CD4049UBE "UBE Screamer" issue...

[gregwbush]

Hello all,

I breadboarded the ROG Ube Screamer last night

http://runoffgroove.com/ubescreamer.html

The behavior of the circuit seems spot on... there's a definite roll off of bass around 700hz, the tone control works, clipping is nice (actual diode clipping coz when diodes are removed it's practically clean), all seems well... Until i turn the DRIVE knob above about 1/4.

Above that, it farts and splutters. I dunno what's up with that. Below the "farting threshold" it is working perfectly in every aspect.

Could i have a dodgy cd4049ube IC?
Could it be a breadboard "no shielding" thing?
This is my first logic IC build; anyone have experience with circuits utilizing 4049's and experienced "farting"?
(I guess it's likely i've got something connected wrong or missing, dare i say it, but i've triple checked! And have a lot of schematic to vero / PCB / breadboard experience)

Any thoughts why it might be acting this way!!??

cheers,
(i feel like a noob atm) gregwbush

[Mark Hammer]

Try making the 330 ohm resistor you see in the upper left hand corner of the diagram a smaller value, like 220R or even 100R, and see if that does anything useful in terms of extending the "range" of the pot.  I suggest this because the issue may rest on how much current the chip is getting.  Don't go lower than 100, for now.

[Derringer]

I've had breadboard problems like this with my pots sometimes. After getting twisted this way and that, sometimes either the solder joint comes loose or the wire just starts breaking.

Happy to see a current hex inverter thread. I just started playing with one because I never have before.

OT: Are they always a bit hissy?

[DavenPaget]

I've had breadboard problems like this with my pots sometimes. After getting twisted this way and that, sometimes either the solder joint comes loose or the wire just starts breaking.

Happy to see a current hex inverter thread. I just started playing with one because I never have before.

OT: Are they always a bit hissy?

OT : Isn't it easy to suppress them ?

[Ripdivot]

I would check that the pot is good.

[gregwbush]

Tried some different resistors as suggested, then a 10k pot, then no resistor at all:
  Still having the same trouble. Different values here seem to change where on the pot a given amount of gain can be found. The "farting threshold" seems to remain the same gain wise (just becomes apparent at different pot positions depending on resistor value)

Tried a different pot. Same $h!t, different pot.

Tried grounding the pot casings: A little less interference noise but same $h!t

So i'm still stumped!

One thing i noticed on the schematic is rather than stating +9V, it just says +V. Got me thinking maybe it's meant to be run on a different voltage. Then i measured the voltage at pin 1 on the IC and it's similar to ROG's measurement so i don't think it's that

[earthtonesaudio]

Make really sure it's an unbuffered 4049.  In my experience the buffered version can sometimes behave as you describe.

4049UBE - unbuffered
4049BE - buffered

The other letters sometimes vary with the manufacturer but the unbuffered version almost always contains a U.

[PRR]

> breadboarded the ROG Ube Screamer

That has a LOT of gain. Over 1,000. I would expect it to oscillate on breadboard. Any sneakage from later stages to early stages gets re-amplified and builds-up to infinity (or 9V swings). CMOS is so fast the oscillations are supersonic. You don't hear the howl, just the "confusion" as the stages try to handle extreme overload with audio under it.

> Until i turn the DRIVE knob above about 1/4.

"Gain"? Yes, if gain is less than sneakage, no oscillation; gain higher than sneakage it oscillates.

Lay-out so that output is far from input. Use SHORT leads on parts. Gain and Tone pots and their wires must not be near each other.

[gregwbush]

Thanks again for the replies!

earthtonesaudio, i'd read about buffered versions somewhere. On this occasion the chip is labelled CD4049UBE... assuming unbuffered

Supersonic oscillation... PRR, that makes so much sense i'll bet that's the problem. I've had a BSIAB oscillate and it was audible as a very high pitch kinda squeal. I'll do a vero layout (that's my thing) of the Ube and see how i go...

Cheers,
gregwbush

[WGTP]

They can get pretty farty if the input cap is too large. 

[gregwbush]

"They can get pretty farty if the input cap is too large"

Would that statement apply to the caps at every stage's input?

[gregwbush]

Ok i made a layout on the PC and built it like this...



and it's worse than on breadboard!

I think it's the oscillation PRR was talking about, but i really don't know for sure. Looking at it now, my layout is pretty dumb. I've got inputs near outputs etc. I've decided i don't like IC's. Anyone built a working ROG Ube screamer? What layout did ya use?

