Boss CE-1 Stereo in Stereo OUT ! HOW ?

[alexibm]

Hi !
I am in the process of assembling boss CE-1 clone that would fit into the rack, not the stombox.

Idea got into me, to make clone with Stereo in Stereo out, so I can use it with my Synth.

Any idea if it is possible at all ?
Thank you.

[Jaicen_solo]

There are two ways of achieving this as far as i'm aware.
Most common 'stereo' chrous units have the chorus and clean signals seperate, so it's not true stero but is a more pronounced effect with a bi-amp setup.
The second method is to have a second BBD section driven from the same LFO as the first, but inverted. That way the chorus effect is 180 degrees out of phase on the left and right channels.
I had planned to work on something similar using the tonepad CE-2 clone as a basis, but alas I haven't had time.
You could also check out the Geofex technology of the Univibe article. In that, RG suggests splitting the output of the phased section to a pair of outputs wired out of phase, perhaps you could try that.

[nelson]

you could use one bbd, however you could repeat the input section and output section, taking the delayed signal from the bbd and inverting the phase then mixing the dry signal from the repeated input stage from the second input with the inverted delayed signal.


Thats how I would do it. BBD's are costly.

[Jaicen_solo]

I agree, it's not quite the same, but it is simpler. BBD's aren't that expensive in the grand scheme of things.
It might be an idea to mosie over to muzique.com and check out Jack's Mos-boost article. There's an interestin technique for getting an inverted and a non-inverted signal from one Mosfet using a minimum of components. I'd hazard a guess at it being possible to fit a CE-2 stereo clone into a BB box, though it would probably need a stereo output jack.

[Plectrum]

One of my pet hates is the way manufacturers describe one paticular method as "stereo":
You have one effected channel, and one effect inverted (reverse phase) channel.
This can sound big on stage with two amps, but if you use this in the studio, the two channels must be hard panned L and R, and is intrinsicaly not mono compatible :- summing the channels results in no effect sound. And there are still (globally) many who use mono playback systems.

The TC pedal (TCF) is a great "clean" chorus/flanger/vibrato, but uses this pseudo stereo method.

For me, the inverted LFO with a second delay line is the most pleasing.

Grant.

[Jaicen_solo]

I agree completely. I really like the Dimension C & D chorus's because they just sound thick, without that pitch wobble.
However, they're not the easiest chorus to homebrew as you have to make sure that the bbd's are clocked in synch, or you're gonna get problems!  And as Nelson says, the extra BBD & clock chip effectively double the price of the components.

[alexibm]

Hi !

Thanks for replay. I guess I had the same Idea in Mind as some of you said

Input L -->  BBD
                  |
                BBD Clock 
                  |(inverted)
Input R-->  BBD
-----------------------------

Dimension D is certainly the way I would go,
but $ Dollar Dollar Bill Yo ! 

I keep great hope I can buy all of the Schematics/Layout/Parts......... and Build one for myself. It has bunch of rare BBD componets that are expensive, and hard to find (managible but still hard to get)

Main use would be for my nice Rhodes Mark I.

[d95err]

you have to make sure that the bbd's are clocked in synch, or you're gonna get problems!

What would be the problems if the BBDs are not clocked in synch? Couldn't you just use one LFO for two independent BBD/clock combinations?

I'm thinking about building a stereo chorus. My plan was to build two complete Small Clones. Then take the LFO signal from the first, invert it and lead to the other. However, if I need to synch the two clock chips it would be a lot more complex.

[DiyFreaque]

I'm thinking about building a stereo chorus. My plan was to build two complete Small Clones. Then take the LFO signal from the first, invert it and lead to the other.

That's how the Dimensions C and D do it.   C=MN3207.  D=MN3007.  I'm not sure I'd want to go there without companding, however.

It's not all that more expensive to add the second BBD and clock.  Small Bear has empowered you to make it a doable thing without giving up a limb or one's firstborn.  The payoff of doing it is huge.

[d95err]

That's how the Dimensions C and D do it.   C=MN3207.  D=MN3007.  I'm not sure I'd want to go there without companding, however.

