AD-3205 delay distortion issue (SOLVED)

[armdnrdy]

I know that there have been threads on this issue, but I have yet to find one with a fix.

I built an AD-3208 with two V3205's which makes it an AD3205. Are you with me so far?

http://www.generalguitargadgets.com/diagrams/ad3208_schematic.pdf

I've biased the BBDs by inserting a 1Vpp, 400Htz sine wave into the input and adjusting the two BBD trimmers and the two balance trimmers.

The problem is...the best I can get the sine wave looking is clipped on the top and bottom for both BBDs. It looks pretty squared off.

When I test the delay with a single coil strat the delays distort on the lower notes and when I play chords as well. I can just imagine how it would sound with humbuckers!

Now I know that a certain amount of distortion is inevitable with BBDs but it is possible to get a much cleaner delay than what I'm hearing. The MXR Carbon Copy uses four BL3208s and produces very clean sounding delays.

I've researched the delay schematics that Scott Swartz based this design off of, but I didn't find anything unusual that might be causing excessive distortion from the BBDs.

I was wondering that if I were to change a few of the cap values in the first filter section if that would cut some of the lows and help keep the distortion at a minimum.

What do you say?

[chi_boy]

This sounds similar to the issues that were had with the V3205 version of the Madbean Dirtbag.  A search over there may yield some helpful info.

Also on a curiosity note, have you probed the signal coming out of the SA571?  Is it clean there?  And full wave?

And does the 571 have a gain control of any kind?  I know there are a couple of resistors that get changed at the 571 based on what BBD you use. 

It sounds like you have a scope.  If so, it would be interesting to know what that signal looks like before it hits the BBD.

[armdnrdy]

It's kind of funny that as you were posting this, I was going through the build scoping the signal.

It looks fine, nice pure sine wave until it hits the first BBD.

[chi_boy]

So then all of the clipping is happening in the BBD.

Is the input signal the same amplitude as the signal just before the BBD, or is it amplified?

As an experiment, is there anything that can be done to decrease the signal and then observe the affect on the clipped signal?  Maybe like increasing the last 10k resistor before the BBD?

[bean]

I have not built the GGG delay, but here are two things I would try: change the two 22k resistors to 10k on the first half of the SA571 to lower the output, and then try a 100k is series with the 1uF before pin7 (input) of the first BBD.

[Dave W]

And lower the supply V to the BBD down to 8V or less as specified in the Cool Audio datasheet.
7.5V is the goal, IIRC.

[armdnrdy]

And lower the supply V to the BBD down to 8V or less as specified in the Cool Audio datasheet.
7.5V is the goal, IIRC.

The Cool Audio V3205 data sheet that I have lists the absolute maximum rating for VDD at -.03 thru +11 volts.
This BBD is an exact clone of the MN3205. There have been plenty of designs with these BBDs running at +9 volts.

[Dave W]

"Absolute Maximum" does not always mean "best performance".
Just making a suggestion from info I read on FSB.
And, while it is an exact clone (on paper anyways) of the MN3205; when I tested a bunch (25 pcs.) of V3205 in my DM2 there was a very noticeable difference in the sound vs. the MN3205. The V3205 were much louder and brighter.

[armdnrdy]

Well... at 9 volts we're not running it at absolute maximum ratings.

As far as your observation of "louder and brighter" that's worth looking into. I'd like to try to tame the input signal a bit and see what happens.

I'll try Brian's idea (thanks Brian!) of changing the resistors on the compressor side of the 571. If that's a no go, I'll try putting a resistor between the first filter section and the first BBD input.

[bean]

If you find that your delay signal gets cut a little too much by putting a 100k before BBD1, then I would compensate by either increasing the two 100ks to ground before the Cancel trim (220k, 330k, etc) or lower the 22k mixing resistor at the output stage. I'm pretty sure some combination of these ideas is going to do the trick. The damn v3205 is a real headache to get just right.

[armdnrdy]

I tacked on a 18K resistors in parallel with each 22K to emulate a 10K for R4 and R5.

No change on the scope.

While I was there I held a .01uf cap in parallel with C12 and the bottom of the wave from the first BBD rounded into a nice sine but the top remained unchanged. I didn't check it for change in tone.

