SAD problem with a ross flanger i fear...

[oldschoolanalog]

Probably needs to be biased correctly.
To my simple mind turning the pots and getting "crackle" is usually indicative of pots that need cleaning.
Aside from that; consider using the SAD1024 in the MXR117 clone. That board fits in a much smaller box than the Ross. There is plenty of room in the Ross for a daughterboard without having to cram things in and getting stressed in the process.

[pinkjimiphoton]

hi mark,

i believe what happened is and electro or a tant cap leaked enough dc to fry the original chip.

pop in the new chip, and the think comes to life, with a good bit of "sweet spot" on the 3 trimmers. but...there is distortion, a muffled crackle i've associated with leaking blocking caps.

so yes, you are right...the problem lay outside the chip. unfortunately, i don't think the chip survived. putting the original chip in with the same settings that make the known good one sing it does nothing. seems to me it would be the chip.

so...i'm off to rat shack, the only big electros i have atm are 1000u @10v...not enough nads to handle the 22 volts coming from the power supply.

gonna start by putting something huge in there...maybe 1000-4700u @ 35 or 50v in first for the filter, then replacing all the 10u electros.

from that point, i'll check it out and see if the crackle remains. if it does, i'm betting on one or more of the tants being open and letting dc thru.

i'll check back shortly...sorry for the confusion.

the weird thing is, the bad sad and the good sad both read about the same, voltage wise.

and mis-clocking the sad? there's a spot at the extreme of the clock adjust where it starts phase shifting. it does a thru zero i never knew these things could do, and could probably be tweaked just a little to self oscillate. i think it's gonna be cool.

dave...thanks bro...
i would like to retrofit this thing and save the sad for a proper project.

[pinkjimiphoton]

dave. the crackle isn't like a dirty pot, it sounds like the crackle from a SHO more..

rat shack here blows...no more metal boxes (good news for small bear, as all the db!!FX pedals to come once my stock runs out will come from there)...no pot cleaner.

sheesh.

hope they have some decent caps..otherwise, i'll have to series/parallel a couple with some resistors so i can use them 1000u ones.

[pinkjimiphoton]

tony...amazing, isn't it?



thanks bro!!!

[Mark Hammer]

Dave, I thought the Ross and the MXR-117, like so many of the Ross and MXR pedals, were  - but for the odd mod here and there - essentially interchangeable.

[analogguru]

Folks, I have one question:
Before you close your nice Ross Flangers, would it be possible that someone takes a picture of the solder side so that this little baby can be traced ? 

analogguru

[pinkjimiphoton]

hell,yah, analog guru.,,

bur you have traces on both sides i thiink,,,

i already changed the filter cap, but other than that is stock.

i have no way to get a macro pic unfortunately. all i got is my sh***y phone or an ANCIENT digital camera.

video shortly, too

[pinkjimiphoton]

[pinkjimiphoton]

dunno if these are useful, unfortunately...probably too small

[analogguru]

If the last picture would have the same size as the others it would be a it easier.....

Anyhow, you have the last version of the Ross Flanger (same pcb as in the black one) and what can be seen from the pictures the schematic of the MXR Flanger (M-117) should match pretty well:
http://www.shredaholic.com/mxrflanger.html

The BBD-bias trimmer (R49) is the upper middle (between the pots), the trimmer beside the recycle pot is the "trim level" (R 48) and the trimmer beside the transistors is the "trim clock" (R43).

analogguru

[pinkjimiphoton]

damn phone pix, bro. they suck.


i'll try and get another...never know what size the dang things will be.

yah, had a "made in taiwan" sticker on the  bottom, definitely the same as the black ones.

very nice, liquid...did ya check the video, analog guru?

[pinkjimiphoton]

ok, took it out last nite, worked brilliantly.

like i said, i never knew these to be capable of thru zero flanging, but it's definitely doing it, and is way more intense (probably due to my intentionally miss-setting the trimmers til it sounds the way i like it) than any of the ross flangers i've had previously...perhaps the chips were tired in them, or they were just set up to be "more polite" at the ross factory.

i like the mix trimmer full tilt boogie.

the sweet spot on the other 20k is gonna be different i'd imagine from chip to chip. around 1:30-2:00 on it (with half way as 12:00) seems to be best..by careful adjustment, you can really dial it in.

the clock trimmer was the most interesting. you can dial in the whole effect right there, and that's where i discovered i could do the thru zero trick. full blast makes some interesting sounds!!! depending on the settings of the front panel knobs, you can get everything from chorused sounds to bending rubber bands.. even a bit of yoy. pretty cool. i think i'll keep it.

so...next phase...

adapting an mn3007 to work in this circuit. i read thru a bunch of the threads, but it's a little (a lot) over my head still..

