Mark Hammers Woody Issue

Started by Scruffie, January 06, 2012, 12:38:03 AM

Previous topic - Next topic

Scruffie

I built Mark Hammers Woody from my own PCB layout a while Back (Mark Hammer verified the board I sent him) and while it works as it should, there is a slight issue.

At high treble settings (and I like a bright sound) it seems to have an octavey/ring moddy/misbiasey sound on the high strings & notes of the guitar.

I've checked the voltages and audio probed it best I could (admittedly the set up for this was poor but better than nothing) but I can't find anything wrong.

The only substitutions I made were 4M7 resistors in place of the 5M1s and B50k Pots with 12k parallel resistors to bring them down to around 10k. Oh and the 100k Output pulldown resistor is missing.

Any clues as to what could be going on? I've checked all the values over and over on seperate occasions, checked my wiring, all the parts are in the right place and as I said, it does work fine apart from that issue. Oh and i've tweaked the trimmer to my hearts content, nada.

I worked from this schematic - http://www.zeta-sound.se/woody/Woody.JPG

Scruffie

#1
Bueller? Anyone?

Tacked 330ks on to the 4M7s to get them closer to 5M1s just to be sure and resoldered all the joints even though they look perfect, no dice.

markeebee

Mark Hammers Woody issue........oh, where do I start?

Is that the same etch as the one I have from you?  I used 4M7 too, and never noticed any glitchiness until you mentioned it. But now, yeah, there's a very faint almost-8-bit thing going on in the upper registers with treble up. Very, very faint. Not enough that I would feel a burning need to do anything about it. Be interesting to find out what's going on, though.

Sorry, not being much help, am I?

Scruffie

Quote from: markeebee on January 07, 2012, 05:12:03 AM
Mark Hammers Woody issue........oh, where do I start?

Is that the same etch as the one I have from you?  I used 4M7 too, and never noticed any glitchiness until you mentioned it. But now, yeah, there's a very faint almost-8-bit thing going on in the upper registers with treble up. Very, very faint. Not enough that I would feel a burning need to do anything about it. Be interesting to find out what's going on, though.

Sorry, not being much help, am I?
Nah this is a different layout, it's a fabbed board.

That is a help though! Makes me think it's not my build but a quirk in the circuit. With mine it can be an issue, if I have the amp treble up high and the pedals treble full it distorts quite nastily if I strum hard. Which is a shame because otherwise it's a damn nice effect.

Sounds like it might be a case for the breadboard.

petemoore

  Assuming there are no mistakes...learned this lesson a while back [until it works and tests 'right' in all respects...don't assume it is].
   High notes only causing the 'anomoly' doesn't lend itself to power supply limitation being the problem', and also suggests it could be in or outside the actual circuit, probably an ''insider'' problem, but swapping 'peripherals' sometimes makes or breaks the 'issue'.
  Kind of waiting for MH to chime in here...
  Probe to the point where the 'issue' begins in the circuit, try a 'dinky caps rolloff' to dampen or shunt the HF's right ~there, or just before there/maybe after [reducing HF content right before where the 'zinger' is generated is probably preferrable because less HF reduction needed to squelch 'it']. 
  ...Assuming...^
  That said, it should work as designed, but as I understand the circuit to be a HF 'character-builder'..placing a character in a workplace with other circuits changes the percieved behavior, perhaps it is wired and working correctly but under certain circumstances exhibits this behavior. Also, being a HF 'accentuator', any adding of LP filters should be done with the notion that it will alter the sound of the HF's of course...perhaps done 'delicately' can improve behavior without dulling too much. Whats fed into it of course alters what it does, trying something right at the guitar [tone knob] might give a hint.
   Thinking maybe that the string/bridge or PU might put in a little HF 'tiz' that gets 'characterized' in the circuit which causes the behavior to surface?
  Passing of all the debugging voltage/probe tests generally shows if it can work', but it doesn't read any component values
Convention creates following, following creates convention.

CynicalMan

My bet's on the distortion stage. It would make sense for it to cause intermodulation artifacts that you wouldn't normally hear with a clean circuit.

Mark Hammer

I have no idea what would do this.  It may well have been there all along and I just didn't hear it because I normally test things out on a very modest battery-powered amp that has limited bandwidth and is played at levels you could talk over.

Note, however, that the cutoff for the highpass filter is a one-size-fits-all cutoff, and it may be that it simply does not complement your guitar and generates some ugly harmonics of harmonics.  That's just a wild guess, though.

