|
Acid Reflux posted:If you're not a member already, scoot over to the Replica Prop Forum and ask this same question. There are a couple of guys over there who do stuff like this all the time and would probably be more than willing to help you out. drat, I didn't even think of them and I've been to that site before too. I've mostly been looking at The 405th, but that's more for the armor. And that blaster is sexy as hell.
|
#
?
Dec 16, 2010 05:50
|
|
|
|
| # ? Jun 11, 2024 11:26 |
|
|
Thanks for the help so far, guys.ANIME AKBAR posted:I've done a fair amount of custom magnetics design, mainly for switching power supplies and current transformers. Anything in particular you want to know? What's your application? I'm mainly interested in power conversion. In particular, I saw a concept for a boost converter that uses a tapped inductor that I find interesting and want to experiment with. Other than that, though, it would be nice to be able to make custom inductors because there's so much variance that it's hard to experiment with a supply of off-the-shelf parts. But yeah, if you have any tips for tapped inductors that would be amazing. Thanks.
|
|
#
?
Dec 16, 2010 06:04
|
|
|
BattleMaster posted:Thanks for the help so far, guys.
|
|
#
?
Dec 16, 2010 06:27
|
|
|
ANIME AKBAR posted:Is that tapped inductor (autotransformer, really) for getting high step-up ratios or something? Yeah pretty much. (I swear I knew they were called autotransformers, it's just I had a brain fart haha) I read a document Maxim put out that talks about one of their boost converter chips. It uses an autotransformer instead of a normal inductor, which avoids subjecting the switch to the total output voltage as with a normal boost converter, which works around some of the disadvantages of high-voltage MOSFETs. It also lets them get away with converting a very low voltage into a much higher voltage without low frequencies (and thus big inductors) or unworkable duty cycles. I thought it was a pretty neat concept so I wanted to try it out myself. BattleMaster fucked around with this message at 06:48 on Dec 16, 2010 |
|
#
?
Dec 16, 2010 06:46
|
|
|
BattleMaster posted:Yeah pretty much. (I swear I knew they were called autotransformers, it's just I had a brain fart haha) Well there's nothing terribly different about using an autotransformer instead of an inductor. It's certainly easier than using a flyback transformer, since leakage inductance is not a big deal. Winding it shouldn't be any harder; just do two sets of windings and connect them in series. If you're going for very high voltages or very high frequencies, you'll want to avoid multiple layers of windings, since that will make insulation faults a concern, and boost the parasitic capacitance quite a bit. Having one end of your inductor capacitively coupling to the other side is bound to cause some bad waveforms. The only really tricky part is probably the saturation current calculations. Just remember that if the core doesn't saturate in during the on time, it won't during the off time either (because core flux is continuous). So if the "primary" inductance doesn't saturate, then the entire autotransformer won't either. So you can take pretty much the same approach as you would with any normal inductor. Do you have a core in mind?
|
|
#
?
Dec 16, 2010 14:52
|
|
|
Thanks. No, I'm just doing some preliminary research before I start coming up with a design.
|
|
#
?
Dec 16, 2010 15:03
|
|
|
From the "oh god what have i done" category - I just started learning electronics recently, mainly from this book http://www.amazon.com/MAKE-Electronics-Learning-Through-Discovery/dp/0596153740 Anyway, I just got my first breadboard yesterday, and after doing some simple circuits to get used to it, mainly just powering a LED, I managed to possibly kill an ac/dc converter and I am not sure how. Everything was working right and then I added a spdt switch into the circuit via some solid core wire and alligator clips (it was too big to fit into the breadboard holes), I flipped the switch to close the circuit, the LED fried, and the ac/dc adapter stopped working. I know why the LED fried, in hooking up the switch I had to move the resistor placement and I do not think it was fully in, I had jumper wire bridging rows so that it wouldn't matter if the resistor was fully in or not. Not making that mistake again. However not sure why the ac/dc adapter stopped working. I can still pull a reading off of it when using my multimeter battery tester mode , so 3v battery test and the ac/dc adapater set to 3v reads 3v. However it won't power anything. If i setup the same circuit again, with a new LED, i can power it fine via batteries so I don't think its anything with the breaboard. In fact if i pull reading from the jumper wires nearest to the +/- connections on the board, it still shows a correct reading. I am going to be picking up a benchtop powersupply soon instead of mangling another ac/dc adapater, but I just want to figure out what would break the adapter in the first place and why it still shows voltage properly.
|
|
#
?