That's kind of what I had in mind. "Small Dimension Clone" perhaps

I'm looking at the schematic for the Dimension C, and I'm having a really hard time figuring out how they are actually mixing the signals in mono/stereo. It's not as simple as just having dry+chorus1 + chorus2. There's a feedback path from the output of each channel which goes to the output mixer of the other channel. This is inverted and contains some of the dry signal, so some of the dry signal is cancelled out. It also seems like some of the dry signal is cancelled out already at the input stage when the effect is on. However, there's a lot of filtering going on, so it's difficult to tell how much real signal is going where... Need to brush up my theory on active filters, mixers and opamps...

[DiyFreaque]

Yes, the crossmixing network - it's to the Dim C what cheese is to a pizza.  Wouldn't be a Dim C without it!

I wrote up a long-winded treatise about it that I was going to put on my Dim C page, but never made it there.

VSAT/Hammer, anyone, please feel free to jump in with both feet for any corrections.  The info I'm most certain about is posted in red:

===========
To review a small bit, the chorus effect is produced when a signal is split down two paths - a 'dry' unaffected path and a path that introduces a constantly shifting short delay.  The two signals are then mixed together to produce the chorus effect.  The constantly changing delay produces two key ingredients - a slight pitch variation on the delayed signal, and also the reinforcement and cancellation of frequencies in the original signal when the two signals are mixed together.  Chorus differs from flanging primarily in the delay time used - the delay of a chorus is usually relatively much longer than the delay of a flanger.  Additionally, most flangers will use regeneration to enhance the signal, whereas this is rarely used in chorus devices.  It might be noted that not all chorus devices mix the delayed signal with the original signal.  Many ensemble chorus devices, which generally employ three delay lines, use the output of the delay lines exclusively to arrive at the ensemble chorus effect.

The excellent audio engineers of Roland/Boss must have really thrown themselves into the Dimensions C and D, because there are more than a few enhancements to this principle that result in a superior chorus effect.  The following is a reflection of my dim (no pun intended) understanding of what is really going on.


The Dry Signals

The dry signals really aren't all that 'dry' - they do undergo some processing by circuitry not shown on this schematic before they arrive at the section shown on this schematic. 

The input signals are pre-emphasized by IC1A - at this point, the emphasized dry signal is split into two portions - one portion is in phase with the original signal, and the other portion is 180 degrees out of phase with the original signal.  When the effect is bypassed, only the out-of-phase signal is passed on to the de-emphasis circuit - Q11 and Q12 block all of the other internal signals to IC12 when the effect is bypassed.  This allows an in-phase, 'uneffected' version of the dry signal to be output from the chorus output(s) - IC12 both de-emphasizes and inverts the out-of-phase signal back to the original phase. 

So - though the signal is 'bypassed', it really is still processed by a portion of the Dimension C circuitry - it is actually pre-emphasized and then de-emphasized, which surely affects the sound of the dry signal.  This is the reason many people prefer true bypass, so that rather than have the original signal processed during bypass, they'd rather route the signal completely around the circuitry in order to leave the original signal truly intact.

As mentioned before, one copy of the dry signal is pre-emphasized and passed in its original phase, and one is pre-emphasized and passed 180 degrees out of phase.  When the effect is active, these two signals are mixed together in some proportion and go to both the delay lines and the cross-mix de-emphasis portion of the circuit.  Mixing the two signals out-of-phase results in subtraction, and because the two signals have proportionate frequency content, I surmise that the result is an attenutation of the signal with the pre-emphasis response intact, 180 degrees out of phase with the original input signals.

The dry signals going to the crossmix/de-emphasis section are split into two paths.  One path is the 'bypass' path mentioned above - it is mixed straight into the de-emphasis IC.  The other path goes through a 'bass boost' subcircuit consisting of a low pass filter and transistor buffer (transistors Q2 and Q302 and associated components).  This 'bass boost' emphasizes bass frequencies that were reduced in the pre-emphasis process, and, from what I gather, provides a complementary signal to the mixing discussed later.  After the bass boost, it is mixed into the de-emphasis IC as well.

