Neovibe debug

[yodude]

Hi folks. I'd appreciate some help debugging a Neovibe. The vibing sounds great, but I'm getting a harsh distortion , which abruptly disappears as the note fades out. Interestingly, I built two simultaneously and they both have the same issue. I get the distortion even with single coils unless they are played very quietly. I know that gated distortion is a known issue for the Neovibe when presented with hot input signals, but these two seem overly sensitive.

So, on to the business at hand.



Version:

I used GGG's excellent PCB, which is marked "Ver. J 10/07" and populated it according to the latest GEO parts list.

Mods/Subs/Choices:

I used 2N5088 except for the lamp driver (Q13) is a 2N3904.
The 78L15 is marked L78L15. I googled and am pretty sure I got the pinout right.
GEO true bypass volume drop mod: R8 is 900R (GEO calls for 910R), R9 is 3.6k.
GEO phasing mix mod: R35 and R36 were lowered to 75k and are fed to legs 1 and 3 of a 50k trimmer. Leg 2 goes to the board.
Power supply: 16VAC 1000mA from Jameco http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&pa=100108&productId=100108
For completeness, the 50k pots are log taper (supposed to be linear tapers).

Tests and results:

I get the distortion with single coil and humbucker EMGs and with tapped and untapped PAF-alikes.
No distortion when true bypass is engaged.
I get the using either of the two power supplies that I bought
The only way to avoid the distortion is to play extremely quietly.
Because I built two and they're both doing the same thing, I don't think a soldering issue is probable.  I did look for bridges and didn't see any.

Voltages:

Voltages measured as follows:
No instrument cables in the jacks.
C, B, E -- black lead on out jack ground.
Vbe -- black on emitter, red on base.

Q1
C   1.98
B   0.82
E   0.90
Vbe   0.48

Q2
C   1.73
B   1.97
E   1.41
Vbe   0.56

Q3
C   10.26
B   4.98
E   4.46
Vbe   0.60

Q4
C   14.93
B   4.23
E   4.31
Vbe   0.42

Q5
C   11.14
B   4.31
E   3.74
Vbe   0.60

Q6
C   14.93
B   4.22
E   4.29
Vbe   0.42

Q7
C   11.17
B   4.29
E   3.72
Vbe   0.60

Q8
C   14.93
B   4.26
E   4.34
Vbe   0.43

Q9
C   11.13
B   4.33
E   3.76
Vbe   0.60

Q10
C   14.93
B   5.32
E   5.41
Vbe   0.57

Thanks for any and all help.

[R.G.]

carefully check polarity on C1.

[yodude]

carefully check polarity on C1.

Thanks for the quick response, RG, and thanks for the project.  Looking forward to ironing out my kinks.

C1 polarity looks good to me (negative side is in the circular pad which connects to R1, R2, and R3).  I believe all the electros are supposed to have negative facing the "top" of the board (as defined by the lettering) and that's the way they're installed.

[yodude]

More info

Here's a clip of my issue:
http://www.aronnelson.com/gallery/main.php/v/yodude/NeovibeDebug.mp3.html
Four bits in there:
- PAF
- PAF with light shield removed in a lit room (to make it easier to hear the distortion)
- coil tapped PAF
- coil tapped PAF with light shield removed

The debugging section of the project pdf says that the collector should measure several volts higher than the base and emitter. Q2 didn't fit that description. Unfortunately, it was just a measurement glitch so no answer there. Re-measured and got:
Q2
C 4.97
B   1.97
E   1.41
Vbe   0.56

None of the other measurements jump out at me as being too far off the mark (B and E sometimes don't look great, but Vbe seems ok).  Also double checked all the voltages and found no other changes.

Regarding Neovibe distortion, RG says:

The input to the Univibe can distort under moderately high signal levels. As to whether this is normal or not - read the DC voltage on all the devices. Where one of them as too little headroom, it will distort. If they're all about correct, then it's "normal" for the circuit.

How do I read the DC voltage? What does "all the devices" mean -- caps, resistors? Have I already done this step?
How do I know how much headroom a device has?
How do I know if a voltage is correct?
Sorry for all the dumb questions.

Meanwhile, DeeBug found that replacing his polystyrene caps in C4 and C12 with ceramics solved the problem. Anyone have any thoughts on this?

Not sure that I wired up the speed pot correctly. I have all three of the non-wired tabs jumpered to the tab that's wired to "C".