Scruffie

Quote from: petemoore on January 07, 2012, 08:09:24 AM
  Assuming there are no mistakes...learned this lesson a while back [until it works and tests 'right' in all respects...don't assume it is].
   High notes only causing the 'anomoly' doesn't lend itself to power supply limitation being the problem', and also suggests it could be in or outside the actual circuit, probably an ''insider'' problem, but swapping 'peripherals' sometimes makes or breaks the 'issue'.
  Kind of waiting for MH to chime in here...
  Probe to the point where the 'issue' begins in the circuit, try a 'dinky caps rolloff' to dampen or shunt the HF's right ~there, or just before there/maybe after [reducing HF content right before where the 'zinger' is generated is probably preferrable because less HF reduction needed to squelch 'it']. 
  ...Assuming...^
  That said, it should work as designed, but as I understand the circuit to be a HF 'character-builder'..placing a character in a workplace with other circuits changes the percieved behavior, perhaps it is wired and working correctly but under certain circumstances exhibits this behavior. Also, being a HF 'accentuator', any adding of LP filters should be done with the notion that it will alter the sound of the HF's of course...perhaps done 'delicately' can improve behavior without dulling too much. Whats fed into it of course alters what it does, trying something right at the guitar [tone knob] might give a hint.
   Thinking maybe that the string/bridge or PU might put in a little HF 'tiz' that gets 'characterized' in the circuit which causes the behavior to surface?
  Passing of all the debugging voltage/probe tests generally shows if it can work', but it doesn't read any component values

I considered it being the guitar, tried it on both a Strat & Tele, same issue, played with the tone knobs too, I can mask it by rolling off treble but that kills the sound for me.

Quote from: CynicalMan on January 07, 2012, 09:20:15 AM
My bet's on the distortion stage. It would make sense for it to cause intermodulation artifacts that you wouldn't normally hear with a clean circuit.

I'd be inclined to agree, when I did my poor audio probing (the signal I had to use was crap so it was hard to hear properly) the bass side seemed fine it's coming from somewhere in that treble/distortion area.

Quote from: Mark Hammer on January 07, 2012, 09:53:02 AM
I have no idea what would do this.  It may well have been there all along and I just didn't hear it because I normally test things out on a very modest battery-powered amp that has limited bandwidth and is played at levels you could talk over.

Note, however, that the cutoff for the highpass filter is a one-size-fits-all cutoff, and it may be that it simply does not complement your guitar and generates some ugly harmonics of harmonics.  That's just a wild guess, though.
As I said above I tried two different guitars and it has the same issue, I tried it on a battery powered amp before I boxed it and didn't notice it but now going back to the Ruby, I can hear it if I listen.

Mark Hammer

Ahhhhh, a Ruby.  Can the 386 handle all that treble?

Scruffie

Quote from: Mark Hammer on January 07, 2012, 11:01:55 AM
Ahhhhh, a Ruby.  Can the 386 handle all that treble?
Tried it on another amp too  :)

gritz

Bear with me on this - it happened a long time ago so I'm fuzzy with the details. It might be relevant though...

Years ago I was working on a homebrew distortion built around a TL074 and I could hear a strange artifact that was completely out of tune with what I was playing. I prodded around with the 'scope and found that one of the opamp outputs was going to maximum excursion and then randomly jumping to the opposite polarity, so instead of a square-ish wave the output was (sort of) random pwm. I since found out that some opamps can be prone to this and it's called phase reversal. The strange thing is that since getting back into analog over the last year I've used plenty of TL07x's and I haven't been able to replicate it. I guess that ic's from different manufacturers aren't created equal.

The only thing in this case is that having looked at the schematic you've linked there doesn't appear to be enough gain for a rail to rail clip (unless something is happening around that highpass filter). Maybe try a different quad opamp?

I see there's a couple of very high resistance values in there. Perhaps there's one of those layout / stray capacitance / rf oscillation things going on there.

Sorry, I guess I'm not being much help...

Scruffie

Quote from: gritz on January 07, 2012, 01:18:49 PM
Bear with me on this - it happened a long time ago so I'm fuzzy with the details. It might be relevant though...

Years ago I was working on a homebrew distortion built around a TL074 and I could hear a strange artifact that was completely out of tune with what I was playing. I prodded around with the 'scope and found that one of the opamp outputs was going to maximum excursion and then randomly jumping to the opposite polarity, so instead of a square-ish wave the output was (sort of) random pwm. I since found out that some opamps can be prone to this and it's called phase reversal. The strange thing is that since getting back into analog over the last year I've used plenty of TL07x's and I haven't been able to replicate it. I guess that ic's from different manufacturers aren't created equal.

The only thing in this case is that having looked at the schematic you've linked there doesn't appear to be enough gain for a rail to rail clip (unless something is happening around that highpass filter). Maybe try a different quad opamp?

I see there's a couple of very high resistance values in there. Perhaps there's one of those layout / stray capacitance / rf oscillation things going on there.

Sorry, I guess I'm not being much help...
I did wonder about the chip, it'll be a pain to desolder them, I think my next step might be just to breadboard the treble/clipping section and see if the same thing occurs and if it can be filtered out.

The chips were TI TL074s... I have LM324s (which would be a bit too noisy me thinks) and OPA4134s on hand if I do change it.

