[sdiy] Generating a large number of CV outputs

[Tom Wiltshire]

That wasn't how I generated it, but yes, the 1's and 0's are scattered across the period instead of being all-the-1s followed by all-the-0s. That massively reduces the amount of noise in the signal by making it way higher in frequency.

One other side effect is that the worst-case situation appears when there are very few 1s or very few 0s, at the extreme values. In this situation it's impossible to avoid outputting long series of 1s or 0s, so the effective output rate drops. This is completely different to PWM, where the worst-case situation is where you get a 50/50% duty cycle for your midpoint value. This makes the "bad bits" of a PDM output much easier to avoid, especially for audio - just reduce the output amplitude a little bit so you don't push the PDM output right to its extremes. This is not possible with PWM.

Tom

==================
       Electric Druid
Synth & Stompbox DIY
==================



> On 8 Dec 2023, at 10:47, Richie Burnett <rburnett at richieburnett.co.uk> wrote:
> 
> I forget, is PDM equivalent to PWM but with the ramp counter bit-reversed before it gets compared with the reference?  So the overall ratio of 1's to 0's in one period of the bit sequence is the same as PWM, but the 1's and 0's are scattered around the period, instead of having all of the 1's grouped at the start of the period and all of the 0's grouped together at the end of the period.  Am I right?
> 
> -Richie,
> 
> 
> 
> -----Original Message----- From: Mike Bryant
> Sent: Friday, December 8, 2023 9:34 AM
> To: mskala at northcoastsynthesis.com
> Cc: SDIY
> Subject: Re: [sdiy] Generating a large number of CV outputs
> 
> 
> Yes the key thing about PDM over other methods is it doesn't have to have a fundamental fairly low sampling rate like PWM or D-S.  This gives it much faster response time to requested changes.  One way to think of it is that PWM and D-S have a low pass filter on their feedback loops, which D-S actually does have, whilst PWM sort of has in the lag of changing the M-S ratio.
> 
> 
> This, and the fact that you get far less energy at the sampling rate, is why many digital microphones have moved to using PDM, and of course most new MCUs now support this mode.
> 
> 
> 
> 
> From: mskala at northcoastsynthesis.com <mskala at northcoastsynthesis.com>
> Sent: 08 December 2023 03:30
> To: Mike Bryant <mbryant at futurehorizons.com>
> Cc: Danjel van Tijn / intellijel <danjel at intellijel.com>; SDIY <synth-diy at synth-diy.org>
> Subject: Re: [sdiy] Generating a large number of CV outputs
> 
> 
> On Fri, 8 Dec 2023, Mike Bryant wrote:
> 
>> I just had a read of this and generally very well put.  However you seem to
>> miss out mentioning PDM  between the sections on PWM and Delta-Sigma. This
>> gives faster settling time of DC voltages than PWM, as well as other
>> advantages when producing audio as its easier to push the noise out of
>> band.
> 
> I'm not sure PDM is meaningfully different from delta-sigma at the level
> of detail I included in the posting.  The things I would want to say about
> it would be the same things I did say about delta-sigma.  Is there an
> important example of a PDM scheme that is not a form of PWM nor
> delta-sigma modulation?
> 
> -- 
> Matthew Skala
> North Coast Synthesis Ltd.
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace at synth-diy.org 
> 
> -- 
> This email has been checked for viruses by AVG antivirus software.
> www.avg.com
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace at synth-diy.org