Author |
Message |
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Mon Jan 10, 2011 3:33 pm Post subject:
40106 clock/gate design |
|
|
hello all!
i convinced myself i 'need' a master clock/gate module to sync different smaller synths together (Doepfer Dark Energy, MFOS Soundlab, Hohner Bass3000), centralising the clocking.
i whipped up this preliminary schematic, the question being if any of the Knowledgeable Folk would shine their light on my humble work?
so far: 40106 oscillator w/switchable caps.
signal buffered by second inverter.
from there to led driver, opamp buffer + outputs
and to third inverter + opamp buffer for inverted outs.
so i had leftover inverters. could build the same thing twice but didn't want a second oscillator on the same chip for stability (? hearsay..)
underneath the main section is an idea for having (a) buffered inverter(s). might be nice for later - i'm already toying with the idea of clock dividers, sequencers and whathaveyou, so could come in handy.
don't know about the power supply yet - might be 9v, or 12v bipolar.
the MFOS runs on 9v bipolar, the Hohner on 9v single, and i don't want to fry either.
thanks for looking!
Description: |
|
Filesize: |
26.79 KB |
Viewed: |
1345 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
diablojoy
Joined: Sep 07, 2008 Posts: 809 Location: melbourne australia
Audio files: 11
|
Posted: Mon Jan 10, 2011 8:07 pm Post subject:
|
|
|
you could use the extra inverters for a one shot and a pulse stretcher see ray wilsons site for examples this could give you a gate out with a variable length
which could be usefull
edit : should have looked at your schematic first
i think you will need to add some components to the second inverter you are using as a buffer to make it into a one shot as it is at the moment it will act as
a 50 % duty square wave oscilator and not provide you with the narrow clock pulse that you were aiming for
cheers hope this helps |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Tue Jan 11, 2011 2:34 am Post subject:
|
|
|
thanks a lot!
and thanks for the tip, one of many pages on the man's site i must've missed..
i'm working with the 50% duty now because it works really well with the Hohner sequencer (high= sound, low= silence in gated mode), and the Doepfer's ADSR.
i was already looking at ken stone's gate to trigger circuit as an option afterwards, but if i can have the same thing on the 40106 with a smaller component count, that's fantastic.
so the one shots are definitely worth having in there.
question on the pulse stretchers though - it seems to me they generate a gate with a fixed time, rather than a percentage of the incoming pulse train, is that correct? looks as if i could vary the gate length by changing the discharge resistor, but it would not be proportional to the speed of the clock.
if that is the case (i'll have a go at it asap) maybe i'd rather have a 555 PWM or some such solution.
while we're at it.
ha! |
|
Back to top
|
|
|
diablojoy
Joined: Sep 07, 2008 Posts: 809 Location: melbourne australia
Audio files: 11
|
Posted: Tue Jan 11, 2011 8:36 pm Post subject:
|
|
|
Quote: | question on the pulse stretchers though - it seems to me they generate a gate with a fixed time, rather than a percentage of the incoming pulse train, is that correct? looks as if i could vary the gate length by changing the discharge resistor, but it would not be proportional to the speed of the clock. |
true , true and true
you would need to play around with the values somewhat to get something that will work to a point after which you would just get a constant gate on
higher clock rates. more fun comes from several such ccts put through logic gates anding /oring mutiple clocks and gates together it was just a thought for something extra |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Wed Jan 12, 2011 5:53 pm Post subject:
|
|
|
diablojoy wrote: | it was just a thought for something extra |
and very useful at that.
if i get the time to try things out tomorrow, i'll be back. |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Sun Mar 06, 2011 1:21 pm Post subject:
|
|
|
after a couple of weeks delay.. finally found some time for testing.
decided to go for +12/-12v supply, to be eurorack compatible.
oscillator, buffer, led driver and opamp buffers all working fine.
then, new stuff:
* gate signal now clamped to 9v with zeners.
seemed better than using a voltage divider, as it's independent of supply voltage differences. it works fine with the doepfer and MFOS, an seems compatible with most systems out there.
the Hohner will need some extra trimming of its own (works best at around 5v) but that one's a long term project anyway
* diodes at the end of opamp buffers eliminate negative voltages.
so far, so good.
and new in schematic, but not tested yet:
* DPDT switch to determine straight or inverted clock out on opamp#3.
* DPDT switch to determine straight or inverted pulse out on opamp#4.
'inverted' pulse would still mean positive trigger, but on falling clock edge.
will test asap (not that that means anything.)
