|
It's a mistake. The 339 is a comparator with an open drain output and the 10k is the pull up. The direct connection from 5V to the output shouldn't be there.
|
#
?
Feb 24, 2016 19:58
|
|
|
|
| # ? May 29, 2024 04:29 |
|
|
How can something have an open-drain output? Isn't an open drain a high-Z input?
|
|
#
?
Feb 24, 2016 21:44
|
|
|
asdf32 posted:It's a mistake. The 339 is a comparator with an open drain output and the 10k is the pull up. The direct connection from 5V to the output shouldn't be there. This. If you look closely, there's no dot signifying a junction, so it appears to be a misprint. reading posted:How can something have an open-drain output? Isn't an open drain a high-Z input? An open-drain output is either shorted (usually to ground) or high-Z. Hence the external pull-up to make it actually work.
|
|
#
?
Feb 24, 2016 22:00
|
|
|
reading posted:How can something have an open-drain output? Isn't an open drain a high-Z input? An open-drain output is the drain of a N-channel MOSFET who's source is tied to GND. When the device puts a logic high at the gate internally and enhances the FET, it will pull whatever is attached to the drain to the source (usually GND) as hard as it can. When the gate is logic-low and the device is not enhanced, the drain floats.
|
|
#
?
Feb 24, 2016 22:19
|
|
|
Open drain lets you do cool things like hook multiple outputs together without things exploding.
|
|
#
?
Feb 25, 2016 00:19
|
|
|
Is open drain like wand or wor?
|
|
#
?
Feb 25, 2016 00:23
|
|
|
It's a wired and; as soon as one output goes low the entire bus goes low. As opposed to hooking together several normal outputs (that aren't in high-Z) where the second one goes low you have a short between the logic voltage and ground. The concept can be useful by itself in simple applications but I like protocols like 1-wire or I2C that use a single open drain serial I/O pin (plus a clock pin in I2C's case) per device to connect several things together without becoming a wiring nightmare.
BattleMaster fucked around with this message at 00:54 on Feb 25, 2016 |
|
#
?
Feb 25, 2016 00:49
|
|
|
I know it's far from the best program for this but it turns out in Illustrator you can get enough precision with all measurements set to mm to make pcb masks for the inkjet transfer method. Here's the surface mount LED array I did earlier today. ed: I just had normal printer paper so the test transfer was really incomplete but that's no big deal. Gonna order some more appropriate stuff off amazon and then get some etching acid at radioshack later. All goes well and I should have a nice 500ish lumen LED panel in an altoids tin this weekend. Parts Kit fucked around with this message at 03:31 on Feb 25, 2016 |
|
#
?
Feb 25, 2016 02:57
|
|
|
BattleMaster posted:Oh, oops; I didn't do much research aside from take the suggestions. I guess that kind of seals the deal for the Microchip one. It's weird how manufacturers are so irate about their USB-UART chips! Figured I'd warn you. For what it's worth it seems that Prolific isn't counterfeited so much anymore, they're now targeting ftdi. I still wouldn't put one in my own designs. They're irate because of the same reason they're targeted for counterfeiting. USB bridges have huge profit margins. Consider the humble attiny10. $1.10 each at individual quantity. People have implemented software USB with enough power left over for protocol translation, which is basically what a 4.50 FT232rl does. Now if you control a fab, and have somewhat questionable ethics, you can probably use some masks you already have (probably an 8051) and implement the same in something like 5c of silicon. Number made up of course, but it's based on just how incredibly cheap tiny micros get in large quantities.
|
|
#
?
Feb 25, 2016 03:41
|
|
|
Parts Kit posted:I know it's far from the best program for this but it turns out in Illustrator you can get enough precision with all measurements set to mm to make pcb masks for the inkjet transfer method. Here's the surface mount LED array I did earlier today. If you're gonna do something silly like that, you should check out PCBmodE. Totally gonna do something with it sometime, it looks awesome.
|
|
#
?