I'll try putting a series resistor in before the first BBD.

[bean]

Hmm, you should at least see a change in the voltage output of the 571. IIRC, the nominal bias is 2.6v and the value of the resistors multiplies that value out. I'd have to look at the datasheet again.

[armdnrdy]

Okay,

I drilled a couple of holes and installed sockets to try different value resistors. The 100K worked best for a good clean sine, but it does cut the delayed signal.

I'm going to try lowering the mixing resistor. I'm afraid if I change the first BBDs 100Ks to ground, it will let too much signal into the second BBD, then I'll have to change those 100Ks as well.

[armdnrdy]

Success!!!

I removed R32 (the 22K wet mixing resistor at the output) and installed a couple of sockets. I tried a 10K first and that sounded pretty close to unity with the dry signal.

This build sounds as it should now!!! It's no longer a delay/distortion box!!!


Thanks Brian! I really appreciate you taking the time!

[bean]

Sweet! Yeah, I've spent way too much time fiddling with analog delays but it comes in handy once in a while. Glad you got it working correctly.

[chi_boy]

Very cool.

So then your final fix was adding an additional resistor that's not in the original schematic between the 2N5088 and the 1uF just before BBD #1, and decreasing the 22k in the mixing stage to 10k?

If that's wrong, you think you could summarize what you did?  These forums are great for future searches and it will likely help someone later.

[armdnrdy]

These forums are great for future searches and it will likely help someone later.

That's why I changed the name to (SOLVED)
Now if someone is searching for a fix for the distortion problem they will find this thread. I searched around before I started this thread and as I stated in my first post, saw that this was an issue but found no fix.

You had the fix correct. Install a 100K resistor between Q1 and C13 and change R32 from 22K to between 4.7K and 10K. I went with 4.7K to give a little bit of room to back off from with the Delay Level control. It seemed the 10K gave about unity between wet and dry signals.

There's one other thing that someone might want to try if they were to route their own board.

I noticed that there was a bit of gain from the output of the first BBD going into the input of the second BBD. I was able to adjust the first BBD for a symmetrical sine wave, but the second BBD fell a little short of that.

I took a look at the Ibanez AD-900 schematic to see how they accomplished the dual BBD design. The AD-900 comes out of the first BBD into the balance trimmer and then a 1uf cap. After that cap, the signal enters terminal three of a 200K trimpot. The signal exits the wiper and terminal one of the trimmer is grounded. The signal passes through another 1uf cap before entering the input of the second BBD.

This trimmer allows for trimming back the signal exiting the first BBD entering the second BBD. With this trimmer in place one should be able to get a fully symmetrical sine wave while adjusting the second BBD.

[chi_boy]

I'm risking repeating myself, but do these mods look like your changes?

I know some of your changes were just resistors, but I thought to put in trim pots instead.  I also added a trim for the feedback.  I've had a couple of delays where that needed adjustment too, so I figured a trimmer could go there too.  All the changes are highlighted.

http://home.comcast.net/~gprause/AD3208_MOD_v1.0.png

[armdnrdy]

Hey George,

Your schematic looks correct.

I routed my own board and added a trimmer and a fixed resistor for the feedback circuit as well. I also used a 100K for the trimmer coupled with a 47K for R6. Try the 10K and see if that will work out.

I built the first AD3205 for a Christmas gift for a friend's son last year. This one, (also a gift) I've been able to spend a bit more time on. On this build I also added a modulation board (modified from the delay modulation board on the other site) which I designed to accept two onboard pots. The pots support the small modulation board above the main board. This modulation add on takes the delay to a whole new level! You can get chorus, tremelo/vibrato type sounds to sea sick modulations. Very cool! If you're interested in the "mod" mod, PM me and I'll send you the Eagle files.

[chi_boy]

For R6, my schematic, I thought 10k because I've seen that resistor between 22k and 100k on other delays.  I figured 10k to 110k would be a good range.  As long as the spot is there, changing values is easy if it doesn't work out.

Modulation seems to be the way to go too.  Sent a PM about the files.

Thanks again,
George