|
I'm working on a nice little exam project that consists of me designing and building a sun tracer, a device that basically follows the sun as it travels across the sky each day. so far i've got everything down on PCB and working, sort of. My issue is that i can only get the darn motor to turn one way, and i've isolated the fault to my control circuit, the H-bridge and sensor PCB appear to have nothing wrong with them.  Click here for the full 1019x589 image.   what i've worked out is that it fails to go high on one out, but it goes perfectly high on the other. i apologise for any weak descriptions, but english isnt exactly my first language and thus technical terms can get pretty iffy to translate. EDIT: once im done i'll release schematics of the whole thing, havent seen anything similar to it on the net, which either means i am a genius or what i am doing is terribly wrong. Frobbe fucked around with this message at 11:51 on May 3, 2010 |
#
?
May 3, 2010 11:13
|
|
|
|
| # ? Jun 11, 2024 13:55 |
|
|
ANIME AKBAR posted:conceptually it's neat, but in practice it's horrible. Spending five hours in a class 100 hundred clean room doing nothing but waiting for things to clean/deposit/etch/dry is mind numbingly boring to the point of being maddening. Having to try the same three hour process three times to get it to work correctly is not fun. Spending four hours prepping a probe station in order to do a half hour of electrical testing is insane. I have never done anything that managed to be so time consuming, yet be so staggeringly boring and mundane and stressful. The sweet satisfaction of seeing a device finally move by a couple micrometers is completely offset by this. I now understand why MEMs researchers are paid the big bucks: because it is grueling and unforgiving, and nobody in there right mind would do it for a living with out huge compensation. I have to agree with you here. I took a semi-conductor fabrication lab where we made a bunch of simple components (resistors, capacitors, transistors) on a wafer. As you said, conceptually it was cool, but after the first lab in the clean room, the novelty wears off fast. Most of the processes are pretty akin to watching paint dry. Our lab was too cheap to use gold, so we just used little balls of tinfoil to sputter on our contacts. Also, laser cut masks were pricey, so we used a high resolution printer on transparencies for our masks. Obviously we did things on a much larger scale, since making something like you did would be impossible with those type of masks. Hillridge fucked around with this message at 15:43 on May 3, 2010 |
|
#
?
May 3, 2010 15:40
|
|
|
Frobbe posted:I'm working on a nice little exam project that consists of me designing and building a sun tracer, a device that basically follows the sun as it travels across the sky each day. First off, I'm not an analog engineer, so take what I say with a grain of salt. How are you testing the circuit? Function generator hooked up to the input? What is the input? A DC value? A ramp? Sin wave? What does the output look like on a scope? Have you looked at the power rails on the chips? Could it be possible that one of your chips is dead? In addition, using op-amps like you are (positive feedback) makes them very vulnerable to oscillations and instability.
|
|
#
?
May 3, 2010 15:48
|
|
|
SnoPuppy posted:First off, I'm not an analog engineer, so take what I say with a grain of salt. testing consists of me covering either of the LDR (light dependent resistors) no function generator involved. The input is 12v DC i haven't checked the chips over all that much, but i did try simply replacing both, to no effect. i'll investigate further
|
|
#
?
May 3, 2010 16:44
|
|
|
I'm not an expert either, but I'm wondering 1) what's with the voltage followers in there and 2) why aren't you using a dedicated comparator like the LM339?
|
|
#
?
May 3, 2010 17:00
|
|
|
Frobbe posted:testing consists of me covering either of the LDR (light dependent resistors) no function generator involved. First step in debugging is to remove all possible unknowns and check ALL assumptions. Use a function generator or power supply or something with known output characteristics. Begin walking through the circuit, checking for any abnormalities. Use a scope to make sure that nothing is oscillating and the supply rails look clean.
|
|
#
?
May 3, 2010 17:23
|
|
|
MATTER SETTLED
Walking Spanish fucked around with this message at 06:04 on May 4, 2010 |
|
#
?
May 3, 2010 17:46
|
|
|
Mill Town posted:I'm not an expert either, but I'm wondering 1) what's with the voltage followers in there and 2) why aren't you using a dedicated comparator like the LM339? i guess im using the 358 because that's what the school had in stock and it fit, i know my circuit is kinda horrible, the H-bridge and control circuit could've been one circuit real easy, eliminated a lot of bulk and wires. gonna revise the PCB in sprint. i've had a walk through the circuit with a scope now, but that didn't seem to yield any usable results. investigation continues.
|
|
#
?
