Something New from Doc Sketchy

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Joe., lisa, luketeaford, Kent

Post Reply
User avatar
thresholdpeople
Veteran Wiggler
Posts: 573
Joined: Sat Dec 21, 2013 6:17 pm
Location: NYC

Post by thresholdpeople » Sun Apr 14, 2019 4:18 pm

woah. this sounds awesome. i am really looking forward to where you take this.

being able to scan between that many filter modes sounds fantastic, i'm especially curious what this will sound like at audio rate. lots of potentials here...!

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sun Apr 14, 2019 6:25 pm

I found another few important modes to complete the set. There were few missing offset notch modes -- basically, I found the missing mirror-image notches. Now the complete set of modes is 32, just exactly as many as I have channels on an 8x4 scanner.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Tue Apr 16, 2019 5:46 pm

So, I've done quite a bit of work on the multimode aspect of the Doc Sketchy Polaris. These are the 28 non-allpass modes that I've discovered.

There are 4 LP modes, 4 HP modes, 6 BP modes, 6 "Centred" Notch modes, and 8 "Offset" Notch modes. (Note that one of the "offset" notches is just a centred notch at 3x higher gain -- this is because it has a mathematical affinity with the double notch.)

Image Image

If you look carefully at the notch mode plots, you will see that they are completely symmetrical -- there are no "missing" modes. I had to derive 3 new ones yesterday to fill in the blanks. This is not as hard as you'd think, if you know the basics. The new modes include: 2/0 NC, 1/1 NL, and 2/0 NR.

If you compare the names of the modes to their respective plots, you will see how the numbers refer to the slopes (in increments of 6dB/octave) on either side of the centre frequency. For the LP, BP and HP modes, these numbers add up to anything up to 4, because 24dB/octave is the maximum slope obtainable from a four-stage filter. You will notice that the notch modes only add up to 2 or less. This is because the notches themselves require 2, so if a notch name adds up to 2, it is actually a four-pole notch.

So, how to order them in an 8x4 scanner? Well, there is really no "perfect" way to do this, because there are 6 BP modes, which doesn't fall either on a row or a column. The scheme I have come up with is given in the following table:

Image

I've put the LP modes across the top row, so that moving the 4-channel fader left to right will scan across these modes, and thus give a variable slope LP filter from 6 to 24 dB/octave. I've also put the AP modes on the bottom row -- these are useful for phasing.

I haven't put the HP modes on one row. Looking at the titles of the modes, and at the patterns of the gain factors, it seems to me that the HP modes have more of an affinity with the BP modes, so I've put the HP modes on rows with their related BP modes. The problem with this is that there aren't enough of these modes to fill four rows. There are only 10 HP+BP modes. This leaves 6 empty spaces in those rows, so I filled those spaces with the Centred Notch modes. This is obviously not ideal, but I don't see any way around it. The next two rows are the offset and special notches.

All 7 non-AP-mode scanner channels (4 modes each) are shown below. You can get some sense of how the modes will morph across the 4-channel fader by going from blue-to-magenta-to-yellow-to-cyan in each plot. Some of the modes make more sense in column-wise scanning, in which case you should look at all of the, for example, cyan curves in the first five plots. Another thing that could be done is to feed the CVs of these scanners from a Planar, or from two S&Hs.

Image

If you look at Scanner Channel B (the second plot on the left), you will see how the 1/0 HP mode falls very nicely into the family of BP modes (1/1, 1/2, 1/3). I'm very happy with this channel. The next two channels mix notch and BP and HP modes, and I'm less happy with those, but they make more sense if scanned column-wise rather than row-wise.

Anyway, that's where I'm at. I've made one of the two 16-mode generator boards, and will be ordering all of the resistors I need to build the various gain ratios (off of 30k feedback resistors) tonight. I should have something built by next week.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon Apr 22, 2019 1:21 pm

So, progress on the crazy morphing Polaris project...

I have designed a 32-channel scanner for scanning the (now) 32 filter modes from the Polaris circuit. I have done it all on boards which are only 2.1" wide, and a maximum of 5" long. They are designed to all stack together, sharing power and other signals from stacking pins. The only problem is that 8 PCBs are required to do the full 32-channel scanning. Here are the layouts.

