I'll try to on my layman's hat. In:
http://groups.yahoo.com/group/DIOmods/files

I have added an updated spreasheet correcting a number of
typos from the original, you can get it here:

http://groups.yahoo.com/group/DIOmods/files/Gary%27s%20stuff/Filter_dio.xls


Sheet 1: Updated/Edited version of Gary's spreadsheet
Sheet 2: Ray's Spreadsheet for advance users
Sheet 3: Gary's original spreadsheet...locked

Most users will want to use Sheet 1. Here there are six different filter
alignments that all line up in terms of response at only one particular
frequency called the cutoff frequency "Fc". At this frequency all filter
alignments (butterworth, bessel, etc) put out 1/2 the power the would near 0 Hz.

The best filter would have flat response below Fc and infinite rolloff above Fc
in the frequency domain.We can go for Filters with lower "eta" values will have
variation/peaking below Fc but more rolloff/(fster rolloff) above Fc.

The flattest response below Fc is Butterworth with eta=0.707 in terms of sine
wave response. If you put a square wave input, the output will ring/overshoot
just a bit on the top of square wave. Some clever person came up with the Bessel
filter(with slight higher eta=0.866) that has no overshoot BUT less roll of
above Fc.

All the Chebyshev filters tradeoff flatness in sine wave response below Fc for
MORE rolloff above Fc. The Art DI/O has an ETA=0.729
and rolls off just a bit slower than the Butterworth but like the Bessel with
have a better square wave time response (but not as good as the Bessel). The
downside to a slower rolloff rate is that digital
noise artifacts could get into your music. With this background in
"eta"/damping factor lets get back to the spreadsheet. For most hobbyist, the
general goal is to preserve the "ETA-0.729" of the orginal deign and second the
Fc (equivalently the Wo) of the original design and altering only the gain.


In the spreadsheets, only the "green" squares are allowed to be altered by the
user. For most users, only "Gain" needs to be altered.
The Fc should be set to 50.3KHz and the Fc, Wo, and eta of the original design
will be preserved for any gain setting. The corresponding resistor values are
given in "DI/O default" which stands for the default characteristic.

The "instructions A" that Gary gave in his original spreadsheet allow you to
keep R4=20K by altering Fc until R4 shows 20K. This is generally not a good
thing to do as we are violating one of the ten commandments of filter design
when we change Fc.

Instructions B illustrate combinations of "gain" and "fc" (along with filter
alignment...ie. eta) that give rise to corresponding resistor combinations that
have been generated in this "filter thread".

The 8.23K resistor for R3 (R1=11.7K R4=20K) gave an eta of 0.52 or a 1.02 dB
Chebyshev alignment. Changing R3 only to 823 will give a Bessel alignment
(eta=0.866) but the the Fc=97.4Khz...which is much too high.

Putting in Gain=-4.255 and Fc=50.3Khz will allow you to see the calulator
generate the default R1=4.7K, R4=20K and R3=4.7K alignment that came with stock
D/IO. Finally, putting in any gain you like in
entry "B2" will show you the resistor values for your new design in the row
labelled DI/O default characteristic. I hope this helps.

PeAK





--- In DIOmods@y..., "Bob McNeice" <ptmconsulting@a...> wrote:
> As Doug stated in his post - can you put this into laymen's terms ...?
>
> The values Ray recommended was to simply swap out the (R4) 20K
> resistors for the (R3) 4.7K resistors. I don't see these values in the
> post below. Does that mean that this approach has been adjusted also -
> or does this still make sense?
>
> > Gain = (?)
> > R1=R14, R15, R20, R22 = 4.7K
> > R4=R13, R18, R19, R24 = 4.7K
> > R3=R16, R17, R21, R23 = 20K
>
>
> ----- Original Message -----
> From: "gbronn" <gbron@w...>
> To: <DIOmods@y...>
> Sent: Sunday, March 17, 2002 6:09 PM
> Subject: [DIOmods] Gain Spreadsheet Updated
>
>
> > I'd like to thank Ray Chau for catching an error in my gain
> > spreadsheet and I'd also like to apologize to everyone for the
> > error. The calculation for R3 (= R16, R17, R21, R23) had an error
> > that gave a number that was a factor of 10 high.
> >
> > Here are the correct numbers:
> >
> > Gain = 1
> > R1=R14, R15, R20, R22 = 20K
> > R4=R13, R18, R19, R24 = 20K
> > R3=R16, R17, R21, R23 = 773 (closest standard value = 768)
> >
> > Gain = 1.5
> > R1=R14, R15, R20, R22 = 13.3K
> > R4=R13, R18, R19, R24 = 20K
> > R3=R16, R17, R21, R23 = 808 (closest standard value = 806)
> >
> > Gain = 1.67 (for doug)
> > R1=R14, R15, R20, R22 = 11.98K
> > R4=R13, R18, R19, R24 = 20K
> > R3=R16, R17, R21, R23 = 821 (closest standard value = 825)
> >
> > I haven't tried putting in the right values into my DI/O, so I can't
> > comment on how important this is to the sound.
> >
> > Once again, I apologize for any inconvenience this has caused.
> > ---Gary
> >
> >
> >
> >
> > ------------------------ Yahoo! Groups
> Sponsor ---------------------~-->
> > Access Your PC from Anywhere
> > It's Easy. It's Fun. - Free Download.
> > http://us.click.yahoo.com/BxtVhB/7XkDAA/_ZuFAA/bDIolB/TM
> > --------------------------------------------------------------------
> -~->
> >
> > To unsubscribe from this group, send an email to:
> > DIOmods-unsubscribe@y...
> >
> >
> >
> > Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/
> >
> >
> >