|
catbread.jpg posted:That's a pretty high efficiency requirement for 10 mW. Sure, it would be important if you wanted to extend it much further than 1 metre. quote:If you were sticking to a 1m design range, surely 0.1% efficiency would be tolerable for that power level.
|
#
?
Jul 15, 2009 06:26
|
|
|
|
| # ? May 22, 2024 03:27 |
|
|
What winding configurations have you been looking at? If you've got directionality, a nice broad pancake winding will give good linkage in the mid-field. A limit of a 5cm diameter will be the biggest problem there... Zero crossing detection could be a good way of pumping an LC tank, I've never done anything like that.
|
|
#
?
Jul 15, 2009 09:19
|
|
|
Why does the efficiency have to be 10%?
|
|
#
?
Jul 15, 2009 14:37
|
|
|
catbread.jpg posted:What winding configurations have you been looking at? If you've got directionality, a nice broad pancake winding will give good linkage in the mid-field. quote:Zero crossing detection could be a good way of pumping an LC tank, I've never done anything like that. Cyril Sneer posted:Why does the efficiency have to be 10%? completely arbitrary, as far as I can tell.
|
|
#
?
Jul 15, 2009 19:58
|
|
|
quote:It's good when you just want to drive the hell out of it with no regard for efficiency and you have decent coupling. And most zero crossing/split phase driving schemes don't actually detect the crossings or use any kind of feedback. You'd normally just run the switches in complementary drive and tune the frequency to give the best output. Yeah, that sounds a bit poo poo. Especially when you have a variable resonance. You'd want actual zero crossing detection coupled with a PLL with phase correction.
|
|
#
?
Jul 16, 2009 00:17
|
|
|
Eh, the more I think about it the more apparent it is that I'm going to have to settle for just driving the series resonant tank with a square wave. Class E amps require switches that are both low resistance and tolerate high voltages, which isn't feasible at the frequencies I'm dealing with. At least with square wave drive I don't need to worry about the voltage requirements. Just need to find good FETs with very low losses.
|
|
#
?
Jul 16, 2009 22:44
|
|
|
So, I just made my first SMT board the other day. It's a 3-channel LED dimmer controlled by an ATmega168 (not an Arduino, straight up assembly, bitches  ).Anyway, in order to connect the various connectors and things to the board, I used 1.27mm pads rather than headers because this board can't have any through-hole components. I'm finding that some of the solder pads are starting to lift, though, due to the various wires bending when open the case for debugging. How can I reduce the likelihood of this happening? I'm thinking larger pads and thicker traces leading up to them. Anything else? Bonus: here's the board design. Sorry it's a bit busy.  Also, C1 is shorting out now so I can't enjoy my shiny new dimmer!  Edit: I'm not asking for any advice on C1, I know exactly why it's shorted and what I need to do to fix it. Mill Town fucked around with this message at 09:58 on Jul 17, 2009 |
|
#
?
Jul 17, 2009 09:46
|
|
|
Mill Town posted:So, I just made my first SMT board the other day. It's a 3-channel LED dimmer controlled by an ATmega168 (not an Arduino, straight up assembly, bitches Tape. Or super glue. Really you just need some type of strain relief so that the wires don't exert any force on the pads.
|
|
#
?
Jul 17, 2009 14:36
|
|
|
Strain relief will help (though it's hard if you can't drill any holes in the board), and so will using stranded wires.
|
|
#
?
Jul 17, 2009 18:19
|
|
|
Yeah, using solid wire is pretty much asking to rip pads off the board.
|
|
#
?
Jul 18, 2009 00:16
|
|
|
When soldering wires directly onto traces or pads, it's always a good idea to tack down a little bit of glue right by the pad for strain relief. Otherwise you're just asking to lift the pad and the trace along with it. Superglue is aces if you're sure you want to keep the wire there. If you want something a little bit less permanent, I'd recommend hot glue. I use it all the time in prototypes because it's somewhat flexible when dry and melts right off. Hot glue is also really good if you ever need to lift a pin from an surface-mount IC and solder to it. Without something for mechanical stability, your chances of ripping the pin right off of the IC increase significantly. Here's what I usually do when I want "just a dab" of hot glue or when I need to put a little bit under a pin I've lifted. I cut off a tiny tiny piece of the glue stick and just put it wherever I want the glue to be. Then I dab it with a fine-tip soldering iron and it melts pretty much instantly. Since it melts at a much lower temperature than solder, you can do it without desoldering anything if you're careful. Seriously, this is some protip poo poo right here
|
|
#
?
