Just got new carpet... found out that the previous install was so cheap that they used something like furniture foam..... \

@wisemonkey ASIC and HDL questions are fair game here, but there's not a whole lot of traffic in those areas. I answer more ASIC centered questions myself

@W5VO yup looks like it in any case answered 2 of them :)

Hmn

Hmn?

HAHAH

meep?

@wisemonkey Sure, that would be welcome! You seemed to have missed our verilog tag - You're right, we're mostly about microcontrollers and C, but we welcome questions about HDLs.

Can you help out with any of these unanswered Verilog questions? Most of them are answered, but these few aren't.

You will find, too, that our users mostly have experience in FPGAs. You're welcome to ask about ASICs, but there's not a lot of experience in that area.

Excepting, of course, W5VO :)

Nearing completion of my first four layer board! i.imgur.com/Hp4Cu.png

@saad whats it do?

@saad There's always something very satisfying about seeing well ordered fan-outs from chips

@saad Let's see...An AVR ISP header, LCD and uSD connectors, a driver and receiver each connected to two 37-pin connectors...guessing it's a handheld cable/wiring tester?

Cheating and googling for 'max v' as seen in the JTAG header...looks like those are spoiler - hover for result Am I close?

BTW - To get a hoverable tooltip, use the normal mini-markdown link format, but add a space and your tooltip in quotes after the url: [displayed link text](http://example.com "tooltip, aka title")

You're on top of it, Kevin! As a newbie, I find it unbelievable that you guessed it! Says a lot about your experience. :)

@W5VO Yes, I think so too!

@saad Working in test and measurement doesn't hurt in my chances on this one. :)

lol

its pretty amazing to me as well though, i had no idea

Just a tip - if you're going to be giving this to line workers to handle repetitively, those connectors and solder joints on your PCB won't hold up from day-in, day-out stresses. You'll want to put it in a box, and get some nice, beefy panel-mount connectors for that. It's OK to design the PCB to use these connectors, this way you can use just the PCB mount and avoid the box and cable expense if you integrate this into an enclosed panel on an automated tester.

We use CPC connectors [pdf] for this; they're remarkably sturdy.

The connectors are my single biggest gripe about this! I had talked to my boss and and he agreed that we ought to find new ones. Looking at digikey and newark I got very lost! And most of the connectors were quite expensive! I'm gonna discuss your suggestion tomorrow morning with my boss. They're a much better choice for this.

@saad Glad to help! There's a lot of connectors out there, it's easy to get confused as to what you're actually getting.

@KevinVermeer Yes. TE Connectivity's website has 3D drawings. They help a lot to see what it actually looks like. I guess cos I'm new, the 2D drawing don't give me that much perspective. Their picture search is very good too!

@saad I also like Digikey and Mouser's adjacent images.

It's just an instant feedback on whether you're looking at something close.

@saad - What's your power supply look like? I see a pair of huge, expensive-looking caps, but no regulator circuitry...

@KevinVermeer We're thinking of keeping the PSU separate. All linear regulators though.

Is there any inherent disadvantage of keeping the PSU separate?

@saad With this design, you need like $3 in caps to filter out whatever might possibly come down the line. A $1 regulator with some heatsinking (you've got 4 layers and extra space, this is easy) can filter just the contents of the daughter card, regardless of what input noise you're seeing.

With an onboard regulator, you can supply 24V industrial power, unregulated input from a wall wart, battery power, or whatever you like, and not worry much about line noise. A 22uF SMD electrolytic 6.3V cap ($0.10), 1uF and 0.1uF ceramics (basically free) and a decent regulator handle the decoupling for low-power stuff like this board pretty easily.

@KevinVermeer Point taken. I assume jelly-bean chips like 7805 will be OK? On a further note, this board is designed to be extendable. As in, there will be daughter cards that can extend this 72-point tester into 72*n where n is the number of boards - max 5. I assume those boards should have their own regulators as well and they should receive power from the wall-wart as well?