Dec 16, 2010 20:20
|
|
|
Whats the mA rating for your ac/dc power supply? Try hooking it up to a 20 Ohm resistor and checking your voltage then. My feeling is that when you hooked up your circuit you caused a short, this did some damage to some part of the device and now it can't maintain voltage under load. If you hook it up and assuming your little power supply is rated higher the 150 mA then you should read 3v across the resistor. If you read less then that, or zero volts then we know you have damaged it and you should throw it away. A lab supply will have a amperage limiting control and fuses. That way you can limit the max amperage it will produce, that should help protect you from damaging anything with overcurrent. Worse case scenario you blow a fuse, but your power supply is saved from damage. E. Changed resistor value. tworavens fucked around with this message at 09:49 on Dec 18, 2010 |
|
#
?
Dec 18, 2010 09:34
|
|
|
It was rated for 1a. I figured it was something like that, already threw it away.
|
|
#
?
Dec 18, 2010 17:11
|
|
|
I'm having problems with LTSpice. I want to make an idealised boost converted, with a 555 as the timer. No feedback or anything fancy. I want the diode to have 0 voltage drop, infinite breakdown. I want the transistor to switch instantly, etc.
|
|
#
?
Dec 19, 2010 01:33
|
|
|
My parents rangehood controller board blew up, and in the process of reverse engineering it I realized the circuit is very similar to something I built once, and I was thinking it would be interesting to document the tear-down and publish it online; there are a few 'interesting' design choices and I'm sure I'm not the only person who would find it educational. However, I'm not sure of the legality of this. My understanding is that the actual circuit layout is protected by copyright, but a circuit diagram explaining its functionality would not be unless it is literally a copy of the original design diagram (and I don't have access to that, of course). There are no patent numbers on the board (and I really wouldn't expect any for this) but even then I'm not producing copies for profit, so it doesn't seem like a big deal even if protected by patents (I don't think it is). There is no encryption or anything, it's straight passives and logic ICs, so there isn't anything that DMCA would protect as far as I know. However, I don't consider myself an expert on this stuff. I think if I post photos of the board, and my own writeup and schematics, I should be okay. Is there something I'm missing? Unparagoned posted:I'm having problems with LTSpice. I want to make an idealised boost converted, with a 555 as the timer. No feedback or anything fancy. I want the diode to have 0 voltage drop, infinite breakdown. I want the transistor to switch instantly, etc. 
|
|
#
?
Dec 19, 2010 09:47
|
|
|
Delta-Wye posted:
|
|
#
?
Dec 19, 2010 13:45
|
|
|
Delta-Wye posted:My parents rangehood controller board blew up, and in the process of reverse engineering it I realized the circuit is very similar to something I built once, and I was thinking it would be interesting to document the tear-down and publish it online; there are a few 'interesting' design choices and I'm sure I'm not the only person who would find it educational. Worst thing that could happen is they send you a cease-and-desist order to take it down. That's incredibly unlikely since a lot of other people have definitely gotten away with similar projects.
|
|
#
?
