BBD differences between chorus and flanger

[Dimitree]

hi everyone
while looking at the block diagram from the Moog Moogerfooger ClusterFlux manual, I noticed that it uses 2 BBD, one "Flange BBD" and one "Chorus BBD".
I read online that it uses MN3007. There's a switch on the pedal that let you choose between Flanger and Chorus.

I presume that one BBD is clocked faster than the other.
If so, why to use two identical BBDs with its own clock and not simply use a single BBD and switch the clocks?

[StephenGiles]

Probably because of our old friend "heterodyning" (very nasty noise) which often occurs where more than one clock is used with BBDs (Puretube where are you!) However it's difficult to speculate why they used this particular design without seeing the circuit diagram.

[GGBB]

The MN3007 cannot do the stated delay time range (0.6-10ms) of the flanger, but it does cover the chorus (5-50ms). The ranges correspond nicely to having an MN3007 for the chorus and an MN3009 for the flanger with a single clock for both.

[anotherjim]

Interesting approach - have just seen the manual diagram.
The 2 BBD's are in series. Range switching is either tapping off the first delay only (flange), or the second (chorus). Clock is common to both so no Het' trouble. The second BBD never stops, but it's only heard in Chorus mode.
Flanging needs short delay range but Chorus needs longer.

[armdnrdy]

This is from an early introduction of the MF-108M.

This will be a limited availability product. The MF-108M Cluster Flux uses BBD (Bucket Brigade Device) chips that were specifically designed for the short delay times associated with chorus / flanger type effects. But like all BBD devices (e.g. the MF-104 Analog Delay) the availability of this chip is very limited. We are constantly looking for more supply but as with the MF-104, the supply will be limited by chip availability.


Under "features" this is noted:

The design includes high voltage MN3007 BBD's for their superior headroom and excellent signal to noise ratio.


I wouldn't put MN3007s in the limited availability category...MN3007s seem to be the most mass produced of the lot.
My vote goes for a 3007 for the flanger and a different BBD for the chorus.

[GGBB]

According to the datasheet, the shortest delay possible with the MN3007 is 5.12ms (http://www.experimentalistsanonymous.com/diy/Datasheets/MN3007.pdf). Although that's in flange range, it's far longer than the 0.6ms range minimum stated by the manual. Unless there's some way of cheating, the BBD for the flanger has to be something other than an MN3007. Of course, that contradicts the "early intro" text, so that's confusing.

[Dimitree]

If Flanger and Chorus mode have its own dedicated and different BBD, then why in Chorus mode does it use the flanger BBD also? Is that small amount of added delay worth the noise of two BBDs in series? Why not simply use the two BBDs in parallel?

[armdnrdy]

I came across this on a synth site:

actually uses a combination of MN3007 and MN3009 BBDs, for better performance across the whole flange/chorus time range

You aren't going to "add" any noticeable noise by adding a 3009's 256 stages to a 3007's 1024 stages.

When you add BBDs like 3005's in series, (4096 stages) that's when the noise gets noticeable.

[Mark Hammer]

While MN3207 clones are indeed widely available, MN3007 chips have not been produced since the mid-90s.  There is probably "enough" to satisfy hobbyists looking for one or two here and there.  That's different than having a large enough assured supply that it justifies sinking money into producing runs of boards and chassis, and preparing all the promotional literature, advertising, etc.

[armdnrdy]

In my extensive searches for BBDs, I have no problem finding MN3007s...in quantity...and with a price as low as a few dollars a piece.

In fact...they are so common that I haven't bothered stockpiling them.

The problem component is the MN3009. I don't think that this model was produced in near the numbers as the 3007.

I've seen small quantities available in the European market, and a few sellers offering them on Ebay.

Like your buddy from Retro-sonic that sourced enough MN3002s to make two different versions of the CE-1....if you look hard enough...you can find it.

[Dimitree]

You aren't going to "add" any noticeable noise by adding a 3009's 256 stages to a 3007's 1024 stages.

