BBD Chorus Debugging

[Bo Deadly]

I found an old Rocktek Chorus pedal and decided to take a dive into it.

Apparently it's just a Boss CE-2 near as I can tell. Same chips, same LFO but with it's own virtual ground (I guess to isolate the trigger amp).

So I thought I would try to fix it up as best I can and play around with it. CE-2 sells for 100+ but this this is literally going for 20. Yeah, it might be a plastic box but it's basically the same circuit. It works fine. I hooked it up to the analyzer and noise performance is surprisingly good.

So I found a schematic (a little wrong / different - no BBD input bias trim - but a good start) and took some measurements:

As you can see, the battery is 8V, Vbbd is 7.4V and the LFO trigger amp out is a square from 0V to 6.2V which is about right for the LA6358. Then, the integrator is putting out a triangle from 1.2V - 4.5V and by the time it gets to the diode feeding OX1 of the BBD clock chip, it's a triangle from 0.7V to 3.3V.

Unfortunately I don't know anything about BBDs so I'm not sure if this is right. The CE-2 schematic claims the LFO virtual ground should be V+/2 but this is quite a bit lower at 3V which is only 40% Vbbd.

So my question is, is this working correctly and / or what can I do to improve things?

In particular, the LFO flops around a bit at the slowest speed (which isn't that slow - only around 1Hz). Is there a way to make this slower and smoother at slower speeds?

 

[Tubetec]

Sounds like the rate dial might need a squirt of cleaner ,  or even a few twists back and forth might  do it .
As far as mods go  you can replace 0.5 to a few uf electrolytics in the direct signal path with foils  if you want, 
Another one I did to my pair is increase the coupling cap into the BBD section , generally in the  'guitar' version I found it to be set a little on the thin side , I cant remember exact values but double or quadruple the input coupler to the bbd and you'll get more low end swirl  ,reference the values in the equivalent boss chorus 'bass' version  , its just gives a bit more of that 'slow leslie' bass rotor effect ,it can really add lots of dimensionality to an other wise simple signal , grab another Rocktek and fit both  into suitable a Eddystone or Hammond alloy box  , add your own home cooked LFO with phase+, phase - outs and some more interesting wave shapes like trapezoidal you'll have something unique yet instantly familar to anyone whos ever used a chorus before.

 

[Bo Deadly]

I don't want to hack it. I just want to know if it's working 100% correctly. I'm starting to think it is and it's just not the correct circuit for lower speeds.

I wonder what sort of range of CV the OX1 input on the MN3102 can use (datasheet says it can accept at least 1V to Vdd-1). As-is the CV is only 0.7 to 3.3V which seems pretty narrow. From playing around with it, it sense that as the rate decreases, you need to increase the depth. Meaning the rate of change needs to reach a certain point before you can really hear it. But because the CV range is only 2.6Vpp, it's just not enough for the slower rate. It's only when the triangle changes direction that I can hear it. But of course that just sounds like it's "flops around" and not sweeping. If the CV was more like 5Vp+ I wonder if it would work with slower rates.

Unfortunately with the circuit the way it is, it's not obvious at all how to get a higher CV range out of it. Even an op amp that can reach the positive rail would probably equate to an extra few hundred mV. Is there any way to get gain out of the LFO integrator without screwing things up?

If I had time to hack I would just make a PCB with a micro controller and ditch the LFO and clock entirely. Then I could explicitly control every aspect of the BBD in code.

I do like the sound of these old pedals though. I have a Lexicon MX200 DSP and it's chorus effect is very lame by comparison. It could be all of the filtering I suppose.

 

[JohnRoberts]

I don't want to hack it. I just want to know if it's working 100% correctly. I'm starting to think it is and it's just not the correct circuit for lower speeds.

I wonder what sort of range of CV the OX1 input on the MN3102 can use (datasheet says it can accept at least 1V to Vdd-1). As-is the CV is only 0.7 to 3.3V which seems pretty narrow. From playing around with it, it sense that as the rate decreases, you need to increase the depth. Meaning the rate of change needs to reach a certain point before you can really hear it. But because the CV range is only 2.6Vpp, it's just not enough for the slower rate. It's only when the triangle changes direction that I can hear it. But of course that just sounds like it's "flops around" and not sweeping. If the CV was more like 5Vp+ I wonder if it would work with slower rates.

Unfortunately with the circuit the way it is, it's not obvious at all how to get a higher CV range out of it. Even an op amp that can reach the positive rail would probably equate to an extra few hundred mV. Is there any way to get gain out of the LFO integrator without screwing things up?

If I had time to hack I would just make a PCB with a micro controller and ditch the LFO and clock entirely. Then I could explicitly control every aspect of the BBD in code.

I do like the sound of these old pedals though. I have a Lexicon MX200 DSP and it's chorus effect is very lame by comparison. It could be all of the filtering I suppose.

I was working with BBDs back in the 70s when we had to make the glue circuitry  ourselves.

If you are going to throw a micro at just do it all in the digital domain... eliminates a lot of tweaking.

JR

 

[Bo Deadly]

If you are going to throw a micro at just do it all in the digital domain... eliminates a lot of tweaking.

Using a micro controller and 100 lines of code to drive BBDs is quite a bit easier than doing a full ADC -> memory -> DAC delay line.

Plus I think the BBD sounds different from my generic multi-effect box. I don't know why but it does. But I think it sounds better. It's clearer. Maybe it's harmonics. I ran the Rocktek through the analyzer and there were some serious low-order harmonics:

Rocktek Chorus Spectrum

green: 800Hz -20dBV
yellow: 1kHz -10dBV

Tone: 12 o'clock
Depth: 3 o'clock
Rate: 9 o'clock

 

[Tubetec]

What you describe about the depth at low speeds is also something i've noticed on the boss versions .

I agree that these humble units have their sound , its just more involving on a psychoacoustic level than your average digital equivalent.

 

[Bo Deadly]

I updated the plot in my previous post because it was incorrect. The rate control was all the way down (the knob was on wrong). So I ran the test again with it at the 9 o'clock position (slow but visibly modulating during capture).

While I was at it, I also did a plot at -20dBV to see how the harmonics persist at different levels (meaning I'm not just overdriving it). Clearly the 2nd harmonic does [1].

Note that the distortion reading is non-sense. That value jumps around wildly during capture. Not sure about the SNR value either. I think the modulation could just be throwing off the analyzer. But from looking at the spectrum as a whole, clearly it is a fairly quiet device. For a $20 guitar pedal it's really kind of impressive. Go Rocktek!

[1] This could easily be because the BBD input is not biased within the correct window. There is no distortion trim.