Boss Dimension C chorus

[Processaurus]

I'm throwing together a poor man's version of the DC2, using two small clone pc boards from this guy that sells kits and good quality PCBs
http://www.buildyourownclone.com/chorus.html

I just made an inverted copy of one of the LFOs to drive the clock of the second chorus, and mixed the second signal back into the mixing section on the first chorus.  I didn't have anything remotely resembling heterodyning from the two clocks when I was testing it out, even with clip leads strewn all over.  Since the PCBs are made to fit in a hammond B size box, I'm sure two can fit in a BB size box.

Of course, its not the same as the DC-2 in terms of complexity, it doesn't have companding, or any pre emphasis/ de emphasis with EQing the modulated and dry signals differently.  It does have a similar sound however, in terms of chorusing without pitch wobble.  Also, the noise with the small clone frankenstein wasn't noticable to me, let alone an issue. 

A downside with this approach is that the parts were somewhat expensive ($15 apiece for the PCBs, $10 apiece for the BBDs, so for two boards its looking to be $50, to box it up, about $75).  You can get a used DC-2 for $150, so an exact clone might not be that rewarding.  A mutant super clone though, would be a different story.

[vanessa]

I used to have the pedal. I really loved it and needed to sell it to pay some bills. In mono mode it sounded to me very much like a Small Clone. But in stereo, wow this thing really shines. I thought I would just build one off a schematic someday. I wish I knew what I know now about the size.
The pedal ate up batteries like they were going out of style. I got it from a keyboardist friend who handed me a Boss AC adaptor and said "you're going to need this", he was right! It might be best to build one with a transformer or wall-wart only. Heck that could save you some more room shoe-horning it into a 1590BB.

[Steben]

Processaurus -> You forgot the BBD driver chips! So sum that up. Yet I think the cost bill doesn't have to be that high. You can use "democratic" BL3207's and BL3102's for the BBD lines ($7-$8 each couple= 15$ for two BBD lines). And your PCB doesn't have to be that expensive as well if you just could deliver some board design yourself. The etching itself can be cheap around the world. Don't forget the real cost is in the standard stuff: switch, enclosure,knobs,pots...) which are at least $20 for every effect.

Vanessa -> Yes a pedal with 2 companders and 2 BBD lines can be left or should be left without a possibility to battery feed it. True enough. Even a 9V clip is not only wasted space but also wasted money if you can use an adaptor. A couple-of-transistors Fuzz however is a true battery friend. It sucks easily 100 times less current than a modulation unit.

[MartyMart]

http://cgi.ebay.co.uk/Boss-Dimension-C-Chorus-DC-2-RARE_W0QQitemZ7368922542QQcategoryZ22669QQrdZ1QQcmdZViewItem

Here's one that's getting a "bit" expensive for a "beat up" version !!

Marty.

[Gripp]

Indeed.
I was bidding on that one (been hunting for a DC-2 here in Sweden for a while), but I'm not going to pay over 80£ (+ shipping) for one in that condition. This past 6 months I've seen two in Sweden going for 87 and 108£ respectively, and they were in far better shape. Too bad I wasn't fast enough for those two...
Best!
Pelle Garpebring

[Steben]

C'mon, a bit of courage and wit is all you need to get a nice PCB design. I admit it would take some time, but 80£(120â,¬) is way more than needed. 40-50£ (60-75â,¬) depending on parts at hand would suffice for DIY IMHO. With this price you could get the stock DC-2 PLUS nice features as there are through zero flanging, controls if wanted...

[Jaicen_solo]

Haha, yeah that's the one I was watching. I like my pedals to look a little beat up so I don't mind all that much!
I found the virtual DC plugin at work today, can't wait to try that thing out!

I don't think I have the necessary skills to make a PCB for this, I don't know enough about the chips really. I'd happily build one into a rackmount unit, that's not an issue for me, so yeah i'd like a DC+ version PCB please!  I have three MN3007's just waiting for a worthy project!
How close to a CE-2 is the actual chorus section of the DC?? It doesn't sound anything like a CE series pedal, it's much sweeter without the sick wobbliness.
Do you have any sound  samples of yours Processaurus???

[TheBigMan]

Yikes!  I paid about £50 for mine and it's pristine.  No box or manual, but even so I was happy with the price cos the DC-2 usually fetches more.

[Steben]

Haha, yeah that's the one I was watching. I like my pedals to look a little beat up so I don't mind all that much!
I found the virtual DC plugin at work today, can't wait to try that thing out!

I don't think I have the necessary skills to make a PCB for this, I don't know enough about the chips really. I'd happily build one into a rackmount unit, that's not an issue for me, so yeah i'd like a DC+ version PCB please!  I have three MN3007's just waiting for a worthy project!
How close to a CE-2 is the actual chorus section of the DC?? It doesn't sound anything like a CE series pedal, it's much sweeter without the sick wobbliness.
Do you have any sound  samples of yours Processaurus???