Feb 25, 2016 05:05
|
|
|
Rad, I'll take a look! Here's a picture of the abortion of a flashlight from last night. It's about on par with an old incandescent mini maglight. 
|
|
#
?
Feb 25, 2016 06:39
|
|
|
I'd like to get a 24V @ ~500mA output DC-DC converter module that is suitable for use in an automotive environment (survives load dumps/etc). It needs to look professional, not like something I got from ebay for $5. Any suggestions or ideas of where I can browse some options?
|
|
#
?
Feb 25, 2016 20:09
|
|
|
Google? I mean, not to be snarky, but if you have the specs and want a particular look, put them into a search engine and start looking at pictures. Include the word "automotive" to get stuff that's hardened for that environment.
|
|
#
?
Feb 25, 2016 20:42
|
|
|
Sagebrush posted:Google? I wouldn't have posted if I hadn't been trying that for a few hours. I have found some stuff that fits the bill from InterVolt, but it is bigger than I would like. Can I use a big TVS on the front of a not-automotive-rated supply and be confident? As in: http://www.littelfuse.com/~/media/e...on_note.pdf.pdf I just don't want to find out this is inadequate a few months from now. taqueso fucked around with this message at 20:50 on Feb 25, 2016 |
|
#
?
Feb 25, 2016 20:44
|
|
|
taqueso posted:I'd like to get a 24V @ ~500mA output DC-DC converter module that is suitable for use in an automotive environment (survives load dumps/etc). It needs to look professional, not like something I got from ebay for $5. Any suggestions or ideas of where I can browse some options? I'm not an expert at all, but couldn't you use an automotive-spec LDO to filter the input to any old DC-DC switching converter?
|
|
#
?
Feb 25, 2016 20:52
|
|
|
taqueso posted:I'd like to get a 24V @ ~500mA output DC-DC converter module that is suitable for use in an automotive environment (survives load dumps/etc). It needs to look professional, not like something I got from ebay for $5. Any suggestions or ideas of where I can browse some options? I assume you're still looking at nominal 12v in.
|
|
#
?
Feb 25, 2016 20:59
|
|
|
Aurium posted:I assume you're still looking at nominal 12v in. Yeah, it is for a vehicle with a 12V nominal system. taqueso fucked around with this message at 21:13 on Feb 25, 2016 |
|
#
?
Feb 25, 2016 21:03
|
|
|
Interesting. I've found that the simple searches are surprisingly useless. Plenty of candidates, little that actually says for certain.
|
|
#
?
Feb 25, 2016 22:54
|
|
|
Aurium posted:Figured I'd warn you. For what it's worth it seems that Prolific isn't counterfeited so much anymore, they're now targeting ftdi. I still wouldn't put one in my own designs. No need to own a fab--you just need to source one-time-programmable micros. It looks like at least one fake FTDI is just a mask-programmable micro: http://zeptobars.com/en/read/FTDI-FT232RL-real-vs-fake-supereal%20this%20deep-down This is a cheap thing to contract out, since it's basically off the shelf and you only need to get ROM mask made. For somewhat lower volumes, you could probably skip that step and use bare die programmable micros, which you would just need to have encapsulated in the lower pin FTDI package. Higher cost + bigger die (maybe?) since you need the flash memory, but if you can get a good deal on counterfeit microcontrollers it might still be profitable.
|
|
#
?
Feb 25, 2016 23:31
|
|
|
meatpotato posted:I'm not an expert at all, but couldn't you use an automotive-spec LDO to filter the input to any old DC-DC switching converter? No good---the switcher wouldn't be automotive rated, so that would be liable to being destroyed by a load dump.
|
|
#
?
Feb 25, 2016 23:33
|
|
|
taqueso posted:I'd like to get a 24V @ ~500mA output DC-DC converter module that is suitable for use in an automotive environment (survives load dumps/etc). It needs to look professional, not like something I got from ebay for $5. Any suggestions or ideas of where I can browse some options? This is not 100% professional, but it will probably meet your needs: http://amzn.com/B00JUHEJJ2
|
|
#
?
