Skip to main content

Messages

This section allows you to view all Messages made by this member. Note that you can only see Messages made in areas you currently have access to.

Messages - teletypeguy

16
Tools of the trade / Re: High speed Pick & Place for $3600
Thanks selectech -- good to know that I can let a few wider pcbs hang out the right side!  That will help a lot, as I would rather have the smaller 220 and manually place a few bits on the wider ones.

I appreciate it.  Gotta get one soon.

Thx.  Gil
17
Tools of the trade / Re: High speed Pick & Place for $3600
Hey folks:

I have been lurking for a while and listening to the comments, and I am close to getting a machine, but I am conflicted about whether to get the cheaper/smaller/lighter 220 or the pricier/bigger/heavier 240.  My boards are under 5 inches deep, but I have some that are 10 to 12 inches wide.  So the 220A only does about 8.6 inches wide but that might be ok (placement-wise) since there are very few smt parts on the extreme right side of those boards.  So this leads to a question:  Can a pcb cantilever past the right-hand extreme edge of the machine, or is there a metal barrier that precludes such a thing?  If there is a limiting extremity, could it be easily milled down to allow a pcb to hang out the right side?  Of course, the machine could not place parts beyond its limits, but I just want to be able to mount a wider board as needed (and will manually place the few smt components that exceed the normal placement area).

So, can a wider-than-max board be mounted in a 220A?

thx,  gil
18
Tools of the trade / Re: Parallel Resistor Network Finder from Personal Stocklist
Hey James:

That is a very handy calculator.  I could have used it many years ago when I was designing active filters for a seismic data acquisition system.  Not only did we need specific resistor values, but component sensitivity analysis showed how some values had particularly critical tolerances.  Way back then, I wrote a fortran program that calculated all of the 2-resistor parallel combinations in the 1% resistor set (over x number of decades), sorted it all and printed the final table out on many, many feet of wide greenbar paper on the old lineprinter.  Crude, but got the job done, and saved me hours of punching numbers into my trusty HP-15 rpn calculator.  It was an FDNR anti-aliasing active filter design, which is a very cool technique using simulated-inductors, laplace transforms, 1/s stuff...  Ahh, you can just look it up if you are interested.  Very few folks seem to really know what analog circuitry to put before their ADC these days.

Now you can add a resistor calculator that does two-resistor voltage division, and then another that adds equivalent source impedance as well.  I recently spent a bit of calculator time figuring the best resistor values for converting a 3.3V logic signal to a 1V signal, with 75-ohms-equivalent-driving-impedance, for a digital-audio output.  But I suppose that is somewhat uncommon.

Anyway, thanks for the calculator -- handy indeed!

gil
19
Tools of the trade / Re: Any eagle library for 1.27mm connectors to 2.5mm connect
Yeah, I'm not sure quite what your question is either, as it sounds like you are just making a tiny board to breakout a smaller header to a bigger header, and well, a pcb doesn't get much easier than that, but I wanted to point out a little header trick I like.

I found this "sneaky-headers" note on sparkfun:
https://www.sparkfun.com/tutorials/114

And I thought to myself "how bloody obvious; why has nobody done this before."  And then I got a dev board from microchip that had the same offset trick on the ICSP header.  Huh!

So I tried it myself:  for a 6-pin 0.1" SIP ICSP header footprint, I offset alternating pins outwards by 5-mil.  I use 50-mil pads with 35-mil holes.  I put a 6-pin male-to-male 25-mil-square header strip in my Pickit3, and it bites nicely into the pads for programming;  no connector required.  Also will work great for retention during soldering of 25-mil-square SIP headers or sockets for an arduino shield or something, so they stay in place when you flip the board over for hand soldering, as the sparkfun fellow pointed out (some sparkfun sip sockets are not 25-mil-square, BTW, so be careful).

For dual-row 0.1" headers (10, 20, and 34 pin), I thought it overkill to offset every pin, as I just wanted them to sit still when the board is flipped for soldering, so I offset the two pads on each end sideways (lengthwise) by 5-mil, and it worked frickin' beautifully.  Header presses in easily, and stays in place when flipped.  Just offsetting the outer 4 pads was enough -- offsetting all of them (in the shorter dimension) would have made larger headers harder to insert.

