[sdiy] 1-quadrant multiplier with 2164

cheater cheater cheater00social at gmail.com
Fri Jan 21 06:11:49 CET 2022

Thanks for the breakdown, that was very interesting.

I wonder, could it sometimes be a good idea - especially if using smd
- to forego using different value resistors, and just use 1 or 2
resistor types and build up the required resistances from
serial-parallel networks of the same resistor? It would definitely be
easier to pick them out and harder to mix them up...

With TH it would definitely be way more work than just using the right

On Fri, Jan 21, 2022 at 5:35 AM David G Dixon <dixon at mail.ubc.ca> wrote:
> First of all, printing, transferring and etching PCBs is a royal pain in the
> ass.  The worst part is that nobody makes a laserjet printer as good as the
> old beige cube HP2100.  The new ones just aren't as good, so the transfers
> are never that great.  That means a lot of correcting of the transferred
> image with a fine-point felt-tip pen.  I make four board sets at a time, and
> just getting them ready to etch takes over an hour, so that's at least 15
> minutes per unit.  Etching of all 20 boards takes about half an hour because
> I've developed a system where I lay four of the boards on the bottom of the
> dish and etch the panel board first, then by the time I get the fourth
> circuit board out, it is already nearly etched.  Plus, the solution heats up
> as I go from the exothermic oxidation reactions, so it gets a little faster.
> So, now I'm up to about 20 minutes per unit.  After that, I build them one
> at a time.
> Drilling each set takes one hour (there are about 1000 holes altogether, and
> I redrill them all from the back to make nice holes), so now I'm up to 1:20.
> Cleaning off the transfer with acetone and tinning them all takes another 10
> minutes, so 1:30.
> Cutting all of the jumper wires takes about 10 minutes, and stuffing and
> soldering them takes about 15 minutes.  I also stuff the panel and
> translator boards, and assemble them, which takes about 10 minutes, so 2:05.
> Next, I install all of the bulk resistors -- there are 4 499R, 5 1k, 2 4.7k,
> 11 30k, 4 100k, 1 511k, and 61 10k to install, so 88 bulk resistors in all.
> Then I have to pull 10 hand-sorted 100k and 4 hand-sorted 200k out of their
> baggies and install them, so that's now 102 resistors to bend, cut, stuff
> and solder.  That all takes about 40 minutes, I'm guessing, so now it's
> 2:45.
> Next I have to pull 60 different valued resistors out of 60 different bags,
> bend, cut, stuff and solder them.  That takes another 40 minutes, I'm
> guessing, so 3:25.
> Next I finish stuffing the top PCB (multipliers and 5V supply), including 10
> ceramic caps, 3 IC sockets, 10 electrolytic caps, and 1 LM336.  15 minutes,
> 3:40.
> Next I install the other 17 IC sockets on the other two boards, then I have
> to consult my computer inventory to figure out which Dome filter film caps
> to stuff -- they are all hand-sorted into 0.2% bins, and I have to make sets
> of 12 and 18 caps from four different values, all with the same mantissa.
> That, then stuffing and soldering, takes 15 minutes, so 3:55.
> Then I do final assembly: pin headers and sockets, plugging the boards
> together, and installing the spacers and screws.  That takes about 20
> minutes, so 4:15.
> Then I do the first tests and adjustments; I test the two Dome filters --
> making circular Lissajous figures from a sine wave over two different
> frequency ranges.  Then I adjust a couple of trimmers to eliminate the
> multiplier signals when there is no CV.  Finally, I test the multipliers,
> looking for symmetrical outputs when CV is applied.  That all takes about 10
> minutes, so 4:25.
> Then I take the unit into the house and install the panel and knobs --
> another 10 minutes, so 4:35.
> Then final testing with my guitar to make sure the output amps work
> (everything else is tested) and I get both frequency shifting and phasing,
> the MOD and SIN/COS inputs both work as expected, and the WET/DRY and
> SUM/DIFF pots work as they should, and all outputs work as expected.
> Another 10 minutes, so 4:45.
> Finally, I print out the mailing label, pack the unit and tape it up.
> Another 10 minutes.  So, grand total, 4:55.  That's an estimate, but it's
> close.  If something goes wrong and I have to do any troubleshooting, that
> can easily add an hour.  Luckily, 4 out of 5 work perfectly, and when there
> is a problem, it is usually obvious (e.g., a solder bridge or broken trace).
> Remember, you asked.
> -----Original Message-----
> From: cheater cheater [mailto:cheater00social at gmail.com]
> Sent: Thursday, January 20, 2022 5:00 PM
> To: David G Dixon
> Cc: Todd Sines; synth-diy
> Subject: Re: [sdiy] 1-quadrant multiplier with 2164
> [CAUTION: Non-UBC Email]
> On Thu, Jan 20, 2022 at 8:28 PM David G Dixon <dixon at mail.ubc.ca> wrote:
> >
> > There are now 87 Freak Shifts out in the wild all over the world I
> > build each one by hand, and each one takes me about 5 hours
> What makes them so time-consuming to build?

More information about the Synth-diy mailing list