The scanner consists of an 8-channel scanner with four 8-channel audio boards in parallel all being controlled from one 8-channel control board and one 8-channel logic board, plus a 4-channel scanner to handle the outputs of the four 8-channel audio boards. Hence, what you get is an 8x4 scanner.

Here's the 8-channel control board. This is where power comes in. This board sits on the bottom of the stack:
Image

Here's the 8-channel logic board. This sits on top of the control board and takes the control signals and converts them into logic signals to drive the switches on the audio boards:
Image

Here's the 8-channel audio board. Four of these are required for 32 channels, and they stack up on top of each other. The 8 logic outputs are all shared on the left-hand end of the board with stacking pin headers. The 8 audio inputs are on the top and bottom of the board. The way this works is that each of these boards gives an independent 8-channel scanner, and the outputs from these boards must be routed to a separate 4-channel scanner.
Image

Here's the 4-channel control board. It may go above or below the stack, but probably below because these boards are longer than the others. It may also provide power to the whole stack, or take power from the header on the 8-channel control board (in which case, the power pins on the left would not be stuffed):
Image

Finally, here is the 4-channel logic and audio board, of which only one is needed:
Image

These layouts are untested, but I have built similar scanners already, and I have gone over the layouts with the schematics many times, so I am 99% confident that they will work perfectly the first time. I'm going to build this monstrosity later this week.

The actual Polaris build will also require the Polaris core PCB (already done, but I may change it), and two 16-mode generator boards (one of which I have made already) with 32 mode chiclets (which require a bunch of special resistor values, which are sitting in a Digikey box at Moog$Fool$'s apartment). Hence, I should have this whole project done next week.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon Apr 22, 2019 1:36 pm

Also, I have looked at the Polaris layout, and I originally made it 4" x 4", so I will be splitting this into two 2.1" wide boards. That, along with the two 16-channel mode-generator boards (and their forest of mode chiclets) will mean that this whole project will require 12 PCBs. That sounds excessive, but I do love making PCBs. That doesn't count the panel PCBs that I will also be laying out for this project.

Finally, this project can also be built in a much simpler fashion if the modes are selected with rotary switches. This would still be 32-channel switching, with one 8-position switch and one 4-position switch. This will require another large pile of DG444 switches, but no "control" circuitry and no crossfader circuitry (so no 2164). I believe that all of this switching can be done from one PCB. It might be better to use 8-channel multiplexers (DG408) for that application, although it depends on whether there is significant switching noise. If that is the case, then that would make a very efficient PCB, and the whole project would require only 5 PCBs (rather than 12). This is obviously a version of the project that more people would be interested in (as it would cost about half as much).

EDIT: Actually, after thinking about it a bit more, I would not use 8-channel multiplexers, because these require binary switching logic, and that complicates the switching. I would just use eight DG444 chips, which are very easy to drive from the rotary switches. It does make the PCB a bit more dense, though.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon Apr 22, 2019 5:17 pm

OK, I have now figured out that this whole approach won't work.

The reason: I can't draw up to 32 different outputs from a single input, because the Thevenin input impedance is just too low. I have to rethink this whole thing. Even putting two buffers for each input onto each mode generating board (to increase the impedance by roughly a factor of four for each input) will still result in levels of each input which are different (because the impedances are really different, and much lower for inputs 2 and 3). That means that the gain ratios won't actually conform to the resistor ratios because the starting levels will all be different. It also means that the buffer amps will run quite warm.

I'll revisit this design in a few days, after I've set and given my final exam...
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon Apr 22, 2019 9:11 pm

OK, I couldn't wait... I had to "revisit" it now. I figured it out.

It turns out that my design will work fine. I just have to buffer the filter output signals (1, 2, 3, 4) before putting them into the mode-generator boards. I'm going to provide a buffer for each signal feeding 8 mode generators (i.e, four buffers per signal). This will increase the overall Thevenin impedance by about a factor of four. I'm also going to divide down the signals going into the buffers by a factor of four. This will decrease the current required from the opamps by another factor of four (a 16-fold decrease in current overall). This will be completely sufficient to ensure perfect opamp conformance, and hence, accurate gain ratios for the modes. The only catch is that the 0 signal (input + feedback) must also be reduced by a factor of four, but since I'm not buffering that signal, I'm simply going to increase the input resistors for the 0 signal from 30k to 120k. That signal only comes in to the mode generators at a gain ratio of 1, so it's easy just to change all of the resistors.