Jul 18, 2009 08:16
|
|
|
Poopernickel posted:When soldering wires directly onto traces or pads, it's always a good idea to tack down a little bit of glue right by the pad for strain relief. Otherwise you're just asking to lift the pad and the trace along with it. Thanks. Everything already had tape on it and was either stranded 26 gauge wire or 30 gauge wirewrap wire so "hot glue" is pretty much the only thing I haven't done. I'll try this on the next iteration of the board. For now, the repaired board with a few extra components stuck in convenient places is pumping out light next to me. 
|
|
#
?
Jul 18, 2009 08:30
|
|
|
This thread has inspired me to learn more about electronics. I am basically starting from zero. I've been reading some basic stuff and want to get dirty with a soldering iron and some simple kits. Would I be better off looking at schematics and getting the individual parts and generic boards or the pre-packaged kits?
|
|
#
?
Jul 19, 2009 21:24
|
|
|
ANIME AKBAR posted:Eh, the more I think about it the more apparent it is that I'm going to have to settle for just driving the series resonant tank with a square wave. Class E amps require switches that are both low resistance and tolerate high voltages, which isn't feasible at the frequencies I'm dealing with. At least with square wave drive I don't need to worry about the voltage requirements. Just need to find good FETs with very low losses. What frequencies are you working at? Again, the solid state tesla coil guys operate in the ~100s kHz range.
|
|
#
?
Jul 20, 2009 02:40
|
|
|
Cyril Sneer posted:What frequencies are you working at? I was aiming for 1MHz, but I'll likely have to go lower to avoid huge losses in my capacitors. And again, tesla coils are completely impractical for the loads I'm driving. Those guys are driving kilowatts into their loads; I'm driving milliwatts. And it doesn't scale down nicely.
|
|
#
?
Jul 20, 2009 22:59
|
|
|
therunningman posted:This thread has inspired me to learn more about electronics. I am basically starting from zero. I've been reading some basic stuff and want to get dirty with a soldering iron and some simple kits. It largely depends on what you are trying to do. If you can find pre-packaged kits that interest you, by all means try them out. If you know what you want to do, I suspect there is very few kits that meet your interest, and getting parts and reproducing schematics that you have found is probably the way to go. What are you interested in doing? Any particular kits interest you? You can build am amazing number of things with a small selection of through-hole components and a breadboard, for what it is worth.
|
|
#
?
Jul 22, 2009 22:20
|
|
|
Thanks for the reply. I hope to ultimately work my way up to interactive light and sound displays. I realize that it is pretty far off at this point. I've decided to start with these http://www.kpsec.freeuk.com/proj.htm small projects. All the parts are cheap as well.
|
|
#
?
Jul 23, 2009 23:28
|
|
|
How do you feel about programming? What you want to do sounds like it would benefit from knowing how to use microcontrollers but those require more programming skill than electronics skill.
|
|
#
?
Jul 24, 2009 00:44
|
|
|
I'm looking for a product, kind of like a heat sink grease, that can be applied to mechanical electrical connections to help improve electrical conductivity and prevent future oxidation of the metal. Does such a product exist, and if so, what is it called and where do I get it? If it does not exist, I could use the names and phone numbers of a good chemist and patent attorney. Thanks for any help!
|
|
#
?
Jul 24, 2009 02:29
|
|
|
BattleMaster posted:How do you feel about programming? What you want to do sounds like it would benefit from knowing how to use microcontrollers but those require more programming skill than electronics skill. If the programming required is something that a reasonably intelligent person could be expected to slowly pick up, then it sounds like something for me. I would be starting at "hello world", so any kind of early direction would be important to be put on the right track early. Looking around briefly I keep seeing the name "Arduino". Good starting point? therunningman fucked around with this message at 03:25 on Jul 24, 2009 |
|
#
?
Jul 24, 2009 03:20
|
|
|
therunningman posted:If the programming required is something that a reasonably intelligent person could be expected to slowly pick up, then it sounds like something for me. Yes, the Arduino is really easy to get started with, has a great development environment and a lot of interested community members who can help you out, and has bits specifically geared towards interactive light and sound installations and other forms of art.
|
|
#
?
