Couple of questions about the Aquaboy

[chromesphere]

Hey guys,
I finally got my aquaboy sounding decent, v3205's can burn in hell for all eternity.  The mn3008's I got are sounding much better.  I had 2 questions some of you guys might know the answer to.
1) Increasing the power to the aquaboy (to +15v) increases headroom.  How much headroom?  the mn3008's (2 of) will have a delay time of approx. 300ms depending on calibration, is the increase only a fraction of that?  Not looking for an exact figure of course, just like....1/4 ish...or is the increase proportional to the voltage? Which would indicate more like 2/3 increase approx. (9v -> 15v) so it would go from 300ms to 450-500ms if that was the case?
2)  The mn3008's are capable of some nice long delays when the clock is set into the audible range.  But the high pitch frequency is annoying.  The first question that comes to my mind is, is there a way of filtering out that whistling?  A very specific filter that would only remove say...4k up?  Or is the problem that doing so will alter your guitars treble response?  If there was a way of getting rid of that clock frequency, I reckon you could use mn3008's at a similar delay time as mn3005's (and how long for mn3005's?  1.2seconds?).
Just a couple of thoughts, thanks for your reading and response.
Paul

[LaceSensor]

Headroom doesn't change the clock time or available delay time, just the chance of overdriving the circuit and introducing noise on the repeats is less at 15v.


You can't filter out the clock noise like you say or everyone would be doing it surely...
The option there is double delay breakout pcbs and 4x 3008

[Scruffie]

Read the MN3008 datasheet, the headroom benefits (S/N & THD) are there.

Regards the second point, yes you can filter heavily, those transistors at the input and output of the BBD are low pass filters or sallen-key to be specific, they smooth out the samples from the BBD and filter out the clock noise.

You can increase the cut off but you will have muddier repeats as highs are lost and also noisier and poorer sampled ones as the delays increase.

If you use the buckets of water analogy for BBDs, one bucket of water pours in to the next in the chain passing the signal along over 2048 buckets, as you stretch the delay past what the BBD is specified for (which we already do, a 3005 is spec'd for about 204mS from memory and we try and pull 250-300mS out of them) then rather than a few drips of water spilling it's more like the buckets being thrown to the next bucket and a lot of water (signal) gets lost, so poor quality repeats.

The vintage MXR delay used filtering that increased with delay time to get the most out of the shorter chips and also give brighter shorter repeats, you could do something like that (although it'd be a whole new build) but you're still probably only going to get 200mS Max (at a guess) of useable delay out the chips... 300mS per chip could probably be done but it wont sound good.

Oh and regards filtering a specific frequency, that's all well and good but frequency changes with delay time, so you have to get the bandpass filter to match the clock frequency at all times and have weird frequency notches in your repeats.

As Lace said, double up on 3008s if you want more delay.

[chromesphere]

Thanks guys, and thanks for taking the time to explain it Scruffie, good analogy!  I might have to try powering it with 15v (and take all the precautions).  Probably worth while doing by the sound it.
I can see how people become obsessed with trying to squeeze as much delay time out of these as they can, its a great sounding 'warm' analogue.  I might try adding a subtle filter to output and...well...squeeze a bit more delay time out of it.  Might fail as well.  I guess you have to make a happy compromise and I agree, 4xmn3008's would make more sense.
Paul