|
Rexxed posted:The goal of the project is to not have to check the mouse traps in the basement every day and also not have dead mice fouling up the traps. So there'll be a RF transmitter with a microcontroller on each trap one and a receiver that can just alert me when the transmitter(s) are doing stuff. The Atmega328 is probably overkill for something that will be using maybe two I/O pins but I got an Arduino Uno as a gift and even "too much" microcontroller is still only like $2. Exactly how mouse-infested is your basement anyway
|
# ? Jan 16, 2013 18:29 |
|
|
# ? May 31, 2024 15:34 |
|
Parallel Paraplegic posted:Exactly how mouse-infested is your basement anyway It may be time to buy a cat and let the problem sort itself out.
|
# ? Jan 16, 2013 18:44 |
|
Parallel Paraplegic posted:Exactly how mouse-infested is your basement anyway It's a rural area that's becoming suburban (there's a farm across the street but it's just kept as a farm so the pharma company that owns it won't have to pay higher taxes for example), so I tend to catch 40 or 50 a year, usually in the fall and summer. Sometimes I'll get a couple a day, sometimes there's no mice for a month or more. The mice are small and the house is old so it's hard to completely stop them from getting in with exclusion products. They tend to run around the ceiling areas and the tops of the walls in the basement so I don't think a cat would handle it (plus I'm allergic to cats). I don't really like having to kill so many animals but before I set traps they'd be having parties under the floor and be chewing and scratching and squeaking and running around too much. I was annoyed by the sound and I'm sure they're doing damage to stuff I can't readily access and possibly causing a fire hazard. For the last few years I've had two traps set in different places and while this project is totally unnecessary I thought it would be nice to have a way to see if the traps had gone off without going and looking at them every day.
|
# ? Jan 16, 2013 19:24 |
|
Okay, one more question. How do I connect stuff to the female headers on an arduino? Do I have to buy an adapter, or do I just stick a wire in it?
|
# ? Jan 17, 2013 02:55 |
Just stick a wire in like it's a breadboard. A breadboard with retarded spacing.
|
|
# ? Jan 17, 2013 03:01 |
|
icantfindaname posted:Okay, one more question. How do I connect stuff to the female headers on an arduino? Do I have to buy an adapter, or do I just stick a wire in it? I find these little pre-made jumper wires pretty handy though. You can order sets from ebay/china for about a buck shipped. The main length of the wire is stranded so it's very flexible, but with a rigid tip on each end and some nice overmolded plastic to hold it together. just search for "jumper wires" or "solderless jumpers" or similar Also you can find these kits which are just solid core wire that are pre-cut, stripped, bent to various lengths.
|
# ? Jan 17, 2013 03:20 |
|
Bad Munki posted:A breadboard with retarded spacing.
|
# ? Jan 17, 2013 03:21 |
Except that one spot along the side where they screwed up the spacing, I think between pins 7 and 8 (on the uno and its fore bearers) and refused to fix it for the sake of backwards compatibility with the various shields. Go ahead and try to stick a block of male header pins across that gap. I dare you. Bad Munki fucked around with this message at 03:32 on Jan 17, 2013 |
|
# ? Jan 17, 2013 03:27 |
|
Alright, thanks for the help.
|
# ? Jan 17, 2013 03:31 |
|
Heyo, if anyone here's interested in making a quick buck putting together a circuit diagram for me for an otherwise electronics unrelated DIY project, I have a buying post up in SA-mart here: http://forums.somethingawful.com/showthread.php?threadid=3528833
|
# ? Jan 17, 2013 09:42 |
|
RE: pinning to Arduinos, I found a good solution if you have a pile of old motherboards hanging around. If you yank off the 2-pin headers, rip out one row of pins and run your wires through the empty holes, you get really nice strain-relieved connectors. Hopefully pix make more sense. http://raspberryhigh.wikia.com/wiki/WiringAndBoards Also, you can see my 40-pin +5/GND extension bus on that page, which I'm proud of.
|
# ? Jan 17, 2013 16:18 |
|
Bad Munki posted:Except that one spot along the side where they screwed up the spacing, I think between pins 7 and 8 (on the uno and its fore bearers) and refused to fix it for the sake of backwards compatibility with the various shields. While it would be nice if this wasn't a fixed in place defacto standard, it really isn't difficult to deal with, even with protoboard. Just get a pair of pliers and bend the header pins. Here's a kit with some pre-bent headers, but it's pretty trivial to do your own.
|
# ? Jan 17, 2013 19:28 |
Duct-tape is all that is.
