CV Quantizer PCB Layouts

[Sine]

The staircase generator for this thing is still on my breadboard, I got the DAC at farnell some time ago ( they sell it now for 15,80 euro, not cheap, but certainly not the end of the world )

This thing is VERY tricky with grounding, you really need to keep the analog and digital circuit as separate as possible, I had a terrible bleedthroug from the clock signal ( and yes, the osc as drawn above will not oscillate )

[fonik]

[fonik]

i am optimistic about using a 4024 and a DAC 0800 or DAC 08EP. the 4024 should be perfect to control the 7bits. the AD7245 has a settle time of 30ns, while the DAC 0800/DAC 08EP show up with 100ns/70ns. these DACs are rather cheap: a DAC for a buck.
we would clock the DAC with a frequency of about 30kHz, which would result in a period duration of approx. 0.033ms = 33us = 33000ns, am i right?
so these DACs should be more than sufficient speedwise!? or am i thinking wrong about it?

scott used the AD7245s reference voltage to center the staircase roughly around GND, and an additional trimmer for fine adjustment. we could use just a trimmer, i think? or wouldn't it be better to use a 5V zener diode for rough offset, and then a trimmer just for fine adjustment?

i could think of a switch to select bipolar (-5V to +5V) and unipolar (0V- 10V) operation.

i ordered the parts today and will start breadboarding as soon as possible.

any comments appreciated.

[fonik]

one more question to all, who are interested in this project: the DAC 0800 needs an external reference voltage of 10V in our case. would a voltage regulator be sufficient, or would you prefer a precision voltage reference (REF102), which would be a little bit more expensive (approx. 4$)?

[Sine]

I would go for both, the pads of a 78(L)10 underneath the dil8 socket for the REF, or something in that direction.

[fonik]

[fonik]

today i received the parts and could not wait to breadboard it. and i get a nice upgoing ramp now.

a schmitt trigger oscillator clocks a 4024 with approx 32kHz, the 4024 drives 7 of the 8 bits of the DAC0800. the DAC0800 is in bipolar operation, so with the 8th bit allways high it outputs a 128 step ramp from 0.4 to 9.6V, which should be fine enough for us!? the ramp will get centered and trimmed to half steps anyway.

i hopefully get the rest o the circuit breadboarded next week. we are on a good way, i think.

[Sebo]

Hi:
For the 10V reference voltage for the DAC I recommend you to use a circuit
around a TL431, is cheap and more precise than a 7810.
I built a Marc Bareille's MCV-628 and a Quantix-8 and both calls for the rare
REF01.
Marc suggest me to replace it by the circuit he used in this project:
http://m.bareille.free.fr/mcv32/mcv32.html wich uses the TL431,
and worked great in both cases.

[fonik]

[Sebo]

Hi fonik:
About temp stability you're right, but I think it isn't a big difference.
About board space, I used a little piece of stripboard of the same dimentions
of a REF01 (DIP8), to fit the TL431 (TO92), two resistor, a trim, and a cap.
I did it that way because I had to replace the REF01 in two already made
PCBs.
So the space it took is the same (I mounted the resistor vertically).
But if REF102 is easy to find and not too expensive, is a good option, and
anybody that can't find it, could do what I did.

[Tim Stinchcombe]

I was just cruising around the forum when this caught my eye:

[fonik]

hello to all who are interested. here is the schematic i am currently laying out the PCB from (the s/h has not been breadboarded yet). i think it will easily fit on a 3x4in board.
again, any comments or suggestions, especially concerning the voltage reference, are appreciated.

[EDIT 6 MAR 09: schematic updated. circuit tested and working - but there will be improvements i am currently working on)

---

Quantizer.gif
Description:
Filesize:	76.55 KB
Viewed:	743 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[fonik]

less success today. i did not get the sH to work properly.

- the comparator works fine.
- inverting of the comparators output works.
- the pulse conditioner works somehow. i am not quite sure. i cant get hand on the pulse with the scope. i can see an continuous line at GND and some small dots at 15V, however when i try to change the X-axis division on the scope the dots fade...
- the output of the LF398 is always low, near the negative rail. the logics reference voltage is set properly.
pin 6 for the hold capacitor shows a ramp between approx. -12 and -14V!?

i tried to clock the LF398 with the original clock signal for the DAC; didn't work either.

i am puzzled for now. will have to give it a rest!?

did someone get it to work according scotts schematic?

[Sine]

I think i have a piece of this still on a breadboard, and IIRC it worked, but very jittery.

I'll see if I can get it to work and post the results here.

[fonik]

i got it working yesterday.

i took a look at chris list's schematic and found he didn't connect the offset input of the LF398 s&h amplifier at all. i disconnected it from GND and it worked.

it actually works quite fine. i recognize some "pops", though. i hope i can post some samples this week.
i still have an email from harry bissell about using latches to improve performance. i will review it carefully but i fear i don't understand what he means. to be honest, i even don't know what latches are . i could tell from the dictionary that latches are switches!? what is meant here in our context? logical gates?

@scott,
in case you are following this thread: you placed a 10M resistor in the feedbackpath of the comparators. what is this for?

[edit: spelling; and i will update the schematic soon]

[fonik]

would a d-flipflop be the right thing? or should i ise the 4516 instead of the 4024 again? or both?

[fonik]

answering this by myself
i will try to put latches (D-flipflops) between the comparators output and the LF398s inputs. the latches are clocked by the counters clock. i was told this would give the DACs output 1/2 clock cycle of time to settle.
if this is not sufficient i will replace the 4024 by two 4516s again. i hope this is not necessary, though. the staircase looks really nice on the scope.

BTW: schematic above updated. circuit tested and working - but there will be the improvements i am currently working on...

[Luka]

matthias will this be much different to the mfos quantizer?

[fonik]

[fonik]

i implemented the d-flipflop today, as once mentioned to me by harry bissell:
the comparators output is fed into the d-flipflop. the flipflop is clocked by the origin clock.
it improved the performance a lot. i don't think it is necessary to replace the 4024 by 4516s (since the staircase is fine anyways). i will have to listen to the output of the current status carefully for a longer time and under different circumstances before i make a final decision, though.
the i will have to try a better capacitor for the hold function (styro?)...

[Sebo]

Those are great news!!!

[fonik]

for now i made three sound samples. the quantizer is currently set up with the d-flipflop latch, a DAC0800, a CA4024 counter, and a 15nF poly cap from WIMA for the hold function. IT IS NOT CAREFULLY CALIBRATED with the trimmers yet. i hope i will have the time to do complement recordings with the 4516s this week so you can judge by yourself...

each sample is about 500kb

sequencer as source, MFOS 12db filter comes in later:
http://www.modular.fonik.de/sound/quant_01.mp3

LFO, 1 step (half note)
http://www.modular.fonik.de/sound/quant_02.mp3

LFO, 3 steps (1 1/2 note)
http://www.modular.fonik.de/sound/quant_03.mp3

[EDIT: minor grammar changes - i try to do the best i can ]

[zthee]

Thanks for the mp3s! I'm really really looking forward to this!

[zthee]

Would it be possible to drop adress lines or having 2 trimmers to flip the scaling, say between major and minor? Or something like that?

[fonik]