Shop OBEX P1 Docs P2 Docs Learn Events
First propeller PCB doesn't work — Parallax Forums

First propeller PCB doesn't work

floflo Posts: 38
edited 2010-06-02 00:46 in Propeller 1
Hey guys,

Today I received my first custom prop PCB but unfortunately my mac doesn't see it appear in the device list like my PropStick does. I followed the PropStick schematic and a guy at work taught me to solder everything. I double checked all the solder joints and quadruple checked my schematic but didn't see anything wrong... then again I'm a n00b at electronics smhair.gif I'm just glad nothing blew up!

You guys have any suggestions for future layouts or even see the mistake confused.gif
PropSchem.png
1086 x 726 - 51K

Comments

  • ElectricAyeElectricAye Posts: 4,561
    edited 2010-05-22 04:37
    You might also want to post the "artwork" of your PCB.

    I'm surprised you learned soldering with a Q44 chip: I'm still trying to get the courage to learn surface mount soldering.
  • Zap-oZap-o Posts: 452
    edited 2010-05-22 04:40
    Looks like the lm2937 output voltage is not connected to anything. Also the propeller needs all the pins 8,18,30,40 VDD and 5,17,27,39 Vss connected as well. Cant use just one pin.


    - Looks like a USB Powered device with a jumper to input power, be careful using both.

    Post Edited (Zap-o) : 5/22/2010 4:51:53 AM GMT
  • T ChapT Chap Posts: 4,223
    edited 2010-05-22 04:47
    Your Prop power pins are not all connected, and your regulator looks suspect in that it doesn't show how the output gets to the Prop, unless the label is missing.

    I would start with a meter to make sure you get power and GND everywhere that it needs to go. A good practice on a the Prop would be to have a .1 cap near each of the 4 power pins to GND.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-05-22 04:59
    Ditto all the above but also noticed you are using a RS485 transceiver simply as a transmitter. R3 is totally redundant as is R4 as it is the receiving end that needs termination on long lines.

    Anyway, even if the schematic was 100% correct and the PCB was connected the same as the schematic it still doesn't mean that it will work. Better to post the pcb layout and photos too as sometimes the schematic and pcb are correct but assembly isn't.

    Just to reiterate what has already been said by others but with a little more emphasis, ALWAYS connect up ALL the VDDs and VSS pins on the Prop or any other chip for that matter as it matters.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
  • T ChapT Chap Posts: 4,223
    edited 2010-05-22 05:08
    If you are using eagle, using wires and labels like that you really have to check that the wire got 'attached' to the part in the schematic. After putting on a wire and pulling the label off and naming it, you should always then take the part with the move tool, drag the part ( ic etc) on the schematic over a little and see if the wire(s) move/follow with the part. If so, then slide the part back where it was. This was always a gotcha when I first started with eagle.
  • PhilldapillPhilldapill Posts: 1,283
    edited 2010-05-22 05:11
    ElectricAye, QFN is actually crazy easy. A tab of solder paste, a skillet, and the magic of surface tension is all you need. I got some PIC samples in(64 pins) and I LOVE soldering these things. It really is magic the way the chip aligns itself.

    flo, as many have said, post the PCB layout and possibly a high-resolution image of the PCB.
  • Pablo1234Pablo1234 Posts: 19
    edited 2010-05-22 05:15
    Todd that one I learned real quick lol. Aside from checking the schematic like that, I do that on the board artwork also, cant tell you how many times i over drew a trace and didnt notice.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    A woman is the only thing I am afraid of that I know will not hurt me. - Abraham Lincoln
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-05-22 05:49
    Philldapill said...
    ElectricAye, QFN is actually crazy easy. A tab of solder paste, a skillet, and the magic of surface tension is all you need. I got some PIC samples in(64 pins) and I LOVE soldering these things. It really is magic the way the chip aligns itself.

    flo, as many have said, post the PCB layout and possibly a high-resolution image of the PCB.

    Do you mean QFP, the one with legs? I think that's what he was talking about. The QFN without legs are a little bit trickier as you really want to have pads that extend out a bit more than the ones recommended so that you can easily check and touch up each contact.