|
|
# ? Jan 17, 2013 19:31 |
|
sixide posted:Anything other than SMA or BNC will hit your pocketbook, but I'd look at SSMA/C for tiny threaded connectors and MCX/MMCX for snap-on. SnoPuppy posted:As long as you have a connector with a reasonable number of pins, you should be able to intermix signals and ground to provide reasonable isolation, so I wouldn't bother to restrict yourself to only RF style connectors. PDP-1 posted:I chatted up one of the folks in our mK lab who runs a LHe3/4 dilution system that sounds like it's similar to the one you use. They use bulkhead connectors to go from room ambient into a header box on the top of the cryostat and then a second bulkhead from the header box down into the cryogenic environment. The header box sometimes has a dry N2 bleed line into it so the inside is non-condensing as the wiring from the cryostat might be cold enough to frost up or condense water on the connections. Once inside the cryostat the cabling is wound around the pumpout line so that any conducted heat is transferred to the cold gas exiting the system (it looks like you have that part covered). On the cold end of the cryostat they had some G10 DIP socket connectors that served as a dual connector port and sample holder. Thanks for the help, y'all! I'm considering two set-ups now: 1) The transmission lines all separate from being closely packed on the way down into the cryostat into their own floating MMCX jack at the bottom of the fridge (like peeling a twizzler), which I can then individually connector to the corresponding plug from the sample holder. After all the connections are made, I tape down the wires to mechanically secure them. 2) A DE-15 connector, with half of the connections as ground to help shield the signal lines from each other. I found one laying around in our lab: And I was thinking about doing something like this: I+ and I- lines just carry DC signals. I'm only likely going to be using 6 of 8 of the pins at once, so another two of those lines will be grounded (the entire system is going to share a common ground). The most important thing is that V+(-)2 and V+(-)1 have minimal crosstalk. I'm doing a cross-correlation measurement between the two of them, so they need to be as independent as possible. Will some grounding pattern like this actually help at all? Outside of this connection, the wires are going to be coaxial until they get to the sample holder. I don't know if I've mentioned this yet, but all of this is contained in a pretty thick copper sheath that acts as our vacuum chamber/faraday cage, so we don't have to worry too much about outside noise. It's mostly the inter-transmission-line noise that I'm concerned about.
|
# ? Jan 17, 2013 23:17 |
|
Insurrectum posted:Thanks for the help, y'all! MMCX is going to have the best SI and isolation, but the DE-15 won't be bad either. That actually looks to be pretty decent pinout for the number of pins you need and a DE-15. Are you treating V+1/V-1 as halves of a differential pair? If not, I would swap the I+ with V+1, as well as I+ with V-2. Otherwise, stick with what you have. As long as the connections to the the connector are coax, you should be fine. Also, neither the MMCX or the DB connector will be vacuum sealed, so that will still need to be dealt with.
|
# ? Jan 18, 2013 07:42 |
|
Speaking of stuff that uses a DE-15 connector, I'm planning to mod a VGA port onto my old Dreamcast, and I have a question about the circuit: I want to try and design it to fit on a custom PCB (from OSH Park of course), and I'm curious about those large electrolytic capacitors. What are they there for? Does the RGB signal really need that much filtering? I know I can just use some surface mount tantalum caps instead, but I'd like to know if any smaller capacitance values would work so I can keep the circuit board small.
|
# ? Jan 18, 2013 17:47 |
|
Those are coupling caps, not sure what their purpose is, IIRC VGA signals are usually DC coupled so unless there's some reason for it I would build it without the caps. DC restoration is standard though, so as long as they're large enough it should work. Nominal impedance is 75 ohms so that should be a very low frequency. I think the lowest frequency component of the signal is the vertical sync, 60 or 50 Hz. If you size the capacitors too small the worst that could happen would probably be that an all-white/grey screen would fade to black on the right hand side.
|
# ? Jan 18, 2013 19:02 |
|
If you do use tantalums, build the circuit and check your voltages on each side of the lines those caps go in, and make sure to orient the + side towards the more-positive DC voltage. They don't like to be hooked up with reverse polarity, at all.
|
# ? Jan 18, 2013 19:38 |
How small do the caps need to be? There are some not-so-bit electrolytic caps of that value. They don't seem that big compared to the connectors you are using, at any rate.
|
|
# ? Jan 18, 2013 19:39 |
|
SnoPuppy posted:MMCX is going to have the best SI and isolation, but the DE-15 won't be bad either. Cool. V+1/V-1 are indeed halves of a differential pair. What do you mean by "connections to the connector are coax"? For DE-15, it's basically a choice between solder and crimp connections. And all of these connectors will be totally contained in vacuum, so there's no worry about vacuum sealing.