Oh btw, i didn't realize you said "portable tester". It's not really portable, though I suppose it's small enough to be but that's not the goal. They'll be put in a box on a table that has connectors for automotive harnesses. Naturally, there will be a mating harness that connects those connectors to the PCB.

@saad Yeah, jelly-bean chips like the 7805 will be fine if you've got them, but newer parts have lower dropout and better transient properties and don't cost much more.

@saad Ah, I was interpolating from the use of the onboard LCD. I assumed that if it was permanent, there'd be an adjacent PC and monitor to display the info on; that's how most of my setups work.

In that case, this board would, I suppose, have an external USB, Ethernet, or serial interface and an API called by the GUI app on the tester.

If you don't have PCs dedicated to each station, then you probably don't need that.

@KevinVermeer Yes, we don't have PCs on the floor. We're gonna keep the UI on the LCD as simple as possible really.

@saad Yeah, we've got several thousand PCs in our production floor. If you don't have a PC on the floor at all, then yeah, that strategy wouldn't make much sense.

You might want to add a breakout connector on the UART of that IC, though, in case you do get a PC for that.

@KevinVermeer Yeah, that should be simple enough!

Having a PC lets you do all sorts of cool stuff, like use nicer, larger LCD monitors than whatever you're connecting to this thing, log pass/fail information to a database and do process control, and update the tester over the network. But that's a big change!

@KevinVermeer What about switchers? The advantage would be that there would be no heat but they'll be more noisy and more fussy about layout. Heat is an issue because ventilation would let dust in and I live in a VERY dusty environment.

@KevinVermeer Yeah, it's just that I live in Pakistan and most of the workers only have a VERY basic understanding of a PC. A PC would make things MUCH betters but it's not really that practical here. :(

@saad Yeah, we're doing embedded system work and don't have much dust. We recently switched to using switchers on some higher-current boards due to heating issues. But they are more expensive...

How much current does this really draw? It looks like you've got a 2"x2" triangle on the left side...you can dissipate a lot of heat if you connect the tab of your regulator to some via-filled copper layers there.

@saad I'm in the US, and many of our workers only have a basic understanding of a PC. Ours are single-task machines; they boot, autostart the app, and the user just swipes their fingers through IR finger-switches to get things moving. Some don't even have mice and keyboards attached; if you need that then VNC into the machine.

@KevinVermeer That's something I have yet to estimate! I did make a small prototype before this and it used 200mA or so on the 3.3V rail (I have two rails, 1.8V and 3.3V). But that didn't have a LCD, SD Card etc. I think the SD Card can consume upto 200mA on its own and the LCD too is a bit hungry - the backlight takes 160mA and the LCD itself about 150mA. As a wild guess, I assume that max. I'll ever need is 1A?

But that is for just one board, not the extensions. But then again, they'll have their own regulators.

@saad With linear regulators, it's all about the input-output drop. We power ours with anything from a 9V battery to 24V industrial. Dropping that to 3.3V does dissipate some power...

@KevinVermeer Indeed. What if I cascade the regulators? Will I run into stability issues? What I mean is: Have a 5V regulator for the LCD, and then a 3.3V regulator that connects to the regulated 5V and a 1.8V that connects to the regulated 1.8V. That may help distribute the heat somewhat. By going with a low-drop out one, I can put in a small voltage like 7 or 8V and hopefully that would help with the heat.

I'm thinking I should test this tomorrow at work. Should be interesting!

@saad No, that should be fine. We cascade 12V -> 5V -> 3.3V.

@KevinVermeer Ah I see. Hmm, if I can manage the heat and I can hopefully nail this! I do have a lot of space left on the board and might as well make use of it. Thank you VERY much for your help.

@KevinVermeer: yup I didn't pay much attention at first but found them later :)

answered 2 of them

I'm not sure about the ones u pointed more towards fpga -- I don't have enough experience to answer them though I use fpgas :D

I don't see any explanation/comment for downvoting?