Follow @contourcorsets
 

Subscribe to Fran on YouTube:

Photo positive artwork for exposing Brooklyn PCB's.

I always prototyped my designs on a regular perf breadboard, and rendered the schematics and circuit layout in Protel.  Circuit Auto-Routing?  I think not.  All of my circuit board layouts and trace routing was done manually, using templates I made in Protel to establish the proper spacing and tolerances for the pads, traces, and lead spacing.  But to extract the design from Protel meant that I had to print out the PCB file to paper, and to make my optical plates to expose sets of PCB's meant that I had to scan the rendered paper artwork at 4x scale, then arrange multiples of the artwork with proper scaling corrections in a photo program to render photographic plates that I would use to UV expose positive resist 4X6 inch bare copper FR4 boards.  In the later years as technology improved I switched to a thermal transfer process to transfer resist to the bare copper in a converted heat press. 

Bubble, bubble, toil and trouble...  Etching!

My ventilated work room, 2008

For etching I used a special ventilated workstation that I also used as my silk screening table.  There I would set up my etching tank that contained an air bubbler system that I made with an aquarium pump and a submerged aquarium heater that would circulate and heat a solution of sodium persulphate to 120°f.  Ventilation and a chemical respirator are necessary to work with this stuff because of the outgassing of sulfuric acid during the etching process, which is quite harmful to breathe.  I could etch up to five 4X6" FR4 plates at one time, and 10-20 plates would make a typical manufacturing run, yielding between 40-100 individual effect PCB's. 

But these etched boards needed to be drilled, and here is where it gets difficult.  For starters I bought a jewelers drill press, which out of the box was intended to be for this kind of work, but it proved very inadequate.  The press was too sloppy and quickly overheated.  I created a forced air cooling system that used a CPU fan to pressurize the motor casing, and I drilled vent holes above the motor to allow the hot air out.  Then I had to accurize the drill, because I was using 30mil high speed tungsten carbide bits that were inherently prone to shatter if the center of rotation deviated more than 4mils.  To get this accuracy it was necessary to balance the chuck with hand filing, then to center the chuck by fitting it to the drill cone using flakes of aluminum foil as shimming material.  With some time and care I had accurized the press to within 2mil (.002") of center deviation, and now it was ready.  I constructed a HEPA filtered vent hood housing for the press within which I could work, to create the minimum amount of hazard.  I still had to wear a respirator while drilling, as the constant inhalation of FR4 dust is very bad for the inner parts of the lungs, causing a form of silicosis with long term exposure, or in the case of sudden high level exposure, can cause sudden death by cardiopulmonary edema.  Nasty stuff.  But with safety measures in place I also had to employ another aquarium pump that I used to blow air to the drilling point with a plastic tube, which would be necessary to clear the dust off of the drill dies so that I could see what I was doing.

To drill the PCB's in volume I invented a die system whereby every board had its own drill die that I made by hand.  The die consisted of a thick aluminum plate that I machined down to a flat smooth surface, with registration pins that aligned a stack of PCB's, and contained enlarged recessed holes for each drill register into which the drill bit would plunge after going through the bottom board of the PCB stack.  Each individual circuit board was cut out of the larger etched plates, and on every board I had selected 4 holes that would be used as registration points through which the registration pins would go to center each board in the drill die.  I had to line up and hand drill each of the four registration holes on every board to start a run.  The boards could be stacked up to 7 high in each drill die, and then on top of the stack was placed a brass drill guide plate that contained the drill register for the PCB, and tightening nuts to hold-down screws mounted in the base of the die created a firm aligned stack of boards ready to be drilled.  With a few bit changes a stack of boards could be drilled pretty fast, even with upwards of 100-200 holes per board.

Depth control was necessary to insure that when the bit plunged through the bottom of the stack into the drill recess that it stopped short of hitting the bottom of the hole.  If the bit were to bottom out - like when I would forget to set it - the bit would shatter, creating a big problem with carbide fragments embedded in the boards. 

PCB assembly station, 2008.

After the PCB's were drilled I stored them in plastic zip bags to prevent any patina from forming on the copper.  I also wore nitrile gloves whenever handling the bare copper boards to keep my prints off as well.  The next stop was the filling station, where I had a board assembly flip-jig in which I could place any number of boards to fill, flip over, and solder simultaneously.  A hold down pad on the top kept components in place while soldering. 

When the boards were completely filled and wired they went back to the vent room to get a bath in methyl alcohol, which dissolved all of the resin flux.  After drying a white powdery film of resin would remain, which was taken off with a dry toothbrush.  Then the clean PCB's headed to the paint room where they were conformal coated.  This added an insulating layer which also prevented oxidation.  The PCB's were now completed and ready for installation.

A run of completed mixed PCB boards.

