Does this kind of counter exist?

[droffset]

Here's what the 4040 does, the ripple carry counter:



See how each output stage is basded on the last, and the pulse length extends?

But that's not what i really want at the moment.

I'd like each output stage to still be based on the clock, but keeping the same pulse length, so:

Q1: up/down on each clock downward change
Q2: up/down on every second downward stage
Q3: up/down on every third downward stage

So the rippled pulses stay the same size as the clock pulses but are more spaced out.

See what I mean? If you know of an IC that does this please let me know.

Thanks

[Blue Hell]

[droffset]

Thanks Blue Hell, I'll look into rate multipliers.

Just to be clear this is the desired behavior.


I was thinking about setting up a 4017 to do the divisions i needed but i love the way that 4040 gives me several outputs ready to go.

Another thing to explore is flip/flops, I understand that's what ripple counters are based on.

The end goal is a monster clock module, that does both ripple carry counting on one column and this sort of Offset Counting on the other column.

All based on a pwm timer.

[Inventor]

droffset,

Q0 and Q2 could be obtained by ANDing of the available signals from the 4040. Q0 is the AND of the 4040's Q0 output with the clock, and Q2 is the AND of the 4040's Q2 with that derived Q0. Q1 requires a divide-by-three counter and Q3 looks like you'd need a divide-by-five counter (plus logic off of the output bits as described.

I'd consider changing the requirement so that the non-binary cascading bit thingie is a binary cascading bit thingie, that way you could derive everything from ANDing of the 4040's output bits. See what I mean?

Les

[slacker]

I think one of Ken Stone's designs uses a counter and some NAND or XOR gates to do this. Can't remember which one it is at the minute though, sorry.

[slacker]

I was thinking of the master divider and unfortunately it doesn't do what I thought it did.
I tried what Les suggested and that gives you the 2,4,8,16... divisions but each one only stays high for 1 clock pulse. They're in sync with other though not offset.
What you could do is run the outputs through individual one shot generators then you could have whatever pulse length you wanted.

[Inventor]

Thanks for checking out what I said, slacker: teamwork in action!

I suppose to get really fancy, we'd need to look at it like this: Those pulses are 2, 3, 4, and 5 cycles long. 2 is a trivial case since 4 is 2x2, so that means we have 3x4x5 = 60 clock cycles before the pattern repeats. So we'd need a counter that counts to 60, then we could derive the desired clocks off of that using logic.

But anyways, I like the one-shot idea if you really really need those cascaded clocks.

[slacker]

[droffset]

Thanks guys, lots of ideas to try!

I'm not that fussed about a lot of the requirements, as long as something interesting comes out.

[amplex]

I don't know anything about this chip but I found this on Ryk's site and thought it might apply here! It looks to be a divide by N counter, might help for some of the odd divisions, not sure really..

http://www.classiccmp.org/rtellason/chipdata/hcf4522.pdf

[RF]

droffset -
I may be missing something, but maybe this is what you are looking for?
It's a portion snipped from Ken Stones Pulse Divider.
It provides divisions of 2,3,4,5,6,7,8.
Does good things with clocks and audio.

bruce

---

pulse_divider_snip_616.gif
Description:
Filesize:	12.86 KB
Viewed:	14144 Time(s)

---

[droffset]

amplex, I have a different divide by N counter, i think it's the 4018, might do something similar, will have to play with it.

Bruce, Thanks for that, i had something similar in mind with 4017, glad i won't have to reinvent the wheel.