can cmos divide to musical 3rds, 4ths, 2nds,etc???

[loss1234]

i have seen a way to do 5ths. and octaves are easy. but can CMOS counters (or any other cmos chip) divide to things other than a 5th or an octave? like a 3rd?? or a 7th?

if a top octave chip can do it, it cant be too hard (since those chips were made long ago)

i assume there is a way using discrete gates?

anyway, i dont need all 12 notes

just a couple

thanks

[Inventor]

loss1234, there you go asking imaginative questions again, will your curiosity never cease? haha. You can make pretty much any kind of counter you want if you use a state machine to do it. Actually a counter is a specific type of state machine if I recall correctly. I'd suggest that you read up on state machines and state machine synthesis. I learned it in college and used it in my career, but I've forgotten how to do it now!

That's about all the help I can provide since I have no music training and do not understand what you mean by a 7th or whatever. Could you explain what you meant by that? Now you are the one teaching me, it works both ways, eh?

[Uncle Krunkus]

Check out this page here,
http://www.electricdruid.com/forums.html?page=info.dds
It explains one way of doing it. Covers the kind of things which matter anyway.

BTW What is a State Machine? (Apart from being a Bill Manganaro that is! )

[jnuaury]

if youd like to explore just intonation you are in luck!
but the conventional tunings we learn arent as easy

i happen to enjoy the sounds of just intonation (which is tuning to whole ratios)

some intervals are will sound very close and some may sound "out of tune"

anyways lets say you have the note E and you divide it by 5 you will get a C that is a 2 octaves and a major 3rd below... if you take the 4th and 5th divisions of your clocking oscillator they will be a major 3rd


so i find it easiest to have a pulse divider set up like ken stones where you have one input dividing into into numerous outputs and trying different combinations of the outputs

an easy way to do this would be to take a few 4017s sharing the same clock and listening to the zeroth output of each counter will changing the which pulse resets the counter

[Inventor]

[loss1234]

krunkus-can you repost that link?

jnuary-i checked out the ken stone circuit. i dont quite get how chaining the 4017's further divides the signal. are they chained somehow other than just sharing the same clock??

very interesting.

from what i heard the problem is that musical intervals like the 3rd of the scale or the 7th are very weird divisions like 451 or something.

thanks a lot

[blue hell]

read the title as "can mosc divide to ..."

The link is working here, maybe it was fixed.

[fluxmonkey]

ken stone's pulse divider: http://www.cgs.synth.net/modules/cgs36_pulse_divider.html

[loss1234]

the DDS link works now but i dont think thats quite what i was hoping for.

i think the pulse divider could be the ticket, or close.

i am surprised that there isnt a counter available that is more programmable.

thanks

[loss1234]

see the pitch makers thread...turns out there is a good answer there that relates to the 4017 question i asked yesterday

[loss1234]

Darn
if you look in the chapter on COunters in the CMOS COOKBOOK

page 312 in my edition, there is a diagram of how to use a HUGHES 0320 with three BCD switches to get ANY division between 3 and 999

too bad the chip is non-existent. no on has even heard of it.

are there other more modern FREQUENCY synthesizer chips??

inventor:
a 7th or a 3rd is a musical interval. ( i might butcher this as i am a self taught musician)

if you look at the semitones in a scale, c c# d D# e f F# g G# A A# B C) the scale for C would be CDEFGABC

there are some that are sharps and some that are flats. (there is no sharp between E/F or B/C)


so the E is the third, and the G is the fifth.

(a major C chord)

for a c minor, you would flat the 3rd,( which would be E flat-sharps and flats are the same btw)


each scale has certain semitones that it uses and some that you can use but are outside the scale. if you start with your first note, an octave would be the 8th. which is 12 semitones (i think)

you can also count intervals by semitones

one semitone is a minor second. two is a second, three is a minor third, four is a 3rd,etc etc

basically, you just count up a certain amount of semitones to get each interval.


do a search for the circle of fifths, or western scales to get more info.


thanks

[Uncle Krunkus]

The Electric Druid link was mainly about getting you to read this:
"The ratio of the frequency of one note to the next note up is the twelfth root of two, or 1.05946." Which shows that the divisions required to map out the notes of a chromatic scale are quite specific, and would be hard to obtain using CMOS division.
But, as that same page explains, it's not as hard if you up everything by 16bits first, (though still not exact). Basically, you can't divide a clock stream by anything but integers. Therefore, to get a ratio like 1:1.05946 (or even close) you would need to multiply by 65536 (2^16) and then divide back down by 69433. As far as Lunetta oscillators go, I think it would be easier to just start 16 octaves up.
But,..... I'm sure that, for some of the intervals, you could get a pretty good approximation by only multiplying up by say 4 to 8 octaves, which is fairly easily done with the 4046 and 4017 (possibly cascaded) in the feedback loop.