Also, to decrease the overall current load of the circuit, I'm going to generate all of the mode signals at 25% voltage gain relative to the raw filter outputs, and only amplify the very final signal from the switching circuit by a factor of four (by using a four-times larger feedback resistor on the final current-to-voltage converter from the 2164 crossfader -- 120k instead of 30k).
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Moog$FooL$
Broke Dick Dog
Posts: 1698
Joined: Tue Jun 03, 2008 2:54 am
Location: Vancouver, BC.

Post by Moog$FooL$ » Wed Apr 24, 2019 4:35 pm

well..... are the changes gonna work??
i trust the order was all there?? digikey don't often fail.
i use to drink & smoke.
that fat cat had to go..... wasn't even mine.

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sat Apr 27, 2019 12:15 pm

...actually...

The Polaris filter signals are already at only 30% of full synth levels (and only brought back to full scale at the final output amps), so I probably don't have to do any dividing down into the buffers. I will leave the option in my layout to add the dividing resistors, but may not stuff them.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sun May 19, 2019 5:26 pm

Hey Team,

I decided that I needed to take a little break from the Polaris project, and that I needed a little diversion project to do in the interim. So, I realized that I always wanted a fixed filter bank, so why not try building one of those?

Anyway, it took about a week of on-and-off effort (because I was doing a whole bunch of other stuff for work, plus cleaning out the house for renovations) but here it is (sorry the picture is so dark -- a photographer I ain't):

Image

You will notice two things straightaway: 1) the frequencies are not the "standard" Moog/Dotcom ones (mine are LP 67, 100, 147, 213, 303, 454, 666, 1000, 1470, 2127, 3029, 4544, 6665, and HP 10000 Hz), and 2) the name of this module is "E6 Fixed Filter Bank." These two things are related. The first person to figure out the connection wins a prize. In any case, my frequencies work very well, although 67 Hz is maybe just a tad low for the LP filter. It tends to generate a low rumble -- maybe that's what you want! The HP is not too high at 10000 Hz, and gives a nice hissy top end. The BP filters are all designed with a stage Q of pi (3.14) and an overall gain of 100% at the peak. This works well.

You'll also notice that there are two input jacks, three output jacks, and a bypass switch. One has to be very careful with the bypass switch, because if the various band attenuators are set to various positions, the output may be extremely loud when bypassing the filters -- this is exacerbated by the fact that the Odd and Even outputs are summed into the All output, so if the bypass is set, the input is sent to the All output at 200% gain -- I'm still trying to figure out a clever way around this.

The filter bank is divided into "Evens" and "Odds" (Evens on the left, and Odds on the right) and these have separate outputs, with a summed output in the centre. This can be used to give cool stereo effects. This is a feature which I copied from the YuSynth FFB. However, I went one step further and added separate Even and Odd inputs as well. The Odd input is normalled to the Even input, so if only one input signal is desired, one plugs it into the Even input jack.

Another little innovation -- if nothing is plugged into the input jacks, there is an onboard white noise generator which is normaled to the Even input jack. When I looked at YouTube videos of FFB demos, I noticed that they were mostly done with white noise, and I really liked the effect, so I thought, why not just put one in the box? So, with nothing plugged in, this thing is a standalone wind generator.

Another thing I noticed when looking at other folks' FFBs was that they typically involved a lot of tedious panel wiring. Mine doesn't, because I made a panel PCB with all 14 pots on it. The jacks and switch are still hand-wired, but it only took me about two hours to complete all of the wiring on this module. I really really LOVE those new 16mm Alpha pots with the 3/8" bushings and the solder pins up close to the body of the pot (check 'em out at Small Bear).