|
# ? Jan 18, 2013 20:01 |
|
Oh and in addition: normal voltage for an RGB signal like that is around 1V, so you can use the smallest voltage available on the caps. Unless there is significant DC, then you'd need to add that to the rating obviously. E: I checked some monitor schematics and as I thought they usually have input coupling capacitors, if you measure the DC potential of the video signal and it's less than a volt or two and not negative I'd say you can skip the caps. longview fucked around with this message at 20:30 on Jan 18, 2013 |
# ? Jan 18, 2013 20:24 |
|
Insurrectum posted:Cool. V+1/V-1 are indeed halves of a differential pair. I just mean that the signals of interest should travel through coax to keep them isolated from one another. It's fine if you strip out the last little bit of coax and solder it to the DE-15 cups, and have all the grounds tied together.
|
# ? Jan 18, 2013 20:29 |
|
Delta-Wye posted:How small do the caps need to be? There are some not-so-bit electrolytic caps of that value. They don't seem that big compared to the connectors you are using, at any rate. There's a few different places I can mount the board, but the easiest one has space for a 2.5cm x 8cm board with about 11mm clearance. There's a couple of spots in that area where clearance goes up to 19mm because the case isn't flat. I would suspect the caps are actually necessary because the commercial solution is a breakout box with plenty of room for the caps. I think I might give it a go without the caps first, if it won't hurt anything, they might be on there out of superstition. I did find a diagram that shows the proper polarity, though. e: Now that I've measured it, I see that there's plenty of surface mount caps that are only about 6mm tall, that'd do just fine. Would 25V rated caps be fine for RGB? Also, 220uF ceramic caps exist. They are not cheap. Silver Alicorn fucked around with this message at 20:46 on Jan 18, 2013 |
# ? Jan 18, 2013 20:34 |
|
ante posted:The problem turned out to be that I didn't have MCLR connected, despite having it configured as an internal pullup. None of my configuration changes seem to be working properly, actually, and I haven't had time to try and figure out why. This was a while ago but it might be helpful to someone. You absolutely need MCLR hooked up for a PIC16/PIC18 programmer or debugger to work. It puts a higher voltage, 12 if I recall, to signal to the chip that it wants to talk. And maybe to erase the EEPROM, I don't know if those still need a higher voltage to flush them out or not. Anyways, any time a device ID reports as all 0s it means there's a communications problem. It could mean the IC is busted but PICs are famously rugged. It usually means something like the data and clock pins were flipped (I forget which colour is which on the breadboard cable all the time) or in this case, it wasn't being put into talkie mode by the 12V signal.
|
# ? Jan 18, 2013 23:02 |
|
Numeron posted:Heyo, if anyone here's interested in making a quick buck putting together a circuit diagram for me for an otherwise electronics unrelated DIY project, I have a buying post up in SA-mart here: This looks pretty similar to the ammunition counter project I'm currently working on. Check out my posts in the Post your projects Megathread. I wouldn't want to sell you anything since it's not completed, but maybe it will give you some ideas.
|
# ? Jan 18, 2013 23:49 |
|
Insurrectum posted:What do you mean by "connections to the connector are coax"? For DE-15, it's basically a choice between solder and crimp connections. Go with crimp connections. It requires special tools or a resourceful mind and a willingness to go with some redneck engineering, but definitely preferable. There's only one elegant way (that I know of) to transition from coax shield to lead, and it's solder sleeves. They cost more than $1 each, but they are definitely worth it. See http://raychem.te.com/documents/webservice/fetch.ashx?fileid=8447&docId=919 The preinstalled lead version is what you should be looking at.
|
# ? Jan 19, 2013 02:35 |
|
Silver Alicorn posted:
Don't get hung up too much on the cap values being exactly that. Edit: I had some math in here but disregard, you may need a bigger one than I thought. I would say that 1 uF is fine but just start small and go up if you have issues. Jonny 290 fucked around with this message at 06:15 on Jan 19, 2013 |
# ? Jan 19, 2013 06:12 |
|
I was using an L7805 5v regulator to turn a 9V battery (which was only putting out ~7.2 volts) into 5V to power a breadboard. I've used these voltage regulators before, but this one got alarmingly hot (as did everything else on my breadboard) and the voltages were ~6V in, ~4V out. Is it totally busted? I replaced it with another L7805 which gives the right voltages, and I also swapped 9V batteries to one that appropriately puts out ~9V. Should I throw out this voltage regulator, or was it all the battery's fault? EDIT: I put the first (possibly broken) voltage reg back in using the good battery, and everything seems to be working just fine, so I think it was the semi-dead battery's fault. Why was everything on my board turning super hot, simply because the 9V battery was only outputting ~7.2V? Kire fucked around with this message at 22:34 on Jan 19, 2013 |
# ? Jan 19, 2013 22:08 |
|
With cheap components like that I almost always err on the safe side and throw out suspected duds. Even working for free spending an hour mucking about because a reused transistor was broken or something similarly stupid is not worth it.