Jul 24, 2009 03:37
|
|
|
The Arduino is an Atmel AVR microcontroller with training wheels. The PICAXE is a similar product but with a Microchip PIC microcontroller instead. Those are the two microcontroller brands that you'll find used the most in hobby projects and even consumer electronics (I've found PICs in joysticks and remote controls) so either one of those will be a great starting point.
|
|
#
?
Jul 24, 2009 04:34
|
|
|
dv6speed posted:I'm looking for a product, kind of like a heat sink grease, that can be applied to mechanical electrical connections to help improve electrical conductivity and prevent future oxidation of the metal. Switch lubricant, or Contact cleaner with lubricant is what you're looking for.
|
|
#
?
Jul 24, 2009 04:40
|
|
|
There should be a game called 'find the PIC'. Just rip something vaugely dealing with electricity apart and see how many PIC microcontrollers you can spot. There's a good reason PICs and AVRs tend to be used a lot. They're incredibly rugged, easy to use processors. The other day, a friend of mine had a board that was powering up AVRs (Arduino Nanos actually) without anything connected on the 5V rail. Turns out he was pushing 24V into one of the analog pins, which was actually powering the part! He fixed the problem, and the analog pin still works correctly! For a lot of applications, you really don't need much more than a PICAXE or Arduino. Both are nice due to their ease of use. Arduino happens to be open source, and I see a ton of projects out there for it. I happen to like the Arduino better, but both are perfectly reasonable starting points. If you get an Arduino, I recommend starting with a Duemilanove. Those tend to have the most useful accessories, and most projects are based around them. The Arduino Nano is another good one, and it has the advantage of being able to fit right into a breadboard.
|
|
#
?
Jul 24, 2009 04:43
|
|
|
clredwolf posted:There should be a game called 'find the PIC'. Just rip something vaugely dealing with electricity apart and see how many PIC microcontrollers you can spot. The AVRs generally have supply-voltage clamping diodes on their inputs, which means you can pull a ton of stupid AVR tricks like this one: http://scanwidget.livejournal.com/32928.html
|
|
#
?
Jul 24, 2009 05:41
|
|
|
I know this is pretty simple, but I just finished my portable iPhone/usb device charger. I'm going to Vegas in a few weeks and my 1st gen iphone is starting to show it's age when playing games or watching videos. I wanted to get some extra juice while on the strip. I know it is pretty crude, but it was my first time using a soldering iron with such small components. I'm wish it could have been a bit smaller, but I'll work on that for the next one. I used 2 9v batteries in parallel, a 100uF capacitor, 5v regulator, .1uF capacitor, switch, usb connector, and a lot of silicon. EDIT: Also, had to use a 12k ohm resistor to get the iPhone to charge. Here is where I got the instructions, however, I think a few things on his diagrams are wrong. http://www.instructables.com/id/SF007H5F5LDFNMM/  Click here for the full 1600x1200 image.  Click here for the full 1600x1200 image.  Click here for the full 1518x795 image.  Click here for the full 1600x1200 image.
|
|
#
?
Jul 26, 2009 15:00
|
|
|
So I've got a neat little project that I'm working on in my spare time and I was wondering if I could get some input on setting up glue logic. What I'm trying to do is interface a flash chip with an old NES along with a pair of SRAMs such that the NES loads it's program ROM data and graphics ROM data from the flash chip into the SRAM (program data in one, graphics in the other). This itself doesn't have much use but if I can get this working I hope to take this into some other directions. The graphics data isn't a problem as it has it's own CPU and is on a separate bus. The program data is where I'm having an issue: The CPU exposes a 16 bit address bus and a 8 bit data bus. On power up the program ROM in the cartridge is mapped into 0x8000-0xFFFF in the CPU's address space and execution begins at 0x8000. I'm going to wire in my flash chip to 0x8000 so that is executes my code on power up. This code will copy itself into a tiny bit of unused internal memory available on the CPU and then proceed to copy data from the flash chip to the SRAM. When it's finished it will begin executing the program ROM now in the SRAM, this will also have to be located at 0x8000. My problem is, the flash chip and SRAM both have to be mapped into the 0x8000 address space. I need a way to read from the flash, disable the flash and enable the SRAM, write to the SRAM and then disable the SRAM and re-enable the flash and start the process over again. When everything is copied I need to permanently disable the flash/enable the SRAM until power down while leaving the CPU's address space unchanged. I think I've got everything worked out except for one detail. I'm going to borrow the 0x6000-0x7FFF address space which I know to be unused by the CPU while I'm doing my copying. Whenever an address in that space is placed on the address bus, I'll latch the current value of D0 which will determine the /CE lines for my flash and SRAM (one will be inverted). I figure I can do this with a pair of inverters, a latch and an AND gate (!A15 && A14 && A13 = latch D0). My problem is that after my copy is finished how do I permanently disable the flash and enable the SRAM? The CPU will need the 0x6000-0x7FFF range back so I can't simply rely on no more data being written there to maintain the value of the latch. Sorry for the massive background for one little question.