Dec 19, 2010 16:20
|
|
|
Unparagoned posted:I'm having problems with LTSpice. I want to make an idealised boost converted, with a 555 as the timer. No feedback or anything fancy. I want the diode to have 0 voltage drop, infinite breakdown. I want the transistor to switch instantly, etc. LTSpice tends to not like ideal components, such as infintely fast switches with 0 resistance. It makes numerical integration impossible to do correctly. There are components for voltage controlled switches and you can change the model file for the diode to make it "more ideal" if you want. To enter a .model command, just press "t" to drop down text and type in your command. Don't forget to specify that you want a spice directive instead of a comment. Then you can Ctrl right click on the device you want to change (the diode or switch or whatever) and change the "Value" field to match your model. Here are some example .models for a switch (2.5 volt threshold) and diode: .model IdealSwitch SW(Ron=1m Roff=1G Vt=2.5 Vh=0) .model IdealDiode D(Ron=1m Roff=1G Vfwd=0) Since you're going with ideal models, I'd also suggest using a square wave voltage source in place of the 555. Just think of it as an ideal 555 circuit  Make sure to read the manual (http://ltspice.linear.com/software/scad3.pdf) and look through the examples that are installed. They cover most of what you might want to do. edit: If you really want to use a 555, they have an ideal model in their library. Just hit F2, go to "Misc" and select the NE555.
|
|
#
?
Dec 19, 2010 18:29
|
|
|
SnoPuppy posted:LTSpice tends to not like ideal components, such as infintely fast switches with 0 resistance. It makes numerical integration impossible to do correctly. There are components for voltage controlled switches and you can change the model file for the diode to make it "more ideal" if you want. Square wave voltage instead of a 555 is the way to go, I only really mentioned the 555 so no one mentioned a special boost chip. Although LT spice can't create a square wave voltage source. Anything close to a square wave makes it feel like I'm running a 486. I'll play about with less square waves, square waves. Should be good enough to use to check if my program models it correctly and then use my program. Edit: Can't seem to get it to do anything useful :/ FSMC fucked around with this message at 20:16 on Dec 20, 2010 |
|
#
?
Dec 20, 2010 19:54
|
|
|
Unparagoned posted:Square wave voltage instead of a 555 is the way to go, I only really mentioned the 555 so no one mentioned a special boost chip. Although LT spice can't create a square wave voltage source. Anything close to a square wave makes it feel like I'm running a 486. How are you creating a square wave? You should just drop down a voltage source, go to Advanced, select PULSE for the function, and then set the parameters you want. It work perfectly fine for me with 0ns rise/fall times. For a 1 MHz 50% duty cycle square wave, I can finish a 1 ms long transient simulation in a second or less for the simple circuit you're describing (single inductor, single diode, single switch, single output cap). Is this similar to what you have? 
|
|
#
?
Dec 20, 2010 21:06
|
|
|
SnoPuppy posted:How are you creating a square wave? You should just drop down a voltage source, go to Advanced, select PULSE for the function, and then set the parameters you want. It work perfectly fine for me with 0ns rise/fall times. Here is what I have http://rapidshare.com/files/438445513/boost.asc Things to note. For the square voltage if you put 0 as the rise and fall times, it defaults it to 10%. I wanted the inductor to charge first rather than trying to charge the capacitor with 0 current through the conductor. Depending on whether I put the square wave as on or off first gives completely different results, one has current as 9kA at t=0 the other the capacitor has 9V at t=0. Both behaviors seem wrong. Any chance you could stick your circuit up on rapidshare or whatever so I can see what I'm going wrong. Edit: I think I got it working. I needed to skip the initial operating point. Nope results still confusing. FSMC fucked around with this message at 00:57 on Dec 21, 2010 |
|
#
?
Dec 20, 2010 23:12
|
|
|
Hey while we're talking about square waves, how does capacitive/inductive reactance work with square waves? The formulas for both have 2*pi*f in them so I assume they're for sine waves. The reason I asked is that I'm doing a little mini-project to make an electronic LED ballast that pulses power through a capacitor to control power to the LED. Yeah it's completely pointless because the extra cost would be disproportionate to the energy savings, but it was something I could put together in a few minutes. Right now I have it working and while my multimeter doesn't want to measure the current (probably because it's pulsed) the brightness of the LED varies as expected as I vary the capacitance and frequency. The only problem is I'm not sure how accurate my math is! Edit: Typo BattleMaster fucked around with this message at 00:01 on Dec 21, 2010 |
|
#
?