Well a Dim C IS in fact just two chorus lines in parallel with one single oscillator. The compander (NE570) trick is just added value. I guess the "wobble" comes from too extreme depth (and even rate) settings, high depth wobble is actually more flanging terrain. You can say the presets of the Dim C exclude any manual misconfiguration? 

[Mark Hammer]

Well a Dim C IS in fact just two chorus lines in parallel with one single oscillator. The compander (NE570) trick is just added value. I guess the "wobble" comes from too extreme depth (and even rate) settings, high depth wobble is actually more flanging terrain. You can say the presets of the Dim C exclude any manual misconfiguration?   

Yes and no.

The wobble comes from the pitch being either sharp or flat, relative to the clean signal.  In the DC-2, the pitch is always both, since the two BBDs are trading off which one is lagging behind and which one is catching up.  While there IS only one LFO, it drives the clocks in opposite directions to provide complementary actions.  The fact that boith flat and sharp deviations are present at the same time eliminates the feeling of directionality in the pitch change.

In some respects, it is probably more appropriate to think of the variable controls on a more conventional chorus as a necessary evil to compensate for its inadequacies.  The presets provide a couple of useful combinations and no mre is really needed...assuming your goal is to sound like a couple of musicians playing the same thing.

The companders are absolutely necessary.  Having two different clock frequencies operating, especially on a board crammed so tightly, introduces the risk of noise from heterodyning clocks, and there is also thefacts that if one BBD is noisy, two is twice as noisy.  You will note that one section of one of the 570 chips is sitting dormant, given that one section is used for compression in, and a second chip is used for two channels of expansion out.

Is high depth wobble "flanging terrain"? Not exactly.  True, bigger sweeps have an aspect of that to them, but generally speaking, pitch deviation needs a bit more delay time than flanging needs.  Doubling the delay time from 1-2msec will produce shifting notches, but doubling the delay time from 4-8msec will produce a more obvious pitch change.

Of course, you know what I'm wondering now?  What happens when you have complementary pitch change produced by two chains of allpass stages.

[A.S.P.]

Hanert explained it here:
http://www.pat2pdf.org/patents/pat2905040.pdf
http://www.pat2pdf.org/patents/pat2382413.pdf
well - artificial delay lines - but to the same effect:
pitch-shift by phase-shift.

[Dave_B]

While there IS only one LFO, it drives the clocks in opposite directions to provide complementary actions. 

Any idea what the Hz is on the LFO?  Just curious how fast the thing switches.

[Steben]

That's a 1/(2*Pi*R*C) thing I guess

[Dave_B]

Really?  ...shows how little I know about BBD clocks.   

[DiyFreaque]

Wish I could answer the question better, but I went back to my notes, and I'd recorded just about everything except the exact LFO sweep freq's and delay times of the four Dim C modes.  The best I could come up with was an email I sent to Antti Huovilainen last February (at the time I was working out external CV to the Dim C):

===
The minimum and maximum delay that the Dim C clock circuit will allow is as
follows:

7.12 ms (min) to 14.3  ms (max)

Obviously the original Dim C modulated the delay over a smaller range at
varying rates, depending on the modes.  The best I can offer are the
settings for a fairly deep, 'motionless' chorus.  That would be from 7.12 ms
to 10.24 ms, with the BBD's modulated antiphase with a triangle LFO of about
0.71 Hz.

Shorter delay ranges (ie from min to max) at higher LFO freq's sound just as
intense, but with different properties.  Longer delay ranges at slow LFO
ranges sound less intense, but still quite pleasing. 
===
Pretty high clock rates for a chorus device (though obviously other choruses share this trait).  IMO I believe it's yet another key among keys to the gargantuanly good sounding chorus of the Dimension C - higher clock rate = higher bandwidth.  Of course the dry bass response is boosted to eliminate the bass wiggle (of which Mark Hammer has pointed out is a factor in making a chorus less wobbly).  Companding is essential.  The equalized crossmixing of delayed signals is another biggie. 

Then there's the clock modulation itself- Mark Hammer again has quite well explained the effect of simultanously rising and falling pitch to produce the relatively motionless chorus.  In no way is half a Dimension C anywhere as good as a normal 'wobbly' chorus.  How do you make the Dimension C sound perfectly awful?  Take away one delay line.  The result is an absolutely terrible, sickly sounding pitch bend that sounds more like square wave modulation than triangle/sine modulation - the clock circuit responds to the LFO waveform in this manner.  The transformation of adding that second delay line in while listening to the Dimension C in this state is magical - it instantly changes from God-awful to heavenly.  One of the best examples of psychoacoustics there is.

I've said it before and I'll say it again - any of you who can shove this thing back into a Hammond box has earned my undying admiration.  For anyone wanting to rack mount or put it in a larger box, don't let that spare NE570 section lie fallow - it's begging to be turned into a second input channel compressor for stereo input (this was a stunt the Dimension D did that the C never did).  Those of you with stereo rigs might enjoy it - it's a bit of airy extra widener.  Of course, the true shining function of the Dimension C is mono to stereo conversion, so no big loss if you don't....

