About the common Boss SD-1 bypass problem

[shredgd]

R.G. wrote:

[quote="shredgd"Thanks for your explanation about jfet switching voltages but, please, read previous posts before answering.
Jfet switching is clearly not involved in this problem (read above), at least in the case of the SD-1.

I'm sorry. I'm getting on in years and my mind isn't as quick as it once was. Can you point out the part I missed, please? [/quote]

Simply, as I also said in my latest post, if I remove the JFET which normally blocks the effected sound in bypass mode (the one you are saying is not turning off hard enough), I can still hear the leakage in the background of my clean sound.
I think this is enough to exclude its malfunctioning!

Read back to know about my other experiments, which led me to the conclusion it must be a layout problem (some components too close to each other).

cbugatti wrote:

Does anyone have a definite solution for fixing this problem without building a bypass box? If so, how, and can you explain it at an easy level to understand?

I previously said I was giving up about this problem, but I decided I will still work on it. Analogguru posted what seems an already good solution, but I have another idea, now, which I'll write about only if it proves good, so just wait for now, I will soon explain you at least how to do the analogguru mod.

Giulio

[R.G.]

Well, as I noted, it's always possible that Mother Nature is trying to teach me something new.

Doctors have a thing they tell new apprentice doctors - "When you hear hoofbeats in the hallway, look for horses, not zebras."

If you remove the single JFET that blocks the effected sound, yes, that is an even higher resistance than a reverse biased JFET.

If you remove both JFETs - which is what I understood you did - you could well get capacitive leakage of both signals by paths which would not be operational with the "dry" signal there.

Given that, I would look to power supply  or ground trace coupling of the effect signal into the output buffer, not mysterious electromagnetic transmission or layout coupling, on the horses-versus-zebras theory. The output buffer, being a single bipolar transistor has 0db of power supply and ground noise rejection.

Does this thing have a decoupling cap on the power line?

If that's not it, I'd go to board contamination.

On the whole, it's really quite hard to radiate audio frequencies between PCB components in smallish, low power effects unless you're using JFET buffers.

[shredgd]

R.G.,
that's of course what I did, and it is the reason of the tests I did (which, again, you can read about above in this thread) after I excluded a jfet malfunctioning.
I understand you don't have the pedal and you probably didn't look at the schematic, so that's why you couldn't follow my previous posts.
Anyway, just have a look at this schematic   http://www.freeinfosociety.com/electronics/schematics/audio/bosssd1.html and consider:

- I pulled out the 10k resistor between the gain stage and the tone stage (to exclude the second half of the circuit and therefore exclude a leakage from what comes after the gain stage)
- I pulled out the leg of the 0.047u cap connected to Q2's drain and connected it to a cable to my amp

and I could still hear the leakage!!

- Then I connected the 0.047u cap of the feedback loop of the gain stage to ground instead of 4.5v (as I was afraid the distorted signal made its way into the clean path through it, via the 100k resistor above, in the image)

Still leaking!

So, today I did analogguru's mod (http://forum.musikding.de/attachements/Boss_SD-1_mod_c.gif) to my pedal and, as expected, it works perfectly!
As I study Medicine, I would say it's not a "causal therapy" but only a "symptomatic therapy", but at least now I can finally use my SD-1!
Thanks, analogguru!

Curt,
as you asked, here is a guide to perform the mod.

You need an SK30A jfet, a 0.1u (=100n) cap and a 1n4148 diode.
Obviously, you have to solder your new components to the top side of the board, because you won't be able to close your pedal anymore if you work on the traces side...!
Watching the front (the flat side) of the jfet, the three legs pointing downwards are named (from left to right) S, G and D (Source, Gate and Drain, respectively).
Yuo have to solder S to D4's cathode (the side with the black ring), G to your new 1n4148's anode (the side without the black ring) and D to one side of the cap.
The other side of the cap must be soldered to the point where R5 and R6 "look each other" (they are connected together at that side, so it doesn't matter if you solder to R5's or R6's leg, choose one!).
Solder the cathode (again, the side with the black ring) of your new 1n4148 to the cathode of D8.
Done!

Giulio

[syzygy]

Good work.. although I'm interested in seeing R.G.'s analysis of your analysis after he sees the schematic.  He seems to know his stuff.

Am I right that this is still not a True Bypass circuit?  With the changes that you've made, is there still the danger of having some signal loss when the circuit is in bypass mode?

[cd]

Am I right that this is still not a True Bypass circuit?  With the changes that you've made, is there still the danger of having some signal loss when the circuit is in bypass mode?

Yup, it's a MUTE circuit.  The output from the clipping circuit is cut off before it reaches the tone control.

The slickest (and best, IMHO) way to get true bypass in a BOSS case is with a latching relay.  I've heard of folks mounting DPDTs and 3PDTs in the battery case but that's ugly IMHO.