Hmm... I don't think it's the layout, it's on a 4 layer board with nice thick tracks, then again stray capacitance seems to be one of those things that'll get you regardless.

gritz

The breadboard thing is a good idea - as you say, desoldering is no fun, especially if you're just swapping out parts on a trial and error basis. Those expensive multilayer boards with through-hole plating make it hard to get things like ic's out without damaging something too. I'm a clumsy oaf with hands like bunches of bananas so I tend to socket all my ic's these days!
I'm not an opamp snob, but I don't think a 324 would perform terribly well. Swapping a fet opamp for a bipolar would also imo mean swapping out the 5M1 (4M7 in your build) resistors for something a bit smaller 'cos of bias current requirements. I'd be tempted to run lower value resistors anyway, but I'm paranoid about noise, stability and whatnot. I'd also be tempted to add a 100n ceramic or similar cap across C3 (the bias smoother cap) and do the same across the supply rails of the opamps (i.e. directly across pins 4 and 11 on the copper side of the board, or somewhere convenient as close to the ic's as possible) as there don't seem to be decoupling caps indicated on the schematic. This may help if it's a stability thing.
Do you have access to an oscilloscope? If not then maybe probe all the opamp outputs with a digital multimeter to make sure they're all sat at half supply voltage.

Hope it works out ok.:)

Scruffie

Quote from: gritz on January 07, 2012, 03:10:01 PM
The breadboard thing is a good idea - as you say, desoldering is no fun, especially if you're just swapping out parts on a trial and error basis. Those expensive multilayer boards with through-hole plating make it hard to get things like ic's out without damaging something too. I'm a clumsy oaf with hands like bunches of bananas so I tend to socket all my ic's these days!
I'm not an opamp snob, but I don't think a 324 would perform terribly well. Swapping a fet opamp for a bipolar would also imo mean swapping out the 5M1 (4M7 in your build) resistors for something a bit smaller 'cos of bias current requirements. I'd be tempted to run lower value resistors anyway, but I'm paranoid about noise, stability and whatnot. I'd also be tempted to add a 100n ceramic or similar cap across C3 (the bias smoother cap) and do the same across the supply rails of the opamps (i.e. directly across pins 4 and 11 on the copper side of the board, or somewhere convenient as close to the ic's as possible) as there don't seem to be decoupling caps indicated on the schematic. This may help if it's a stability thing.
Do you have access to an oscilloscope? If not then maybe probe all the opamp outputs with a digital multimeter to make sure they're all sat at half supply voltage.

Hope it works out ok.:)
Yeah, desoldering 4 layer boards is definitley no fun, I only have one of these boards spare now so I don't really want to waste this one by messing it up.

No, a 324 would do poorly I think, with all the treble amplification I think it would hiss far too much.

I'll give the added caps a go, I hate a messy board but if it solves the issue it'll be worth it and that's something I can do without too much hassle. Then it'll be to the breadboard, as Markeebee seems to be experiencing the same thing I don't think it's my set up or build, just a quirk that perhaps shows up more so in certain settings.

No scope available but I did probe all the pins and they all seemed within reasonable voltage range, except I couldn't test the inputs as my multimeter couldn't deal with the impeadance.

I hope so too!  :) This is one of my favourite effects i've built other than this issue.


gritz

I've used harmonic exciters before (on bass and vocal recordings mainly) so I know the principle is sound. It's easily possible to overdo it, but you defo shouldn't be getting "ring mod" type sounds. If you have the patience to breadboard the whole thing then perhaps you'll be able to duplicate the problem and fine-tune it out before having to reach for the soldering iron.
Re. the decoupling thing - I'm not promising that it will solve the problem, but it's good engineering practice imo and it's a non-destructive mod too.

Let us know how you get on! :)

gritz

Me again...

I just had a thought regarding the highpass built around ic1d. I see it has a bit of loop gain - maybe it has enough resonance around it's cutoff frequency that it's ringing when passing audio and that's interacting with the signal. Perhaps someone could run a Spice sim on it?

Perhaps it may be worth temporarily shorting the 22k feedback resistor to lower the loop gain to unity in order to test this (there are two resistors marked R15 around ic1d so I'm referring to the top one). If the ring mod distortion lessens then this might be the problem.

Does this sound feasible or am I talking out of my hat?  :icon_lol:

evogel99

I have this circuit up in ltSpice. I recommend lifting one end of R16. If it is larger than R15, you may be getting oscillation. (I'm not sure why R16 is there anyway!)

Scruffie

Quote from: evogel99 on January 07, 2012, 07:21:47 PM
I have this circuit up in ltSpice. I recommend lifting one end of R16. If it is larger than R15, you may be getting oscillation. (I'm not sure why R16 is there anyway!)
R16 as connected to V.Ref? (I know it looks like there's two R15s or 16s but one is 15 and one is 16).

If so, thanks for the tip! I'll give it a go.

evogel99

Meant to add - R16 is R15 in the original schematic. And I should have said if R16 is LESS than R15 it may be oscillating! :-\

evogel99

Yes - the one connected to VR.  it is hard to read, but I take it to be R16.