|
# ? Jan 19, 2013 22:23 |
|
Kire posted:I was using an L7805 5v regulator to turn a 9V battery (which was only putting out ~7.2 volts) into 5V to power a breadboard. I've used these voltage regulators before, but this one got alarmingly hot (as did everything else on my breadboard) and the voltages were ~6V in, ~4V out. Is it totally busted? I replaced it with another L7805 which gives the right voltages, and I also swapped 9V batteries to one that appropriately puts out ~9V. Hopefully this isn't an issue, but if your components ever get "alarmingly hot" in a breadboard, check the breadboard afterward. I once wired up a temperature sensor incorrectly (which made a very small heating element) and within about five seconds it had melted part of the breadboard. I only noticed it was getting hot because it started to tip over when the holes softened and enlarged.
|
# ? Jan 19, 2013 22:55 |
|
BattleMaster posted:This was a while ago but it might be helpful to someone. You absolutely need MCLR hooked up for a PIC16/PIC18 programmer or debugger to work. It puts a higher voltage, 12 if I recall, to signal to the chip that it wants to talk. And maybe to erase the EEPROM, I don't know if those still need a higher voltage to flush them out or not. Yeah, I eventually got a rig together to put a jumper on MCLR to signal programming mode. Part of the problem was that there was over six months between design time and build time, so some design decisions had to be figured out anew. Also e/n time: after graduating two years ago and working in a completely different industry, I've lost way more knowledge than I'd like. It's actually really concerning how much I don't remember.
|
# ? Jan 20, 2013 04:28 |
|
Anyone have any tips or tricks for cutting perfboard? I tried to score it pretty deeply (first one side then both sides) and snap it in half on the edge of a table and it just flexed a little and refused to budge. I finally wound up using pliers to crack it off (which was very visceral and snapped in a weirdly disturbing way, like how I'd imagine spines would snap ), and while it did crack right on the score I had made, half of the piece I was cracking off twisted and broke apart. I managed to save enough to still put my circuit on but it wasted more than I'd like since I had to throw some away. Do I need to score it even more (I think I scored it pretty drat deep but I could be misjudging) or should I just invest in a good Dremel tool? Any other ideas I missed?
|
# ? Jan 20, 2013 04:57 |
|
Is this classic phenolic perfboard or some sort of FR4 circuit board in a perfboard layout? The phenolic stuff can be cut with a boxcutter and snapped, it's pretty trivial. If that doesn't work you could try coming in from the edges with a sidecutter, it should be no problem to cut hole-to-hole one at a time.
|
# ? Jan 20, 2013 06:01 |
|
sixide posted:Is this classic phenolic perfboard or some sort of FR4 circuit board in a perfboard layout? It's "Epoxy Fiber". I guess I'll look for the "classic" kind next time.
|
# ? Jan 20, 2013 06:14 |
|
If the board is relatively thin and is only one or two layers, a papercutter will do the job too. It's sometimes a bit tough to get the cut started but once it is going it clips right through. One of these:
|
# ? Jan 20, 2013 15:30 |
|
I generally just use a fine metal saw to cut my boards.
|
# ? Jan 20, 2013 16:12 |
I've gotten really lazy and just use a dremel cutoff wheel. It makes a mess but the last few times I've done that I've had a shop to work in. I also tried a scroll saw for some small boards and I can't recommend that to anyone.
|
|
# ? Jan 20, 2013 19:45 |
|
PDP-1 posted:If the board is relatively thin and is only one or two layers, a papercutter will do the job too. It's sometimes a bit tough to get the cut started but once it is going it clips right through. This is what we used at my old job to chop up Vector board, worked great.
|
# ? Jan 20, 2013 20:09 |
|
How I cut my boards, with really good success rate: -Score on both sides with X-acto, lining up as close as I can -Place on edge of lovely MDF desk (you want that hard 90 deg edge) with cut lined up with edge -Use something like a book to hold down the part on the desk -karate CHOP! Don't ease it, snap on that thing. You've given it the weak point. Sometimes I go over it with a flat file afterwards to pretty it up (outside, of course) I can do anything from perfboard to double-sided glass epoxy board with this method.
|
# ? Jan 20, 2013 20:55 |
|
|
# ? May 31, 2024 15:34 |
|
Jonny 290 posted:How I cut my boards, with really good success rate: Yeah I basically did this and the fucker cut my hand open but didn't break. I guess I just got the "extra tough" stuff or something
|
# ? Jan 20, 2013 22:15 |