"jnuary-i checked out the ken stone circuit. i dont quite get how chaining the 4017's further divides the signal. are they chained somehow other than just sharing the same clock??"
loss1234, the 4017 in the feedback loop of the 4046PLL divides down what is coming out of it, and then uses that as it's reference. This "tricks" the 4046 into locking onto a frequency which is multiplied up by however much the 4017 divided it down. Does that make sense?
To cascade 4017s (so you can get 1 pulse out for say every 15 that go in), requires more than just sharing the same clock, you need to add the AND gate like in the example at the bottom of their datasheet. Basically the second 4017s clock comes through the AND when the first gets to it's 10th step. This "state" also inhibits the first 4017. When the second (or third, or fourth) 4017 gets to the "last step" this output resets all the 4017s and the count starts again on the first.
Remember though that you can realise the same division by using a binary counter and a few OR gates, because in this situation we're not that interested in using all the intermediate steps. (but we could be! )
Whether you could actually use this method to make a 4046 multiply the incoming clock up by 16 octaves would be an interesting thing to try. It may have trouble locking on at such high frequencies. (for a 1khz input, 16 octaves up would be about 65Mhz!) No I don't think the 4046 will cut that kind of mustard.
The question is, are there any other intervals apart from the 5th which are relatively simple ratios? Or close enough to be approximated?

[blue hell]

All simple ratios like 1/2, 3/4, 4/3, 4/7 etc. etc. are very useful musically and more easily so than the 'standard piano notes', they are easier to obtain as well.

The idea of having a multiplier followed by a divider is a great one! Don't go over-complicate it by trying to make a 12 tone equally tempered scale out of it, it's a lot of work and the result will just be less good than simple ratios.

Remember that the 12 TET scale is a compromise really that was derived from the simple ratios just for making the piano a manageable thing to build.

[loss1234]

cool

i think this is plenty to get me started. you are right, i dont need the 12 tone scale, just musically useful intervals other than octaves.
i will let you all know how tomorrow goes with the breadboard


thanks

[loss1234]

blue hell:

("ratios like 1/2, 3/4, 4/3, 4/7 etc. etc. are very useful musically")

Could you possible explain which DIVISIONS on Something like the 4017, these equate to? is it simply that the 2 output= 1/2? and the 3 output equals 3/4? or is it more complicated. I suppose one of the problems i have is that when we start talking about divisions that arent exact intervals but also arent octaves, i dont undertsand how to translate the pin labels to fractions.

but i will start with my ears.

thanks

[jnuaury]

if you use one 4017 and compare it to the clock you get a ratio of 1/x
if you divide it by two its 1/2 by three its 1/3 etc

thats why having multiple divisions of the same clock is useful
you could do this with by feeding multiple 4017s and then using a rotary switch to select the division of each chip or whatever....

so you have your main frequency you divide it by 4 on one chip and 5 on another to get a ratio of 4/5

[loss1234]

krunkus...no i got how the 4046 circuit worked. it was this circuit which confused me:

http://www.cgs.synth.net/modules/pic/schem_cgs36_pulse_divider_r1-1.gif

unless i missed something, he DOESNT seem to be using the cascade system that they highlight in the datasheet. from what I Can see (and i am not too good at understanding this stuff) it just looks like they all share a common clock but are not actually cascaded. so how is he getting different divisions from the same pins?

his Q0 seems to have divide by 7 on the first chip, on the second its a divide by 8 and a divide by 2 (which maybe he is doing with those diodes?) and on the third and 4th its by 6 and 5.

can anyone explain this?

thanks

sorry for being a pest

[loss1234]

thanks jnuaury

that makes sense

well my wife is finally gone for the day so i AM going to test this stuff right now...

will report later

[loss1234]

actually, hold on

this last line confused me

"so you have your main frequency you divide it by 4 on one chip and 5 on another to get a ratio of 4/5"

you mean you logically AND the two divisions together to get the 4/5?

i understood you up till that line.

sorry for being so slow

[blue hell]

When you do nothing all of the outputs of the 4017 will simply divide by 10.

The idea is to take one of the '0' to '9' outputs and connect that one to the reset input which will make the device divide by that number + 1 .. so connect the 0 out to reset and it will divide by 1 ... output 9 back to reset will make it divide by 10. Audio output can be taken from the '0' output, as that one will pulse always. ... hmm I might be one off with the division ratios, not sure if reset is immediate or clocked.

Anyway, it will not on it's own make fractions like 3/4 as for that you need a division by 4 and a multiplication by 3, so you need a rate multiplier as well. A rate multiplier can be made with the PLL thingie mentioned before.

[jnuaury]

no logical AND
you have two signals one is 1/4 the input's frequency the other is 1/5
when you listen through both of these (through mixer or wherever)
you will have an interval with a frequency ratio of 4 to 5

[loss1234]

ok so its as easy as a mixer. wow.

thanks...now i will read blue hells reply

this is going to be a good saturday.

[blue hell]

Sorry, the PLL rate muliplier is not in this thread but in the other

[Uncle Krunkus]

[jnuaury]