Ok its an amp question… jc 120 noise

[Locrian99]

Audio probe shows it at one side of r59 and one side if r38. 

So i am going to replace r38 i guess.

[Rob Strand]

470n cap betweeen q5 drain and ground channel 2 is he same.  However its not passing guitar signal which im
assuming is expected.  didnt notice any difference of note in the noise.  With the 4k7 across r94.  I also tried with a decibel meter app and they were the sameish.  The noise while steady seems to have some variance. 

OK two possible conclusions:
- The added 470nF cap is shorting noise from Q5.  The Q5 area of the circuit is
  the source of noise.
- The added 470nF cap is shorting noise from getting into the gate of Q16.

The problem with 4k7 across r94 not doing anything is it may or may not muddy up the conclusions.  However if we accept that test as being valid it opens up another possible conclusion:
- The DC bias on the gate of Q16 is noisy.

You could do a secondary test by placing the 470nF (or something larger) across R100 (220k).  If that fixes the noise then there has to be something wrong with R102, or R100 or the 12.5V supply is noisy.  You could even try tack soldering a cap across the 12.5V supply, maybe 100uF to 470uF.

As an alternative test to changing R94, you could lift R94 or C53 to see if the noise *drops*.  What that does is leave the bias along and drop the gain of the stage. It's a tiny bit risky as the stage could oscillate.  If there is no change in the noise then I'd be thinking the DC on the gate of Q16 is fluctuating.

[Rob Strand]

Audio probe shows it at one side of r59 and one side if r38. 

So i am going to replace r38 i guess.

I'm still looking for those parts.

R38 is expected to behave like that.  No problem there.  The problem is close to the input.

Can't see R59.

[Locrian99]

Audio probe shows it at one side of r59 and one side if r38. 

So i am going to replace r38 i guess.

I'm still looking for those parts.

R38 is expected to behave like that.  No problem there.  The problem is close to the input.

Can't see R59.

R38 comes from
Channel 1 and meets r59 right before c26.   Right before the chorus comes back. 

The point where those two meet is the start of noise.   Replacing 38 and r59 did nothing.   But the noise is present at both sides of r49. 

[Locrian99]

Could the problem be in the pcb trace between those two components?

[Rob Strand]

But the noise is present at both sides of r49. 

Can't find that either.  R59 looks like R55 to me.

If it's outside of the region I posted I suspect you are in the wrong area.

[Rob Strand]

Could the problem be in the pcb trace between those two components?

Anything is possible.

You need to isolate the problem first.

[Locrian99]

But the noise is present at both sides of r49. 

Can't find that either.  R59 looks like R55 to me.

If it's outside of the region I posted I suspect you are in the wrong area.

The 5 is a 9




The metal fil on the top right is r38, the metal film is the middle there is r59.   I have continuity (well i think the best it read was .1 ohms) between the top of r59 and r38.   I also have the noise at both of those points.   Was hard to probe the film cap so i jumoed to the next resistor and noise is present at both sides right before q6.   

As you can see i have replaced both 38 and 59.   Which made me think trace maybe, not sure.

[Locrian99]

I guess lift one side of r38 and run it via jumper to r59.   If it goes away its the trace.

[Rob Strand]

If a cap to ground at R80 (12k) removes the noise it means the noise originates earlier on in the circuit.   The region of the circuit around R80 cannot physically affect the circuit around R38.
The noise is already present at Q15's collector and is just passing the noise on to R38.  Also noise at R38 doesn't quite add-up with the symptom that chorus removes the noise.

If the noise just happened to disappear when you did the R80 test then that could give a false result.

[Locrian99]

Ok lifting r38 did nothing.   But...




R59 there the 100k that meets up
With r38 and c26 is new and c26 is new i lifted r38 and the noise presists.   But that is the first place noise shows up in the signal path.

Any ideas?   I feel pretty confident that is where its showing up.    Could i lift r59 leg that connects to c26 and lift c26 and jumper to see if that changes anything ?

So in this image the two red marks there is no noise.    The two black marks have noise, and have continuity they are where r38, r59, and c26 meet. 

[Locrian99]

Oops forgot to ass image

[Rob Strand]

Any ideas?   I feel pretty confident that is where its showing up.    Could i lift r59 leg that connects to c26 and lift c26 and jumper to see if that changes anything ?

What about through the path C42, R70, Q10, C81.  All just below the area you have been looking at.  The idea being noise follows the path C81 --> C26 etc.  The noise source could be around those components.

Also, try switching the chorus in and out,  Does the noise change at any of those points?

[Locrian99]

Any ideas?   I feel pretty confident that is where its showing up.    Could i lift r59 leg that connects to c26 and lift c26 and jumper to see if that changes anything ?