Dec 20, 2010 23:56
|
|
|
BattleMaster posted:Hey while we're talking about square waves, how does capacitive/inductive reactance work with square waves? The formulas for both have 2*pi*f in them so I assume they're for sine waves. A square wave is an infinite sum of odd multiples of your fundamental frequency. The amplitude of the harmonics is 1/n so in practice, after the 5th harmonic, amplitudes drop off rapidly and can be neglected. See: http://mathworld.wolfram.com/FourierSeriesSquareWave.html
|
|
#
?
Dec 21, 2010 00:41
|
|
|
Unparagoned posted:Here is what I have Well for one, you're trying to do a 2 second long transient simulation, so that will take a long rear end time on any machine, under most conditions. The reason you see different values for different initial conditions is because the simulator first attempts to work out the initial DC operating point. It solves the system by treating capacitors as opens and inductors as shorts. So when you start with 5v, the tool realizes that the transistor is turned on, so you are shorting the inductor to ground. Since you didn't specify any series resistance with the inductor, it defaults to 1 mOhm, resulting in the 9kA you see (it assumes a diode drop across the transistor). A similar thing happens with the capacitor when you start at 0v. Additionally, you have no load resistance - an ideal capacitor will continue to store all the charge you give it and will just keep increasing in voltage. Add a load resistor to bleed off some of this charge and let you get to steady state. Take the inductor again - remember that the voltage across it is equal to L*dI/dt. That means that if you have a 1 uH inductor and try to switch 1 amp in 2ns, you would get a voltage of 500V! This is why doing "ideal" simulations leads to very strange results that obviously don't match reality. I'd just pick some reasonable components and go from there.
|
|
#
?
Dec 21, 2010 01:21
|
|
|
SnoPuppy posted:Well for one, you're trying to do a 2 second long transient simulation, so that will take a long rear end time on any machine, under most conditions. I've tried a few variations and the results match up fairly well to my excel based model. I realise there is no load, I never intended to analyse a boost converter it just turned out the circuit I was looking at was really close to one. I do intend adding a load but wanted to get my head round this simpler circuit first. If you let me know how I can buy you a plat for your help.
|
|
#
?
Dec 21, 2010 02:36
|
|
|
BattleMaster posted:Hey while we're talking about square waves, how does capacitive/inductive reactance work with square waves? The formulas for both have 2*pi*f in them so I assume they're for sine waves. It depends entirely on the circuit itself. As mentioned, a square wave is really a collection of an infinite sum of sines across frequency. Feeding a square wave into a lowpass filter (RC circuit) will attenuate the high frequncy components producing an output waveform with rounded edges. Feeding a square wave into a highpass filter will give a weird kind of sloped-pointy waveform as the low frequencies are attenuated.
|
|
#
?
Dec 21, 2010 05:52
|
|
|
Thanks guys, I think I have it figured out now. I guess it won't be exact because I'll need to pick a limited number of frequencies to sum, but it will be more accurate than what I have.
|
|
#
?
Dec 21, 2010 06:04
|
|
|
Unparagoned posted:Thanks for the help. The option "skip the initial operating point" seemed to fix the weird initial conditions problem. You're welcome, glad I could help
SnoPuppy fucked around with this message at 23:07 on Dec 21, 2010 |
|
#
?
Dec 21, 2010 18:38
|
|
|
I was asking a bunch of questions a couple months ago about hacking into my car's starting mechanism and controlling it with my Android phone. It worked, and I ended up going with relays and transistors instead of transistors only (loving Ahab right here). Video: http://vimeo.com/18066415
|
|
#
?
Dec 22, 2010 00:28
|
|
|
I cant believe I am a tard and forgot to snag a picture before I left work but basically we have some old CNC mills which are prone to blowing fuses on certain axis when they rapid 100%/over-travel. My boss orders the fuses by part number and pays out the butt that way. I am betting they're some generic fuse that we can find cheap once we know the name of it. Ill update with a picture when I can, however in the mean time maybe you all could help. The fuse: -Has three prongs (with continuity across two) -When it blows a little metal thing flips up indicating it has blown -Measures about 1"x.375" from a top view -Comes in many common amperages like 3.0, 5.0, 7.5, 15. Googling types of sues or CNC fuses, etc. didnt help. EDIT: Annnd that was fast.....theyre called Daito fuses. AnomalousBoners fucked around with this message at 01:04 on Dec 22, 2010 |
|
#
?