May 3, 2010 19:27
|
|
|
Walking Spanish posted:Alright electrical goons, here's my situation and I'm hoping you can help me with this. I recently purchased an industrial sewing machine head for around 500 dollars. The problem is that it is not sold with the motor or table. The table I can build but the motor is my primary concern. Here it is: You don't need a servo motor, just an AC motor with a speed control. A quick search on Ebay shows tons of sewing machine motors under $20. Then you just need to find a pedal (Ebay, once again, has a ton of decently priced ones). Or if you're feeling really cheap, hit up harbor freight: http://www.harborfreight.com/router-speed-control-43060.html
|
|
#
?
May 3, 2010 21:57
|
|
|
Keebler posted:You don't need a servo motor, just an AC motor with a speed control. A quick search on Ebay shows tons of sewing machine motors under $20. Then you just need to find a pedal (Ebay, once again, has a ton of decently priced ones). Or if you're feeling really cheap, hit up harbor freight: awesome, thanks man
|
|
#
?
May 3, 2010 22:59
|
|
|
My Dryer decided to blow a couple of Diodes. It is an easy enough fix to replace them. Problem is I am unable to locate a place to purchase them. I think they might not be made anymore. Does anyone know where I can get these: 4745AP color is red 1N4745A 948 there is also what looks like a Motorola logo with a K next to it. Color is silver. Any help would be appreciated.
|
|
#
?
May 3, 2010 23:51
|
|
|
dv6speed posted:That would seem like the sensible thing huh? A rotary phase converter is a 3-phase generator hooked to a single-phase motor. The output looks like any other 3-phase generator waveform.
|
|
#
?
May 4, 2010 00:43
|
|
|
babyeatingpsychopath posted:A rotary phase converter is a 3-phase generator hooked to a single-phase motor. There are two types of rotary phase converters. I'm not intersted in the generator based one. http://en.wikipedia.org/wiki/Rotary_phase_converter posted:A rotary phase converter may be built as a Motor-generator set. A motor-generator set has the advantage that it isolates the generated three phase power from the single phase supply and that the three phase voltages can be balanced. However, because of weight, cost, and efficiency, most RPC's are not built this way. I just want to see exactly what that waveform looks like, before and after rectification. AbsentMindedWelder fucked around with this message at 00:55 on May 4, 2010 |
|
#
?
May 4, 2010 00:52
|
|
|
JEEVES420 posted:My Dryer decided to blow a couple of Diodes. It is an easy enough fix to replace them. Problem is I am unable to locate a place to purchase them. I think they might not be made anymore. Does anyone know where I can get these: Typing 1N4745A into Digikey yields some responses, they are 46 cents a piece (plus a few dollars shipping, but what can you do?). As for the other one, Digikey says it's a non-stock item, meaning they could get it from the OEM, but it's probably gonna take a while and/or cost more. They are both 1-Watt zener diodes and their specs are almost identical. I would suggest getting two of the 1N4745A kind and using them. Here's the link: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=1N4745ADICT-ND
|
|
#
?
May 4, 2010 01:08
|
|
|
Corrupt Politician posted:Typing 1N4745A into Digikey yields some responses, they are 46 cents a piece (plus a few dollars shipping, but what can you do?). As for the other one, Digikey says it's a non-stock item, meaning they could get it from the OEM, but it's probably gonna take a while and/or cost more. Thanks a million.
|
|
#
?
May 4, 2010 02:00
|
|
|
Hillridge posted:Most of the processes are pretty akin to watching paint dry.
|
|
#
?
May 4, 2010 05:54
|
|
|
JEEVES420 posted:Thanks a million. No problem. Unless something is really specialized or esoteric, Digikey, Mouser, and Jameco are great first places to look for pretty much any electrical hardware.
|
|
#
?
May 4, 2010 06:47
|
|
|
as it turns out, my issue turned out to be two malplaced resistors that needed to be switched around. i'll release the schematics after my exams are done, it's a neat little cheap project that is incredibly useful for increasing usefulness of solar cells. edit: picture of the completed assembly  while i have the latest post, allow me to post this, even though it might belong in a dedicated microcontroller thread?  http://picasaweb.google.dk/lh/photo/UEBeBKSheatsyiC_iP2AHQ?feat=directlink found this in the trash pile at my school, apparently it's a kit designed by some professor at the university. the wirewrap is impressive. posting in hopes of info and because people might get a kick out of it, the PCB on the right has a date of 1994. Frobbe fucked around with this message at 16:40 on May 4, 2010 |
|
#
?
May 4, 2010 09:27
|
|
|
ANIME AKBAR posted:That pretty much sums it up. And it also bugged my that the quality of our devices are pretty much completely dependent on the quality of the instruments and equipment, and has little to do with any skill or creativity. Oh, your structural layer delaminated and cracked? Whelp, should have spent 5 million dollars on your furnace instead of a measly one million! You need a layer of platinum for your temperature sensor? Too bad, we only have gold! What a joke. That reminded me of something else. In pretty much every pre-lab lecture, the professor spent a good 50% of the time just telling us how much poo poo cost in the lab.
|
|
#
?