[Earthscum]

One thing I would do is add a .1 uF cap (preferably ceramic or mono) between row 2 and 4, right over the first cut in row 3. You pretty much want to keep around a handful of caps just for this purpose. It's generally not shown in schematics, but is something you should do, like shielding high impedance flying signal wires, star grounding, etc. It helps keep oscillation down, and it is recommended to strap it straight across the chip's VDD and VSS (V+ and Gnd). I noticed a couple 4049 projects in the past that started to oscillate (hiss) worse when I put in a 220R resistor between 9V and chip VDD. I finally started using a cap like I mentioned, and all but one project I was able to cure a major amount of hiss, down to what I feel is normal for CMOS, and quite tolerable, even on stage.

[WGTP]

I have built lots of these type of circuits on a bread board and rarely have oscillation problems.  My input and output are on opposite ends of the board.  I have seen suggestions of using large filter caps to help oscillation, might try a 220uf.  Also, try a 47/100pf cap across the 1M resistor in the feedback loop.  This should help.  The input and all caps between stages have the potential to reduce bass depending on size.  I usually adjust the one at input to keep the bass out of the first stage if it is too much.  When I first started messing with the 4049's it was a little different and frustrating, so don't give up.  Once you get it working try different diode combos, lots of possibilities. 

[gregwbush]

Earthscum...

You make me wonder what is goin on here... i did as you suggested and soldered in a 100n cap (square edged yellow one, whatever type that is) and it did clean things up a bit noise wise... but the thing still farts and carries on!

After that i took out the resistor in the "clean" signal path, rx on the ROG page, so all i'm hearing is the "overdriven" part of the circuit (with the diodes 'n' EQ etc)... and what i hear is MUD. Waaaaay too much bass for a TS.

That of course got me thinking about what WGTP has said about too much bass in the first stages of a 4049 circuit. I looked at the schematic, i read the notes... and the 700hz cutoff is apparently happening with the 2n2 and 100k before the second stage. The mud i hear is crazy low.

So i touch the 100k resistor ... and kkkchhhk, signal cuts out. Either slight buzz or nothing, until i disconnect and reconnect the power supply to the circuit... then signal (albeit muddy) is back again. I can repeat this every time. Touch the resistor terminal: circuit dies (that stage at least, dies), reconnect the power and it's back again. It's not a dodgy solder joint...

What is it, static???
If i touch the input to U1B (4049 inverter 2)... it goes silent until i disconnect/reconnect power!
Man, trippen, i said that twice

Oh btw WGTP, i tried different caps when it was still on breadboard and it's just not working right no matter what, don't know what i've done

Right now i feel like closing this thread and saying FFFK cmos... but i'm just annoyed with it and had a few 2many drinks so i'll leave it for now and let readers brilliant minds suggest more suggestions

cheers,
gregwbush

[PRR]

I notice the pin-numbers on runoffgroove's scematic and see you have alloted the inverters differently?

Individually they are all identical. However in the larger context ROG's layout seems to split the "glue" from the "effects" on opposite sides of the chip, and the in-glue from the out-glue by skipping the middle of that side.



I wonder if ROG's basic around-the-chip allocation of sections may have been well-considered and may be good to copy? (Yes, I know, I too hate to trash hours of work.)

[gregwbush]

Oh

Hmmm yeah i totally overlooked that!
The pin numbers are clearly noted on the ROG schematic. I didn't think twice about it.. just started laying out components referring to this image...



I have spent a while on this project, but i'm not mad... it's all a learning experience!

I'll post an update when i get around to doing a new layout

Thanks for everyone's input thus far!
cheers,
gregwbush

[WGTP]

I too have said kiss my ube and gone on to other stuff, only to return later upon seeing some cool new 4049 project.  Might try listening to it after each inverter and see where the treble is getting lost.  The first stage is a buffer and shouldn't sound any different, the second is a gain stage and should sound the same with some bass roll off (maybe with some distortion at the high setting), the third is where the distortion starts realling happening, the forth is the tone control and the fifth is a 2X gain recovery stage.  

[gregwbush]

Yep WGTP... when it was on breadboard i disconnected the "drive" section and sure enough the clean unaltered signal was just that, pretty well transparent, a buffer, if you will. (well 2 buffers, first inverter and last inverter). No problems there.

I'm really thinking that PRR is on the money here. It's quite likely the way it's layed out.

It was bad on breadboard,
worse on Vero

I didn't follow the RoG pin numbers (it seems wrong saying pin numbers, personal identification number numbers, but in this context it's actually legit)

Anyway i need parts, I've got perfboard... i will build the bloody thing on perfboard exactly as shown on RoG...

Then, i'll report back

cheers,
gregwbush

[earthtonesaudio]

"PIN number" is always wrong, just like "ATM machine" and "RSVP please".