Dec 22, 2010 01:02
|
|
|
So I'm planning my first proper electronics build - 8-bit R-2R DAC with shift register input - and have some dumb idiot questions: 1. An R-2R is just a bunch of overlapping 2:1 voltage dividers, hence the R and 2R values. But does the actual ohmage of R make any difference to the performance of the divider? Would 1Ω/2Ω work as well as 10MΩ/20MΩ? 2. I think I need an output buffer to lower the impedance, but I have no idea which op-amp would be suitable (ignore the model number on the schematic, EAGLE doesn't have a generic op-amp). Any tips on choosing parts like that? 3. did i do it all wrong  Thanks a lot.
|
|
#
?
Dec 22, 2010 20:46
|
|
|
The resistance should depend on how much current your LS595 can source. Using too low a resistance will cause obvious problems. I am sure a buffer will be necessary. Buffer ICs are cheap. You select one the same way you select an opamp. Frequency response/slew rate are important, as are the various biases. I can't say I know much about this DAC configuration or what you plan to use it for, but an antialiasing filter seems necessary.
|
|
#
?
Dec 22, 2010 20:54
|
|
|
I think most opamps connected in that configuration will have a nearly infinite input resistance, at least for FET based models.
|
|
#
?
Dec 22, 2010 21:17
|
|
|
repiv posted:So I'm planning my first proper electronics build - 8-bit R-2R DAC with shift register input - and have some dumb idiot questions: 1. The size of your resistors will depend on how much current the shift register can source, and how much noise your output can tolerate. Since this is a homemade 8 bit DAC, you shouldn't worry about the thermal noise much. Just stick with 1K/2K 2. How much current do you want to drive? That will determine what op amp you select. I'd also throw a 1nF cap (or less) on the + input of the opamp, to slow your edge down. It creates a low pass filter that should help to prevent ringing and keep the opamp stable if it has high bandwidth. You should also make sure the opamp is rail-to-rail capable, otherwise you need a negative supply (if you care about DC accuracy near the rails, you need this anyway).
|
|
#
?
Dec 22, 2010 22:39
|
|
|
I have a question for those of your who are power experts. This is mainly to satisfy my own curiosity. What is the most efficient way of rectifying 115 volt RMS at 10 amps? I know that a bridge rectifier is the most traditional, but at 10 amps, we're talking ~32 watts (assuming 0.8v drop). I'm not so concerned about the actual power being lost, but cooling a 32 watt bridge rectifier won't be simple. I suppose I could try to get a center tapped transformer and get away with two diodes, but I'd still have to cool 16 watts. I had also though about running an H bridge in reverse (monitoring the incoming AC waveform with an ADC and control the switches with a micro), but that also seems to have problems. All the MOSFETs I can find that tolerate the voltage have 0.3 ohm of on resistance or more, leading to 30 watts per FET! Any ideas?
|
|
#
?
Dec 27, 2010 22:58
|
|
|
Been awhile since I messed with electronics... Hinted that I wanted this: http://www.transistorclock.com/ for Christmas and the GF obliged. Its a clock built entirely with nothing but diodes, resistors, transistors, and capacitors. No ICs at all. About 1500 discrete components altogether. Several hours of work and I am about half way in, everything works, no mistakes and I've become a pro at soldering...  lovely cellphone pic.. Only have the seconds and tens of seconds chains finished thus far. Its a bit on the expensive side, but if anyone else wants to go for this kit I recommend it. It comes nicely packaged. The components are identified by custom symbols. Each component and value has its own symbol which is silk screened on the PCB and printed on the small bags the components come in. You really can't miss place a part unless you are completely not paying attention. The instruction book is very detailed and includes a lot of trouble shooting info should you need it. stevewm fucked around with this message at 02:47 on Dec 28, 2010 |
|
#
?
Dec 28, 2010 02:36
|
|
|
Oh god why did they use through-hole  that would take me a week to complete, but I could knock out an SMT equivalent in probably 8 hours.
|
|
#
?