When you add BBDs like 3005's in series, (4096 stages) that's when the noise gets noticeable.

Ok, but what is the advantage of using them in series instead of switching between them in parallel? Is it more pratical (reduced part count, for example, or other reasons) or is it more sonically desiderable?

[armdnrdy]

I believe that Jim answered your question in reply #3.

The 3007 is used for the flanging...and the 3009's additional 256 stages are added in series for a longer delay time for the chorus.

[Dimitree]

Most chorus fxs uses a single MN3007, additional 256 stages would make a difference? Longer is better in a chorus?

[robthequiet]

Yes, chorus is by definition longer than flanger. As I understand, you basically have a 256 + 1024 = 1280 stage BBD with a tap at the 256.

[Mark Hammer]

When it comes to short delays, there are a few "tipping points", mostly based on what draws our attention.

When the delay can get VERY short  (well below 1msec), what we notice most is how the signal becomes "infected" with notches as the delay time gets longer and the notches move from being only way up high to being distributed throughout the spectrum. 

As the delay time gets longer, and results in many notches at all times, we tend to notice the comb filtering and modulation of tone.

As the delay time gets longer still, the movement of notches fades to the background and our attention is captured by the perceptible asynchrony of the two signals (wet and dry) and especially by their pitch difference.

All of that happens within a surprisingly narrow range of delay times.  Indeed, clearly audible differences between various commercial chorus pedals is often on the order of 2 or 3 msec, and differences between flangers is often less than that.  A chorus that sweeps between 2-6msec will sound more like a slow Leslie than one which sweeps from 4-12msec.  So, designing a flanger around an MN3007 and then extending its delay range by 25% with an MN3009, can be expected to yield chorus effects.

[anotherjim]

My instinct would be to use the 256 stage for Flange. Many chorus have 1024 stages, I doubt the extra 256 makes much difference to chorus - maybe it was easier to do the changeover switching by leaving the Flanging element in circuit?
You would have to over-drive the clock spec to get Flange from a 1024 stage?

[GGBB]

The 3007 is used for the flanging...and the 3009's additional 256 stages are added in series for a longer delay time for the chorus.

Based on the manual's stated delay time ranges, it would have to be the other way around. The MN3009 used for the flange and extended by the MN3007 for the chorus.

[Mark Hammer]

That would be my working assumption as well.  As numerous demonstration have shown us over the past decade, an MN3007 can be coaxed with additional buffers into being clocked up to 1mhz for ultra-short delays and dramatic flanging.  But tacking on a mere 256 stages on top of that, using the same clock, wouldn't add much more delay time.  Far easier to use a basic clock for the both and simply thicken up the sound by adding 1024 stages to an initial 256.

[ElectricDruid]

You would have to over-drive the clock spec to get Flange from a 1024 stage?

Yes, you would. I can't speak directly about the MN3007 or 3009, but I've got experience with the low-voltage versions. My recent Flangelicious design can use either MN3207 or MN3209, but both are over-clocked. The datasheet specification is based on driving them with the MN3102 clock chip, which doesn't provide much power. Both BBDs will go much faster is driven by a more powerful clock. Using multiple CMOS buffers in parallel to increase the clock drive is a common trick you see in old flanger schematics. The datasheet claims 200KHz as the maximum clock, but they both run easily to 500KHz or more (signal level drops off above 500KHz, but it runs fine). I would expect the MN3007/3009 to be similar in this respect.

Tom

[thermionix]

Apologies in advance for derailing the thread a bit (I'm bad about that).

I have an MN3007 in my parts drawer, no idea when or why I bought it.  Thinking maybe I should build something with it.  I know virtually NOTHING about BBDs and related circuitry.  What are some good, relatively simple pedals that could be built around a single MN3007?  I don't care for chorus, too 80's sounding for my taste.  What about a short slap-back type delay (echo) like would be used for chicken pickin' type stuff?  Is one MN3007 sufficient for such a thing?

Any other ideas?