Homebrew PCBs

On Sun, 15 May 2005 14:25:35 -0400, Alan King <alan@...> wrote:

>   Sizzling almost guarantees you still have liquid turning to gas and  
> escaping
> through the other parts of the material.  Dry solids just melt.   
> Silicone is
> relatively inert so unlikely it even has a melting point that low, I'd  
> find it
> very hard to believe that the gas temp is anywhere near toner melting  
> temp.
> Melting hardly expands anything, almost has to be gasses if you're  
> hearing
> sounds, which implies remaining liquid.
>

Ok, here what i see:
Some liquid is created, which i think also sizzles.
When i remove the board quickly the liquid is still there, and it can be  
used to wipe partially toner off the board like a oil or solvent. i dunno  
what it is, it is not there with high-temp silicone, no difference how  
long you let cure, so i do assume it comes from non-high temp silicone.

>> Do not put silicone on the first 7-10cm of the page or you will from  
>> then
>> on have a paper coated drum.
>> (By the way, drums and light-sensitivity, how bad is it really? which
>> precautions must be taken when working with drums?)
>>
>   LOL Been there done that.  If it's even sticky enough to do this  
> though I
> think it's still maybe too sticky.  If the only reason it's coming off  
> the drum
> is because that first 7-10 cm is pulling it off with other rollers, then  
> little
> bits are still probably sticking to the drum.  May be a livable  
> situation though
> and just consider a new cartridge now and then a part of the ease of  
> transfer,
> it'd be $20 well spent for perfect transfers with little work.  I can  
> make mine
> near perfect with the other papers, but the process is more tedious and I
> usually say screw it and just touch up any problems.

I do not agree it is _still_ to sticky. it is too sticky, without still,  
it will not get better.
Note i do not use uncured silicone, only fully cured. Silicone stays  
sticky even when fully cured.


>    And the answer for light is very, very sensitive.  600X600 is 33  
> million
> pixels per page.  So 10 seconds per page that laser is sweeping about  
> 3+ dots
> per microsecond.  Fractions of millionths of a second per pixel, times  
> the 2500
> or so pages the cartridge is rated for, that's all the light that's  
> supposed to
> hit the drum.  Any exposure to light degrades the whole drum evenly  
> though, and
> I think there is some recovery effect too.  The lasers are red or  
> infrared, so
> that's what the drum works off of.  The color it mainly reflects (green  
> or blue
> depending on type) should be more safe at low intensities than just  
> white.  I've
> thought about making a green and blue single LEDs safelight for working  
> with them.
>    I think the degrading though is not 'black image' but leaky drum so  
> less dark
> image from less attraction.  Since it only takes a tiny bit of plastic to
> protect the copper I doubt a poorer drum that makes lighter prints would  
> pose a
> problem really, whereas for prints you can't make your blacks look dark  
> enough.
>   Still best to avoid of course.

Yes, best to avoid. the lexmark has a uncovered drum, and it can be  
installed by stupid people, so i assume normal, short, low intensity light  
will not damage it.

>   The coatings are organic, so organic solvents are a no-no.  I'm trying  
> to
> find the inorganic solvent they use to clean them, it should let me  
> recover my
> wrapped drum since what's remaining isn't stuck too badly.  Had a friend  
> in a
> copier shop that knew all the details, but haven't gotten in touch with  
> him yet,
> and I don't remember which solvent was for what they also stripped  
> typewriter
> drums etc.
>

Let me know what solvent you find. I have so far only needed a dry paper  
tissue, since i wasn't quite silly enough to put in uncured silicone ;-)

>> I have a problem with creases forming on the last few cm of a page, but

>   Sounds like the back end of the paper is no longer in rollers, and  
> this lets
> it shrink.  Maybe try a legal sheet with another 7-10cm on the end with  
> no coating..

I do not have legal paper, but yes, for smaller boards it is not an issue.  
I only noticed because of the big geometry test board.
I do not think it is because there are no rollers, the printer is pretty  
badly made and there is no wide roller before the drum, like on a HP, just  
a small transport to the left. The creases are there at the drum, 'cause  
there is no toner in them. they are not really creases (folded/crumpled),  
more like vallies in the paper, bent down in a u-shape, and not touching  
the drum. Maybe it will go away when i sort out the feed skew. i think the  
left edge runs against the guide, and the silicone is too sticky to allow  
the page to turn slightly halfway in.

>   What printer are you using?  There is a small amount of distortion in  
> my 6L's
> but it's so small I don't worry about it much.  Because of the rotating  
> to
> linear conversion it'd be hard to be 100%, but it should be possible to  
> map the
> distortion and precompensate the original before printing.

A lexmark, and yes, i can compensate in X and Y, but the right Y is  
different from the left Y. probably because of feed skew and laser  
misalignent, the service manual (such a crap, not even a schematic),  
describes how to fix it. I must move the printer from under the table to  
on it before tho, and the table is full of ****, you know the story.


> And what for the
> CNC?  If you're just talking drilling alignment etc, then use two  
> opposite
> corner holes and manually locate them.  From where they are to where  
> they should
> be will tell you exactly what your scaling error in X and Y is, far  
> easier to
> compensate by math for the drill hole locations and simply match what  
> your
> printer puts out.  The absolute error is small, doesn't affect any  
> components
> I've seen.
>   Of course if you have to have 100% exact dimensions then it much be
> precompensated.  Just for a general circuit board though a percent or  
> even a few
> doesn't matter as long as the holes are matched..
>   And with 3 corners (two opposite and one of the others) you can get  
> accurate
> scaling and rotation information.  Align the three manually for their  
> actual
> locations and you can find all the other holes without even aligning to  
> be
> parallel to the axes.  Some basic trig is all it takes..  But it's so  
> easy to
> move side to side and use two holes to be exactly aligned to X, then go  
> up to
> the corner and get scaling, that I only do the 2 corner scaling and  
> align to
> parallel manually.  LOL really just lazy and haven't written the code  
> for 3
> point with rotation, wouldn't take 15 minutes to do it.

you. me it would take a month. I fear i can not use such code and must  
compensate before print.

>   I am starting to work on the CNC again some for cutting more than  
> drilling,
> so maybe I'll write it up soon.  Hardly even used the CNC for anything  
> since I
> started just making everything single sided and mount all DIPs etc live  
> bug
> style without any holes.  So much faster to just not drill anything at  
> all.  But
> nibbling out circular boards is a pita, so time to get it back up for my  
> board
> routing.
> Alan

Customers do not like to use bug style.
Use sheetmetal shears for round boards.


ST