Feb 25, 2016 23:44
|
|
|
I was thinking that it was kind of interesting that Microchip made a USB to serial chip when they make micros with USB modules and free open-source firmware for the USB emulated com port class. I mean why make a special chip when you can just sell the customer a PIC to do the same thing without having to make a totally new product. But maybe the Microchip MCP2200 is actually a PIC with different markings, rejiggered pins, and mask ROM firmware 
BattleMaster fucked around with this message at 23:51 on Feb 25, 2016 |
|
#
?
Feb 25, 2016 23:48
|
|
|
Slanderer posted:This is not 100% professional, but it will probably meet your needs: tbh I wouldn't find that out of place in an automotive installation. It looks like any other automotive gizmo. I'm more concerned with the crimps and such that're going to go on the ends -- that's what will make or break the professionalism.
|
|
#
?
Feb 25, 2016 23:56
|
|
|
Slanderer posted:This is not 100% professional, but it will probably meet your needs: I saw that, and considered suggesting it, but taqueso said what he's currently looking at something "from InterVolt, but it is bigger than I would like," I figured something that was 5.3 x 4.9 x 1.8 inches was out of the question. Slanderer posted:No need to own a fab--you just need to source one-time-programmable micros. It looks like at least one fake FTDI is just a mask-programmable micro: It was things like that I was speaking from. By saying control, I trying to be a bit more ambiguous than outright ownership, though I suppose I should have just been direct. Somewhere along the line you'd still need someone in charge questionable ethics to mark them FTDI after getting them made, or buy the silkscreen/laser etcher/engraver yourself. Or just dupe someone into thinking you are FTDI. Slanderer posted:No good---the switcher wouldn't be automotive rated, so that would be liable to being destroyed by a load dump. I think he's suggesting using the LDO as a preregulator, which should work. They are designed to take the spikes and give a clean voltage out. It'd just be less efficient than using a dcdc setup that could take it natively. BattleMaster posted:I was thinking that it was kind of interesting that Microchip made a USB to serial chip when they make micros with USB modules and free open-source firmware for the USB emulated com port class. I mean why make a special chip when you can just sell the customer a PIC to do the same thing without having to make a totally new product. According to http://dangerousprototypes.com/docs/MCP2200_breakout_board#MCP2200 the MCP2200 is a PIC 18F14K50. EDIT: Just going to post this too, cause it's interesting: http://dangerousprototypes.com/2011/01/18/hack-open-source-usb-stack-on-mcp2200/ Aurium fucked around with this message at 00:39 on Feb 26, 2016 |
|
#
?
Feb 26, 2016 00:36
|
|
|
Aurium posted:According to http://dangerousprototypes.com/docs/MCP2200_breakout_board#MCP2200 the MCP2200 is a PIC 18F14K50. Ahaha. I had actually considered just using a USB PIC and loading it with the CDC demo code as an option for a USB-serial converter but I decided I'd rather have a drop-in solution. I guess the MCP2200 meets those choices halfway  Apparently it doesn't even use mask ROM for the code and straight-up has the same pinout complete with working debugger interface. Edit: Though one of the other K-series PIC18 has a USB module that can use the inaccurate internal oscillator as a clock source instead of a crystal by continually resyncing it with signals from the host so I might use one of those instead so I can get away with less crap on the board Edit 2: The PIC used as a USB-serial converter on Microchip's evaluation board for the RN4020 Bluetooth module is one of the crystal-less ones, wonder why they didn't use it for the MCP2200 because not needing a crystal would be a good selling point BattleMaster fucked around with this message at 01:25 on Feb 26, 2016 |
|
#
?
Feb 26, 2016 01:16
|
|
|
The 14K50 is the cheapest and mostly commonly available one with USB I know. I use it in a ton of stuff, so I assume they already have the processes in to crank them out faster than demand. It's also been around since 2010, and I imagine the host-correcting ones are newer tech.
|
|
#
?