    So Phil, a skillet hey? Do you save time if it's early in the morning and fry your eggs at the same time? Or maybe if you are British it would be more like fish n chips!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
  • LeonLeon Posts: 7,620
    edited 2010-05-22 06:58
    Fish and chips are deep-fried! We don't have them for breakfast.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Leon Heller
    Amateur radio callsign: G1HSM

    Post Edited (Leon) : 5/22/2010 7:06:10 AM GMT
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-05-22 07:03
    Leon said...
    Fish and chips are deep-fried!

    English breakfast kippers are deep-fried? Certainly many Propeller chips are routinely deep-fried smile.gif

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
  • LeonLeon Posts: 7,620
    edited 2010-05-22 07:05
    Kippers are usually grilled.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Leon Heller
    Amateur radio callsign: G1HSM
  • heaterheater Posts: 3,370
    edited 2010-05-22 07:13
    I think fish'n'chips for breakfast might not be so strange in Glasgow[noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    For me, the past is not over yet.
  • LeonLeon Posts: 7,620
    edited 2010-05-22 07:18
    heater said...
    I think fish'n'chips for breakfast might not be so strange in Glasgow[noparse]:)[/noparse]

    With a deep-fried Mars bar to follow! It has to be coated in batter, like the fish, of course.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Leon Heller
    Amateur radio callsign: G1HSM
  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2010-05-22 08:43
    This is exactly why I am so afraid to have custom boards made.
    I just know they would come back and be terribly wrong :-(

    I have had some made that other people worked up from a schematic
    I drew...they worked fine, but I got no feeling of accomplishment.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-05-22 12:05
    No pain, no gain. Holly, you have got to give it a go. There is a safety net here in the forum, just submit your design files to the forum before rather than after as many do. Do you think if we can find a problem (or many) after it's too late that we couldn't find it before hand?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    *Peter*
  • LeonLeon Posts: 7,620
    edited 2010-05-22 12:25
    The better PCB packages tell one about inconsistencies and missing connections in the schematic, check the schematic against the PCB, and also check manufacturability. Learning how to drive them effectively takes some time and effort, of course.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Leon Heller
    Amateur radio callsign: G1HSM
  • Zap-oZap-o Posts: 452
    edited 2010-05-22 12:57
    HollyMinkowski

    I use National Instruments miltisims / ultiboard and its never steered me wrong. Over at advanced circuits you can have them make up a few boards for under $140 and if you are a student its even less expensive. If you are especially tight on cash or need to make a board real cheap (not for production) they will make it without silkscreen for around 30 bucks.

    When I solder small ic's and had too much coffee. I make sure to gob solder all over the legs on the part, then using solder wick I remove what is need.
  • sylvie369sylvie369 Posts: 1,622
    edited 2010-05-22 14:00
    I've made several useless boards due to a variety of stupid newbie mistakes (and some outright absent-mindedness). I consider the cost to be the tuition one pays to learn to make boards. Compared with what you'd·pay in a real university, it's pretty cheap.

    My most recent mistake? I ordered a board with holes too small for the header pins connecting a GPS - this after taking great care in using the datasheet to set up a footprint with the right·positions for those pins and some other connectors. I was so proud to have gotten everything right...until I tried actually putting the GPS on the board. Oops. Two minutes to revise the footprint, and a few weeks to wait for the new boards to arrive. Lesson learned.
  • floflo Posts: 38
    edited 2010-05-22 17:56
    Thanks for all the input everyone!

