AD-3208 issues

[Wazoo]

Well I finally built my first effect, the AD-3208 analog delay from generalguitargadgets.com, and miraculously, it works the first time I test it out. 

But it's kinda distorted, especially if I play chords. I don't have an actual oscilloscope to bias it with but I have been trying to use Visual Analyzer software oscilloscope. However, having never used a scope before, I don't know how to use it.

If anyone knows about Visual Analyzer, or can even tell me which pins I need to touch the probe to on the ICs to set the bias, I would appreciate the help.

Thanks.

[Mark Hammer]

Pin 7 is the magic pin.  Please note that "pin 7" is essentially where pin "13" would be if this 14-pin sized chip had all the pins to fill up the empty spaces.

Biasing of essentially all BBDs works the same way.  As you start from the extremes of the trimpot (hard clockwise or hard counterclockwise), there is no sound whatsoever from the BBD.  As you get closer to the ideal bias voltage, the BBD starts to pass signal, but it's distorted (where you're at now).  At the correct bias voltage, sound is as clean as it is going to be.  Should you overshoot, or keep rotating the trimpot, the sound gets distorted again, and as you overshoot more, you lose signal once again.

Ideally, you CAN tweak it by ear, though the AD-3208 uses a pair of BL/MN3208 chips, which means that you need to be able to audibly separate the consequenceas of setting up the one BBD right from whatever state the second one is in.  That approach would benefit from having a little 386-based or similar headphone amp handy, to use as an audio probe monitor.  If you use a scope or scope software, then you need a signal generator of some sort to feed a steady tone in so you can tweak until the scope says the waveform is kosher.

[markm]

I set mine up by-ear and it works and sounds Great! I won't say it didn't take some time as well as trial and error but,
once it was done it sounded Fantastic.
If you follow the "by ear procedure" as described on GGG, you should be fine.

[newbie builder]

Long live o-scopes 

[Wazoo]

I think I used the software o-scope correctly and readjusted the trimpots to get as close to a perfect sine wave as I could using the little signal generator I had built a while ago. It sounds MUCH cleaner now. My only problem now is it oscillates a little too easily for my tastes. Setting the repeats knob much farther than halfway is about as far as I can go with the oscillations. Is there anything I can do to fix that?

[Jaicen_solo]

Looking at the parts layout, (can't view the schem at work!), i'd guess that upping the 100K resistor connected to the wiper of the repeats knob would do the trick.

[Mark Hammer]

I think I used the software o-scope correctly and readjusted the trimpots to get as close to a perfect sine wave as I could using the little signal generator I had built a while ago. It sounds MUCH cleaner now. My only problem now is it oscillates a little too easily for my tastes. Setting the repeats knob much farther than halfway is about as far as I can go with the oscillations. Is there anything I can do to fix that?

Yup.  Solder a 10k-22k fixed resistor in series with the input of that pot.

[Wazoo]

Thanks I'll try that. I'll try a few different values in that 10-22k range to see what the minimum is needed to stop or control the oscillations so I won't lose too many repeats.

The odd thing is, it oscillates mainly when I play high notes, and the effect is supposed to reduce high frequencies on each repeat.

[Mark Hammer]

The odd thing is, it oscillates mainly when I play high notes, and the effect is supposed to reduce high frequencies on each repeat.

Well it doesn't really do that on its own.  There IS some lowpass filtering on the delay path, but it,s just the normal type found in analog delays that is intended to  keep clock noise under control.  If you wanted to introduce some further trimming of treble on subsequent repeats, you'll want to take that 100k resistor that goes to pin 12 of the NE570 and do some tricks with it.  If you split that 100k up into two or three fixed resistors that sum up to the same resistance (100k) when in series, you can run some caps to ground from their junctions that will add some further lowpass filtering over and above what is being done by those 3 subcircuits built around the 2N5088 stages.