Feb 26, 2016 04:19
|
|
|
Partially crossposted (a lot of the minutiae is added just for y'all) from the TFR DIY thread, the person AWOG was replying to was talking about having designed a PCB for a grid of small LEDs as an all-purpose portable light: Well, now I have a project. I have one of these:  (Sadly, in similarly rough shape.) It uses two of the big 6V lantern batteries to power a couple of old-school automotive bulbs -- I'm fairly sure the main one is a taillight, and the little one for less light/longer battery life is a dome light). Unfortunately the case is pretty form-fitted to the batteries, so I'm not sure where I'd put the driver, but random Google results from flashlight forums suggest I should get decent life out of the batteries (opinions vary between "10-25" and "50+" hours, but either way is a hell of a lot better than the incandescents for the light output -- the LED is as bright as a 70W incandescent, as opposed to whatever relative drip in the underpants the brake light puts out (I got bored before I found an answer relevant answer from [brake light lumens], most of the results are talking about LED conversions). Maybe make the head/leads removable so I can swap the light part to a magnetic base and it can double as a hell of a work light for poking at the car.  Do they make rechargeable 6V lantern batteries? I can't find any. But I know they're four F cells, and I found this that should fit in the can. And at 13 amp/hours per cell and eight of them wired in series, that gives 9.6V, which the rig AWOG linked should handle, and 104 hours run time, if I'm doing the math right (calling it 1A instead of .9A draw to simplify things and make the numbers nice and round). Or maybe this R/C car 4-D pack that I think has more amp-hours in a pair, and maybe even have enough space to put the driver in the battery case (which puts the kibosh on the detachable-head-to-plug-into-the-car-battery-as-a-magnetic-worklight thing; otoh I could just buy a second driver module to mount in the car, it's cheap enough).
|
|
#
?
Feb 26, 2016 06:35
|
|
Bad Angus! Bad!
|
Dude this is what the 2x 3S 5200mah Multistar packs on hobbyking are for. I got three for $45 shipped on Black Friday. Add a balance lead splitter, and the $7 charger that uses the balance leads, and you've got 10ah @ 12v, and it's likely smaller and lighter than the original batteries.
|
|
#
?
Feb 26, 2016 06:47
|
|
|
Delivery McGee posted:Do they make rechargeable 6V lantern batteries? I can't find any. But I know they're four F cells, and I found this that should fit in the can. And at 13 amp/hours per cell and eight of them wired in series, that gives 9.6V, which the rig AWOG linked should handle, and 104 hours run time, if I'm doing the math right (calling it 1A instead of .9A draw to simplify things and make the numbers nice and round).
|
|
#
?
Feb 27, 2016 00:50
|
|
|
Parts Kit posted:It's important to remember that wiring batteries in parallel will add to your capacity and that in series it will not. So you'd still be at 13 amp-hours with them in series. Dang, forgot about that. Still, it'll be tons brighter for the same battery life at worst -- I can't imagine that little taillight bulb drawing more than an amp. It just occurred to me that a single flat LED like that doesn't really need a reflector, or at least not a parabolic one like a bulb, does it? Removing the parabolic reflector and putting it right near the lens would give me plenty of space to mount the electronics in the head. I like the idea you mentioned in the TFR thread about using a three-position switch for a lower-current/longer battery life option. Could probably use the stock switch, it had a similar functionality originally. Although it might need to be a special switch for what you were talking about. On the other hand, gently caress it, I'll use the stock switch and have a secondary array of low-power LEDs like the ones in a normal flashlight. Edit: alternate option, I could mount enough lesser LED(s) and all the batteries to run them to far outshine a normal flashlight just in the head (like, at least three or four of those little 3xAAA flashlights would probably fit in there using a shared set of AAs) and use the battery can to hold first aid kit/survival supplies/emergency toolbox. OTOH, go big or go home, right? Hell, I'm pretty sure the guts of two 18v cordless power tool batteries would just about fit in the battery box with room to spare, and the lens is 6" wide ... maybe I'm not thinking big enough. Edit again: what kind of heatsink does this thing need? I'm about to start ordering parts. I think I'll go with the slightly different one linked here, it's ~6000K, like sunlight in the shade; the other one's ~3500K like a tungsten bulb, and as a photographer, I loving hate yellow lights. Do I need a resistor, or does the constant-current 900mA driver linked in the earlier post cover that function itself? I'm new to LEDs, I learned on kits with tiny lil' incandescent bulbs. Chillbro Baggins fucked around with this message at 04:37 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 02:51
