[sdiy] Hand selecting capacitors

David G Dixon dixon at mail.ubc.ca
Sat Apr 10 08:55:41 CEST 2021

```Well, it's a fairly big pain in the ass as it is just to grab the 60 or so
unique resistor values needed for the 12-stage Dome filter and the 18-stage
Dome filter from the 60 different bags.  Just stuffing those PCBs takes
about an hour, even though I have a box with the two sets of resistors all
organized in little baggies in the order in which they are stuffed in the
PCB.  And that assumes that I'll have sets of 4+4+4 (102, 103, 104) caps for
the 12-stage filter and 4+4+4+6 (102, 103, 104, 105) caps for the 18-stage
filter, all with more or less the same mantissa (to 0.2%).  I have a big
inventory table of all my film caps with the four different values separated
into 46 different mantissa bins, from 968/9 all the way up to 1050/1, so it
is easy to assemble sets from my inventory for the two filters.  My
inventory table is in the same Excel spreadsheet as my phase displacement
simulations, so if I have to mix mantissa values to make up a set, I can see
what effect this will have on the phase displacement error, and even figure
out the optimum way to arrange the caps within the filter so as to minimize
the error.  For example, if I have to mix two adjacent mantissa values to
make up a set (for example, in my last 12-stage filter I mixed 974 and 976
caps), I have found that it is far better to mix the values so that the
average mantissa is roughly equal between the two filter halves.  For
example, for the 974/976 mix, I used the following sets:

Filter half A: 97400-97600-9740-9760-974-976(nF)  Filter half B:
97400-97400-9740-9760-974-976

The limiting factor there was that I only had one 976 104 cap and three 974
104 caps, but I had quite a few of both mantissas in the 103 and 102 values.
Putting them in that specific order gave a phase displacement error which
was scarcely different from the ideal (about 0.2 degrees off of 90 max).  As
you can see, I'm fairly serious about building good Dome filters, with the
result that the Freak Shift is probably the best-sounding analog frequency
shifter that money can buy right now (plus, it's cheap!).

Anyway, bottom line: if I had to customize the 60 resistor values for every
Dome filter, I'd kill myself.

_____

From: Synth-diy [mailto:synth-diy-bounces at synth-diy.org] On Behalf Of
Bernard Arthur Hutchins, Jr via Synth-diy
Sent: Friday, April 09, 2021 6:26 PM
To: synth-diy at synth-diy.org
Subject: [sdiy] Hand selecting capacitors

[CAUTION: Non-UBC Email]

Oren Leavitt said Fri Apr 9 17:17:01 CEST 2021

“Would it be easier to take, say 4 1.03uF caps and nudge any frequency

determining resistors up/down as needed?”

Indeed – TRUE – see my frequently referenced, (even just in January):

“Design of Wideband Analog 90° Phase Differencing Networks Without a Large
http://electronotes.netfirms.com/EN168-90degreePDN.PDF

Using this state-variable approach (outside in), you can use a handful of 1%
caps (ALL the SAME value) and trim the poles with the resistors.

-Bernie

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20210409/89487a08/attachment.htm>
```