14 Doc Martins to stomp slight sideways as well as down and break shafts off the knobs. Mostly this broke just the pots and the PCBs were OK. Sometimes a truly nuclear stomp could get from the pot shaft to the PCB, but not often.

And as a historical perspective, I've been involved with PCB mounted pots on effects in commercial pedals with sales in the thousands for over a decade. We have remarkably few PCB stress problems, and with that much statistical history, I would expect to see problems if they were prevalent.

As I said earlier, you have to go do the learning and work at it. You have to think about and possibly modify your designs to to it well, but it is possible to use board mounted pots reliably. Good designs fix the pots individually to the panel, so the stress is held by the pot bushing, not the pot-to-PCB solder joints. I guess that's another way of saying that you have to learn about what you use, think about how to use the training, and then keep working on doing it well.

[mistahead]

And parts count for me also reduces the possible points of failure (rather than specifically cost), the M's don't offer functional redundancy by adding parts - just more things that can pop/wear/etc.

A 3PDT contains three switches, so the chances of one wearing out are higher than for two switches (DPDT) and the M bypass.  Putting the three switches into one plastc package just creates the illusion of fewer things to fail. Solid-state solutions are always more reliable long term.

Very true - but as a "Black Box" its one "module", remember my mindset comes from PC world where we don't work out which chip on the stick of RAM is the problem, I look at it as a module to replace...

[mistahead]

Perhaps I've missed something but there is little difference between 3PDT and DPDT with relation to stress on a PCB (assuming mounted, not wired). Further to this is you are using a PCB mounted stomp switch and stress is being transferred to the PCB you've failed to design a box that distributes the most common stress it will see to its strongest points. Similar is true for PCB mounted pots, switches, whatever - if a little thump from a stray toe on my big black boots on the gain knob applies stress to the PCB instead of the supporting structure of the box something is wrong with the way the mechanical stress has been planned for.

As far as fly wires vs. direct PCB mount - what is the point here about "better" - it's possible to get away with cheating more on (not) planning mechanical stress when its off-board? True - but is that really a better practice or just a resilient workaround to poor practice?