[sdiy] Mixed-signal problems (dsPIC digital delay project)

[Ben Bradley]

I've seen comments I like, especially about making the common ground
point at the processor, and especially NOT having the ground between
the analog and digital sides having to go offboard to the power
supply. That's the signal return path, it's in series with any signal
going between the analog and digital side, so any voltage between
these two "grounds" will be ADDED to the signal.

I see two holes labeled "link" between DGND and GND near the bottom of
the board. If you haven't changed your grounds yet, have you tried
running a wire between these two (and maybe removing one of the
grounds going to the power supply)? I bet just doing that will improve
the situation.

For further bypassing, everywhere you have a 0.1uF, you can add
another 0.1uf in series with a 1 ohm resistor, that series combo in
parallel with the original 0.1uF. This will damp the RF ringing caused
by fast switching activating tuned circuits made up of stray
inductances and capacitances. This sort of thing is covered in the
book "High Speed Digital Design" which is a great resource even if
you're not making digital circuits that go above 100MHz or 1GHz. It's
an expensive book, but certainly worth reading if you can borrow a
copy from a library.

On Thu, Jan 26, 2017 at 5:02 PM, Tom Wiltshire <tom at electricdruid.net> wrote:
> To clarify:
>
> The op-amps provide the input/output mixers and filtering. This is all run at 9V with a 4.5V Vref virtual ground. Well within spec for the TL072.
> The analog 3.3V supply is used by the dsPIC's ADC and to feed the control pots that it reads. The rest of the chip is run from an independent 3.3V digital supply.
> The analog supply and the 9V circuits share the same analog ground. The digital part uses a separate ground.
>
> Tom
>
>
>
> On 26 Jan 2017, at 20:03, jays at aracnet.com wrote:
>
>> Maybe there is something I don't get but I don't see how the analog section of this works at all. Don't under stand why the analog section is ran at 3.3V. Couldn't find any minimum voltage for the TL07X chips in the data sheets but the web site says 7 volts. Check the min volts in the parametric section.
>>
>> http://www.ti.com/product/TL072
>>
>> Next issue with the analog section is Vref at 4.5 volts and running the opamps at 3.3V. If low voltage, rail to rail, non-latching/non-inverting opamps are used on the input I might see it working but not the output. Just doesn't seem right having a 3.3V opamp with Vref 4.5V and 3.3V signals. Output doesn't seem to be AC coupled either. Don't know if that is a problem or not.
>>
>> Grab the Microchip AN682 app note and take a look at Fig 7.
>>
>> Personally would not recommend doing a ground plane fill. This should be left to people that understand ground planes and current flow over a circuit board.
>>
>> Other things mentioned that are a good idea are more bypass caps, ferrites and make sure the digital and analog traces don't run parallel to each other.
>>
>> Jay S.
>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy