DAC0808 4 Channel 10V Half Step Quantizer

[fonik]

from the samples i made ith the 2005 version and from the samples scott uploaded for the new version i would say the new version is much more accurate, say less glitches - or non at all!?

[Peake]

[xpmtl]

in for one or two.

def in need of a quantizer.

[henrikgo]

Second that! I'm also in for a PCB or two.

[wackelpeter]

....cough, cough...

taking the dust off of this ancient pre-historic thread...

To me it looks like a long forgotten treasure as i'm looking for another quantizer for my synth boxes... already build the MFOS CV quantizer which i'm not getting properly adjusted at the moment...

My plan is to try this nice design out and put in onto stripboard.
so i would like to ask is it designed for +/-15V Operation or 12V?
Also i would like to ask you Scott, can you upload a less compressed schematic? I do usually print the schemtaic out and then colour all traces and parts i soldered so i can see what's left to do or not...
This schematic is a bit much compressed so i can't figure out any Details when i've printed it.


And the most important question at last, i don't have to Programm the dac like i have to do in most Quantizer designs using PIC, correct?

Good thing is all components i can get from my supplier for reasonable costs, i.e. DAC0808 for 1,20 Euro and LF398 for 1,10 euro...

[wackelpeter]

I have just built this on stripboard and i now think i need some help to get it working as mine didn't.

First of all, i recognized that in the schematic on side one of the thread it says 80kHz clock is formed by the inverter. From my measuring with a 1,2nF cap and R6 33k i measure a 25kHz clock signal. To get 80kHz i have to reduce R6 to 10 or 12k.

But that shouldn't be the most siginificant problem.
What i have is the upwards going ramp (0-10V) out of pin7 from U6.

A bit mysterious to me it occurs that i always have an pulse output at my comparators output, unless i put a CV in or not. Adding CV to the input changes the pulse width slightly, but not really the pulses at the logic input of the LF398.

I have figured out that it's very sensitive to capacitive loads. When i touch my fingers either at both resistors at the comparator input or the divided outputs from the CD4024 i get some sampled voltage steps out of the LF398.
Adding some "voodoo" circuitry like a 470pF cap and a 470k resistor between both comparator inputs produces also some steps, sounding somehow "musical or quantized" when connected to the 1V/oct imput of a VCO, but not covering the whole range of the CV input.

Is this right that there's always a output of the comparator even with no CV inserted? Theoretical if the pulses would be in phase with the ramp when the latter is at 0V the LF398 would sample the 0V from the ramp, so no problem, right?

Anyone with some hints or tips where to look at, i don't want to abandon this because i'm in need for a quantizer desperately and have already invested a lot of time in soldering this together plus hours of inspecting the board and trying to fix it...

Cheers and thanks in advance for any enlightening insight or tip.


Edit:
Well, don't know if i should call it a small step forward, but i noticed that the pulse output of the 4042 latch behaves as i would think it should, with some cv in this case a triangle 0-5V the width of the output pulses is variable, depending on the CV... but after the cap i couldn't spot anything like that... mostly fully loaded with some sharp spikes towards 0V...
so i desoldered the 560pf cap and put a trimmer inside... trimmed it until i got a staircase CV out of my triangle...
At the logic input of Pin8 of the LF398 i don't have those narrow pulses anymore and there's significant glitch and ringing in my VCO when i hook the output to it's V/OCT input but it steps in from what i hear equivalent steps... Seems if i'm right only a matter of the pulse shaping for the LF398, correct?
Before that i tried different cap values between the latch and inverter including 10,47,56,150,220,330,470,680pf,1nF, 1,5nF & 2,2nF without success...
Maybe some more pulse shaping (cap,narrowing down the pulse) after the last inverter before the LF398 would do the trick... but will have a look at this again tomorrow.

Still any idea is welcome... perhaps i'm all the time messing on the wrong boat...

[wackelpeter]

Think i got it now finally...

At least one of the channels is working. Now i have to stuff all the other channels again correctly as i changed parts and tried different settings on my stripboard massively... but i have a working reference channel and there's some light at the end of my debugging tunnel...

Will write later what i did to solve my problems if anyone is interested...

[wackelpeter]

Okay here's what i had done... getting my build working...

first of all i simply put a trimmer in parallel with the clock timing 33K resistor... after several attempts i ended with 77,8kHZ as clock frequency...

I also changed the arrangement of the staircase a bit... ended with a falling ramp with slightly negative offset so i can't reach the full 10V quantizing scale... and this also affects the +/- quantizer section a bit and moves it's range a bit down... roughly +4,3/-5,7V now...
But i can live with that because made Matthias/Foniks version a +/-5V quantizer...
So i ended up with 3 bipolar quantizer and 3 normal 0-9V quantizer sections...

I simply inverted the staircase into a falling ramp by routing this over a 10k from pin1 of U5 to pin pin6 instead of 5 and putting 10k in the feedback path between pin 6 and 7. Pin5 goes to gnd and i added a 100k trimmer through 20k to pin 6 too with it's outer legs going to + and the other to -.
Now it works and samples nicely without glitch and jitter...