Anushri & MCP4822
  • I have a question to Pichenettes…
    Did you ever considered / tried to use a classic solution for signals distribution, I mean DAC123x + 4051 + S&H instead of MCP482x?

  • I did not use a parallel DAC because this would have required a bigger MCU with more data lines.

    Serial interface + internal reference + widely available + DIP + not so expensive -> MCP4822 (I couldn’t even find DAC1232 or DAC1230 at Mouser…)

    4051 + S&H increases the part count, board space, and possibility of people doing things wrong. The simpler the circuit, the easier troubleshooting is.

  • I see :)
    Besides, I wonder if DAC + 4051 + S&H wouldn’t be also too slow for Anushri..?

  • No, speed wouldn’t be an issue here.

  • I want to refresh this thread, because I have a problem with my design.

    I’m trying to design a digital control to my analog synthesizer, similar to Anushri.
    2*VCO + VCF + VCA, the rest (LFO, 2*ADSR) will be digital. AVR must generate 8 CV signals. Having four MCP4822 is a bit expensive, so I did some testing with one 4821 + 4051 + S&H.
    With small capacitors, sawtooth wave looks (and sounds) quite bad. Too big capacitors make glide effect. I’m lost…

  • Schematics? Waveform traces? Which sawtooth wave are you referring to?

  • Here is a simple schematic…

    I’ve just corrected one mistake: setting too many binary control inputs (ABC) of 4051 at once.
    Now, the sequence is as follows: changing one input of 4051 and then sending data to 4821.


    Now it looks promising. 4 kHz sawtooth looks good for C = 380nF.
    But I’m afraid what will happen when ATmega48 will have more to do? Because now it only switches port bits and sends data to the 4821…

    709 x 263 - 14K
  • 380nf seems a big capacitor for a s&H circuit … What happens is that takes some times to have this capacitor charged completely and what might occurs is that you are switching the 4051 faster than the time it tales to charge the 380nF capacitor, I guess.
    I guess you already check some services manual (like the one for the prophet5, memory moog, oberheim xpander etc) and you saw that usually 10nF is the maximum capacitor size used in such synth.
    What I would recommend is not switching the CD4051 too fast … If I remember correctly the prophet5 update the S&H every 1ms or 10ms.
    What I would do first is making a simple CD4051 + S&H circuits works out of an arduino and make sure that when sending a midi value, I get the right voltage quickly on an output without disturbing other outputs. (I did the test – running in the main loops with a 16Mhz chrystal – and I can tell you that at such speed, output of S&H is horrible. I add to put a “wait 10ms” statement before selecting one of the CD4051 ouputs … timers interrupts might be your best friends …).

    Oh, that’s only my 2 cents ;-)

    Good luck ! I let us know your progress ‘cos I’m quite interested …

  • Any progress MaxZorin ?

  • Well, I recently started to play with 74HC595, 74HC165 & MCP3202 chips…
    And then I discovered that ATmega was clocked by… the 128kHz internal oscillator all the time!
    This is my first experience with AVR. Afterwards, I messed up the fuses and bricked ATmega :(
    Now I’m waiting for a package of new chips :)
    I’ll back to S/H testing soon.

  • argh …;-) If I have some time in the next few weeks, I’ll give it another try. I did have something working but S&H output was flickering as I was updating S&H too often. Also, as I said earlier, the C value is important in the S&H circuit as, I think, it relates directly with the speed at which you update the them.

  • It seems to be ok now, even with 10nF caps :)
    CV signals are updated at 1,22 kHz.