Hi guys! Sorry I've been a bit silent on this board lately -- I've gotten caught up with "Real Work" and not had time to play with this much, plus I'm sort of stuck at an awkward stage of this project.
I'd hoped to give a couple of close friends WBs for Christmas, but there's no way that can happen this year. :/ I have five WBs in various stages of functioning -- all of which are using the PCB ("internal") antenna.
* My "best" WB has a range of 2-3m, maybe. This is encouraging but not ideal. The bigger problem is that it has some sort of glitching issue where the ATMega will usually hang right after booting ... unless I have it plugged into my computer via USB, which leads me to suggest that it's not getting enough power -- but the battery is good.
* Two more seem to function correctly, but with very small range -- less than a meter.
* One has a dead 3.3v regulator -- the 15th one I've blown -- so I've been experimenting with using a 2.8v regulator instead. My idea there is that maybe I could drive both the MCU and the PLL off the same supply, and remove the 3.3v regulator entirely. This seems to start up correctly, but it does not recognize any input from the serial port -- I get "invalid command" for anything I type. The FTDI chip is probably overdriving the RX line of the MCU.
* The last has a bad +5v supply -- I used this as an opportunity to install the more powerful RF amps (SXB-4089 or something?) to see what that would do, but I'm going to have to hook the 5v up to a bench supply to test it, and I haven't gotten around to it.
I've decided that I need to try to replace the PCB antenna on the first "best" unit with real antennas using SMA connectors -- which I currently have on order with Mouser. (Those things are *horribly* expensive!) That way, I can rule out any antenna problems, and I'll actually be able to probe at the radio board while it's running (because it won't be covered up by the antenna PCB). (If you have not yet built one of these, I'd have to recommend you put antennas on your first one -- even though they're expensive -- because it will make it easier to debug.)
Assuming I get all of my current units working, I have several ideas that I think would improve the project. The biggest problem I have, so far, is the cost per unit -- and the difficulty of obtaining some of the parts doesn't help.
* As I stated in an earlier thread, I think we could shave $16 in parts costs off of each unit by switching to Z-Comm VCOs (which are about $5 cheaper, each). Those VCOs use a +5v supply, so we could omit the +12v supply and simplify the circuit and save another $5. Z-Comm has a minimum order of 5 pieces, and the last time I check they were out of stock of one of the VCOs we need. However, I was able to get them to send me a free sample of the VCO they have listed as back-ordered (?!), but I can't really try it until I have at least one unit that works well with the MiniCircuits VCOs.
* I've had so many TPS79333 LDO 3.3v regulators blow that I can't possibly give a built WB to a friend in good conscience, knowing that it might die in a puff of smoke any time they turn it on. I replaced a couple of them with Fairchild parts, and have not had further problems, but I can't reproduce this consistently enough to know if they fixed the problem.
* I have trouble getting clean connections with an iron, so I want to move to paste soldering. I have access to a Epilog laser cutter for $30/day at Techshop.ws -- I'd love to go there with a stack of overhead transparency stock and make 50 stencils and sell them here for $5 a piece, but that's yet another project.
I haven't given up on this project, but I've spent a *lot* of time on it, and over $1000 in parts and PCBs. I have learned a lot, but I still don't have anything to show for my efforts. I can't afford to buy more parts (for example, the RF amps or VCOs) until I can finish some of the units I already started.
Hopefully, I'll have some more free time in January, and by that time I will at least have mounted the SMA connects and see if that helps. Until them, I'm kinda stuck.
I'll try to drop by here more often; if anyone has any other ideas about making this project more reliable, easier to build, or cheaper to build, I'd be open to suggestions.