Cliche / Notakliche - Are diodes really necessary?

S

sammo303

Member
Hi all. I recently finished a Notakliche build and as part of the final process, I auditioned a bunch of diodes using tacked in croc clips to permit fairly quick changes. I found a few elements of this to be slightly odd.

I began by placing germanium diodes per the schematic and trying various diodes in the position controlled by the switch. After a short while of confusion and looking at the schematic, it became obvious why I was hearing no difference between diodes given the germanium diodes have a lower forward voltage and essentially prevent other types of diodes from doing any of the work. WDJD makes sense now!

I then tried some Schottky diodes given their lower forward voltage but again, I didn't hear anything different when bringing them in with the switch. I decided to change the germanium diode location such that they would be brought in by the switch, allowing me to audition different diodes to go in their place.

I tried LEDs, normal silicon diodes, etc. and the differences were fairly subtle. That's when I decided to listen with no diodes at all and there really wasn't much of a reduction in distortion. Things were a bit wilder, but the character remained very similar. For now I have therefore landed on no diodes in the normal position and germanium diodes in the switch position.

Given all the talk about magic diodes, and them being in a hard clipping arrangement, I really wasn't expecting it to sound dirty without any diodes. Is this normal? I presume this is the sound of overloaded opamps?

With that in mind, I did one last check before quitting which was a bit of a voltage survey across the board. To my surprise, the voltages all seem to sag a fair bit. Maybe this is normal given the 9v supply runs through a 100 ohm resistor? Here's what I measured:

+9v rail is +7.5v
+18v rail is 13.3v
-9v rail is - 6.7v
Vref (ideally 4.5v) is 3.7v

Removing the opamps results in the voltages rising up much closer to the stated values on the schematic, less the diode drop. On this basis I think the 100 ohm resistor might be the culprit which makes me question its purpose.

Please note I got the optivisors out and checked the entire board for soldering issues, all seems to be in order. I also removed and reseated the opamps and charge pump IC. I also built it using parts from Mouser so fake ICs isn't a concern.

Sorry for the long read but if you have any thoughts on whether my pedal is behaving oddly or if this is totally normal, I'd love to hear it.

Sorry also if this should be under troubleshooting. I suspect my pedal is behaving normally and I'm just not quite understanding how it works, hence why I posted this in General.
 
My experience with this come mostly from that I modded a EHX Soul food with an On/Off/On switch, meaning I had the option of no clipping, 1N60(?) (stock) and russian D9E germanium diodes that are close to the originals in forward voltages.

Ofcourse it distorts at higher gain with diode clipping off, and the difference between the silicon originals and the D9E isn't huge.

IDK about those voltages, but they sound a bit low? To me it sounds like you should check continuity and shorts for all the components around that charge pump.
 
I have my Aion klon setup with ge (1N34A) and Bat46 on a toggle switch. The middle position is no diode, but I also have a little 1/8” jack connects to the clipping point to test other diodes with alligator.

I find the bat46 and Ge to be close, though I think the Ge is a bit warmer and the bat46 a bit harsher. But it’s pretty close, could be biased listening.

Silicon, LED, and no clipping are all progressively very noticeably louder and less clipped, in that order. Especially with the gain knob past noon. Notakliche results should be the same
 
I’m not sure about the voltage drop. Sometimes it depends on the multimeter impedance on the voltage setting. I notice a drop on the 4.5v in+ pins often (pins 3&5), due to the 1M or 470k resistor from Vref to the pin, creating a voltage divider with my cheap meter’s input impedance. But I don’t think I’ve noticed it that drasiically on the pin8 and pin4
 
Ah! That 100ohm resistor is dropping your voltage! Most Klon pcbs don't have that. Not sure why Robert added it, I think he re-worked the power section to protect for both reverse polarity and overvoltage (like if the user plugs in 18V, normally it would fry a klon, but in this design the Zener protects the rest of the circuit, and the 100ohm resistor protects the Zener.

Let’s say the whole circuit draws 15mA ( pretty close, might actually be a bit more).

V = i*R can be used to calculate the voltage difference across that 100ohm
V = 0.015A * 100 ohms
V = 1.5V (which explains the drop from 9 to 7.5). All the other voltages are in the ballpark of what you expect with 7.5V input.
 
Last edited:
OK, I've jumpered the resistor and voltages are looking healthier.

+9v rail is now +8.5v
+18v rail is now +15.5v
-9v rail is now -8.2v
Vref is now +4.3v

I think these look OK now. 18v rail seems a bit low but it passes through at least 3 diodes so it's probably fine, and should provide more than enough headroom for the opamp. I'll do a listening test next.

Just thinking out loud. Given this is based on a kit made for non technical people, maybe the resistor in the power supply was seen as some cheap insurance to avoid polarity related problems and returns headaches.
 
Playing test seemed to show a bigger difference between diodes in / out of the circuit now. Hard to tell it that's because I was listening with fresh ears but it seems to be doing things as it should now.

One thing I'm still not 100% on is the change in resistor when flipping the WDJD switch. I guess it influences the amount of clean blend? In the original JHS circuit it might have been added to offset perceived additional signal loss due to the use of another pair of clipping diodes? Given I have this arranged with no diodes / switched germanium diodes, I'm not sure the clean blend is quite right.

I think I'll experiment with resistor values here. I'm starting to see how this hobby sucks you in. 😅
 
Here’s from the NOTAKLÖN page. There is a slight boost and mid bump, but I take issue with the description of the additional hard clipping. The LEDs shouldn’t clip or change the sound at all since the Ges (or Schottky) clip at a way lower threshold.

Besides Robert directly tracing the notaklon, there’s been other discussions / pictures on here about the shamrock mod on KTRs showing the same led on the switch, so I think it’s accurate, just a not well thought out mod. The Resistor is doing all the changing.
 

Attachments

  • IMG_0854.jpeg
    IMG_0854.jpeg
    91.9 KB · Views: 7
Well I've now done a playing test, adjusting the switched resistors R15 and R16 using trim pots. After measuring where I landed with them, I reset them and then adjusted them using a scope to try and achieve volume parity across the switch.

Results:

- Stock values: 47k & 27k
- Values determined by ear: 47k & 10k
- Values determined by scope: 44k & 14k

I think I'll go with 47k and 12k.
 
One thing I did notice with the scope was that if I had the level very high, with a hot input signal, it was possible to induce oscillation in the circuit. I'm not at all sure if that is normal, or even possible under normal circumstances, but it did make me wonder if there is something slightly unstable in the circuit.
 
Nice! Probably more effort and study than the original shamrock mod.

I thought part of the point was more volume boost with the lower 27k resistor. But you got equal volume going even lower. So I guess there’s a peak volume as you adjust R16?-
 
Yes, as you decrease R16, the switched voice gets louder. I don't have diodes in the circuit until the switch comes in, so I suspect my values do need to be a bit different than the original mod. I now see an increase in distortion and a slight volume rise when I flip the switch, which gives two quite different and usable options. Prior to adjustment I saw an increase in distortion but a drop in volume.
 
You must log in or register to reply here.
Back
Top