This is all at the mercy of how accurate your pcb house is in maintaining the specified interior diameter of a finished plated hole, which can vary a lot, so we generally use a 35-mil or so finished hole for a 25-mil-square pin (with a diagonal of 35 mil), and it is often loose.  Tweaking a few pads fixes that, not just to help with hand soldering, but also for retention when a board is run over a wave.

gil
20
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
Hey Peter:

Sheesh, any improvements to speed things up would be awesome, of course.  But we're all getting some very nice, accurate, stencils, and we are all real happy just to improve our prototyping workflow and move one more step in-house. 

You really got something started here, and lots of folks, well, all around the world, have gotten a big-old-grin on their face when their first stencil came off the cameo -- which, I imagine, put a big-old-grin on your face too :)

And, I now have a reason to learn more about python (and not just monty python, which has a classic syntax in which I am well-versed).  Python looks like a fun thing to get my 13-yo son into as well -- this summer is his first exposure to electronics and programming, as I am teaching a little class for him, and some buddies, using the simple-yet-powerful arduino (and some unusual devices, like a 1920s candlestick dial phone for digit-input to the arduino, using a pulse-counting state machine, while looking at the signal on a 2012 logic analyzer).  And then Sparkfun rolled through town, on their National Tour, getting the local kids even more excited and ready for robotics clubs in the fall.  All-in-all, it's been a great summer for electrical crapola so far.

Anyway, thanks much for getting us all moving in a good direction Peter,

gil
21
Tools of the trade / Re: Group buy in Shenzhen?
Hey mplafleur:

Which machine did you get?  How much did your machine end up costing on aliexpress, with extra fees and shipping etc?

I am still on the fence, and was waiting to see what seeed would do, but they have not offered them yet.

gil
22
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
One option to minimize the travel time after training cuts, and to speed up in general, is to scatter training cuts around the perimeter of the board, just outside the final boundary.  Make a training cut, then cut a half-dozen nearby cuts in that orientation, train another orientation, keep cutting, and just work around the board.

gil
23
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
Hey Jesus:

Wow, a gui front end -- I can't test it yet, as the cameo is at the office, but that's frickin-awesome.  Thanks!

It seems to only want to open .txt files, but it's easy enough for me to change xyz.spt to xyz.spt.txt.

And thanks to lots of other folks for the optimize stuff, tests of different materials and so on.  We finally figured this machine out pretty well.  Sheesh, this thread got a lot bigger than I expected.  A lot of folks are buying machines and cutting stencils now!  We should all be getting commissions from the cameo folks -- and sending them to Peter!

One note on cut quality:  while it is, of course, important to get reasonably-square, straight clean cuts, and even more important to get very-accurate positioning over the entire board area, I really don't feel the need to obsess over perfectly-clean cut corners and stuff, because, quite frankly, the paste will smear and ooze anyway, and even short-out pads -- but the good thing is that during reflow the solder sucks up nicely onto the pads and only leaves the occasional short where the paste was heavy.

Anyway, a big thanks to everyone for all your contributions.

gil
24
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
Hi reserve:

If you go back to page 1 of this thread, and scroll down to the Apr-3 post by jesuscf, the command lines to type are there -- that post of his has all the details.  Open a command window, cd to the proper directory, and type in the commands (if you have everything named the same as he did). 

Jesus had his files all in one place, but I put g2g in a subfolder, shortened the name, and wrote a little batch file so I could just double-click it and type in the filename of the gerber (plus remind myself to set blade depth etc).

You can still test cut a file with a bit of cardstock or file-folder paper.  Might want blade depth of 2.  Make a REALLY SMALL test gerber with just a few parts.

gil
25
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
Hey Jesus:

The little eyedropper-like gizmo I got for manual pick up is many years old -- forgot where I got it, but it might have a name on it -- I'll look when I get in on Tuesday.  Kinda similar to this very-expensive tool (almost two dollars!) on ebay:
http://www.ebay.com/itm/190776608479
or, an even-more-expensive variant:
http://www.ebay.com/itm/330566776703

Here is another tool I like a lot:
http://www.circuitspecialists.com/csi825a.html
It is a hot-air rework station ($109) that has a hot-air gun and a few nozzles, and also includes a vacuum pickup wand with a few tips.  I have not used the vacuum pickup part, as my little eyedropper pickup works quite well.  The hot-air section is great for both removing chips and re-flowing a replacement chip.  It comes with some wire doo-hickey to fit under the part when you are trying to remove it, but that didn't really fit so I don't use it.

