About the common Boss SD-1 bypass problem

[shredgd]

Some time ago I asked for help with this common problem, consisting in a subtle but noticeable bleeding of the effected (i.e. distorted) sound through the bypassed sound (the noise level is maximum when the drive knob is at max).
I was told to check the Q1 jfet for leakage, because that transistor blocks the passage of the effected signal to the output buffer in bypass mode, so I replaced it with a new one, but the bleeding through was still there, so the problem wasn't Q1's fault (I also verified this by playing the pedal with no transistor in Q1). I later noticed that the level of the distorted signal bleeding through the clean sound doesn't change with the position of the level knob, so I could have just used this information to exclude the role of Q1...
Someone suggested to lower the two 22k resistors which are bitween the two jfets Q1 and Q2 in the schematic, but this didn't work either (I also can't understand the logic of trying this, to be honest).

Today I decided to test the circuit with the famous "audio probe" and I found out that the bleeding through must be the IC's fault (at least in my case), because the noise is already present immediately after the input buffer and even at the input jack (although reduced by the internal resistance of the input transistor in that direction), while there is none at the +4.5V and +9V points (which are the only two other possible ways for the noise to reach the uneffected audio path).
I don't know much about electronics, so this is only what I suppose, but I'm pretty sure I must replace the IC. I'll do that as soon as possible and I will tell you if I was right!

Giulio

[shredgd]

I socketed the IC and tried different (and compatible, of course) ones, but the leakage is still there. I now think that maybe every SD-1 has got this problem, just not everyone notices that.
Any other suggestion?

Giulio

[Boofhead]

I also can't understand the logic of trying this, to be honest

It's to do with how leakage gets attenuated around that part of the circuit -  it takes some visual circuit analysis to see the leakage mechanism.

bleeding through must be the IC's fault (at least in my case), because the noise is already present immediately after the input buffer and even at the input jack

You have done pretty well so far.  Because it's at the input jack the audio probe could be picking-up stray signals, so your result might be false.  Assuming for now that it is in fact real I would suggest two options - you should try both:

1) For the sake of experiment try changing the opamp to a TL072 (no others).  

2)  Try adding a 10k resistor in series with C2 (18n) at the input of the opamp.  You will have to lift the leg of C2 out - so be careful no to lift the tracks off (especially if you try to bend C2).  After you try 10k, if the fizz is still there try a higher value like 47k, or lower like 1k.  You could get away with 10k as a solution but higher values are really only useful to identify the mechanism.

[analogmike]

I socketed the IC and tried different (and compatible, of course) ones, but the leakage is still there. I now think that maybe every SD-1 has got this problem, just not everyone notices that.

Yes, they all have the problem. But most people don't notice it.

[shredgd]

Sorry for taking so long to reply: I didn't get the emails to notify me of your replies, I don't know why.
I tried a TL072 as well, but it is the same. I still didn't try the resistor before the IC, but I'll keep you informed. I just wonder why tubescreamers, which have a very similar circuit, don't have this kind of problem (I did the experiment using an audio-probe with my Son Of Screamer).

[MartyMart]

I had the same problem with an old Sd1 from about 1986, removed and jumpered the pair of 22k resistors and there was some improvement.
Then fitted two new 2SK30A's to the flip-flop circuit and it was fine, no distortion in the bypassed signal at all.
I got them from Olaf Nobis at Banzai Effects,  www.banzaieffects.com

Marty.

[nooneknows]

I have an SD-1 (well, a bit modified...) and I'm quite sure the clean mode is.. clean. I can't hear any distorsion and, imho, I have a trained hear.
Are you sure the first bjt works correctly? I changed the input stage with a jfet, it also rises the input impedance to 1M. it's a simple mod, just desolder the original tranny and insert a bf245 (or 2n5457, j201...) fet (keep an eye on the pin out, it's different, you have to bend to legs) then change the 510k R with a 1M or higher resistor. cheap and effective, and you can exclude another potential source of problem.
bye

[analogmike]

I have an SD-1 (well, a bit modified...) and I'm quite sure the clean mode is.. clean. I can't hear any distorsion and, imho, I have a trained hear.