Here are a couple more pictures of the back:

Image Image

There are three PCBs: 1) the panel PCB, which just has 14 pots on it and some pins, 2) the LP-HP-Outputs-Noise PCB, which is in the middle, and 3) the BP PCB, which has all 12 BP filters on it, which is on the top. I'm super duper happy with my layout of the BP filter board, which is 2.1" wide and 6" long. It is my best layout ever (which you would be able to appreciate if it wasn't covered with those shitty green Mylar caps). The BP board is completely stuffed, but the other board has lots of empty space on it -- I made it 6" long to mate it with the other board. The panel PCB is also 6" long, so it all fits together. The LHON board plugs into the panel PCB, and the BP board plugs into the LHON board. This makes assembly a real snap. The two circuit PCBs are not in the centre of the module, because the pins on the panel PCB are not centred. I may change this in a future layout.

You will notice that I mostly used crappy green Mylar caps in this build (which are all different sizes, ugh!) This is because my local shop (Lee's Electronic Components on Fraser St. in Vancouver) doesn't carry all of the values of those lovely little rectangular film caps with the 0.2" lead spacings which I intended to use, and I didn't get around to making a Digikey order. In any case, my next one will be much prettier because I'll use nicer caps, but this one works perfectly nonetheless.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Rex Coil 7
Super Deluxe Wiggler
Posts: 6274
Joined: Thu Nov 29, 2012 11:29 am
Location: Captain Of Outer Space

Post by Rex Coil 7 » Mon May 20, 2019 12:41 am

You're a legend. Skills of a Jedi master, armed with a soldering pen.

I'm not kidding around.

:hail: :hail:
Never Quit, Die Falling Forward
5U PROJECT - (skip pages 4 through 6, boring junk) ... https://www.muffwiggler.com/forum/viewt ... highlight=

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon May 20, 2019 1:31 am

Rex Coil 7 wrote:You're a legend. Skills of a Jedi master, armed with a soldering pen.

I'm not kidding around.

:hail: :hail:
Oh, stop it! You're making me blush.

Actually, I live for doing this stuff. It's pretty much all I can think about anymore. I guess you could say I'm addicted. And I can sit here in my easy chair, or on my bike while I'm riding around town, or in my car while commuting, then I can conceive of a complete module in pretty much full detail, and then just throw it together.

Also, I just figured out how to "solve" the too-loud bypass problem. I'm just going to divide the two inputs in half using a couple resistors, and so I'll be sending half the input to each side, and only 100% of the input will come out the "ALL" output. I'm gonna go fix it right now.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
BugBrand
Knowledge of Bugs
Posts: 7202
Joined: Tue Jul 15, 2008 11:59 am

Post by BugBrand » Mon May 20, 2019 2:12 am

Dr. Sketch-n-Etch wrote:.. the name of this module is "E6 Fixed Filter Bank." These two things are related. The first person to figure out the connection wins a prize. ...
Even if we read the SDIY list?!

Good work!

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon May 20, 2019 2:39 am

BugBrand wrote:
Dr. Sketch-n-Etch wrote:.. the name of this module is "E6 Fixed Filter Bank." These two things are related. The first person to figure out the connection wins a prize. ...
Even if we read the SDIY list?!

Good work!
Cheater! I didn't say what the prize was -- maybe it's a big wet kiss on the lips!

In any case, I just kludged in the divider resistors to cut the bypass gains in half, and it worked perfectly. Now, if several of the bands are turned up, and the bypass switch is flipped, the levels don't change too much.

I wasn't quite prepared for how "ringy" the bands would be. The Q values are 3.14, but it seems to be quite resonant even at that relatively low value. It sounds good. The wind sounds really good, and it's pretty easy to make a pulse wave sound pretty reedy. It's cool to cut out all but a couple of bands, like low, medium, and high -- that gives a weird sound. Also, panning the Odd and Even outs to left and right sounds really good.

One thing I'm thinking, based on playing with this thing tonight... It would really benefit from voltage control. I designed the circuitry specifically to be able to break into each band easily with a VCA. I'm just not sure how I would control that many VCAs simultaneously. A 14-channel sequencer would be nice! A timbre sequencer.

