symmetrical waveshaper from 4022

[trav]

Well, first I had a lot of fun with the circuits in the CMOS waveshaper thread: shift register + resistor ladder makes for many interesting wave forms. Then I played around with a 4029 and finally got the up/down count to ping pong, giving me a nice triangle out of the r/2r; but this required an extra flip-flop and AND gate, making three chips for a stepped triangle wave (perhaps someone has a better way to rig the 4029).

So I thought about it and came up with this. No extra logic, just a bunch of diodes :) You could use a 4017 if you wanted more steps, but I prefer the frequency division to be by a power of two. I just tweaked the resistor values until it sounded good. Not sure how close it gets to ground, but with the glide pot cranked and a slow clock into the 4022 I get some nice smooth sweeps.

protip: hang an LED off the 4022's carry out (pin 12) to indicate the output frequency.

[edit: removed useless pot and added extra resistor after more tweaking]

---

ws.png
Description:
Filesize:	8.92 KB
Viewed:	694 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Cynosure]

Cool. I like the 4029 ping-pong idea too.

Got some sound sample of this in action?

[RingMad]

Whoa! By "got the 4029 to ping pong" do you mean you got it to count up then back down again, over and over? If so, could you please please tell me how? (or a schematic). I tried to get that to work for a couple of weeks and finally gave up in frustration.

Thanks,
James.

[trav]

here's a sample, first clocked at a few different slow rates and fed into a VCO, with some twiddling of the glide pot, then I bring the clock up and listen directly to the output.

---

ws.mp3
Description:		Download (listen)
Filename:	ws.mp3
Filesize:	1.69 MB
Downloaded:	1095 Time(s)

---

[trav]

[trav]

Here we are. The carry out (pin 7) dips low at the end of the count (1111 when counting up, 0000 when counting down), so invert that and you have a pulse which can kick a flip-flop everytime you hit one end of the count or the other. The flip-flop then tells the 4029 to reverse direction.

Note that this only counts the endpoints once, ie. it counts from 0000 up to 1110 then from 1111 down to 0001, so there are only 30 clock cycles before the count repeats, not 32. This may or may not be what is desired. It may be possible to get it to repeat the endpoints by doing something with the carry in.

Now, I forked out $5 on a 4514 this week (24 pins!), the next job is to rig up a sequencer.

[edit: messed up my counts; put them in binary for clarity]

---

4029.png
Description:
Filesize:	17.29 KB
Viewed:	735 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[RingMad]

Thanks for the ping pong schematic! I did try something very similar, with a flip flop, but it seemed that the carry out pulse didn't always trigger it. In the datasheet, one can see it's a half-pulse, so I thought it was because of that.

I also tried with a 4040 running with the same clock, using the /32 output (I think) which would trigger the up/down pin. Something like that, but there was a problem wherein it would count up to 15 then dip to 0 then count down from 15, and for my purposes it was a deal-breaker.

Anyway, I shall try your circuit. Finally, I used a PICAXE microcontroller to do what I wanted (and counting 0-255... but I suppose one could maybe cascade two 4029s to get that range).

James.

[trav]

[Uncle Krunkus]

You can do a similar thing by putting all the outputs of a binary counter into their own XOR gate, except the MSB. The other input of all the exclusive XORs goes to the MSB. Send the outputs of the XORs to an R/2R and the wave rises as long as the MSB is 0. Then as the MSB goes high, the same count on the other bits is converted to a down count, and the wave falls. I did it with a 4 bit counter, and you get 8 steps going up, and 8 coming back down. That's only 2 chips. No flip-flop.

Sorry, that might have been 3*NAND gates, I'll have to dig it out and check.

[trav]

Uncle Krunkus: yes! That is a neat trick. XOR sounds right:

[richardc64]

[Uncle Krunkus]

Yeah, I actually liked the flat top and bottom, if you put it through an LPF, the flats take it closer to a sine wave anyway. Smoothing out the Lunetta sound is always a nice option to have if you can do it so cheaply.

[trav]

As for a visual indication of these stepped voltages I have been doing something like this with a non-inverting buffer and duplicate R/2R. Excessive, perhaps?

---

R2R_with_LED.png
Description:
Filesize:	11.74 KB
Viewed:	524 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[RingMad]

Thanks folks for the XOR + Counter thing... looks good... for my previous application, the flat top & bottom would have been desirable. Anyway, I don't want to derail this thread too much with the up-then-down counter stuff.

trav... maybe a VU meter is a simpler visual indicator of the voltage?
On the last box I built, I had a more abstract representation of an 8-bit number by running the bits to a 7-segment display (plus the dot).

James.

[trav]

breadboarded up the XOR version. How about through a VCA this time?

---

vca_up_down.mp3
Description:		Download (listen)
Filename:	vca_up_down.mp3
Filesize:	348.38 KB
Downloaded:	935 Time(s)

---

[trav]

I haven't tried it yet, but I think the circuit Uncle Krunkus described could be made a little bit more versatile with the addition of a simple ON-OFF-ON toggle switch

---

ramp_or_triangle.png
Description:
Filesize:	22.22 KB
Viewed:	393 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Richarius]