Description: |
|
Filesize: |
44.52 KB |
Viewed: |
785 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Tue Mar 08, 2011 5:03 pm Post subject:
40106 master clock |
|
|
breadboarded some more.
i'd forgotten to add pullup/pulldowns to the inverter inputs on the switches.
since the clock would be running anyways, i probably only needed them on #5 and #11, but in case anything goes wrong (never know) i thought it'd be good to have the gate outputs go low by default. so i added pullup to #13 as well.
however, this created a hypothetical situation where - supposing the clock wasn't running due to whatever reason/malfunction - i could create a voltage divider through pullup on #13 and pulldown on #11, depending on the position of the switches, thus creating possible instability.
may not be much of a realistic situation, but it puzzled me a bit nonetheless.
then tried moving the whole inverted clock array to be coming straight off of #2 (oscillator out) instead of the buffered #4, since it includes at least one 'regular' inverter IC1F, with much the same function as that buffer.
(as opposed to the pulse generator around IC1D.)
it might theoretically not be the best of ideas to have a switch sitting straight on the oscillator's output, but i tried it and it worked fine, no glitches or duty cycle variations or whatever to be seen on the scope.
the pulse generator package can then still come off of pin #4 and both switches can be set any which way without interference.
any thoughts/improvements on this are welcome as always!
to me though, at the moment, it seems this thing works. yay
Description: |
|
Filesize: |
81.86 KB |
Viewed: |
873 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
diablojoy
Joined: Sep 07, 2008 Posts: 809 Location: melbourne australia
Audio files: 11
|
Posted: Tue Mar 08, 2011 5:43 pm Post subject:
|
|
|
great to see this is working
will look it over properly when i get the chance
I have shelved my own 40106 odd clock project for a while
to get rid of some of my backlog |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Thu Mar 10, 2011 10:11 am Post subject:
|
|
|
Hi I've stripboarded the first version of the clock .. just wanted to say thanks for the design .. i needed a clock for the baby 6 and hopefully some other seq like .. just wanted to ask there are 3 - diodes on the outs is it on every out a diode .. something like that .. ? i build a 555 clock but was not happy with how it's working ..
thanks again
Vlad |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Thu Mar 10, 2011 12:48 pm Post subject:
|
|
|
hi vlad
you're actually faster than me in building it - shame on me..
i'm working on a pcb design, but i don't have that much time, so it might take a while.
anyway, glad you tried this thing - hope you like it.
there's diodes to bring the output voltages down to 9.1v (around 8.5v on the outputs), those are the zeners to ground (paralleled by the caps.)
the diodes on the outputs of the opamps are there to eliminate the negative voltages that the opamps might introduce. they might not be totally necessary, but they won't hurt either.
so which version did you build exactly? |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Thu Mar 10, 2011 2:13 pm Post subject:
|
|
|
well actually the stuffed stripboard is waiting for output jacks so i haven't tried it yet .. but just when i saw that you've changed the design .. i thought .. ok should i add that thing with the electrolit there now that it's done ... i build the very first design since right before that i had a strugle with tuning a 555 clock .. just thought i ask .. that electrolit there ..does it shortens the pulse ?? cause i saw you're talking about those circuits from mfos .. i was thinking on driving a baby 6 or similar patern gen or something with it .. my psu is 12 v so these changes make it easier for me but i've added the voltage regulators .. anyway good luck with the pcb .. i too hope that will work .. |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Thu Mar 10, 2011 2:15 pm Post subject:
|
|
|
Anyway in so many circuits here a saw there i saw not a single accurate clock with many outs so to say master clock ... |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Fri Mar 11, 2011 3:47 am Post subject:
|
|
|
i'm not really sure which cap you're referring to..
the electrolytics (C5, C7, C8 in version 2 and 3) are there to stabilize the zeners (D2, D3, D4). not entirely necessary maybe.
C4, the cap before IC1D, helps turn the clock gate into a pulse.
all the clock outputs are at 50% duty cycle - on and off equally long.
if you choose the right value for C4, you can reduce the ON time to a couple of ms, turning the gate into a trigger/spike.
and yeah, that part of the circuit was borrowed from MFOS
but actually, if you want to drive a 4017 with it (like in baby 6, no?), both ways should work fine - the 4017 advances on positive trigger iirc, so you could hook up clock or gate without any difference i would think. |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Fri Mar 11, 2011 4:53 am Post subject:
|
|
|
thanks for the explanation .. i actually was refering to the electrolit after the zener diode but i missed that you've actually incorporated that one shot - from mfos further in the cit\rcuit .. i think i'll do a simpler version as you say 50 % and maybe later do all of those mm logic circuits on a different panel , icouldn't completely understand what is a one shot .. simply because of the name one shot ?? .. but it's there to shorten the pulse i guess |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Fri Mar 11, 2011 5:04 am Post subject:
|
|
|
hehe, yeah sometimes i get confused as well..
things like 'pulse', 'spike', 'trigger', 'gate' or 'clock' seem to overlap a bit in conversations like these.
but it'll work out in the end, no worries
off for 3 days holiday - after that i'll prototype the pcb, hopefully.