Also, to answer somebody's question from the SDIY list: Based on the fact that I used linear-taper pots, I would say that linearized VCAs would be the best. That would be a helluva lot of circuitry, though. Or, four Quad VCAs with two channels to spare (to be used for overall loudness, I guess). I'm probably not going to do that right away, cuz I have bigger fish to fry. (I really wish that Intellijel still made those picoVCA PCBs, which were 1-inch-square plug-in chips with a complete dual linearized VCA circuit on board. I coulda used a few of those for this, and they only cost a few dollars each to make.)

Anyway, the final verdict: I'm very happy with this little toy, and I'm very glad that I decided to build it.

Tomorrow I'll try to post a video (if my wife lets me back into the workshop -- the home renovation preparations went pretty slowly today...).
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
indigoid
Super Deluxe Wiggler
Posts: 1252
Joined: Thu May 29, 2014 7:52 pm
Location: Sydney, Australia

Post by indigoid » Mon May 20, 2019 3:01 am

Dr. Sketch-n-Etch wrote:I really wish that Intellijel still made those picoVCA PCBs, which were 1-inch-square plug-in chips with a complete dual linearized VCA circuit on board. I coulda used a few of those for this, and they only cost a few dollars each to make.
I know you and SMD really don't get along but if Danjel would share the Gerbers and BOM with you, you could get a bunch professionally assembled... eg. with Seeed Studio's PCBA service. I doubt they stock 2164s but they could install all the other components for you, leaving you with a single SOIC (?) 2164 to solder. Not as bad as SMD passives, right?
Zap him again! ZAP THE SONOFABITCH AGAIN!

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Mon May 20, 2019 2:04 pm

indigoid wrote:
Dr. Sketch-n-Etch wrote:I really wish that Intellijel still made those picoVCA PCBs, which were 1-inch-square plug-in chips with a complete dual linearized VCA circuit on board. I coulda used a few of those for this, and they only cost a few dollars each to make.
I know you and SMD really don't get along but if Danjel would share the Gerbers and BOM with you, you could get a bunch professionally assembled... eg. with Seeed Studio's PCBA service. I doubt they stock 2164s but they could install all the other components for you, leaving you with a single SOIC (?) 2164 to solder. Not as bad as SMD passives, right?
Actually, I'd just have them made at the same factory where they were made originally -- Dena Technologies, where Intellijel HQ is located. I would just pay for the run out of my own pocket.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
BugBrand
Knowledge of Bugs
Posts: 7202
Joined: Tue Jul 15, 2008 11:59 am

Post by BugBrand » Tue May 21, 2019 1:46 am

The (?forthcoming?) AS3364 might be useful for multiple linear VCAs - it is a quad linearized VCA chip, though not sure if it is out yet & haven't delved into specs: http://www.alfarzpp.lv/eng/sc/AS3364.php

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Wed May 22, 2019 3:58 pm

BugBrand wrote:The (?forthcoming?) AS3364 might be useful for multiple linear VCAs - it is a quad linearized VCA chip, though not sure if it is out yet & haven't delved into specs: http://www.alfarzpp.lv/eng/sc/AS3364.php
Yes, that would be perfect for something like this. It would require four chips, and there is enough room (with a small bit of rearranging) to put these on the LP-HP-Outs-Noise board -- it probably wouldn't even have to get any longer. I could use the two spare VCAs to rig up a crossfader for the Even and Odd outputs.

I'm sure that lots of things would be voltage-controllified once those bad little boys come out.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sun Jun 16, 2019 1:08 am

Sorry I haven't posted anything lately. I'm very nearly ready to start building my massive morphing Polaris filter project (which will involve 12 PCBs plus the panel PCB, but will fit nicely behind a 2U panel). However, we have been having home renovations for the last two weeks, and my house is a total shambles -- the kitchen is in the garage, there are no floor coverings, most of the ceilings were destroyed (although today a bunch of drywall went up in the kitchen) and I've had to move almost everything we own into a box on the driveway. We're huddled into the family room, which was left mostly intact and which is separated from the rest of the house by a big sheet of plastic. Hence, my ability (and energy) to complete electronic projects has dwindled to virtually nil. However, I'm going to get something started next week, so please stay tuned.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Rex Coil 7
Super Deluxe Wiggler
Posts: 6274
Joined: Thu Nov 29, 2012 11:29 am
Location: Captain Of Outer Space

Post by Rex Coil 7 » Sun Jun 16, 2019 1:26 am

... good time to partake in that Canadian legal weed .... :drunkhomer:


:mrgreen:

:hihi:

:whistle:

:poke:

:hide:


:trampoline: ... kidding, of course ...

