another board

On a sunny day (Wed, 30 Jun 2010 13:50:12 -0700) it happened John Larkin wrote in :

We ordered the boards from Phase 3

who apparently brokered them to Allfavor. We ordered four boards, quick turn, electrically tested, 6 layers, screened both sides, 8/8 rules, for $175 each.

I guess it's cheaper to email the gerbers to China and fly the boards back, than to make them here.

That's a pretty good photo, for hand-held. The Allfavor bug is pretty small. That's not even the original full-res picture. The silkscreen reference designators are 60 mils high and are nice and sharp.

It seems to work so far. No assembly problems, just one complaint about a bottomside part that can't be glued so has to be hand soldered. The twelve isolated channel power supplies are fine, and we're talking jtag to the ARMs and blinking LEDs. We'll test the analog stuff today.

To power the 3 volt ADC reference bandgaps in each channel, I fudged the 3.3 volt supply up to 3.5 and just used a resistor. I did suggest a beta-limited PNP current source, to predictable squealing, but that would have been one more part for about the same performance. LM4040s aren't very picky about operating current.

I wrote the manual (21 pages so far, more to come) and a 16 page design notes document, so my software guy has all the input and math he needs to program the ARMs. We'll release the design doc to the library for future reference.

John

On a sunny day (Thu, 01 Jul 2010 09:44:48 -0700) it happened John Larkin wrote in :

That is a good price, and it helps to have a contact in your own country.

I had trouble reading the text on the IC packages, could hardly make out 'Xilinx' in the big one at top pright. Did you erase the text or all all those chips that way?

Does it run hot?

Yes sometimes he documentation is even as much work or more then the design, I remember one requirement from gov: All parts must be addressed in the docs.

I am impressed by the 'fill rate', I mean how much you got on that bord. Hope it works,

'Xilinx'

That's the way most chips are these days. I think it's laser marking on the epoxy. If the illumination is straight-on, the markings are just about invisible.

There are three SSRs, upper left, that actually have silkscreened part numbers, and whose markings are visible with most any illumination.

It shouldn't, at least until we get the 4-20 mA stuff running. Each channel can dissipate up to 700 mW of loop power. With just the CPUs and stuff running, everything is cool. VME crates have lots of fans.

One thing I do on a new board is feel all the parts to see if anything is warm.

It has parts on the back, too!

Yeah, it's always a little tense before it works, hoping there's no fatal blunder.

John

'Xilinx'

Yes, with my eyesight I need to photograph parts in order to read the marking, so I take several shots at different angles hoping to catch some contrast with the marking vs the package.

Yes, impressive, looks like a bought one ;)

No smoke so far, can't be too bad?

Grant.

'Xilinx'

Low-angle side illumination is usually best.

One old trick is to swipe the top of the part with a q-tip dipped in isopropyl alcohol. There's a brief flash of time, as the alcohol evaporates, when the markings are more visible.

Grrrr. Why can't they apply a coating to the top of the epoxy that laser-marks better? What would that cost, 1/100 of a cent?

And why can't they mark capacitors?

John