For chip removal, I just put the blade of an xacto under the chip, apply a SMALL amount of torque for lifting force, and start heating the snot out of the chip (I set the air to about 320C for leaded solder and 370C for lead-free).  It only takes a minute or so before the chip pops off the board.  If you apply too much lifting force, you will either crack and split the chip, or lift pads, or both (and then you will call yourself an f-ing asshole, and be really pissed that you ruined a board), so just be gentle and wait for the chip to release nicely.  Practice on scrap boards.  Clean the pads with solderwick when the chip is off.  Get some good solderwick from mouser or digikey (I got some stuff from dx.com that was complete crap and I threw it away).

For chip replacement, I made some solderpaste chip stencils on the cameo out of that adhesive vinyl that I used for chassis overlays -- first results were poor, as I just cut the chip patterns as I would for an original mylar stencil.  The little chads got stuck everywhere and the mask was unusable.  So next, before I peeled the cut stencil from the backing, I used an xacto to cut a rectangle out for the array of pads on each side of the chip.  Sheesh, I can't count the number of times a day I use an xacto knife -- what a handy tool.  With this "rectangular-pin-array" mask, I spread paste over this larger area which included paste between pins.  I got a plastic putty knife from home depot and cut it with scissors to make a nice little squeegee for just fitting the chip pins.  I was able to use the hot-air tool to reflow it and it worked pretty well, except for some shorts between pins (which I removed with solder wick). 

For my next chip replacement time, I plan to make some special rework stencils that provide a rectangular area over the pins, but only on the OUTER-HALF rectangular area of the pins -- this will reduce the total amount of paste and it will hopefully reflow without shorts.  Will keep you all advised.  I think I will still use that adhesive vinyl -- even though it is a bit "stretchy," I think that adhesive is important for keeping the solderpaste contained well.  The stretchy adhesive vinyl is also good for fitting into a chip area that has adjacent components sticking up (a flat mylar stencil would never work).

So why did I have to replace an smt chip recently?  Well, sometimes I have a need to rebuild a board with a dead micro or something, but here is a recent scenario you do not want to repeat:  I had a new proto, freshly-built, fully-tested, a code-coming-to-life sort of board sitting on my bench.  My new golden board -- you know, one of those special boards that seems to talk to you, giving you little insights along the way as you tweak code and such, helping you get it running.  One of those magical, special first boards.  Or maybe that's just me, but anyway, for me, a scenario not unlike many, many over that last thirty years.  And then I did a stupid, stupid thing:  I forgot that I had the board powered-up on the bench as I went to solder a cable onto an lcd module or something -- the solder fell across the board, hitting headers and other stuff.  There was an audible arcing noise -- that's never a good thing.  And there was that smell of burning electronic-something -- that's never a good thing either.  No, not something as simple as a backwards electrolytic cap that blew its load;  this was more insidious.  I realized that I had let my solder fall across the board, probably connecting 12V, 5V, 3.3V and various logic-level signals in an orgy of gate-oxide-vaporizing, protective-diode-frying, silicon-melting destruction, the likes of which I had not seen in many years.  It was nasty, but the survivors, who should get honorable mention, were an Intersil HIN202, a TI 74LV08A, and an Atmel AT25256 -- on the destroyed side were an AT45DB45081 flash, a PIC micro, a 25AA mac/eeprom, a 23xx sram, an ENC28J60 ethernet chip, and an NXP 74HC04.  It was just sad.  But with a little bit of hot air, an xacto knife, and some good solderwick, my board came back to life.

ps:  when I got the hot-air tool, I also picked up these extra nozzles, all of which have proven useful:
A1125 - for QFP-44 etc (10mm x 10mm)
A1131 - for SOIC/SOP (4.4mm x 10mm)
A1132 - for SOIC/SOP (5.6mm x 13 mm)
A1260 - for SOIC/SOP (8.6mm x 18mm)

FWIW,

gil
26
Tools of the trade / Re: Making Chassis Overlays with a Cameo
Etching circuit boards?  Really, don't waste your time.  I say that as an old fart who actually did etch a lot of my own boards.  In fact, my first board layouts were on a light table at 2x size with sticky dip patterns and black tape (and my first programming class was fortran with hollerith punch cards).  A few of us got together and copied the Apple ][ PCB this way (late 70s) and made a dozen or so before the clones came out (it was easy to copy the apple proms).  Shot the 2x artwork with a big graphic camera to 1x negative for contact printing to photosensitive copperclad.

