Author |
Message |
françois
Joined: Dec 23, 2006 Posts: 139 Location: Paris (France)
Audio files: 1
|
Posted: Tue Jan 01, 2008 4:25 am Post subject:
Subharmonic generator revisited Subject description: Request for suggestions |
|
|
Happy New Year to all ! Including these who do not use the Gregorian Calendar
I have been SPICE-experimenting with my idea of a sub- and super-harmonic generator for sawtooth waves. It seems to work quite well for ×2, ×3 and ×5 waves, as well as for ÷2 and ÷3. But I am a bit stuck for the ÷5 wave.
I have as input a rising sawtooth, ±5V peak to peak, which a simple comparator will easily turn into a square wave (or a pulse wave) that can be used as a clock signal. What I need is a pulse wave one third or one fifth the frequency of the original. One third is easy : two D-type flip-flops and a NOR gate (divide-by-three counter). But one fifth is trickier.
I need five outputs : the 1st one active during the first period, the 2nd during the 2nd period, and so on. Or... the 1st during the 1st & 2nd periods, the 2nd during only the 2nd period. Or... the 1st during only the 1st period, the 2nd during the 1st & 2nd periods, the 3rd during the first 3 periods... I do realize a shift register will do the job, but I'd like a less straightforward design (why make things simple when you can have them intricate ?). I mean, I do not want to have many multiplexers or switches on the board since they produce awful glitches. The less transitions there will be, the better.
If someone has ever designed a good divide-by-5 frequency converter for square waves, I would greatly appreciate suggestions. BTW, googling for that does not produce significant results beyond the classical divide-by-N counters.
Thanks in advance,
-- françois |
|
Back to top
|
|
|
Pehr
Joined: Aug 14, 2005 Posts: 1307 Location: Björkvik, Sweden
Audio files: 2
|
|
Back to top
|
|
|
guitarfool
Joined: Feb 26, 2007 Posts: 160 Location: Maryland
Audio files: 8
|
Posted: Tue Jan 01, 2008 5:20 am Post subject:
|
|
|
Pehr wrote: | The CD4017 could be used to divide by 5 I think. With the carry out signal (ouput 6 connected to reset). |
Indeed. Take a look at the pre-scaler of Ken Stone's CGS-01 Sub-Oscillator: http://www.cgs.synth.net/ (go to "PCBs for sale" and scroll down to the sub-oscillator) |
|
Back to top
|
|
|
françois
Joined: Dec 23, 2006 Posts: 139 Location: Paris (France)
Audio files: 1
|
|
Back to top
|
|
|
Pehr
Joined: Aug 14, 2005 Posts: 1307 Location: Björkvik, Sweden
Audio files: 2
|
|
Back to top
|
|
|
JovianPyx
Joined: Nov 20, 2007 Posts: 1988 Location: West Red Spot, Jupiter
Audio files: 224
|
Posted: Tue Jan 01, 2008 11:54 am Post subject:
|
|
|
As far as "superharmonics" go, I did this 4046 PLL frequency multiplier device for my FatMan. It produces seven outputs at 1x 1.5x 2x 3x 4x 6x and 12x the frequency fed into it. Yes, a PLL will have some delay, this is controllable with the loop filter I used and I've found it not to be objectionable - in fact, I've found it useful.
The divide by 3 circuit is glitch free and I designed it for providing a 50% duty cycle output. I used divide by 3 because what I wanted was a perfect fifth generator. This circuit can also be used to provide a divide by 3 subharmonic. Output from my circuit as well as a similarly implemented subharmonic will be a square wave.
I'm certain that 5x is possible with just a few JK flip flops and some logic. I'm not sure what a divide by 5 interval produces.
http://home1.gte.net/res0658s/fatman/4046pll.html |
|
Back to top
|
|
|
françois
Joined: Dec 23, 2006 Posts: 139 Location: Paris (France)
Audio files: 1
|
Posted: Wed Jan 02, 2008 3:42 am Post subject:
|
|
|
Hello Scott, & Happy New Year !
I used a '4017 divide-by-ten counter to generate a divide-by-5 square wave. Or, alternately (I'm still SPICEing, that costs nothing but time), I had a look at the '4017 datasheet and reproduced the divide-by-five section with JK-type flip-flops and a few gates. That works well and is glitch-free as far as square or pulse waves go.
But when it comes to sawtooth waves glitches are a concern, especially at higher frequencies (above about 1kHz). For example, the divide-by-three sawtooth output is given (from a ±5V riding sawtooth) by the formulae :
Vout = (Vin - 10V) / 3 during the first period,
Vout = Vin / 3 during the second period,
Vout = (Vin + 10V) / 3 during the third period.
An analog selector/multiplexer selects from -10V/0V/+10V and is driven by a divide-by-three counter. But there are many glitches ! Ian's suggestion of using a "slow" opamp with low slew rate proved invaluable there (the '741 strikes back !).
The very same principle can be used for a divide-by-five sawtooth generator. My idea for having a multiply or divide by three generator was also to have a perfect fifth generator (from the "driving" note C, you get F->C->G). The multiply and divide by five versions give perfect thirds (A->C->E). Using 7th intervals is also possible but seems to be quite controversial as for everything I've read.
-- françois |
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006 Posts: 924 Location: Silicon Valley
Audio files: 11
|
Posted: Wed Jan 02, 2008 7:30 pm Post subject:
Subharmonic generator revisited |
|
|
Hey Francois,
How about using a window comparator (or an LM3914) to switch in a smoothing cap just during the switching periods? I suggest a window comp because you'll want the cap inline just before the actual level switch occurs, but you don't want the cap in there during the saw reset.
Tim (how much is that doggie in the window comparator) Servo |
|
Back to top
|
|
|
françois
Joined: Dec 23, 2006 Posts: 139 Location: Paris (France)
Audio files: 1
|
Posted: Thu Jan 03, 2008 3:41 am Post subject:
Re: Subharmonic generator revisited |
|
|
Tim Servo wrote: |
How about using a window comparator (or an LM3914) to switch in a smoothing cap just during the switching periods? I suggest a window comp because you'll want the cap inline just before the actual level switch occurs, but you don't want the cap in there during the saw reset. |
Hello Tim,
This sounds like a good idea. The glitches on the frequency tripler occur at the end of the "master" sawtooth period and exceed 12V (see 1st attached plot). Instead of filtering them out with a capacitor (too much smearing) a simple clipping circuit gets rid of them.
But for the divide-by-3 generator this does not apply (see 2nd atteched plot) so the idea of switching in a filtering capacitor at these precise instants is interesting indeed. I'll give it a try very soon.
Thank you !
-- françois
Description: |
|
Download |
Filename: |
MultiplyByThree.pdf |
Filesize: |
21.72 KB |
Downloaded: |
153 Time(s) |
Description: |
|
Download |
Filename: |
DivideByThree.pdf |
Filesize: |
49.17 KB |
Downloaded: |
142 Time(s) |
|
|
Back to top
|
|
|
|