|
|
|
From what I've seen people have been using CPU coolers of various types to keep those 100W modules from burning up. You might also be able to get away with cramming a small case fan in there to help keep it cool if you're okay with say putting a few holes in the back of the housing to let air in and out. As for your switch if it has 3 positions chances are pretty good you could rig up a high-off-low circuit with piss all effort. If not beefy switches like that are easy as poo poo to find. Similarly now your project has me wondering if it would be possible to cram all the circuitry and cooling stuff needed for one of those 100w modules into a 3 or 4 D cell maglight. I'm betting the answer is yes. ed: hm, maybe not. Worth a shot though. Parts Kit fucked around with this message at 04:44 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 04:36
|
|
|
Parts Kit posted:From what I've seen people have been using CPU coolers of various types to keep those 100W modules from burning up. You might also be able to get away with cramming a small case fan in there to help keep it cool if you're okay with say putting a few holes in the back of the housing to let air in and out. I do have a couple of dead desktop CPUs that I could loot the coolers from if it needs to be that big. I was hoping for something closer to the 1" cube onboard-GPU heatsink. I could do the fan thing, but I'd also kinda like to waterproof at least the head (ideally the whole thing, but not sure how to work that with my wish to be able to pull the head off and run it off the car battery, and putting a gasket in the battery case will be difficult at best). The switch is a standard SPDT (one wire in, two out, off detent in the middle) and probably should be replaced anyway because it's 50 years old, but it's a standard toggle switch, IDK what I was thinking wanting to keep it. I momentarily forgot they still make them to fit the same hole, I guess. Here's the innards:   I love the clever (but expensive, you don't see that these days) way they made it so it doesn't matter which way you put the batteries in (well, they have to be right-side-up, but you don't have to put the + terminal on any specific corner; nowadays I'd assume you have to orient the battery perfectly and maybe the fancy ones have a dot at each corner with the bare minimum of wire connecting them, back then they just put a bigass brass ring as the + contact.) Any comment on the resistor vs. driver question I edited into my previous post? If it's either/or, which is better for this application? Edit: holy poo poo, the company still exists and is still making the things. Pretty sure the H266M is an updated version of my Model 211 from God knows when (though when you click, it's the page for the one-battery model). I wonder how flexible they are with the lifetime warranty, mine's a bit rusted. (Sadly they don't have email and I hate making phone calls.)
Chillbro Baggins fucked around with this message at 06:15 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 05:24
|
|
|
I think it'll be either or since the main thing is to prevent it from drawing too much current to burn itself out. If you go resistor you should look at the chassis mount resistors, they handle a lot more wattage. On the other hand I was playing with my 10W module earlier and got my specs mixed up (bought a 30v, .3amp one instead of 10v, 1amp by mistake) and turned my power supply up to three times the current it should max at and it still didn't draw more than what it was supposed to. So it's possible (emphasis on possible) that these modules may have some built in deal already. I do remember the guy in the video I linked in TFR didn't put any kind of current controls on his, though I think he was using too small of a step up module to get to max draw anyways.
|
|
#
?
Feb 27, 2016 05:58
|
|
|
Heya Electronics thread, I've got a few DIY repair projects that have piled up, which require a soldering iron. After doing a bit of research, seems like a soldering station is the way to go, are generic Chinese ones on eBay OK? Looking at stuff like this and this.
|
|
#
?