|
|
#
?
Jul 28, 2009 13:25
|
|
|
Would you be able to post a diagram of the physical connection of the hardware? I think one solution might lie with the use of more data pins (and an appropriate section in the software) to encode a value which sets up another latch (with power reset logic), that overrides the first latch through more logic (OR, before going to your /CE inverter section).
|
|
#
?
Jul 28, 2009 22:46
|
|
|
Why does the NES need to copy the data out of flash and into SRAM? You could use a small microcontroller to look at the address lines and move data around as needed. This would probably be the same price as buying some logic gates if you want DIP chips. IIRC some larger NES games had special memory mapping chips to do just this.
|
|
#
?
Jul 28, 2009 23:32
|
|
|
catbread.jpg posted:Would you be able to post a diagram of the physical connection of the hardware? I think that will work just fine, thanks! I dusted off my digital logic book and it looks like if I add a second SR latch with it's reset tied to ground, input tied to the second data pin and output put in OR with the input to the second latch, it will act as a one shot. Once it's set it will always force the second latch to 1 enabling SRAM until a power cycle. All my notes are in my notebook right now and I don't have a scanner but if you want to get an idea of what I'm working with you can view the CPU/PPU and memory schematic here (not shown is the power and video circuitry):  I'm hooking into the cart connector shown in the upper right. PRG refers to the program ROM and CHR the graphics ROM. CIRAM is the internal 2k RAM tied directly to the PPU. The CPU also has 2k internal but it's /CE signal is not exposed on the cart connector. quote:Why does the NES need to copy the data out of flash and into SRAM? You could use a small microcontroller to look at the address lines and move data around as needed. This would probably be the same price as buying some logic gates if you want DIP chips. IIRC some larger NES games had special memory mapping chips to do just this. That's the other way to do it. My ultimate goal is to have ROM's loaded off a compact flash cartridge via a menu system. Other people have done this so I'm basically reinventing the wheel but it's a great project for me since it covers a lot of things that I've wanted to learn about. I did think about using a micro controller but I thought it would be more elegant to have the NES CPU do all the work (plus I know it's possible as someone else managed to pull it off). The mappers themselves are a whole other beast. Any game worth playing uses a mapper. The only ones that don't are Super Maro Brothers, Excite Bike and a handful of single screen arcade conversions (Burgertime, Joust, etc). I initially toyed around with using a micro controller to simulate the mappers through software emulation. Ultimately I decided to use a FPGA instead, mostly because my experience with them is very limited and I wanted to learn more. It also potentially could do away with some of the inaccuracies I hear exist with software emulation of the mappers.
|
|
#
?
Jul 29, 2009 03:41
|
|
|
Keebler posted:I did think about using a micro controller but I thought it would be more elegant to have the NES CPU do all the work (plus I know it's possible as someone else managed to pull it off). The mappers themselves are a whole other beast. Any game worth playing uses a mapper. The only ones that don't are Super Maro Brothers, Excite Bike and a handful of single screen arcade conversions (Burgertime, Joust, etc). I initially toyed around with using a micro controller to simulate the mappers through software emulation. Ultimately I decided to use a FPGA instead, mostly because my experience with them is very limited and I wanted to learn more. It also potentially could do away with some of the inaccuracies I hear exist with software emulation of the mappers. Cool, if you want to look into emulating mappers in an FPGA you should check this (http://tripoint.org/kevtris/Projects/console/) out and maybe contact the author.
|
|
#
?