    Things I'll do from your input:

    - always post board layout ( thanks Philldapill)
    - connect all power/ground pins on ICs (thanks Zap-o and Peter)
    - Add Cap to all power pins to ground (thanks Tod)
    - Insure all connections are actually made in Eagle (thanks Tod! Thankfully this was the case)
    - Check board artwork no matter how lazy you were that day (Thanks Pablo1234)
    - Fish n chips, kippers, and deep friend Mars bars are the sh**! (hahah thanks for the laugh guys!)
    - Manually fix autoroute traces in eagle cause some are silly
    - move vias that are too close to unrelated pins that make it harder to solder


    I've attached the board artwork and again nothing jumps out at me as being wrong. I went back to the magnifying glass at work to double check all the solders and everything is clean. So, I'll wait for Monday so that my coworker can show me how to solder a jumper wire on the power and gnd pins of the prop. He has unbelievably steady hands.
    PropBoard.png
    1087 x 718 - 88K
  • T ChapT Chap Posts: 4,223
    edited 2010-05-22 18:05
    If you go to radio shack and find some enameled magnet wire, you should be able to easily solder to the Prop pins. Just tin the wire and tin the Prop pins, then run the wires as needed, test and see what happens. Did you put a meter from the regulator output to all the places the 3v3 needs to be showing up?

    There is an unrouted trace on the Max chip, maybe you saw that. On eagle, the error check feature is very useful both on the schematic and board side. The board error check gives lots of errors for things being too close, but it will show some obvious things too.

    Take a pic of the finished board, that always helps too.

    Post Edited (Todd Chapman) : 5/22/2010 6:10:46 PM GMT
  • floflo Posts: 38
    edited 2010-06-01 00:52
    I've been trying to figure out this problem and did some test cases to narrow the field of search. I made a couple different PCBs: one with just the propeller chip (and associated resistors, voltage regulator, and caps) and one with the FTDI FT232 USB chip.

    1 - Prop only board: all voltages seem normal ok
    2 - FT232 only board: maintains USB connection and computer sees it fine
    3 - Fully soldered board: resets USB connection periodically and continually every ~3 seconds.

    I'm suspecting the DTR pin on the FT232 isn't playing nicely with the BOE and RES pins on the prop. Other people seem to have trouble with this:

    http://forums.parallax.com/forums/default.aspx?f=25&m=445025

    After some googling, I found this:
    http://www.embeddedrelated.com/groups/basicx/show/22147.php said...

    As supplied, the Plug's /RES pin yields an open-collector active-low
    pulse when the onboard FT232R DTR# pin goes high. Even with a pullup,
    that won't successfully act as ATN. It's easy (with small tools),
    though, to remove a 10nF series capacitor and connect the DTR# output
    through an (optional) 150ohm resistor to the module's /RES pin. That
    provides a level output for ATN control.

    I don't understand why the transistor schematic I have now doesn't work and what "ATN" stands for in the quote above. Any ideas?

    Post Edited (flo) : 6/1/2010 1:43:53 AM GMT
  • T ChapT Chap Posts: 4,223
    edited 2010-06-01 01:12
    Did you get the 3v3 on all the Prop pins?

    Try cutting the trace to the input of the regulator from the USB power, let the USB power only the 232rl. Find another voltage source and tie it to the input of the regulator, tie the other voltage source GND to the board GND as well though. See if that helps.

    Post Edited (Todd Chapman) : 6/1/2010 1:19:50 AM GMT
  • floflo Posts: 38
    edited 2010-06-01 01:46
    Yes, 3.3V on all prop pins.

    I'll try this tomorrow evening. What are the implications if this works or doesn't work?
  • T ChapT Chap Posts: 4,223
    edited 2010-06-01 02:09
    If the prop is constantly resetting, it could be the regulator not having enough current to run everything off the USB rail. I have the exact schematic you have but I power the Prop from another supply, the 232rl gets it's power from usb, works great.
  • floflo Posts: 38
    edited 2010-06-02 00:01
    I feel stupid guys. It was a cold solder on the FTDI chip.

    I guess my advice if anyone runs into this problem is probe the USBBDM and USBDP pins on the FTDI chip. Make sure they're at normal voltages. Mine was at a funky 2.7V and thats how I discovered the cold solder. In my defense I didn't solder it; I was being shown cool solder tricks on small pins :-p

    Thanks to everyone. I guess my first PCB does work after all!
  • hover1hover1 Posts: 1,929
    edited 2010-06-02 00:46
    Good Deal! Congrats!

    Jim
Sign In or Register to comment.