The Delayed Signals

When the effect is unbypassed, the same copy of the pre-emphasized, 180 degree out-of-phase signals are sent to their respective delay lines.  These signals pass through five different stages before being presented to the crossmix/de-emphasis section of the circuit. 

First, the dynamic range of the signals are compressed by the two sections of IC2.  Second, the signals are low pass filtered to eliminate any high frequency signals that can result in aliasing noise from the BBD's.  Third, the signals pass through the BBD's themselves in order to be delayed - with the delay time moving in opposite directions between the two BBD's. Fourth, they pass through another low pass filter, this time to eliminate the switching noise imparted by the BBD's.  And fifth, they pass through the two sections of IC11 to expand the dynamic range back to the original range of the input signals.  Going through the compression and expansion is, of course, called 'companding' and, along with the low pass filtering, results in the reduction of noise imparted by the BBD's.

The Cross-Mix De-Emphasis Circuits

So, we're back to where we started, the nexus of the Dimension C's sound quality.  Each section of IC12 acts as both a mixer and a de-emphasis circuit.  Each section mixes the 'dry' signals and the 'wet' signals for each channel.  But - there's an added twist.  Each section also mixes the high pass filtered output of the other section into itself - R81, R83 and C57 and R80, R84 and C56 form the high pass flters.  Remember that IC12 re-inverts the internal signals so that it is in phase with the original signal, so this high pass filtered signal is 180 degrees out of phase with the other signals applied to each mixer!

So, into each de-emphasis mixer we have four inputs (actually five when considering IC12B, but a bit more on that later).  The phases mentioned are in relation to the input signal:

1.  The pre-emphasized 180 degree out-of phase delayed signal.
2.  The pre-emphasized, 180 degree out-of-phase 'dry' signal.
3.  The bass boosted version of the pre-emphasized, 180 degree-out-of phase 'dry' signal.
4.  The high-pass-filtered, de-emphasized, noninverted mixed output of the opposite de-emphasis mixer.

So, what is happening here?  Admittedly, I can really only guess at some of what's going on.  Mixing the bass-boosted and non-bass boosted signals emphasizes the bass response of the signal when mixing with the delayed signal, which has a much more treble-weighted response.  I believe this makes a more harmonically balanced signal. *

Introducing the noninverted, de-emphasized and high passed output of the opposite mixer is really the same as *subtracting* this signal from the mix of signals 1,2 and 3 listed above.  Being high pass filtered, this will result in preserving the bass response, as less of the low frequencies are being subtracted.  And - recall that the delays are moving in opposite directions, so the frequencies being subtracted belong mostly to the domain of the opposite delay.  All I can positively say is that it all comes out sounding very, very good.

========

*Since this was written around a year ago, Mark Hammer has brought up the very good point that wobbly bass frequencies do not a good chorus make.  It's an unnatural sound.  I believe most likely the case is that the dry low bass frequencies are boosted so that they come through with less chorus effect than the higher frequencies.

The Dim C has so many layers to it, putting two chorus units together most likely won't get that precise sound; there's the actual 'ugly' detuned modulation that is made beautiful when mixed with its twin, the cross-mix network on the output, the treatment of the bass signal - all of it thrown into one wonderful unit.  I'm not saying two choruses put together would sound bad, but probably a bit different than a Dim C.

I've posted these before, but here's another one:

On this page:

http://mypeoplepc.com/members/scottnoanh/birthofasynth/id9.html

there are a couple of samples to illustrate what I'm talking about at the bottom of the page.

Dim C Stereo In/Out Pulsed Modulation (2 MB) is a sample of the Dim C being externally modulated by a pulse LFO.  The interesting thing is that the bass frequencies do not come across as being as radically modulated as the mid-register and treble frequencies.

Dim C One Hand Clapping (867 KB)  Plays a sequence unmodulated, then modulated by Mode 4, then modulated by Mode 4 - but with only one of the delay lines being heard, un-cross-modulated with the other delay line.  It's a glimpse of the actual delay line modulation; an ugly duckling waiting for its opposite twin to make it whole.

Cheers,
Scott

[d95err]

The info I'm most certain about is posted in red

Thanks a million! Exactly the info I needed (not just the part in red...) !