Cheerio,
Scott

Edit: Dry bass is boosted, not knocked down.

[bwanasonic]

Of course, you know what I'm wondering now?  What happens when you have complementary pitch change produced by two chains of allpass stages.

I think the classic *Metheny* chorus sound is two complimentary (static) pitch shifts of about ±15 cents, from two MXR DDL units. One of the main *chorus* sounds I use, is a blend of straight signal with +15 static pitch *detune* (Quadraverb). Very nice with a hint of delay and reverb. I would compare it to *button #1* on the Dim C.

Kerry M

[Steben]

7ms to 14ms is very "close range". It comes close to my believe a flanger should be deep range (>20:1) and a chorus "shallow" (2:1). Mark didn't agree partly, but for me it's clear. Is it "high" clocked? Perhaps. Anyway it means it doesn't matter whether you have 512 or 1024 stage BBD. That's for sure...

[DiyFreaque]

7ms to 14ms is very "close range". It comes close to my believe a flanger should be deep range (>20:1) and a chorus "shallow" (2:1).

That's true.  Beause the Dim C was not built as a flanger, any TZF benefits are simply opportunistic.  Rather than run the risk of unbalancing the primary function of the Dim C, I'd start fresh if I wanted a box that had the primary purpose of producing TZF.  In that case, I'd incorporate elements of the Dim C that 'worked'.

With that abbreviated sweep range, the longer flange delay effects are obviously not going to come into play.  What is available, though, that a standard 'non-TZF' flanger can't provide is a range of delay through absolute zero (7 mS to 0).  With a low sweep speed, that particular range adds a good degree of psychoacoustical tension. 

A 'normal' flanger takes a long run and stops just short of jumping off the cliff.

The Dim C, when tapped as a flanger, takes a shorter run and jumps right over.

A TZF flanger takes a long run and jumps over the cliff.

A Flanger Hoax takes a long run, spins in figure eights, dances on its hands, and jumps over the cliff.


If you replace the 1024 stage BBD with a 512 stage BBD, it will make a difference.  If you don't adjust the clock frequency range, your delay range will run from around 7 ms to around 3.5 ms- an even smaller ratio.  This would certainly jack with the dimension chorus effect.  Halving the clock frequency would alleviate that, but then you would have to limit your audio bandwidth accordingly, which would still jack with the chorus effect. 

Note these delays are not the actual Dim C operational delays when used with the internal LFO - it's the max range you can run the clock with the original, unchanged clock circuit when driving it with something other than the internal LFO.  The actual LFO delay range, depending on mode, falls somewhere within these two values.

Cheerio,
Scott

[puretube]

A Flanger Hoax takes a long run, spins in figure eights, dances on its hands, and jumps over the cliff

if you want it, it`ll jump right back up, again...

[SeanCostello]

Of course, you know what I'm wondering now?  What happens when you have complementary pitch change produced by two chains of allpass stages.

The problem is that allpass chains do not produce "pitch" change. They produce a fairly complicated FREQUENCY change.

"Pitch" change assumes that the amount of frequency shift is proportional to the frequency being shifted. Higher frequencies will be shifted by a greater number of Hz than lower frequencies. Delay line modulation results in true pitch change, such that for all frequencies, the result is originalFreq*modulatingFunction. The harmonic relationship of the input signal is maintained

As a comparison, a frequency shifter adds a constant offset to every frequency in the signal, such that for all frequencies, the result is originalFreq+offset. The harmonic relationship of the input signal is NOT maintained.

A standard phase shifter is not really close to either of these. In phase shifters, the amount of frequency change differs with frequency, but in a very nonlinear fashion. For cascades of large numbers of allpass filters with the same coefficient, the results can get close to a delay line, but you might as well use a delay line in this case. Generally speaking, a phase shifter will produce a vibrato sound that is similar to a delay line vibrato over a certain frequency range, but outside of that frequency range things start getting "out of tune" and the harmonic relationship of the input signal is NOT maintained.

I tried to create a digital ensemble effect years ago, that used the basic Solina ensemble configuration (2 3-phase LFOs, 3 delay lines), but with 3 parallel 6-stage phase shifters. It sounded really sour and bad. Part of this was undoubtedly due to the nonlinear relationship between coefficient and phase shift frequency, but even when I modified the phase shift network to have Hz as an input, it still sounded flat.

This is not to say that having antiphase swept phase shifters would sound bad at all. However, I don't know if you will get the cancellation of pitch change that you do with the Dimension C implemented with modulated delay lines.

BTW, once you work in the DSP domain, all of these issues with heterodyning clocks of modulated delay lines just go away. I have a chorus running on the SHARC that implements a simplified Dimension topology (antiphase swept delay lines, but no crossover filtering), and it sounds really nice. You can use a single delay buffer with 2 modulated taps, or use 2 separate buffers, depending on your intended applications.

Sean Costello