The Bobbler

[RingMad]

The Bobbler is a circuit I came up with (sortof) and breadboarded and had a lot of trouble with. I consider it kindof a failure, in that there are too many chips and strange things with, but I thought I'd post about it before I clear it off the breadboards, because it has produced some interesting sounds that I really really like, despite everything.

If you'd like to hear it right now before a bunch of verbiage, check out these 3 short sound snippets (024, 025 & 026) on my soundcloud starting from here cmosnippet-024.

I was reading about analog-to-digital converters, and was inspired by the "tracking" kind. Then I found this Tracking ADC page on allaboutcircuits and I saw an opportunity to produce sound if I split the output of the DAC part and fed a VCO. Reading about the "problem" known as "bit bobble", I thought it might produce nice semi-random variations in the sounds.

The circuit uses two 4029 cascaded up/down counters which feed an R/2R, which goes into an LM339 comparator as well as a 4046 VCO. The other side of the comparator is fed by an input signal that has been amplified. The comparator decides if the counters go up or down.

As usual, I had a load of trouble with the opamps (since I was using line-level sources as input, they had to be preamped to work with the comparator).

Another weird thing is that originally, I had used a 555 VCO, but it made the counters go crazy and not count properly. Problem solved by using a 4046 VCO instead. WTF?

I found that using a very fast clock on the counters and very high pitches on the VCO, some interested sounds resulted. It sounded even better when I added the triangle waveshaper circuit that Cynosure posted about in this thread: http://www.electro-music.com/forum/topic-52443.html .

After getting some help in the chat room, I fixed one of my opamp problems, but ironically, this made the circuit boring. i.e. the tracking action worked much more accurately, but the interesting sounds were gone! So I went back to whatever wonky way I was doing it before.

---

The_Bobbler_final_breadboard_RingMad.jpg
Description:	The Bobbler as represented by a very messy breadboard after many changes.
Filesize:	186.19 KB
Viewed:	213 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[JingleJoe]

The problem with the 555 timer may be that it draws too much current, the resistance of the CV input is less than 10k.

Sounds very good though! With some tweaking it may become even better!

[JovianPyx]

There is an additional problem with a (bipolar) 555 timer and it's also a current related issue and I believe it was the source of "counters going crazy" - that being the TTL totem pole output structure of the 555 timer. The output is made from 2 transistors, one from +rail to pin 3, the other from pin 3 to ground rail. When the 555 switches states, BOTH transistors are ON for an instant which causes a severe power supply current spike and can cause the power supply voltage to be severely lowered for just an instant - long enough to cause problems with things like counters. It should be noted that the 555 timer was designed as an industrial timer.

There are a couple of ways to fix this problem, the best IMO is to chuck the bipolar 555 timers into the trash bin and use 7555 which is a CMOS version and does not have the crowbar problem.

As for the CV input, first, it's not really a proper CV input in that it controls timer frequency (assuming astable) by changing the threshhold at which the timer discharges the capacitor. This has a side effect of also changing the output amplitude (assuming the output is taken by high impedance monitoring of the capacitor voltage). Since most industrial applications of this timer use the pin 3 output, this is not a problem for it's "intended" use. The 7555 suffers similarly. The other thing which JingleJoe points out is the 10 K impedance - this is because the timer is an industrial timer and as such is expected to be fairly immune to noise.

In fact, a 555 timer (or preferably a 7555) can be used as a VCO, even as an accurate one if the oscillator is designed to operate between ground and -V and a current source is used to charge the capacitor. Both linear and expo responses are possible this way. The PAiA Fatman VCOs are just this kind of VCO.

[JingleJoe]

[JovianPyx]

http://www.paia.com/fatman.asp is the Fatman page

http://www.paia.com/manuals/docs/9308-fatman-manual.pdf is the documentation, schematic is in there too.

I used 7555 instead of 555 to fix the soft synch problem (a known problem).

[RingMad]

Thanks for the eye-opening info on the ol' 555. Although, I never seemed to have trouble with the 555-based VCO in my still-on-the-breadboards Lunetta.

I put it in the circuit at first because it was the closest thing on my worktable and it was easy to build.I haven't gotten around to getting the CMOS 7555... I guess I should, when I finally get some serious work done on my suitcase-synth Lunetta. It'll use less power in general than the old 555.

James.

[JovianPyx]

This is an issue I learned about with the Fatman. The crowbarring would cause the VCOs to softsynch lock together when they were close together in pitch - like when you want them to do the phasing thing. It won't affect all circuits - but when I read "555" and "counters going crazy" it jumped out at me. The 7555 will not do that and I'd bet that your circuit would have worked just fine with a 7555 instead of the 555. I should also mention that the 7555 is pin compatible with the 555. Resistor and cap values will not need changes either.

[JingleJoe]

While we are on the subject of VCOs, you probably have enough electornic junk lying around to build this one, Ringmad (see attachments). Its high CV input impedance will stop it interfering with things .

To get a square wave from this VCO I put the saw wave through a comparator, you can get pulse width modulation then too I have also had best results when using a 1uF polyester capacitor, but any cap near that value will do.

Also you can use a 4093 chip instead of the 40106, or any other inverting schmitt trigger.

---

VCO.bmp
Description:		Download (listen)
Filename:	VCO.bmp
Filesize:	7.06 KB
Downloaded:	262 Time(s)

---

[JovianPyx]

Another technique for things like the 555 timer low input resistance for CV is to simply buffer it using a nice TL07x opamp. The buffer provides a high input impedance that we like and also provides a nice stiff voltage for the 555 or 7555 (or whatever low impendace circuit we are driving).

[JingleJoe]

I'd recommend an LM324 because unlike the TL07X series, it goes all the way to the negative rail (or ground).