[sdiy] Issue with CD4024 Ripple Counter
magnus at rubidium.se
Mon Sep 20 22:09:17 CEST 2021
Put a logic chip as "driver" of the clock before the 4024, and then try
having various wide loops of cables on the ground and VCC to see if that
replicates the problem. Also try to jump a 100 nF ceramic across the driver.
It is good to trigger the scope on the generator edge and then monitor
the clock input and also the first DFF output.
I still suspect ground-bounce.
The "sharper" the "driver" is, the more ground-bounce I expect.
On 2021-09-20 19:11, Tim Stinchcombe wrote:
> Hi list,
>> I give up. I'm going to put all of this shit away and build a few more
> Freak Shifts.
> Well it seems overnight I've been a little overtaken by these events, but
> last night I started breadboarding a little test circuit for the sole TI
> CD4024BE that I have, to see if I could provoke it into misbehaving by
> gradually slowing down the clock edge rate, and having gotten so far decided
> to carry on regardless. Executive summary: I can't, on the face of it, it
> basically 'does exactly what it says on the tin'.
> My very simplistic circuit and scope traces attached.
> My signal generator is no great shakes, but the square wave it outputs has
> rise/fall times of 200ns, so plenty fast enough to not need to worry about
> further. Feed this (+/-5V signal) into a simple integrator with a 4k7
> resistor in series with a 500k trimpot (R1 just shown as 4k7 in the
> schematic); thus the output of this OA is slamming nearly rail-to-rail, and
> I can control the slew rate with the trimpot; 2nd OA scales this down by
> about half, so the signal swing is about 10V, and the resistor to -12V adds
> about 5V of offset so the signal is near enough unipolar 0-10V to
> comfortably feed the clock pin of the 4024. I'm triggering off of Q1 out of
> the 4024, and the 'divide by two' action is clearly seen.
> With the trimpot at zero, so R1 at just 4k7, traces annotated 'fast ...',
> the slew rate is about 6.5V per microsecond, i.e. not far off as good as a
> TL072 will go, and the clock falling edge time is less than 2us. The chip is
> clearly reacting to the falling edge of the clock, and there was absolutely
> no hint of the 'CLK' trace 'flipping' as it would have if the chip were
> occasionally responding to the rising edge of the input. (Note the changed
> vertical and timebase scalings for the 'zoom'ed trace.)
> With the pot at 500k+, traces 'slow ...', the situation remained exactly the
> same, apart from the more sluggish clock edge: the slew rate is now around
> 0.15V per microsecond, with a fall time of more than 70us.
> So on my sample size of 1, the chip seems to behave as it should, with no
> hint of any of the weirdness that David is/was experiencing in his circuit,
> suggesting that there is either some odd noise/coupling/whathaveyou issue in
> his PCB or it just so happens his TI chips are 'wonky' in some unspecified
> David: I'd be more than happy to stick one or more of your chips into this
> set-up if you felt inclined to post them to me, but it is so simple a means
> of testing it probably isn't worth the postage, i.e. you could probably just
> as well lash something very similar up yourself...?!
> Tim Stinchcombe
> Cheltenham, Glos, UK
> email: tim102 at timstinchcombe.co.uk
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> Selling or trading? Use marketplace at synth-diy.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Synth-diy