Flangers, delay time and modifying Boss flanger

[Kevin Mitchell]

The A/DA is turning into a rabbit hole of new tricks (for me at least). I've just learned about parallel multiplexing and my head's spinning

So I was wondering about the sample rate differences between running two 512 stage BBDs in parallel compared to running a 1024 stage BBD at twice the clock speed. Effectively they both give the same amount of samples within a given time - it's just that the 512 x2 in parallel can do this at half of the clock speed.

It took me a while to wrap my head around that 

So are these resolutions matched between PM 512x2 vs a 1024 at x2 clock speed?
Or is the 1024 at x2 clock better sampling?
Or does the PM configuration provide better resolution with *dual sampling*?
Uht my head's spinning again.

Drew this up trying to visualize it I believe they'd be matched for sample rates but I can't wrap my head around if it's the same deal or one is better than the other in terms of signal resolution.



Also - I have a potential source for some MN3010 BBDs

-KM

[Scruffie]

Yeah 512 parallel multi & 1024 doubled clock have the same sampling.

The differences come in at various areas, BBD's have individual clock frequency vs. gain responses, for a good flanging you want 50/50 wet/dry mix. Some BBD's will start to lose quite a lot of signal by the time they're reaching high clock frequencies (The Panasonic 512 stage BBD's are actually all quite poor in this area). The MN3007 is actually one of the better chips for this for general flangers but as we're having to use it at a very high clock frequency here, it is getting pushed to the very edge of its usable limits. So swings and roundabouts as to the choice in that regard.

Then there's high frequency content, by nature of having more stages, higher stage chips tend to shave off more high frequencies in general (not a hard rule, this was a developing area and improvements were made along the way) this isn't something you'd probably notice in a real life band situation or in a highly filtered chorus pedal but at home on a lightly filtered flanger, it may be apparent.

There's more to it but that's a good gist.

If you have a MN3010 source and do intend to get some boards, keep it to your self, or that source will be gone very quickly, like the last time there was an A/DA clone PCB that used them (actually, you could ask if someone still has one unused, was a while back but you never know!). But if you do have some made, I have a ton, so I'll grab one

[Mark Hammer]

One way of addressing potential signal loss when running a BBD at too high a clock frequency is to co-opt the trick used for through-zero.  That is, the "clean/dry" signal is delayed by a fixed amount that is largely imperceptible, like 0.5-1msec.  It is the relative delay between dry and time-modulated that produces the flanging effect.

[Kevin Mitchell]

Thanks for the info! So the 1024 stage is better because the signal is being spat out twice as fast - therefore less signal loss.

Is it worth it to parallel multiplex two 512 sections? Would there be a noticeable change if only running one section?

I'm actually drawing up a board in kicad. Already have the 3010 to 3210/3204x2 connections settled. I figured since MN3204s are at the least available to some people this would be an ideal project. And perhaps a coolaudio distributor could inquire on their v3204 devices (they make them but not listed). Such a board could home these 3 or 4 devices for a classic/more true A/DA circuit.

-KM

[Scruffie]

I never said the 1024 is better, no and as for signal loss it really, really depends and is nowhere near 'faster clock, less signal loss' infact the opposite. For the most part, the 3007 is very available and does a good if not great job, so we use it.

In a design that uses parallel multiplexing, trying to use just a single stage would be a mistake, it'll be very noisy (sampling theorem) the A/DA in particular has lax filtering.

You have remembered about the BBD source follower resistor regards the power/ground connections? Easy thing to forget.

I asked Steve @ smallbear about CoolAudio's 3204 a few years back, at least then, it was a no go.