Sujet transformer le CE300 en "Dimension DC2" pour synthé

INPUT LEVEL AND IMPEDANCE MODS:

The CE-300 originally may have been aimed at guitarists, and so there is a 1M input impedance. For synthesizers or general line-level use, R2 can be reduced to 10K for lower input impedance and less noise. To retain low frequency response, C1 should be increased to at least 4.7uF. (Thanks Margus for this info!) In my case, I used a 4.7uF Panasonic nonpolar electrolytic, along with the original .047 film cap bypassing it on the trace side of the PCB.

There's a fair amount of gain after the LEVEL control, and again for line-level use, this gain isn't necessary. Also the gain-setting resistors around IC1b don't need to be as high a value as they are, since input buffer IC1a can drive a fairly heavy load. Both R4 and R3 can be changed to 10K, which will give slightly lower noise and unity gain at full CW on the LEVEL control. With those values, a typical synth output should have no problem driving the level meter into the red even at 50% of the LEVEL control. If you want to keep the gain of the stock unit, the input noise still can be lowered by changing R4 to 10K and R3 to about 22K.


LFO SYMMETRY ADJUSTMENT:

Trimmer RT8 is for the purpose of matching the inverted and noninverted outputs of the LFO. This is important because a suitable "Dimension" effect requires fairly tight symmetry between the two BBD outputs. If you have an o-scope make sure the LFO outputs at TP8 and TP9 are as equal as possible before moving to the next mods. . . .


LFO SYMMETRY MODS:

Listening to the way the LFO modulated the clock and BBD on my CE-300, it seemed to me the sweep of the LFO was off-center in terms of rising and falling equally on either side of a "zero" point. The resistive dividers consisting of R44/R69 on the "A" side and R64/R65 on the "B" side serve the purpose of setting this "zero" point, as well as reducing the LFO level somewhat. For testing purposes, I tacked in a 1M pot in place of the "B" divider network and adjusted it for a more symmetrical-sounding LFO sweep, then measured the pot. The ratio was about 1-5, so I replaced R65 with a 200K and R64 with a 1M resistor. Because the symmetry was slightly different on the "A" side, there I replaced R44 with a 1M resistor and R69 with a 500K ten-turn trimmer and adjusted it to match the other side. (On my particular unit, after adjusting the trimmer by ear to match the sweep symmetry of the "B" side, the resistance measured about 235K.) This mod could also be done effectively by replacing both sets of resistor dividers with 1meg trimmers and adjusting to taste.

As a footnote here, the higher resistance of the dividers also seemed to reduce the PS fluctuations I noticed on the PS pins of the clock and BBD chips. Previously they were varying more than a volt in sync with the LFO, but now they were down to a couple hundred millivolts. Whether this has any effect on the overall sound is open to question, but it does seem like a strange design glitch.


DELAY TIME MODS:

The stock delay time is a bit on the short side for a chorus device in my opinion, somewhere in the range of 3 or 4 milliseconds on my unit. The clock caps C27 and C41 can be increased from 47pf to 100pf or even 220pf. I prefer the value of 220pf, which in my particular unit gave a clock frequency of about 33K, or about 15 ms. This gives plenty of range in the DEPTH control, everything from subtle to warped, and the longer delay also diminishes the flanger/comb-filter effect apparent in the CE-300, which I find inappropriate.

Polystyrene caps are recommended here for more accurate timing of the clock and hence smoother sampling in the BBD. To keep the delay times of the "A" and "B" sides equal, capacitors C27 and C41 should be matched as closely as possible, and then the frequency of the side "A" clock can be adjusted by RT5 so that both BBD delay times are identical.

A sidenote: any change in the 500K symmetry trimmer will slightly change the overall clock frequency of IC8. Therefore, a bit of recursive adjustment is necessary between the 500K trimmer and RT5, until the "A" side clock finally matches the "B" side clock.


OUTPUT LEVEL MOD:

It's easier to get a decent amount of input headroom if the output gain is increased a bit. With the all previous mods instituted, S/N for my CE-300 was measured at about 115dB with no effect, and about 100dB with effect. Substituting 200K metal films for R105/106 on IC13a/b boosted the output gain about 12dB over stock without adding too much extra noise. With those values in place, S/N with no effect was about 100dB, and with effect about 85dB. For my purposes, a worst case S/N of 85dB is nothing to worry about, and I think the improvement in ease of adjusting input level for max headroom while still having plenty of output gain is a worthwhile trade-off.


POWER SUPPLY DECOUPLING MODS:

Some extra PS decoupling is optional, but it never hurts, especially where spurious high frequency clock signals might dirty-up the PS lines. The CE-300 has a bare minimum of decoupling. The "digital" supply rail caps C12 and C22 can be increased to something like 470uF. On the trace side of the PCB, anything in the range of .1 to 10uF decoupling caps can be added to PS pins of all op amps, compander chips, BBDs, and clocks, as well as the BBD bias, Vref, and Cancel supplies. Some might call this overkill, but heavy PS decoupling always seems to makes op-amp circuits sound much smoother and cleaner to my ear.