[B Tremblay]

I started to disassemble mine to provide the trace side photos as AG requested, but I don't have a deep enough socket to remove the pot nuts. Sorry!

[Mark Hammer]

Technically, it can't do TZF, and neither can the MXR.  There is certainly an interesting "turnaround" in your demo, that has a similar feel, but it's not TZF.  Here's an example: http://www.youtube.com/watch?v=y94SFk3CID0  As one reaches the through-zero point, the signal appears to get "sucked away" and then return.

In theory, an SAD1024 could be used for TZF, since it is configurable as two independent 512-stage sections.  But that would mean a separate fixed clock for one section and a variable one for the other, which is not found on either the Ross or MXR.  There is also no delay imposed on the dry signal.

Haing said that, one of the reasons why the Reticon chip is loved for flanging is because it is very easy to get it to clock so high that it is very very close to the zero point (though as a delay chip, it would always lag behind an untampered clean signal by some amount).  The PAiA Hyperflange gets it down to 170usec (i.e., 0.17msec), and the A/DA is somewhere in that neighbourhood as well, a few more octaves of upward sweep than the BF-2.

Not intended as a dressing down, bro.  Merely a friendly reminder that there are still new worlds to conquer.  As always, if it sounds good, it IS good.

[pinkjimiphoton]

far out, thanks mark, no dis taken, it helps me learn.

so to actually do true "thru zero" you have to start with a delayed signal, correct? so you can move the two side independently and make them cross over, right?

it sure sounds pretty close...the ones i had before never sounded like this. 

i've gotta check back later...shower and gig...early tonite, downbeat for 8. ugh.

[12Bass]

Technically, it can't do TZF, and neither can the MXR.  There is certainly an interesting "turnaround" in your demo, that has a similar feel, but it's not TZF.  Here's an example: http://www.youtube.com/watch?v=y94SFk3CID0  As one reaches the through-zero point, the signal appears to get "sucked away" and then return.

In theory, an SAD1024 could be used for TZF, since it is configurable as two independent 512-stage sections.  But that would mean a separate fixed clock for one section and a variable one for the other, which is not found on either the Ross or MXR.  There is also no delay imposed on the dry signal.

Haing said that, one of the reasons why the Reticon chip is loved for flanging is because it is very easy to get it to clock so high that it is very very close to the zero point (though as a delay chip, it would always lag behind an untampered clean signal by some amount).  The PAiA Hyperflange gets it down to 170usec (i.e., 0.17msec), and the A/DA is somewhere in that neighbourhood as well, a few more octaves of upward sweep than the BF-2.

Was going to mention how it comes close to TZF, but cannot go through the null with a single delay line.   I wonder how much trouble there would be if using two different clocks on an SAD1024A or MN3010?  Given the close proximity of the delay sections, I would suspect that aliasing between the two clocks would be a major concern.  Of course, the static delay could be set rather short, so the static BBD clock could be well above the audio range (> 100 kHz).  But it would seem that the two clocks would interfere, especially near the through-zero point.

FWIW, my A/DA clone is presently set for a minimum delays of of around 0.45 ms - much shorter than that and there's not much audible flanging, even on full-bandwidth signals.  That would put the first notch of the comb filter around 2200 Hz.  0.4 ms results in the first notch at 2500 Hz.  Also, I've found that the SAD1024A's gain is reduced by a couple of decibels when it pushes up close to 1 MHz, which results in shallower comb filtering, and less intense flanging.  I would suspect that the signal loss at 0.17 ms would be even greater.   On mine, I've set the dry/delay mix for maximum cancellation through the critical midrange, say from 200 Hz to 1 kHz, which roughly corresponds to the 5 ms to 1 ms range, or a clock rate of 51.2 kHz to 256 kHz respectively.  The BBD gain stays relatively linear through this range.

Also, the Ross sample and a few others have a pronounced "boingy rubber band" effect at the bottom of the sweep that I don't hear from my A/DA.  I wonder if the LFO waveform is somewhat different?

Oh, and I'm wondering if there might still be some faint hope for that old Reticon chip.  If you're lucky, it might work if the rest of the circuit is brought back to spec.

