Electronics > Projects, Designs, and Technical Stuff
Home Brew Analog Computer System
GK:
Hmmm... The whole tin plate idea is looking less and less attractive. I'm probably being a bit paranoid about PCB leakage current in my integrators anyway and would be hardly worse off just using a conformable coat. JFET leakage currents will likely dominate anyway, and you shouldn't expect the computer to hold your solution in the "Hold" mode with the utmost stability over extended periods (eg. while going off to lunch) anyway.
That said, now just thinking aloud (way too late at night), how much luck would I have evenly/uniformly solder-coating a bare copper board if I melted a hefty roll of solder in a frying pan over a gas burner and somehow just skimmed/floated the board over the surface?
notsob:
If you are looking at plating systems and supplies, I recently found these guys when looking for liquid electrical insulation (look at their plasti dip range), anyway I haven't used them but it may be worth contacting them.
http://www.caswellplating.com.au/store/store.php
mamalala:
--- Quote from: GK on February 18, 2013, 12:39:47 pm ---That said, now just thinking aloud (way too late at night), how much luck would I have evenly/uniformly solder-coating a bare copper board if I melted a hefty roll of solder in a frying pan over a gas burner and somehow just skimmed/floated the board over the surface?
--- End quote ---
Virtually none. You will always get an uneven solder coating that way. Just look at consumer electronics that is assembled using a solder-bath. There is a reason why you almost never see big chunks of copper plane solder coated that way there. They either go fo a "grid" type copper-pour instead of a solid copper plane, or they restrict the available tinable area with a solder mask.
If anything, you have to recreate the HAL process to get a somewhat even surface. This is basically just applying the molten solder to the PCB, and then using hot air to "level" (i.e. blow away) the excess solder. HAL = Hot Air Leveling.
So, if you don't care about an even surface, you can do it the way you proposed. Otherwise that method is good only for etched boards, i.e. when tinning only the traces instead of a whole bare copper area.
So, i would think that conformal coating would probably the best way to go. Cleaning in an ultrasonic bath, with suitable solvent. Maybe then dry the boards in an oven for a few hours before coating, to make sure that no moisture is trapped under the coating to be applied.
How were the boards "finished" in the commercial analogue computers? After all, these machines are a rather old thing.
Greetings,
Chris
GK:
--- Quote from: GK on January 10, 2013, 01:09:04 pm ---I vaguely recall seeing it done as you describe but couldn't say if it was Pease or Williams.
--- End quote ---
It was Williams:
http://cds.linear.com/docs/en/application-note/an124f.pdf
bottle cap was epoxy filled too.
I'm almost finished loading/testing the Time Base PCB. After that I'll be doing the Log Amp boards. I've already collected 10 coke bottle caps. I think I'll drill a ~5mm hole in to the top of each and inject araldite with a syringe.
GK:
The Timebase board is slowly nearing completion. I decided in the end that I wanted much more comprehensive sweep waveform (x-axis) output options and provisions for re-trace blanking (z-modulation) for use when viewing problem solutions on an X-Y oscilloscope in the repetitive operation mode. These functions will be performed on a separate circuit board. As a consequence I modified the circuit a bit. The revised schematic is attached.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version