May 4, 2010 21:41
|
|
|
I've got a toughie. I'm looking for a way to control 4 separate bipolar stepper motors for a budget project. Step speed is slow, but independent, current is modest (10v, 500ma). I can't switch to unipolars because the devices are all-in-one units. I'd like to spend a total of $50 or less controlling these. I have access to 3 PIC16F688's, and an ATmega for brains. I have a pile of (and am willing to acquire more) 74xx and 40xx series logic chips. Discrete transistors and MOSFETs aren't an issue either. I'm very familiar with breadboarding, and have DIY'ed a few single-sided boards before. I'd like to eschew SMT, but it's not ruled out. Eventually, I'd like for a computer to talk to the steppers via RS232 or USB. Surprisingly, this isn't for a CNC machine 
|
|
#
?
May 6, 2010 01:56
|
|
|
What's the question? It's pretty trivial to write a stepper driver in C, then buffer the output with your preferred transistor(with a small heatsink). Serial communication is super easy to get working, too. What's the project? Please say killbot.
|
|
#
?
May 6, 2010 02:24
|
|
|
ante posted:What's the question? The question is "how do I cheaply & easily drive these". Remember, it's a bipolar stepper, so the driving electronics is a full H-bridge per coil, of which each motor has two. If they were unipolars, I'd be fine. That's a lot of transistors if I'm implementing it myself. It's much more boring than a killbot, it's a dosing pump for an aquarium. I got killbots out of the way a long time ago.
|
|
#
?
May 6, 2010 02:34
|
|
|
I was replacing a dying hard drive in an xbox... and a little orange resistor started on fire and quickly went out. Now the xbox doesn't start. Is it worth the effort of replacing this resistor... or is it likely that something else horrible happened? The xbox was unplugged at the time of fire... but only unplugged for a couple minutes.
|
|
#
?
May 6, 2010 02:55
|
|
|
Either of these look decent. One's a little cheaper overall, but you'll need two per motor. Depends on your priorities. http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=TA7291P5-ND http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=TB6561NG-ND I still stand by building your own 30-odd array of transistors, though 
|
|
#
?
May 6, 2010 03:26
|
|
|
FreshShoez posted:I was replacing a dying hard drive in an xbox... and a little orange resistor started on fire and quickly went out. Now the xbox doesn't start. Is it worth the effort of replacing this resistor... or is it likely that something else horrible happened? The xbox was unplugged at the time of fire... but only unplugged for a couple minutes. Depends on how much of the magic blue smoke you released. It's hard to tell without a multimeter. If you can figure out a local place to buy the resistors, you'll be able to buy a pack of way-more-than-you-need for a couple bucks. Follow the rails on both sides, though. Look for bulging capacitors or sensitive components. Hey, does anyone know whether resistors burn open or short? I just realised that I have no idea. That might need some experimenting.
|
|
#
?
May 6, 2010 03:35
|
|
|
Hm... none of the capacitors look blown, but the only real experience I have is replacing the ones on my graphics card that were actually busted and oozing. The resistor is orange and says R5 next to it... I know that the color of stripes really matters, but they're all burnt black now. I'll find my multimeter and experiment a little and see if I can find some answers.
|
|
#
?
May 6, 2010 03:40
|
|
|
Take it out of the circuit, first.
|
|
#
?
May 6, 2010 03:41
|
|
|
ante posted:Either of these look decent. One's a little cheaper overall, but you'll need two per motor. Depends on your priorities. To be honest, I entertained the idea of the ginormous transistor wad. However, at $4/ea the dual full bridge looks good. I can pair that with a PIC and some address lines, and get all 4 of them controlled on just 6 output lines.
|
|
#
?
May 6, 2010 03:48
|
|
|
insta posted:The question is "how do I cheaply & easily drive these". Remember, it's a bipolar stepper, so the driving electronics is a full H-bridge per coil, of which each motor has two. If they were unipolars, I'd be fine. That's a lot of transistors if I'm implementing it myself. Seems like a fine job for a SN754410. One per motor, and the datasheet has an example for wiring up a bipolar motor. Should be pretty easy from there. quote:Hey, does anyone know whether resistors burn open or short? I just realised that I have no idea. That might need some experimenting.
|
|
#
?