:tu:
Never Quit, Die Falling Forward
5U PROJECT - (skip pages 4 through 6, boring junk) ... https://www.muffwiggler.com/forum/viewt ... highlight=

User avatar
Dcramer
Super Deluxe Wiggler
Posts: 5401
Joined: Fri Mar 29, 2013 12:11 pm
Location: Canada

Post by Dcramer » Sun Jun 16, 2019 1:49 am

That morphing Polaris sounds killer :love:

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sun Jun 16, 2019 10:53 am

Dcramer wrote:That morphing Polaris sounds killer :love:
Well, it'll probably kill me, but that's hopefully not what you meant.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Moog$FooL$
Broke Dick Dog
Posts: 1698
Joined: Tue Jun 03, 2008 2:54 am
Location: Vancouver, BC.

Post by Moog$FooL$ » Sun Jun 16, 2019 11:56 am

how about pics of your latest (house) project???

:razz:
i use to drink & smoke.
that fat cat had to go..... wasn't even mine.

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Fri Jun 21, 2019 3:21 pm

No, too depressing. In any case, the drywall taper is here today and has taped all of the seams, so things don't look quite as disastrous as they did a few days ago.

I finally was able to start building things again yesterday. I bought a new drill press, and it is much nicer (and runs much more quietly) than my old one, but the damn chuck wouldn't take the tiny numbered bits I use to drill PCBs. After much consternation, I finally just swapped out the old chuck, so I was back in business. I managed to make the panel and the panel PCB for the crazy Polaris project. I also put all the parts in the PCB just to make sure that everything fits in the panel (it does). I've just finished cleaning up my workbench, and so now I'm going to solder up the panel. Before I could do that, I had to lift one fader off of the PCB because it was going to be too low, and the lever wouldn't stick out of the panel enough. I just stacked up some spare PCB material. I discovered that one can extend the pins of faders by forcing on machine socket pins. Then, I just drill a few holes in the PCB pieces and they went on nice and tight, and now the fader will be at a more appropriate height, and will sit nice and firm on the PCB. It would be nice if they made all that stuff the same height...

More soon. Stay tuned.
For every bullshit job, you need a bullshit education -- Brian Eno

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7617
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sun Jun 23, 2019 10:45 pm

Now I've made and stuffed the Polaris core boards (2) and, after looking at them and the panel board, I realized that I should probably do a translator board because the panel board has groups of connector pins (about 70) placed randomly all over, and there is no obvious way to attach the other boards to it. So, with a translator board which plugs into the panel board, I can rearrange the connections into a more rational arrangement and put on mounting holes to screw the other PCBs to it. I'm in the process of laying that out now. Once that's done, I'll be able to put the basic filter together (without the multimodes or the scanner) and test everything. I should get there tomorrow.

EDIT -- After staring at the layouts for a couple of hours, I now see that making a translator board for this system would be a lot more work than just wiring directly to the panel PCB pins. In both cases I'd still have to wire up all the pins -- they would just move to more "convenient" locations with a translator board. This is simply not enough extra convenience to justify the major effort of laying out this new board.

All I have to do is use a "stooge plate" -- this is like a stooge bracket, but its only function is to allow for the changing of positions of the mounting screws for the PCBs. The mounting screws on the PCBs are all in the corners, but these cannot be connected directly to the panel PCB because there is stuff in the way. Hence, I need to put the mounting screws on the panel PCB wherever they will fit between things, and as long as there are three screws, not in a straight line and not in the corners, then this will provide for stable mounting of the stooge plate to which the lowest circuit PCB can be attached. All I have to do is to mount the plate far enough away from the panel PCB that the female MTA connectors can be slipped onto the pins underneath the plate and the wires bent over with little or no stress -- about 3 cm. Ideally it will be far enough away that I can slip the connectors on without actually removing the plate or bending over the pins.
For every bullshit job, you need a bullshit education -- Brian Eno

Post Reply

Return to “Music Tech DIY”