Also not sure if I wired the power correctly. The jack is a Jameco 151590. Two of the tabs show continuity, so I wired one of those to "K" and the non-continuous to "J". It didn't work at all, so I switched the "K" wire to the other continuous tab and it fired up.

Any thoughts?

[yodude]

More info

Tried powering it with two 9v batteries in series supplying 19.12V. No cure.

Measured voltages on C1 using same procedure as above. Positive side was 0.8V, negative side was 0V.

Also measured 0V at input (A1) and outputs (M & N).

Any and all input welcome...

[novotm]

I used ceramics for C4 & C12 and I'm having the same problem.  Otherwise, my build is almost identical to yours.

I don't know if this will be of any help, but here are my voltages:

Q1
C 1.94
B 1.31
E 0.89

Q2
5.00
2.00
1.42

Q3
10.25
5.00
4.40

Q4
14.91
4.73
4.38

Q5
14.91
4.38
4.13

Q6
14.91
4.68
4.29

Q7
11.22
4.3
3.7

Q8
14.91
4.68
4.29

Q9
11.21
4.29
3.70

Q10
14.91
5.96
5.45

Q11
22.12
9.3-11.7
10.9-13.7

Q12
21.46
10.9-13.7
10.3-13.0

Q13
17.2
3.6
2.92

[yodude]

Audioprobed according to:

I think that from here on you'll need to use an audio probe to figure out exactly which stage is doing the breakover. Probe the collector and emitter of Q3, 5, 7, and 9, and the emitter of Q10, listening for the first stage where you hear the breakover. You will have to use a capacitor-isolated probe.

location / breakover?
Q3E yes
Q3C no
Q5E yes
Q5C no
Q7E yes
Q7C no
Q9E yes
Q9C no
Q10E yes (most breakover here)

I've triple checked caps for value and orientation.  Any ideas?

[yodude]

 

Had Es and Cs backwards above.  Should read:

Q3C yes
Q3E no
Q5C yes
Q5E no
Q7C yes
Q7E no
Q9C yes
Q9E no
Q10E yes (most breakover here)

[R.G.]

Two things come to mind. You might be having some problem with oscillation. That would be odd, but it would account for the way it seems to break up with higher signals.

The other is more probable. Clip your meter leads onto ground and the output of the 15V regulator. Watching the meter, play softly and then more loudly. Does the 15V output drop when it's doing distortion?

[yodude]

Clip your meter leads onto ground and the output of the 15V regulator. Watching the meter, play softly and then more loudly. Does the 15V output drop when it's doing distortion?

I did this measurement with a looper playing a low A over and over, and the pedal output going to a little practice amp.  With the looper volume turned way down and the amp turned way up there was no audible breakup.  With the looper turned up, the breakup happens just like the recording posted upthread.  I get 14.95V at the output of the voltage regulator regardless of the volume of the looper.

I also measured I and G and got I=24.1V and G=0V for both cases.  Is 24.1V weird for a 16VAC 1000mA wall wart?

As stated earlier, I have L78L15 instead of 78L15.  I'm pretty sure they're compatible and that I got the pinout right (the flat side lines up with the drawing on the GGG board).

I appreciate the help.  Oscillation ... ?

[R.G.]

Could you verify again your audio probe results as listed below - that is, you get breakup/distortion on the collectors of Q3, 5, 7, and 9, but not on the emitters?

Q3C yes
Q3E no
Q5C yes
Q5E no
Q7C yes
Q7E no
Q9C yes
Q9E no
Q10E yes (most breakover here)

[yodude]

Could you verify again your audio probe results ...

Good call.  Probed again with headphones instead of crummy practice amp.  Results have some subjectivity, so I made a recording.  Used the looper as above.  Warning: volume is all over the place.
http://www.aronnelson.com/gallery/main.php/v/yodude/NeovibeDebug_01.mp3.html

In order:
1. Bypassed signal at output -- A little fret buzz at the beginning.
2. Processed signal at output -- Same old breakover.  (Also note volume drop.)
3. Q3C -- Faint breakover near end of note.
4. Q3E -- Clean (?)
5. Q5C -- Faint breakover near end.
6. Q5E -- Obvious breakover near beginning of note.
7. Q7C -- Faint breakover near end.
8. Q7E -- Faint breakover near end.
9. Q9C -- Obvious breakover near end.
10. Q9E -- Obvious breakover over middle.
11. Q10E -- Obvious breakover over middle and end.