For example, if the 100k component is swapped for a 33k and 68k in series (33k between the 68k and the stock 1uf cap), you can run a .0039uf cap to ground from the 68k/33k junction.  This will provide a 6db/oct rolloff around 1.2khz, which should permit an audible effect which still permits some treble left in the signal to trim on repeats 2, 3, 4, etc.  That ought to prevent yur unit from oscillating wildly.

Given that such treble-trimming also reduces overall amplitude of repeats, you can consider using a 10k, rather than 22k ahead of the 50k pot.

[oldschoolanalog]

I'll try a few different values in that 10-22k range...

50k trimpot maybe?

[Wazoo]

50k trimpot maybe?

My plan was to connect a pot in there with jumpers and see what value of resistance would be best, then solder it in.

[Mark Hammer]

Something you often see in commercial analog delays is a trimpot.  At the point of calibration, the regen/repeat/feedback pot is set for maximum, and the regen trimpot is eased back until the point where the oscillation stops.

[Jaicen_solo]

I like the look of this pedal, but i'm not overly keen on the short max delay. Any idea what it would sound like overclocked? I can live with dirty sounding repeats, as long as there's no clock whine. I mainly use delay in the 500-700ms range, so would it be possible to mod this to do that?

[oldschoolanalog]

Overclocking (clock rates >100k) will give you shorter delay times. For more delay, try 2X3205. More stages = more delay, assuming the same clock rate is used. With a 10k clock this will give you a bit over 400ms delay. This is all explained in the article at GGG that accompanies this project. I realize some claim >400ms using two 4096 stage BBDs and a 10k clock. This is wishful thinking. Do the Nyquist equations. The math doesn't lie. BTW, 400ms is a nice amount of delay for an analog delay. You could try clocking at less than 10k for longer delay times, but then you run the risk of aliasing and clock bleed-through.
osa

[StephenGiles]

Yes, but consider for a moment the hideous noise made by guitars through fuzz boxes - but a nice delay box distorting at long delaytimes is something else

[Jaicen_solo]

Of course, I did actually mean clock it at a lower frequency. As I say, I can live with some aliasing. I really like tape type delays (don't we all!) and they often create distortion as the sound decays, so that's cool. I have a modded DD-3 for cleaner stuff, so a dirty delay would do nicely! Anybody done this?

[Wazoo]

Update,
I finally changed the 100k resistor to a 150k and it helps some. I can turn the knob up a little more before the unit oscillates, but I did lose quite alot of maximum repeats for only a little more room to twist the knob. I'm just not quite sure there isn't some other issue. I mentioned before that it oscillates much more easily if I play higher notes than lower ones, and I also notice that if I hold a note or chord and let it decay naturally rather than stopping it abruptly, it does not oscillate, even on high notes.

Is this normal for this effect? Or could there be some other problem such as the compander, or the clock frequency not set right (i couldn't figure out how to check the clock freq and set it to 10kHz max)?

Thanks again for all the help.

[newbie builder]

Not to completely de-rail the thread, but some how I'm missing the spot where the polarity of the 1uf caps is noted for this project. I've got my entire board populated except for those 1ufs because I only have electrolytic 1uf caps and I see no polarity markings on the schem or GGG pcb/layout.... what am I missing here?
Does it require non-electro caps?

[caress]

yeah non-polarized.  you can use polarized caps if you don't have any NP...

...from the DIYWiki:

From R.G.
"You can make a quick and dirty NP cap by tying together the negatives of two equal-sized polar caps.
In the series-NP connection, the capacitance value is funny. Normally caps in series are a smaller capacitance than either cap by itself. If you had two 3.3uF polyester caps, then the expected value for two of them in series is 1.65uF. However, electrolytic caps actually conduct in the reverse direction, so two 3.3uF polarized aluminum electrolytic caps act like they each have a diode in parallel with them that conducts when the voltage is backwards for that one cap. So two 3.3uF caps hooked up as series non polar (i.e. negative to negative) look like a single 3.3uF NP cap.
... except for tiny region near zero volts where they withstand a tiny reverse voltage, so they look like 1.65uF there..."

[newbie builder]

I shoulda thought to look there...thanks!