A completed Sandwich PCB.  The black part behind the LED is my proprietary precision optocoupler, also manufactured in house.

Completed Sweet PCB's.

So, why did I do it this way when anyone can outsource PCB manufacturing and get great results?  Well, mostly because it was cheaper and I could make changes at any time - which often happened when a component would change size, or a revision of the circuit was required.  It was also part of the philosophy of complete vertical integration that I established for Frantone in 1999.  I machined, painted, and silk screened every case in house - and every PCB was also designed, etched, drilled, and filled by hand.   I wanted every Frantone pedal to be exactly what it was said to be: Completely Made By Hand, With Love.  :) 

Another cool project in the works is a converter for Soviet Vacuum Fluorescent Displays like this one on my bench:

 These larger USSR made displays use a lower filament voltage (+1.5VDC) and higher grid voltage (+25VDC) than most VFD's you would find, which makes using common 12V VFD clock drivers impossible.  I will have to make up an array of 30V PNP transistors for a step up display driver stage.  A lot of work, but ain't they perdy??!

Here is a handy schematic for my USSR-VFD power supply which you can customize for any VFD display by changing the R1 and R2 values on each LM317 voltage regulator to get the desired output you require for the filament and grid supplies.  I used a 115v/24v center tapped 2amp transformer, and keep in mind that these tubes do draw some current so best to give a good 100ma of overhead in the supply for each tube you want to drive.  This supply has a 1.5v regulated filament supply, but if you are using multiple tubes you can also stack the filaments in series to divide a higher combined supply voltage (example: 5X 1.2v in series = 6v supply)

This is the start of a side project I have been doing in my spare time, in my new and still coming together science lab.  I wanted to create a universal 4-bit controlled nixie display driver that could be used as a module for any application.  Here I have breadboarded an automatic counting circuit with a clock, and also a separate high voltage board for the nixie display that I have hooked up to a manual rotary switch for demonstration.  The ultimate goal is to reduce this down to a single condensed board.

I leave out some details in this brief video demonstration, but I will post the schematics and technical data for the project when I get it all together. 

Update: Phase 2!
Here I demonstrate the completed breadboard, with power supplies and the interface for the logic to the high voltage Nixie displays.

This is a 'foxhole radio' that I made years ago for an R. Lee Ermey challenge in 2003 to show that in fact you can make a radio receiver from common stuff - that is, so long as you have a crystal earset lying around.  The razor should be blued for greatest diode effect but they do not make blued razors anymore, so I substituted a modern hardened steel blade.  The graphite pencil point acts as the anode and you move it around the the razor to find the point of contact where the diode is detecting the best.  You tune the circuit by adjusting a wire which you tie to earth ground to contact on the coil, where the insulation has been shaved away - presumably by the razor before you put it in your now fully functional radio!

When I worked at an AM radio station in my 20's the bathroom was almost right under the broadcast antenna tower, so I made a modern tuned diode detector with a tiny speaker that I hung on the wall in there, and it played the station out loud 24/7 with no electricity.  Just the power of the rectified radio waves - which was at a strength of about 10 volts per cubic foot in the bathroom - and a good earth ground on the cold water pipe was sufficient.  There is also the story of the oven in a trailer park that played country music near another radio transmitter I worked at that was local legend..... and true!  Tuned resonance at work.

Part one of making the Frantone Fretboard.  The audio is bad in this video due to  the noise canceling feature of my old webcam. But you can see what I am doing.

I recently completed my second kick press project.  I bought a used industrial sewing table which I customized to mount my hand press to it, then constructed the foot pedal with the necessary leverage to amplify the force of the hand press from 10:1 to 18:1.  This new kick press delivers 650lbs of force to the die with just 36lbs of pressure on the pedal, and all completely hands free. 

I then wanted to make this kick press more than just a grommet press, so I designed a series of practical die sets for other press operations that I do frequently.  I contacted my good friend Bud Mohrman at TAPE Inc. to make the dies for me, and fortunately he is one of the last hard core machinists that can cut dies directly from high strength tool steel.

I made some initial sketches from my mental images to better visualize the manufacturing tolerances for these dies.....

I then made a series of mechanical drawings that contained my calculations for tolerance and proper contouring, finish, and measurements for each die.  Mechanical drafting is a specific language which communicates the design concept in a way that another person can manufacture the part to exact proportions and have the parts interact with proper tolerances.  The Machinist will not make judgment calls, it is up to the designer to be very clear on all parameters.  If the designer makes a mistake or miscalculation then the part will either not fit, or not function as needed.  These drawings were made with a requested manufacturing tolerance of ±.005 inches.  No computers here, I do everything in pencil on paper at my drafting table....

The results were very good, and Bud did a magnificent job.  There were a few small tweaks, and the dies work great, as expected.  Here are the actual dies....