Feb 27, 2016 12:15
|
|
|
Ethics_Gradient posted:Heya Electronics thread, I've got a few DIY repair projects that have piled up, which require a soldering iron. After doing a bit of research, seems like a soldering station is the way to go, are generic Chinese ones on eBay OK? Looking at stuff like this and this. If you're already looking at things in the $86 range just spend the extra $10 and get an Actually Good Iron, the Hakko FX888: http://www.amazon.com/Hakko-FX888D-23BY-Digital-Soldering-FX-888D/dp/B00ANZRT4M
|
|
#
?
Feb 27, 2016 12:24
|
|
|
Parallel Paraplegic posted:If you're already looking at things in the $86 range just spend the extra $10 and get an Actually Good Iron, the Hakko FX888: Those were Aussie dollars (I'm in Australia), looks like the Hakko would run me AU$125.
|
|
#
?
Feb 27, 2016 12:39
|
|
|
Ethics_Gradient posted:Those were Aussie dollars (I'm in Australia), looks like the Hakko would run me AU$125. Oh I didn't even notice  I mean I still say you should get a good one like the Hakko because it's totally worth it, but for a couple of years I got by with this Weller one which works fine too. My dad's an electrical engineer and actually bought one of these Weller ones before I was born and it's still fine and he still uses it regularly 
|
|
#
?
Feb 27, 2016 12:46
|
|
|
Ethics_Gradient posted:Those were Aussie dollars (I'm in Australia), looks like the Hakko would run me AU$125. I think that if you only plan on soldering a small number of things then you should just get something much cheaper than what you linked, but if you plan on doing kits or electronics in general as a hobby you should get a quality one like the Hakko FX888D.
|
|
#
?
Feb 27, 2016 15:20
|
|
|
Random question, what do people use for calculators, graphs and especially algebraic solvers? I just discovered Microsoft Mathematics and it's fantastic. But given that its years old I'm now wondering what else I've been missing. I've also long used wolfram alpha, but it can sometimes be hit or miss in terms of its willingness to solve your problem (and it's not super fast), though it's still the most complete solver I've found (it does laplace transforms and inverse laplace transforms and has many circuit tools etc). Is there anything else out there? Mathmatics lets you leave things in symbolic form and do most of the manipulation that way. For example I started with a circuit and used it to simplify the S domain transfer function and find the roots all in symbolic form. A neat detail is that 'C1' is treated as C subscript 1 (as you can see in the picture) which is really well suited for circuits. At any point you can assign numeric values to C1 and re-evaluate to simplify or find the numeric solution. Or you can assign a symbolic expression to a variable as well, such as z1:=1/(s*c1). Although it appears to run out of steam when you hit complex numbers. Although you can substitute s:=wi and i is seen as complex, it appears that it can't continue to find symbolic solutions in most cases when an i is involved. So I can't get the full symbolic representation of the gain/phase plots out of the transfer function. 
asdf32 fucked around with this message at 19:10 on Feb 28, 2016 |
|
#
?
Feb 28, 2016 19:06
|
|
|
|
| # ? May 29, 2024 04:29 |
|
|
asdf32 posted:Random question, what do people use for calculators, graphs and especially algebraic solvers? I LOVE Mathcad 14. It has a slight learning curve for which equal signs do what but once you spend a few hours with it, it is easy and theres a good base of knowledge online. As a lazy mechanical engineer it is my best friend. In undergrad for those rare classes we got formula sheets or note cards I could have beautifully printed ones in very small font including worked out problems. Keyboard shortcuts are very important. I still use it every now and then in my professional life instead of writing a matlab script to solve some matrix or plot something. It looks quite a bit more advanced in capability than MS mathmatics. CarForumPoster fucked around with this message at 04:27 on Feb 29, 2016 |
|
#
?
Feb 29, 2016 04:24
|
|