Note on C/E: I think I got it, but just in case ... If the big caps define the "bottom" of the board and the Q1-Q10 line defines the "top", then collectors are on the top side and emitters are on the bottom side.

[R.G.]

See my pm (for obvious reasons)

[yodude]

ok ... so

The problem being discussed above was so weird that RG asked me to send him the unit. He did a bunch of stuff which culminated in replacing Q1, Q2, and Q3 (all 2N5088). I have a twin board that I built at the same time that had the same problem. The only thing I did to that one was to replace Q1, Q2, and Q3 and that solved the problem on its own. The original transistors I had were from Smallbear, the new transistors came from RG's basement.

Here's RG's post on his debug:
http://www.diystompboxes.com/smfforum/index.php?topic=70757.0

And here's a clip of a happy unit:
http://www.aronnelson.com/gallery/main.php/v/yodude/NeovibeDebug_02.mp3.html

Thanks a million to RG and all you other debuggers from another mother.

[R.G.]

Hmmm. It really was the transistors, two cases in a row. I think I'll write Steve Daniels and let him know to be suspicious about some of his stock.

Congrats on the repair work, and have fun with the newly-working pedal.

[pold]

"The original transistors I had were from Smallbear, the new transistors came from RG's basement"

2 in a row! That sounds really bad. I don't think it's just coincidence. Anybody else had faulty transistors from Small Bear previously? Do you think it's better to order Q1 Q2 and Q3 from somewhere else then ?

[R.G.]

"The original transistors I had were from Smallbear, the new transistors came from RG's basement"

2 in a row! That sounds really bad. I don't think it's just coincidence. Anybody else had faulty transistors from Small Bear previously? Do you think it's better to order Q1 Q2 and Q3 from somewhere else then ?

Dear deity.

Please don't start any more internet rumors like "all the transistors from Small Bear are bad".

There was a SINGLE sample of ONE kind of transistors from Small Bear. ALL six of the "failing" transistors in the two Neovibes were from one order from Small Bear. It is not clear where they came from or what brand they were. Simply buying somewhere else is NOT a guarantee you won't have the same problem because the other place may have bought their stock from the same place as Small Bear. Or you could get even worse, given the counterfeiting situation in the world today.

Here's another tidbit for you to think about. "Anybody else had faulty transistors from Small Bear previously?" is not even an appropriate question. I can promise you that 100% of every source of transistors has sold faulty ones. It's impossible not to, even for the manufacturers. What you do care about is whether the seller stands behind the product for the few times when they are faulty.

[Filament]

I've jumped on the SmallBear gripe train a time or two but in all fairness, I just completed my Neovibe populated with transistors I got from SmallBear and it fired up the first time.  Seems like an unfortunate coincidence and probably not a widespread problem.   

[yodude]

On the Smallbear issue, I just wanted to pass along the info for the sake of completeness. I've got nothing but good things to say about Steve and the rest of the Smallbear gang.

Now that my Neovibe is sounding good, I want it to sound great. I seem to be missing some complexity in the vocalization.  The clips provided by John Lyons (thanks John!) have a ye-wow ye-wow ye-wow sound, where mine just goes wow-wow-wow.

John's clips: http://www.diystompboxes.com/smfforum/index.php?topic=54665.0
Mine: http://www.aronnelson.com/gallery/main.php/v/yodude/NeovibeDebug_02.mp3.html

John's recording is of SteveB's Neovibe.  Downthread, SteveB describes some mods.  Is this where the difference lies?

I have a 10k resistor in place of the 4k7 that is connected to the base of the bulb driver. 47k was too subtle, & 4k7 was too choppy for me.

I think he's talking about R49. I tried jumpering it and putting a 47k in parallel with it with no noticeable effect.

[kurtlives]

There is a very little-known secret to getting good, deep phasing out of phasers in general and the 'vibe in particular.
The depth of phasing depends on getting a good mix at the output. Resistors R35 and R36 mix the dry and phasedelayed
signal together. The stock values may not be the right one for a good mix. If you want the best performance
out of your Neovibe, leave these two resistors out and use a 200K pot to adjust the mix for deepest phasing. Then
measure the resistance on each side of the wiper and put the nearest 1% resistor values in for R35 and 36.

Maybe try that?