Okay, I'm back at this again. In the interim, I found this, which looks=20 really promising for low cost and reliability... http://www.dotmana.com/weblog/2015/08/microsd-card-reflow-quick-and-dirty-p= cb-with-kicad/ I still want to try a handful to see how repeatable it is. To make the data easily transferable to a PC, I'm also considering using=20 a second SD card in a socket. Data will be held in RAM, streamed out to=20 the soldered-in uSD, then the user can insert a card in the other uSD=20 socket and copy the data to that. Cheers, -Neil. On 6/17/2016 5:41 PM, Neil wrote: > Memory can be volatile, as the data will be retrieved after the logs are > done, and the memory does not need to be retained after that. I will > probably have to put in a battery backup or similar, but that's > relatively easy. > > "rugged" means it will get shaken quite a bit, so the > connector/connection to the microSD card is what I'm worried about. I do > have some push-pull microSD connector samples on the way, but will have > to test. > > This will be for for a portable/mobile device. WiFi is not an option, > as well as directly sending the data to some other device. It's going to > have to be quite compact. FWIW though, I recently investigated various > wireless options for another project and WiFi modules (especially the > lower-cost ones like the ESP8266) are really power-hungry. For on-going > data transmission, you should look into the nRF24L01 which can be had > for less than a buck nowadays. I understand it's good on power > consumption, but have to investigate that still. > > Up to this point, I've been assuming that if I use SD/microSD it would > be removable for retrieval of the data, but perhaps I should consider > just soldering and attaching permanently to the PCB. I've seen some > hobbyist 3DP control boards do this and thought it looked hokey/amateur, > but I can probably do that and 3D-print a cover to hide it. > > Re: wear leveling, reliability is important. Shouldn't I be > uncomfortable logging data for a couple hours over 4GB or 8GB w/o any > wear leveling? FWIW, my pallpark logs will be just under 1K bytes per > second. > > Cheers, > -Neil. > > > > On 6/17/2016 3:35 PM, Denny Esterline wrote: >> Your question is a bit confusing. Are you looking for volatile or >> non-volatile memory? >> PC RAM is of course volatile and dynamic - you would need to refresh the >> data quite often, I don't think that's a practical option in the PIC wor= ld, >> but you didn't mention what you're using. >> You mention "low-volume" and "rugged", both of which are vague enough to >> make any advice possibly meaningless. >> >> Various versions of SD cards can be quite rugged - probably not "artille= ry >> shell" rugged, but certainly "automotive" rugged. (GoPro cameras?) And >> depends more on the connector used. For more robust mechanics in low vol= ume >> I would start with a gob of hot-melt glue on and around the card after >> install (micro SD better here because of low mass). Other option is no >> connector at all, solder wires directly to the card fingers then bond th= e >> card to your product with a suitable adhesive. May complicate retrieving >> the data, but in "low volume" it may be an option >> >> Either of these options require you to manage the file system which may = or >> may not be an issue. >> There are also several commercially available "dataloggers" that basical= ly >> have some flash memory (SD card slot usually) and a processor preprogram= med >> to allow you to send it the data via serial. It may be helpful in a low >> volume situation. I've seen some of those used in model rocketry, so >> there's one measure of "rugged". Here's a couple links from a cursory >> search: >> https://www.adafruit.com/product/1141 >> https://www.sparkfun.com/products/12772 >> >> There's also the FTDI Vinculum parts that do something very similar but = use >> a USB flash drive: >> http://www.mouser.com/new/ftdi/ftdivinculum2/ >> Options from bare chips to breadboardable modules to panel mountable ful= ly >> enclosed units. >> >> One of the other emails mentions "wear leveling" - absolutely critical i= n >> long life products. Completely meaningless for some short lived >> applications. I've done datalogging during development projects that's o= nly >> ever had to be written a couple times - didn't bother with any of the >> "normal" file system niceties, just spew data into the chip and parse it >> back out later with PC side tools. >> >> The other option worth mentioning - can you upload the data realtime wit= h >> some type of telemetry? I've recently been digging into WiFi modules and >> it's been an eye opening experience. If you're in an environment where >> wireless of some type is an option, it wouldn't be "too much work" to >> uplink the data to a more capable system. WiFi is almost trivial to >> implement nowadays, but dozens of other wireless options exist with vari= ous >> benefits and drawbacks. >> >> >> Hope this helps - if it doesn't, we may be able to do better if we bette= r >> understand your application. >> >> -Denny --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .