Memory man with slight "fizzle" on long delay times

[nirvanas silence]

I have a Memory Man with a MN3005 IC and the longer delay times have a "fizzle" form of distortion on the repeats.  I have a schematic just wondering if there is a well known fix, or if anyone has biasing info for the IC.

[aron]

It's an analog delay, I wonder if it's just the normal "distortion" on the delays???

[nirvanas silence]

Well is there anyway to lower the clock noise?  I can hear the whine at anything over 50% delay time and as soon as it becomes audible it adds the interference/distortion.  Thanks!

[Tim Escobedo]

It's analog mojo. People pay big money over digital for that extra noise.

[nirvanas silence]

I'm not about to write this off as mojo since I've never heard this on any other properly calibrated analog delays.

[aron]

Actually I have heard it on many analog delays when they are capable of longer delays. The shorter the delay the less noise usually. Most of my analog delays are very noisy.

[squidsquad]

I was able to lessen that noise by carefull adjustment of one of the trimmers inside.
Mine were older models....and the position (which trimmer) changed over the years...so proceed with caution.

[markphaser]

The biasing Trimmers need to be calibrated to the right voltages from the data sheets

[Mark Hammer]

Said it before and I'll say it again.  BBDs have lots of teeny-tiny capacitors and FETs onboard the chip.  Those caps leak and have limited capacity to hold a charge for a long time.  As a result, the longer you oblige them to hold the charge (by slowing the clock rate), and the more stages you force the signal to pass through at that clock rate (and keep in mind that 3 repeats beyond the original first one is tantamount to sending the signal through 4 x 4096 stages, each leaking a little), the less the output starts to resemble the input.

If a design aims for the least such sample/error, then it will usually either have less than the max delay possible (i.e., clock doesn't go as slowly) or use a lower filter cutoff frequency and steeper filtering, or both.  Some designs in the 70's and 80's went the other way, and used less than optimal filtering so that you could get bright short delays, and "fix" the longest delays with the treble control on your amp.  The MXR Analog Delay (and its Ross successor) used a clever approach in which the filter frequency was set by the master clock used to drive the BBDs.  So, as you slowed the clock down to produce longer delays, you also lowerd the cutoff frequency for a muter sound.  Whipsmart, as they say.

My recommendation is to simply install some sort of shallow treble cut in the regeneration path.  This leaves the first echo with as much bandwidth as the pedal allows but shaves off a bit more on each repeat.  Not only does this produce much more natural-sounding echoes (in the real world, reflections lose high-frequency energy faster than they lose low-frequency energy), but it also has the strategic advantage of taking out more of the fizzle as it accumulates.

[aron]

I read that post where you described this mod. Sounds cool Mark!