What about through the path C42, R70, Q10, C81.  All just below the area you have been looking at.  The idea being noise follows the path C81 --> C26 etc.  The noise source could be around those components.

Also, try switching the chorus in and out,  Does the noise change at any of those points?

Funny i was just going throuh there.  Its there at c81, the not banded side of d6, source and gate of q10, one side of r56, and the side of r70 that connects to the source of q10

[Rob Strand]

Funny i was just going throuh there.  Its there at c81, the not banded side of d6, source and gate of q10, one side of r56, and the side of r70 that connects to the source of q10

OK, that's progress.

The question now: is the noise coming from switch JFET Q10, or buffer JFET Q6, or other.

My bet is the Q10 area since the noise disappears in Chorus mod.  You could lift Q10. See if the noise is still there.  Then perhaps replace Q10 with a short across S and D to make sure it's still gone - with D & S shorted the amp will appear to work but channel 2 will operate as a pedal chorus instead of a vibrato mixing with channel 1.   The aim isn't to fix it,  the aim is to see of Q10 is the cause.

Oh forgot, try chorus on and off.

[Locrian99]

Funny i was just going throuh there.  Its there at c81, the not banded side of d6, source and gate of q10, one side of r56, and the side of r70 that connects to the source of q10

OK, that's progress.

The question now: is the noise coming from switch JFET Q10, or buffer JFET Q6, or other.

My bet is the Q10 area since the noise disappears in Chorus mod.  You could lift Q10. See if the noise is still there.  Then perhaps replace Q10 with a short across S and D to make sure it's still gone - with D & S shorted the amp will appear to work but channel 2 will operate as a pedal chorus instead of a vibrato mixing with channel 1.   The aim isn't to fix it,  the aim is to see of Q10 is the cause.

Oh forgot, try chorus on and off.

With q10 lifted or jumpered there is no noise.   In any of the chorus switch positions.  Does that mean the 2sk246 is bad?   And then youre going to say in that spot i can put pretty much any thing to sub! 

[Rob Strand]

With q10 lifted or jumpered there is no noise.   In any of the chorus switch positions.  Does that mean the 2sk246 is bad?   And then youre going to say in that spot i can put pretty much any thing to sub!

Way to go!

Sure seems like Q10 (2SK246-GR) is bad.   You could do some DC measurements but since it's noisy that might not show up the main problem with it.

Well there's plenty of scope for subs but I'm always a little wary of JFET switch subs.  They work but you can get little delays where the Q9 and Q10 switching times don't quite match up.  It can be very subtle.   If you think about it: Why did Roland bother using different JFETs for Q9 and Q10?!?!   I'd be looking for a JFET with a VP of around 2V to 2.5V, a little lower like 1.8V might work.   On the schematic the text for Q9 looks like Q5 but there is clearly a Q5 as well.
(FYI, there's a couple of threads with obscure problems related to JFETs used for JFET switching in Boss pedals.)

[Locrian99]

With q10 lifted or jumpered there is no noise.   In any of the chorus switch positions.  Does that mean the 2sk246 is bad?   And then youre going to say in that spot i can put pretty much any thing to sub!

Way to go!

Sure seems like Q10 (2SK246-GR) is bad.   You could do some DC measurements but since it's noisy that might not show up the main problem with it.

Well there's plenty of scope for subs but I'm always a little wary of JFET switch subs.  They work but you can get little delays where the Q9 and Q10 switching times don't quite match up.  It can be very subtle.   If you think about it: Why did Roland bother using different JFETs for Q9 and Q10?!?!   I'd be looking for a JFET with a VP of around 2V to 2.5V, a little lower like 1.8V might work.   On the schematic the text for Q9 looks like Q5 but there is clearly a Q5 as well.
(FYI, there's a couple of threads with obscure problems related to JFETs used for JFET switching in Boss pedals.)

Sounds like its time to breadboard a jfet tester!  Awesome again thank you, this was a bit more comolicated than i anticipated for sure.   

Edit so VP is all i really need to worry about?

[Rob Strand]

Sounds like its time to breadboard a jfet tester!  Awesome again thank you, this was a bit more comolicated than i anticipated for sure.   

Edit so VP is all i really need to worry about?

If you get VP close you probably don't need to worry about the other parameters.  The 2SK246 is closer to a lot of the common audio JFETs.  You might even get away with a J112, although the on resistance is a bit lower than generic JFETs.

Problems which are intermittent or change with warm-up are always a little more difficult.

[Rob Strand]

There is something unique about the JFET circuit.  It's operating at 30V.  So you might need higher voltage ratings on the JFET.  The J112 is only 35V which is close to the limit.  The 2SK246 is 50V.