Jul 30, 2009 01:45
|
|
|
Poopernickel posted:Just to check - you do mean 40-60 _milliseconds_ right? That's pretty reasonable based on the system you're describing. I would bet that 90% or more of the latency lies within the operating system and driver. I'm sorry it took so long to get back to this. I appreciate all of the information. In fact, I wound up installing real time Linux and have gotten latencies down to around 10 microseconds which is pretty drat good. There's still some issues to work out though with various poo poo causing massive latency fairly randomly (think DMA, TSC). For that, I'd like to move to FPGAs but it doesn't seem like it will happen. There is in fact a DAQ made by National Instruments that has an FPGA built into it. Unfortunately, since my project mainly exists to suck up the extra scraps of my professor's budget, I doubt I'll ever see that being purchased.
|
|
#
?
Jul 30, 2009 22:09
|
|
|
Keebler posted:I did think about using a micro controller but I thought it would be more elegant to have the NES CPU do all the work (plus I know it's possible as someone else managed to pull it off). The mappers themselves are a whole other beast. Any game worth playing uses a mapper. The only ones that don't are Super Maro Brothers, Excite Bike and a handful of single screen arcade conversions (Burgertime, Joust, etc). I initially toyed around with using a micro controller to simulate the mappers through software emulation. Ultimately I decided to use a FPGA instead, mostly because my experience with them is very limited and I wanted to learn more. It also potentially could do away with some of the inaccuracies I hear exist with software emulation of the mappers. The only problem I can think of with using a microcontroller to software-emulate a memory circuit is speed. For instance, one time I designed a parallel ATA hard drive interface for a Z80 system but I was unhappy with the fact that it required several 74-series logic gate ICs. I tried to replace them with a single low-pin PIC16. The idea was that the input pins would change and wake the microcontroller up, which would then read the value of the input port as a number, feed it into a lookup table, and write the resulting number to the output port. It worked perfectly but even though the software was written completely in assembly and only needed to execute about 6 instructions every time the inputs changed, which is as efficient as you can possibly get it, it was still too slow to interface with a Z80 without adding wait-states. Now that was a PIC16 running with the internal 8MHz oscillator, so maybe it would be fast enough if I useda PIC18 with proper table read instructions and an external 40MHz crystal. I just ended up scrapping it and plan on redoing the design with a proper programmable logic device, though. BattleMaster fucked around with this message at 02:48 on Jul 31, 2009 |
|
#
?
Jul 31, 2009 02:45
|
|
|
Anyone know where I can get a 0-5k potentiometer with a 90 degree sweep?
|
|
#
?
Aug 4, 2009 17:57
|
|
|
withdrawn
Canary Yellow fucked around with this message at 12:10 on Aug 5, 2009 |
|
#
?
Aug 5, 2009 08:57
|
|
|
Slung Blade posted:Anyone know where I can get a 0-5k potentiometer with a 90 degree sweep? you mean a 90 degree range of rotation? can't say I've seen one, but one likely exists.
|
|
#
?
Aug 5, 2009 15:02
|
|
|
Here's a nifty little chip I just started to use in some designs: NC7SZ57 and NC7SZ58 http://www.fairchildsemi.com/ds/NC/NC7SZ57.pdf It's a 6 pin configurable 2 input logic gate. Rather than having a bunch of different bins for simple gates, keep a few of each flavor of these around. Here's what it can do: code:The NC7Z57 uses this logic: Y= (I0')•(I2')+(I1)•(I2) The NC7Z58 uses this logic: Y= (I0')•(I2')+(I1')•(I2)
|
|
#
?
Aug 5, 2009 15:49
|
|
|
ANIME AKBAR posted:you mean a 90 degree range of rotation? can't say I've seen one, but one likely exists. Yep, exactly. I'm thinking I'm going to have to find a gearing system to translate the 90' into the 270' that most pots offer.
|
|
#
?
Aug 5, 2009 15:53
|
|
|
Just attach a head to the pot that has a peg hits something on the chassis at 90'.
|
|
#
?
Aug 5, 2009 16:03
|
|
|
|
| # ? May 22, 2024 03:27 |
|
|
Alternately, depending on whether your measurement circuit is end-to-end resistance based (and depending on how it treats the polarity of low vs high resistance), or 3-terminal voltage based with an input-wiper-ground, you could: a) use a bigger value pot and sweep less of it b) use a higher input voltage to compensate electrical solutions to an electrical problem 
|
|
#
?
Aug 5, 2009 22:50
|
|