[MartyMart]

SHREDGD,
That sounds like a great fix from analogguru  
Well, no its not "true bypass" but if you read Roger Mayers explination in
the Dave Hunter book, then you wont place too much significance on TB
anymore !!!
Yards of cable, un-buffered between gtr and amp .........  :shock:
Or just use some 70's "curly" leads, that will remove all your top end 4 ya !!

Marty.

[syzygy]

OK, I'll bite.  Are you saying there is something wrong with True Bypass?  By using stomp pedals you're increasing the distance between instrument and amp anyway, so the way I see it, True Bypass will at least prevent you from the signal loss into effect boards of those stompboxes, assuming your wires are shielded properly, and all connections are in top shape.

[cd]

Before anyone goes fruit bonkers crazy with an argument for/against true bypass, let's all take a deep breath and use the search function above

[MartyMart]

OK, well I'm no where near as "experienced" ( excuse the pun ! ) as Roger Mayer, so I'll quote roughly from his comments :
Manufacturers ( boutique ) seem to place importance on "true bypass" as a feature.
The best signal path from guitar to amp is a short piece of wire, usualy not possible/impractible, so you go through an fx pedal, bypassed though a piece of wire and through a 20 foot cable, in most cases you will be better off using a buffer, to eliminate noise maintain the signal etc.
Amps like to be driven from a low source impedance.
Amps seeing a long cable attached to a pickup can start oscillating.
So "buffered" outputs can be very useful here.
It depends on how many pedals are in the chain and where the buffer is in that chain.
He talks of his "vision octavia" where he has TB outputs and buffered outputs, on short cables there's no difference but with longer "real world"
cable runs, thats when buffering becomes very important.
He reckons it was a 3DPDT switch marketting con, which stems from the old days of crappy pedals/components.
Modern pedals should not have that problem and hence NOT require TB.
I've always been told that a few "Boss" buffered pedals in a chain can
help out with this problem.
So, a whole line of "TB" pedals ...... perhaps not a great idea !!!
as that can add up to a lot of "tone sucking" cable....
Hope that makes sense ?

Marty.

[cbugatti]

it works....

I'm amazed...I really am...I modded my pedal months ago and found an unreal fuzzy/overdriven/mild distorted sound...used a mpf 102 transistor in d6, 2 1n34a diodes in series in d5 and a led and 1n34a in parallel in d4...along with changes in alot of other places...

now I can finally use the pedals drive knob...

thank-you everyone: analagguru, shredgr, rg, cd and everyone else who contributed to this....the problem has finally been solved...

I also tested to see if there's a loss of high's b/c of the lack of TB and I didn't notice any...

so stoked...thanks...

curt

[MartyMart]

Before anyone goes fruit bonkers crazy with an argument for/against true bypass, let's all take a deep breath and use the search function above

CD, you just beat me to it  
no argument, just interesting to get "RM's" take on it, as I said he's much more qualified to have a valid opinion on this than me.
I know it depends on a lot of things, but the more pedals/cable there is
the more I can see how buffering rather than "TB" comes into its own !!

Marty.

[cbugatti]

oh yeah...I also used a J201 instead of a SK30a....seems to work flawlessly...thanks again

[analogguru]

As I study Medicine, I would say it's not a "causal therapy" but only a "symptomatic therapy", but at least now I can finally use my SD-1!
Thanks, analogguru!

You are absolutely right...  

The cause may be capacitve crosstalk between pcb tracks, this would require a complete new pcb design....like if you make an organ-transplantation e.g. a new eye...for what, when its possible to wear glasses or contact lenses to achieve the same result... :roll:

analogguru

[vanhansen]

Can a MPF102 be used instead for the FET?  I don't have a J201 or SK30a.  Mine is having the same issue but since I have started using the ZW-44 it's no hurry on my end.

[syzygy]

MartyMart,

Thanks for the extra info; If that is the case at those long cord lengths, it would be much better to build one True Bypass box around your whole stompbox chain to reduce the amount of wire the signal has to go through when bypassed.

Luckily for me I don't have that many pedals to lengthen my signal, so I'll be building True Bypass in every one

[cbugatti]

Can a MPF102 be used instead for the FET?  I don't have a J201 or SK30a.  Mine is having the same issue but since I have started using the ZW-44 it's no hurry on my end.

hmmm...I was going to try the same thing as I only had MPF102's lying around but since I had to go get the 1n4148's I picked up something else...

the gate-source voltage max is higher at -7.5V than the J201 or SK30a...

I'd try it see what you get and if not just pick up a J201 or something similiar..

[Fret Wire]

thank-you everyone: analagguru, shredgr, rg, cd and everyone else who contributed to this....the problem has finally been solved...