May 6, 2010 06:37
|
|
|
found this microcontroller board in the dump pile as well. now, it is fitted with a Atmel AT89C4051 but the board is labeled Atmel AT89C2051, though i guess that doesnt make much of a difference, gonna investigate data sheets after im done with this exam project. to my eye the circuit and such looks to be simply the supporting electronics for the chip itself, with pins at the bottom for bread boarding, is it possible to rig something that can flash the chip using just a breadboard or do i need a dedicated burner? i already have the Keil programming environment.
|
|
#
?
May 6, 2010 16:42
|
|
|
Frobbe posted:found this microcontroller board in the dump pile as well. Those pins are spaced incredibly far apart so I wouldn't be so sure that it's intended to fit into a breadboard even though that's what the orientation would suggest. But you're right, the rest of it is just a microcontroller, a crystal, a voltage regulator to power it and filter capacitors. No clue what that potentiometer (?) is for though.
|
|
#
?
May 6, 2010 17:44
|
|
|
ante posted:Either of these look decent. One's a little cheaper overall, but you'll need two per motor. Depends on your priorities. That's still a viable solution considering you can get some darlington arrays which simplifies and organizes everything really well.
|
|
#
?
May 6, 2010 21:06
|
|
|
BattleMaster posted:Those pins are spaced incredibly far apart so I wouldn't be so sure that it's intended to fit into a breadboard even though that's what the orientation would suggest. But you're right, the rest of it is just a microcontroller, a crystal, a voltage regulator to power it and filter capacitors. No clue what that potentiometer (?) is for though. yeah i might be able to fit in some normal pins or do some other fooling around with that regard. the big knob isn't a pot, it's simply a giant pushbutton, i assume it can be wired up for reset or is wired up for a reset. either way, i need to figure out how to program these chips, i managed to dig out 4 of them, and there arent any bad soldering and all the components appear to still retain their smoke.
|
|
#
?
May 6, 2010 21:17
|
|
|
Frobbe posted:found this microcontroller board in the dump pile as well. gently caress the poo poo out of the keil compiler/linker/entire toolchain
|
|
#
?
May 6, 2010 21:22
|
|
|
Otto Skorzeny posted:gently caress the poo poo out of the keil compiler/linker/entire toolchain well, okay, i learned ASM using this poo poo, but that didn't help me get much further in knowing how to program these bastards.
|
|
#
?
May 6, 2010 22:03
|
|
|
I'm just saying, in 2010, it chokes on spaces in filenames When I was using 8051s it was more productive to run a buggy Perl script on Keil headers to make them work with SDCC than to use the Keil toolchain
|
|
#
?
May 6, 2010 22:26
|
|
|
I hesitate to post something so basic here, especially when the controller is an Arduino but I just can't seem to solve this on my own. What I've got is just about the balls-simplest thermistor circuit you see in every example on the internet: The thermistor is actually from a Maverick ET-7 remote-read thermometer, connected to a 2.5mm mono headphone jack. The problem is that the probes appear to use the metal cord shielding as ground. What this means is that when I physically touch the cord, the voltage I'm reading at A0 goes all over the place. When two probes are in use (with an identical circuit except attached to another analog input), if the two probe cords touch each other, I get completely invalid data (the ADC reads right about 512/1023). What can I do to prevent this from happening, because it clearly doesn't occur when the probes are used in their original Made-in-China device. I've tried swapping the fixed resistor and the thermistor with no luck. Also, what am I supposed to connect the third leg of the 2.5mm headphone jack to? I've got the sleeve connected to ground and tip to the VCC side.
|
|
#
?
May 7, 2010 15:54
|
|
|
CapnBry posted:I hesitate to post something so basic here, especially when the controller is an Arduino but I just can't seem to solve this on my own. What I've got is just about the balls-simplest thermistor circuit you see in every example on the internet: I suspect that's not a thermistor. According to your circuit diagram (and I think it's right) you would need three wires (a stereo plug) to make a voltage divider circuit like that. Are you sure it's not a thermocouple?
|
|
#
?
May 7, 2010 20:37
|
|
|
Delta-Wye posted:I suspect that's not a thermistor. According to your circuit diagram (and I think it's right) you would need three wires (a stereo plug) to make a voltage divider circuit like that. Are you sure it's not a thermocouple?
|
|
#
?
May 7, 2010 21:11
|
|
|
|
| # ? Jun 11, 2024 13:55 |
|
|
CapnBry posted:I'm fairly certain it is a thermistor, considering I use the voltage divider circuit there to read values, calculate its resistance and plug that into a Steinhart-Hart equation and get a reasonable temperature value at both room temperature and in a cup of boiling water. Then again, I'm not at all familiar with thermocouples so I wouldn't know if these traits would be indicative of a thermocouple as well. Can you take resistance measurements between all the different points (tip/sleeve/outer conductor)?
|
|
#
?
May 7, 2010 21:17
|
|