Dec 28, 2010 03:36
|
|
|
madpanda posted:From the "oh god what have i done" category - Just ordered the book with some gift cards from the holidays. Looks pretty awesome.
|
|
#
?
Dec 28, 2010 04:28
|
|
|
SnoPuppy posted:I have a question for those of your who are power experts. This is mainly to satisfy my own curiosity.
|
|
#
?
Dec 28, 2010 04:37
|
|
|
I'm going to have access to a woodshop for the next little bit, so naturally I'm going to be making a clock from scratch that contains a shitload of blue LEDs. What are the estimates for lifetime if I use batteries? Say 3 LEDs x 12 places x 2 rows = 72 LEDs. I'm thinking not so good. Writing it out puts it in perspective, so I guess I'll have to go off AC. drat.
|
|
#
?
Dec 28, 2010 04:57
|
|
|
SnoPuppy posted:I have a question for those of your who are power experts. This is mainly to satisfy my own curiosity. I actually had to do some research because this one intrigued me. The easiest/most robust is just to deal with the heat properly, but MOSFETs controlled by comparators seems to be the most efficient. I imagine you'd have to use a very high resistance divider or a secondary transformer to get your comparator voltage, then switch your FETs with the comparator outputs. ante posted:What are the estimates for lifetime if I use batteries? Say 3 LEDs x 12 places x 2 rows = 72 LEDs. I'm thinking not so good. Keep in mind you'll have to multiplex the LEDs to control them, they won't all be on all the time. I don't really know a fair estimate for an individual LED's power consumption, probably 25-50mW? sixide fucked around with this message at 08:01 on Dec 28, 2010 |
|
#
?
Dec 28, 2010 07:58
|
|
|
ante posted:I'm going to have access to a woodshop for the next little bit, so naturally I'm going to be making a clock from scratch that contains a shitload of blue LEDs. As mentioned you'll have to multiplex them, most battery powered LEDs use a PWM controller to allow dimming and significantly save power by only having them lit for short intervals as well.
|
|
#
?
Dec 29, 2010 12:17
|
|
|
I am wondering again about playing with a microcontroller that has Wifi capability (add-on or not) but is generally powerful enough to run Linux like a SBC. The reason being is that I would need something with a nice web UI, and I don't expect a smaller microcontroller with a built-in web server will be able to provide it. Nor would I want to be responsible for writing web server code to do that itself. So perhaps an SBC with Linux and Apache with some modules built in so I can create a web UI like I might expect to do on my PC. I was hoping for something with at least 8 A/D pins, "some" serial I/O pins with some doing PWM, and generally some of that I/O goodness that you'd expect from a microcontroller. It looks like Atmel has some stuff--too much really. Between assessing the individual products and the kits I could then buy with them, I am overwhelmed. Does somebody have some general recommendation with which I can focus on this? If this brings on any deja vu, I did bring this up many months ago but it languished for awhile.
|
|
#
?
Dec 31, 2010 23:32
|
|
|
|
| # ? Jun 11, 2024 11:26 |
|
|
Rocko Bonaparte posted:I am wondering again about playing with a microcontroller that has Wifi capability (add-on or not) but is generally powerful enough to run Linux like a SBC. The reason being is that I would need something with a nice web UI, and I don't expect a smaller microcontroller with a built-in web server will be able to provide it. Nor would I want to be responsible for writing web server code to do that itself. So perhaps an SBC with Linux and Apache with some modules built in so I can create a web UI like I might expect to do on my PC. I was hoping for something with at least 8 A/D pins, "some" serial I/O pins with some doing PWM, and generally some of that I/O goodness that you'd expect from a microcontroller. I'd probably go with either a Beagle Board with a USB wifi controller or a GumStix module. To do what you want (run linux, apache, use wifi), you will need to have something quite a bit more powerful than a simple micro. If you don't really need linux, and just want a web server, I know there are some for the Arduino+ethernet. And at least one company makes a wifi module for arduino (http://asynclabs.com/)
|
|
#
?
Jan 1, 2011 00:10
|
|