Maybe not. There seems to be two different problems with the same symptom, bypass leakage. When you read the whole thread over again, you'll see that myself and MartyMart have fixed this symptom a couple of times by lowering/removing the 22k resistors. So maybe we have two possible problems with the SD-1 that can both show the same sympton, bypass leakage. :?:

[cbugatti]

When you read the whole thread over again, you'll see that myself and MartyMart have fixed this symptom a couple of times by lowering/removing the 22k resistors. So maybe we have two possible problems with the SD-1 that can both show the same sympton, bypass leakage. :?:

yeah that's the first thing I tried...removing the 22k's...didn't help at all though...I still don't understand the whole theory as to why there is the leakage...

I'm finally satisfied though...I've tried fixing this for at least 4 months on and off...roland should just fix it in the first place...

[BJF]

Hi,

I found this thread through another message board

For the BOSS SD-1 it would be clearly noted the the designers did not think the gain of the circuit would warrant a J-FET switch as used in BOSS DS-1 e.g.
At maximum settings there could be bleed through and is partly through the design of the PCB and the pedal as a whole. Easily a J-FET switch as used on the BOSS DS-1 at the input of the OD section could be used, but it would require outboard parts.

Also the ON resistans and OFF resitans of a J-FET would neither be 0 Ohms nor 1M Ohms respectively but just as RG pointed out in the order of  maybe 100 Ohms ON and maybe 100M Ohms OFF.
It would be possible to use a MOS transistor to get higher OFF resistans.
Still as in any high gain circuit with a side chain of low gain such as amplification 1, would warrant silencing the high gain chain in some way.

This is the reason a Fuzz Face has the kind of bypass it has- grounding the input-if you don't input will hang as an antenna with the maximum gain of the fuzz circuit and you will hear a fuzz sound riding in bypass. Now the BOSS  SD-1 is nowhere near that high gain so the economics department would say use just the minimum parts needed.......

Right so, there can also be this problem that the incoming flank triggers the switching J-Fets if amplitude of incoming flank is higher than the Ugs OFF and this can be due to a leaky J-FET device, and substitution for one with higher Ugs OFF could solve the problem.

Now it is also possible to install a J-fet switch into the feedbacknetwork as in series with the resistor going to pseudo  ground and hooking up the gate of which to bypass output of the flip/flop............which would have no influence on sound; it is ofcourse also possible to short circuit the feedback loop by shorting the output to inverting input resistors baut that would leave the J-FET used exposed to the outputswing of the OP-Amp and it could 'falsely' trigger, depending on the device......and really just  circuit knowledge would tell this.

Don't sweat it use the connections as suggested by Analog Guru.

No I can not see it and never have, but I can guess what it does and have employed it before I knew there even was an Analog Guru and it will work as suggested by RG and evrybody will be happy as suggested by Analo Guru.

Please respect Analog Guru's property rights to his drawings- just in case that is not enforced

Gentlemen have fun
BJ

[reggyboy1970]

R.G.,
that's of course what I did, and it is the reason of the tests I did (which, again, you can read about above in this thread) after I excluded a jfet malfunctioning.
I understand you don't have the pedal and you probably didn't look at the schematic, so that's why you couldn't follow my previous posts.
Anyway, just have a look at this schematic   http://www.freeinfosociety.com/electronics/schematics/audio/bosssd1.html and consider:

- I pulled out the 10k resistor between the gain stage and the tone stage (to exclude the second half of the circuit and therefore exclude a leakage from what comes after the gain stage)
- I pulled out the leg of the 0.047u cap connected to Q2's drain and connected it to a cable to my amp

and I could still hear the leakage!!

- Then I connected the 0.047u cap of the feedback loop of the gain stage to ground instead of 4.5v (as I was afraid the distorted signal made its way into the clean path through it, via the 100k resistor above, in the image)

Still leaking!

So, today I did analogguru's mod (http://forum.musikding.de/attachements/Boss_SD-1_mod_c.gif) to my pedal and, as expected, it works perfectly!
As I study Medicine, I would say it's not a "causal therapy" but only a "symptomatic therapy", but at least now I can finally use my SD-1!
Thanks, analogguru!

Curt,
as you asked, here is a guide to perform the mod.

You need an SK30A jfet, a 0.1u (=100n) cap and a 1n4148 diode.
Obviously, you have to solder your new components to the top side of the board, because you won't be able to close your pedal anymore if you work on the traces side...!
Watching the front (the flat side) of the jfet, the three legs pointing downwards are named (from left to right) S, G and D (Source, Gate and Drain, respectively).
Yuo have to solder S to D4's cathode (the side with the black ring), G to your new 1n4148's anode (the side without the black ring) and D to one side of the cap.
The other side of the cap must be soldered to the point where R5 and R6 "look each other" (they are connected together at that side, so it doesn't matter if you solder to R5's or R6's leg, choose one!).
Solder the cathode (again, the side with the black ring) of your new 1n4148 to the cathode of D8.
Done!

Giulio

Wow... nice one... just did analogguru's mod to my SD-1... no bleed - all clean in bypass with drive up full!  thanks analogguru