I never heard it either but it's there in the background if you listen VERY carefully with a loud clean amp. I have not tried the above methods to fix it yet.

[wampcat1]

I have an SD-1 (well, a bit modified...) and I'm quite sure the clean mode is.. clean. I can't hear any distorsion and, imho, I have a trained hear.

I never heard it either but it's there in the background if you listen VERY carefully with a loud clean amp. I have not tried the above methods to fix it yet.

It's actually much more apparant when you turn the gain knob all the way up (with the pedal in bypass mode), and even more so when you use a guitar with hot pups. Every sd-1 has this problem, and boss is actually aware of it, but to be honest, its not enough of a problem to them to change anything, since most of the folks that buy them simply can't hear it.

[Brian Marshall]

my boss dd3 has a similar problem...... if i turn the gain up all the way on my amp, i can hear the delay signal faintly.

[shredgd]

Finally I found the time to work on this issue again...
I didn't try to put a resistor after C2 as Boofhead suggested because I wanted to solve the problem radically... but I didn't manage to do it!

I really can't see where the wet signal enters the bypass path. I tried everything:
using a probe I could hear it (after the input buffer) even after I desoldered both the jfets (I did this test because, as my TS clone doesn't have this problem and it is an almost identical circuit, I thought the flip-flop bypass mechanism should be involved).
So I thought the wet signal might make its way via Vr (the 4.5v points) through the 470k resistor that connects the base of the input transistor to Vr or the 100k that connects the input of the opamp to Vr, but the audio probe has stated Vr is "clean" from any signal.
If anyone ever finds out what I couldn't, please write it in this forum!!

P.S.: by the way, MERRY CHRISTMAS!!

[Fret Wire]

Did you try lowering the 22k resistors?

[12StringStratMaker]

Did you try lowering the 22k resistors?

So did anyone try this or did everyone get lazy over christmas like me?  

[wampcat1]

Did you try lowering the 22k resistors?

So did anyone try this or did everyone get lazy over christmas like me?  

I did it, didn't make a difference at all with my setup, unfortunately.

[Fret Wire]

I've had the resistor method work two out of three times. It's a first step, that's all. If the problem can be fixed at the lowest level, that's where you start. If not, it's jfet replacement time.

[MartyMart]

I had more of a problem with my own early SD-1 ( 1986 ) and had to change out the two "flip/flop" Jfets on that one. ( 2SK30A'a )
A newer one ( bought to try various mods on ) was rescued by just reducing the two 22k resistors to 1k's.
On all of the "Brand new" SD-1's that I've sold as "MOD" versions (15 !!) there has been a very small amount of "break through" and I've also reduced those resistors to cure that problem.
In those cases I went with 2k2's and it worked fine.

Marty.

[cbugatti]

so has anyone found a solution to this problem?

I've tried replacing the 22k resistors with a straigh thru connection and it didn't change anything...

this pedal is only usable at a really low gain setting now and as soon as you turn up the dirt it is really noticable when playing loud...

does anyone know an actual solution??

thanks
curt

[MartyMart]

The "solution" is in the above post ........  new 2SK30A's  i'm afraid !!

Marty.

[Joep]

I re-soldered all the joints of the switching fet's for a SD-1 from a friend of mine and that solved "a" switching problem. I can't remember if it was bleed-through, because it was some time ago.....

[GFR]

Search GEO (RG's site) for an article called "more fun with the millenium bypass"

http://www.geofex.com/Article_Folders/Mill2extn/mil2plus.htm

It explains how to shunt the input of the FX when it's bypassed, using the millenium. You can easily adapt this concept to the boss Jfet switching. You just need an extra transistor that can be driven by one of the FF outputs, and find an appropriate place in the circuit where the shunting can be done.

If you shunt the input to the fx, there will be no fx to bleed through and the bypassed tone will be clean.

The Marshall Blue Breaker (at least the original) also uses input muting when bypassing (well, not exactly the input - it shunts the signal between two stages).