TX <--> TX. Did it again

I pin connectors (male/female) as DCE/DTE, accordingly. Then, whenever I want to mate two connectors, I simply look at the gender: if complementary, they will mate 1:1; if the same, then I need a "gender swapper" (that swaps the outs and ins, accordingly -- so the new, "effective" connector is now complementary to the mating one)

Life is too short to fuss with pinouts! :>

I used to get it right all the time. I had a port monitor/configure thingy

- I just kept flipping switches until I got the maximum number of lit leds and whoosh went the data. Checking the switches then told me how to wire up a permanent cable. One of the best bits of kit possible for RS232 and about $10!

This is on a board. Chip out to chip out, chip in to chip in. Cut traces and little blue wires now abound.

I sometimes use different words, like POLL and REPLY, to avoid the tx/rx ambiguity.

Yes.

I just refer to the Annabooks PC handbook.

Cheers

Ah, well... you're stuck, then -- unless it was a trivial component whose schematic symbol was misrepresented. Nowadays, with pins having a multitude of optional uses, you can't realistically draw each chip with the "right" pin orientation (gazintas and cumzoutas).

There's a reason we have Xacto knives with *tiny* points! :>

I get it right every time. My naming convention makes it pretty easy to keep them straight. What I f*ck up is I2C clocks and data. I don't know why but it happens *way* too often. That, and forgetting the pullups. They usually don't find it in design reviews, either.

I was in a design review meeting once where people were getting a bit confused about what was in and out at the board/chip level and what was xmt/rcv at the system level. In the end they decided to use names for the system level. One guy was so confused by the whole thing that he was trying to suggest that every such signal should have two names since each signal was an input on one end and and output on the other. lol

Then shortly after that I ran into an actual description of SPI and found signals labeled MOSI and MISO. Can't get any more clear than that unless you don't know your master from your slave!

Tim Wescott:

time.

I get it right every time (of course) but make up for it my mixing up genders of D connectors so you have to solder them to the other side of the PCB. I do get into problems when interfacing with other boards - my TX is TX but they think that their TX is as well !

I once worked a lot with an engineer (he was the boss) who reckoned that the odds of getting it right were so little improved by thinking about it that it was a waste of effort since it took so little time to cross the wires in the cable. (They made bespoke production lines.)

Michael Kellett

It's a variation of Murphey's law - you have 50% chance of getting this right, and 90% chance of getting it wrong.

I feel with you. The board I got finally today has the same problem. Fortunately it is just an USB-serial adapter for programming. Having a cable between with Rx/Tx swapped makes everything much more flexible. ;-)

Th P-FET in sot23 lying upside down does not disturb anybody, either.

Without thinking about it, the odds are 50% for right and 50% for wrong. Thinking about it moves this more to the wrong side.

I still have a Smart Cable that does that. For my own stuff, I used to rely heavily on Jan Axelson's "Serial Port Complete".

Cheers

Phil Hobbs

Just have three or four people each check it five times.

Except they guy who has to put it there.

Ah, the oops register bit :-)

So far, yes, knocking on wood. I am surprised that there isn't an oops chip for this. Like a crosspoint gender changer SMT thingie that you can buy in a through and a cross version. After all, they used to make oops transistors in the early days of SMT when people messed up the SOT23 footprint. Not sure if they still make those.

That'S me for the five prototypes. As far as I know my customer I expect him with change requests after seeing the prototypes. So he will pay for fix. There is an old saying attributed to IBM sales guys: The customer will pay for it, he just may not know when and how.

Hopefully this customer doesn't read Usenet :-)

Ofr prototypes, you made the mistake - own it. ;-) Production frowns on such things though. BTDT.