hope your stripboard version works! |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Wed Apr 06, 2011 5:01 am Post subject:
|
|
|
it's a good looking prototype ... about the caps ... wanted to ask .. i build it on stripboard and it's working .. but it wouldn't trigger my s/h .. still have no seq to try it .. so it's wise to remove the elkectrolits after the zeners right .. more square wave ? i have only scope on my pc so it's not reliable ..
greetings
vladosh |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Wed Apr 06, 2011 11:35 am Post subject:
|
|
|
yeah, the caps after the zeners have a big influence on the square wave - mostly slowing down the transitions (making them more trapezoid style, and ultimately triangle at higher speeds).
i can imagine slower rising and falling edges are easier on the zeners or whatever, but for this purpose it's not a good thing.
you could try with smaller caps, but chances are that at high speeds it will still induce problems.
on the scope everything looked perfect without the caps, so i would just go with that.
do you know what kind of trigger the S/H requires? |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Wed Apr 06, 2011 12:16 pm Post subject:
|
|
|
thanks for the reply ..
i removed the caps and it's working as it was .. i think it's the s/h that wouldn't trigger with this it's not the clock .. as the s/h is working good with the lfo's i guess it need the negative portion of the signal somewhat .. it's a very simple s/h made of only one cmos chip from Osamu Hoshuyama design ,anyway i needed a clock for the sequencer i'm building at the moment so will see how that goes .. it's that patern generator from bug brand maybe you know it .. and thanks for the advice for the caps .. i hope it would trigger the pattern gen
greetings
vlad |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Thu Apr 07, 2011 1:57 am Post subject:
|
|
|
is that the one based around the 4052 chip?
that should work, i have a 4052 running the sequencer in the Hohner.
i don't know what voltage you are running the Bugbrand on? |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Thu Apr 07, 2011 3:06 am Post subject:
|
|
|
yes .. that's the one .. i'm making it a dual but still need to get outs and ins so it'll take me a while i guess .. seemed like a good project for a first seq low parts count .. it will be running on +12 v -12 v .. and the cmos on 12 and ground .. actually my PSU is a bit in lower range it's 11.3 v .. but so far it's working with these few modules .. |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Thu Apr 07, 2011 5:07 am Post subject:
|
|
|
ok, i don't think it should be a problem.
if i interpret datasheet correctly, the 4052 triggers:
Lo below 4.50v, Hi above 5.50v at 10v supply. (=1v hysteresis around Vdd/2).
marginally higher for 11.3v.. so 0-8v should still work fine.
if the 4052 doesn't trigger (which i don't expect), then it's possible the trigger voltage was dropped too much - i just aimed for 8v, since that seemed a middle-of-the-road kinda voltage.
assuming you have the clock running on the same supply, you could then try losing the zeners altogether. or having the opamps amplify, but that seems overkill.
i thought the bugbrand one was designed for 2 or more different clock signals though? |
|
Back to top
|
|
|
vladosh
Joined: Aug 02, 2010 Posts: 659 Location: macedonia
Audio files: 46
|
Posted: Thu Apr 07, 2011 1:50 pm Post subject:
|
|
|
thank you for the insight ... the last time i measured the voltage on the clock outs it was around 7-8 v something like that .. i was thinking on clocking the bugbrand gen with lfo's too as it has protection diodes so i thought it wouldn't be a problem and maybe making that pulse delay from yusynth .. as we're on the subject if you don't mind me asking .. can i use the same method that bugbrand used in his pattern gen for the other modules he did like this one http://www.bugbrand.co.uk/bugbrand_old/images/cmos/Schematics/4089_RateMultiplier.jpg my plan was to add protection diodes and a op amp as a buffer output just so i can use it with lfo's and make it suitable for the rest of the modular ?
thanks
vlad |
|
Back to top
|
|
|
remork
Joined: Aug 02, 2009 Posts: 25 Location: bhellgium
|
Posted: Fri Apr 08, 2011 2:05 am Post subject:
|
|
|
yup, running LFO's through the diodes should work.
when all your 4000-family ic's are powered with the same voltage, i don't think you can do anything wrong - they were meant to be chained together. diodes won't hurt if you're mixing signals though.
might wanna check the datasheets just to be sure (that's a disclaimer right there..)
the buffered output might not be necessary. depends on what you plan to hook up after your circuit. if you plan to drive a lot of other inputs, your logic IC might not cut it - so buffering (+series diode against negative spikes) is always the safer option.
by the way, i'm not sure if all of those IC's you want to use have hysteresis on their inputs.. it could cause some erratic up-and-downing when your lfo sits around half the supply voltage, especially with slow tri/saw/ramp waves, where the input doesn't know if it's being pulled hi or low.
if you don't mind, fine. if it bothers you, easiest is probably a 4050 buffer or some added Schmitt Triggers. |
|
Back to top
|
|
|
|