[oldschoolanalog]

It is possible one side of the "bad" SAD1024 is still good. That'll give you a 512 stage BBD which is still very useful. Don't toss it yet!
@ 12Bass: When I finally get some free time to do some meaningful work I plan on trying to clock the 2 sides of an MN3010 (I have plenty of them) differently. This will require a lot of careful layout planning, ground planes, etc. This has been on the "to do" list for waaaay too long. Just need to get my life back first. 

[armdnrdy]

Dave,

MN3010s? I have plenty of them?

Okay boys.....Get Him!!!

[oldschoolanalog]

Okay boys.....Get Him!!!

Get in line. There's lots of "get him" directed at me lately. 

[Mark Hammer]

Was going to mention how it comes close to TZF, but cannot go through the null with a single delay line.   I wonder how much trouble there would be if using two different clocks on an SAD1024A or MN3010?  Given the close proximity of the delay sections, I would suspect that aliasing between the two clocks would be a major concern.  Of course, the static delay could be set rather short, so the static BBD clock could be well above the audio range (> 100 kHz).  But it would seem that the two clocks would interfere, especially near the through-zero point.

FWIW, my A/DA clone is presently set for a minimum delays of of around 0.45 ms - much shorter than that and there's not much audible flanging, even on full-bandwidth signals.  That would put the first notch of the comb filter around 2200 Hz.  0.4 ms results in the first notch at 2500 Hz.  Also, I've found that the SAD1024A's gain is reduced by a couple of decibels when it pushes up close to 1 MHz, which results in shallower comb filtering, and less intense flanging.  I would suspect that the signal loss at 0.17 ms would be even greater.   On mine, I've set the dry/delay mix for maximum cancellation through the critical midrange, say from 200 Hz to 1 kHz, which roughly corresponds to the 5 ms to 1 ms range, or a clock rate of 51.2 kHz to 256 kHz respectively.  The BBD gain stays relatively linear through this range.

Also, the Ross sample and a few others have a pronounced "boingy rubber band" effect at the bottom of the sweep that I don't hear from my A/DA.  I wonder if the LFO waveform is somewhat different?

Oh, and I'm wondering if there might still be some faint hope for that old Reticon chip.  If you're lucky, it might work if the rest of the circuit is brought back to spec.

One way to get around the potential heterodyning of clocks might be to use different capacity BBDs for which the required clocks would be so different as to yield inaudible sideband products.  So, something like a 2048-stage BBD used to produce a 2msec fixed delay, and a 512-stage unit to provide the swept delay, that wouldn't have to go much lower than 1.5msec to produce a robust through-zero effect.  The clock for the 2048 stage unit would be VERY high, such that adding or subtracting the clock from the swept delay would still ikely yield something out of hearing range.

The starting point in the spectrum for comb notches is often inaudible at the high part of the sweep.  Particularly with guitar, the signal bandwidth does not extend upwards all that far, so there may well be no audible notches at the top end of the sweep.  As I'm fond of reminding people, the benchmark flanging sounds that send all of us running to buy or make one, or at least try one out, are frequently flanging applied post-production to multi-source mixed-down tracks that cover an enormous chunk of the spectrum.  Under those conditions, the comb filtering is more obvious, because there is more content to filter.  When you throw a rather spectrally-limited guitar at a flanger, there can be big chunks of the spectrum where the flanger is producing potential notches, but there's nobody home at that address.  The result is that it tends to disappoint us.

Nothing wrong with the flanger, just with the program material.

The same, incidentally, is true about phasers as well.  Sixteen-stage phasers will sound pretty dang fine with a full drum kit or white noise, but won't deliver much more sonic bang on a guitar than a 6 or 8-stager will, despite what seems like it oughta be fabulous in theory.  Again, if a notch is hypothetically created i the spectrum, and there is no source material there to be affected, was a notch truly produced?

Steve Giles has intermittently waxed eloquent about the "bounce" circuits found in some flangers.  This is a circuit intended to emulate the little momentary tug that a tape take-up reel would provide as one took one's thumb off the flange.  The old Eventide Instant Flanger had this feature, and the late and greatly missed Jurgen Haible incoroporated it into his Storm-ide and Son of Storm-Tide flanger designs.  I don't know enough about the inner workings of the MXR/Ross LFO circuit to know whether anything sort of like that is being produced.  It could simply be an asymmetrical waveform that has a jarringly quick turnaround at one end of the sweep.