Then things got high tech with computer layout, but it still went to 1x negative, expose, develop, etch...  Nasty crap.  And even then, you needed to plan ahead so you could solder pins on both sides of the board, or put eyelets in -- and the resolution is crap too.

My current fav board shop is http://www.myropcb.com -- do a quick online quote for 2, 4, 6-sided or whatever boards, with nice 8-mil or better mins, 15-mil or smaller holes -- office in Canada, fab in China.  Quality has been superb.  Gotta convert to millimeters for some stuff, but I am grudgingly getting more used to mm.

Even on small quantities, Myro lets you spec a laminate and stackup.  So when I wanted controlled-impedance (without paying the extra for it), I could spec the final rohs laminate I'd be using, and the prepreg and core thicknesses, so I could layout for specific impedances (with appropriate trace widths, for those who may have never thought of a pcb as the electromagnetic component that it is).  I did test trace coupons for 50-ohm, 75-ohm, 90-ohm-diff, 100--ohm-diff, etc -- the boards came back from myro and the impedances were very close on my HP TDR.

Here's a little QC tip if you are specing a stackup and a wondering whether they actually did it or just paneled you into some other guy's generic 4-layer job -- put a little triangle or square on each layer at one of the board corners.  When the board comes back, file a bit off the corner to expose the copper, get out a good magnifier and look to see if the prepreg really looks to be 9.5 mil or so (or whatever you spec'd).  I made a little "ruler" with a pcb scrap that had 8-mil-traces and 7-mil spaces.

gil
27
Tools of the trade / Re: Making Chassis Overlays with a Cameo
That's nice royco -- I want to experiment with color overlays next.  I like the idea of the clear on top as well.  Have to get some.  What was the base material on which you printed?

What is that plastic enclosure?  Ya, know, the one thing we need from sources like seeed and sparkfun is a variety of decent-looking, reasonablly-priced plastic enclosures.  Pactec's line is pretty dated-looking and pricey.  There are some options from Hammond and others, but I would like to find any other sources anyone has to recommend.

thx,, gil
28
Tools of the trade / Re: Making Chassis Overlays with a Cameo
The inkjet material I used is made by Papilo, and I got it at http://www.texascraft.com.

It is a "white, waterproof, adhesive-backed vinyl, inkjet-printable, and must have some UV protection as well.
I measured it to be about 4-mils thick, including adhesive.  After I printed a sheet, I adhered it to a second sheet before cutting -- this gave me a final overlay that was about 8-mils thick, which works better for me covering unused holes.  I used a blade depth of 3 to cut this 8-mil final thickness.

It is a "permanent" (not removable) adhesive, though it is quite removable and re-positionable for a while.  It sets up with time like lots of permanent adhesives.

I DID NOT get the "contouring" vinyl, which is thinner so it can be applied to non-flat surfaces.

WVF1117KH  11x17 sheets, $1.44 ea in 100s
WVF1117G  11x17 sheets, $2.38 ea in 10s

WVF8511KH  8.5x11 $0.72 each in 100s
WVF8511G  8.5x11 $1.19 each in 10s
WVF8511S  8.5x11 $1.62 for one.

RCWVF1230 is 12"-wide by 30-foot roll (if you have an inkjet that can take it.

gil
30
Tools of the trade / Re: Cutting mylar stencils with a silohuette cameo
Hey Jerry:  Glad to see you are getting some good cuts now.  Not sure why blade depth of 1 did not penetrate for you, but maybe it is related to your mylar -- jesuscf had Canon mylar that was the same thickness as Apollo stuff, but was apparently a tougher plastic and did not cut well.

You can also use the sil-studio program to test material, even little scraps.  Load a bit of test material (anywhere on the mat really), fire up sil-studio, and in the send-to-cameo dialog there is a button for a test cut.  Use the blue cursor buttons on the cameo to move the knife to where you want to cut, and click test cut.  It cuts out a little square with a triangle inside -- you can move the head out of the way (or unload the mat) and dig out the square with an xacto, or peel of your samlple.  This is a good way to try different blade depths for different materials.

Are you sure you shrunk your cream layer?  I thought I had at first but got the same results -- then I used gcprevue, imported just the top-copper and solderpaste layers (which gcprevue assigns different colors).  You want the layer list on the left to have paste above copper (if not drag one layer above/below the other).  Zoom in and then you can see clearly whether your paste is shrunken (?), shrinked (?), ahh, reduced-in-size, as you expect.

So, is this little round board going in the nosecone of a rocket?

gil