|
Alright, I'm working on my kegerator control scheme and have run into a snag with the wiring. Here is a diagram of how I have it wired right now (yellow wires are actually white in my wiring scheme). I have a table lamp plugged into the outlet box right now for testing. Eventually the freezer will get plugged in there. I tried it without the outlet box earlier today and the SSR seems to work. If I set the setpoint higher than room temperature, the light on the SSR turns on, so it's getting power. With everything wired as it is in the drawing, the lamp is on all the time. I checked across L1 and L2 of the SSR with my meter and got continuity (no power applied, red wires disconnected). Shouldn't that path be open if there's no control signal telling it to close? Did I get a bum SSR? Edit: Pretty sure my SSR is toast and it might be my fault. If I pull the controller out of the scheme altogether the lamp still turns on. I put a meter on the input side of the SSR while the controller was telling it to turn off and read 54V - the maximum input voltage on my SSR is 32V. I probably would have caught that with a more complete read-through of the manual. I'm guessing the over voltage killed it. I did figure out how to limit the output on the controller, so at least I won't kill the next one. LordOfThePants fucked around with this message at 21:19 on Jul 7, 2012 |
#
?
Jul 7, 2012 16:25
|
|
|
|
| # ? May 9, 2024 11:59 |
|
|
edmund745 posted:Sounds nice. Where can I get such a thing, without paying industrial prices? You're making this harder than is has to be. A hand-built 12-axis controller has precisely the same problems synching as a pair of six-axis, or three four-axis, or four three-axis. You're the one writing the software to control the thing, and their synchronicity is your problem. What feedback does your system have to keep everything in line? RAMPS firmware already has fairly advanced interpolation to get everything to the same spot at the same time. I'm pretty sure I could take three RAMPS boards, plug them into a USB hub, and get them to enumerate in a unique way so they're detected as a 12-axis controller. From there, messing with control software writing g-code is just a software issue. You're saying "I know for sure that this won't do what I want" when you don't seem to have a good grasp on what it can do.
|
|
#
?
Jul 7, 2012 17:01
|
|
|
Updates on the PSoC ADC issue. On Tuesday I tried using a TPS7A49 as the Vref, but got roughly the same results as with the internal Vref. This was somewhat discouraging, but since the TPS7A49's noise floor at low frequency isn't super duper low I wasn't quite defeated yet. On Friday, I got a shipment of ADR130 voltage refs in. They have <6 uvolt of .1-10 Hz noise. After jerry rigging some boards up (we actually didn't have anything on hand that could accomodate a TSOT-6 without some mangling  ), I got to testing. Before using one as an external Vref, I tried one as an input to one of our boards with a PSoC 3 swapped in, and got about 14.5 bits of stability. A software IIR LPF that I wrote did very little to improve this, implying that HF noise wasn't the problem.I swapped one in as a Vref, using one of our sensor modules rigged to output .998 volts, and got 16.0 to 16.3 bits of stability without any filtering, and 18.0 bits of a stability using a software IIR LPF that I wrote. Bingo. Using an EDC voltage standard, I tested stability from 0 to 1v in .1 volt increments, and found that this stability was consistent throughout the input range. I tested further to 105% of full scale and still saw >17 bits of stability. At this point I believe that the internal Vref was 100% of the problem. I then tested an identical board that had a PSoC 5 swapped in. With the ADR130 as the vref, I got 14.7 to 15.0 bits of stability without filtering and 16.0 to 16.3 with (although this was only in level-shift mode; rail-to-rail gave horrible results on the PSoC 5). So it looks like the PSoC 5 will be usable for beta testing of the product we're developing, but we'll need the LP for real production. And it won't be available even in small quantities for a good while  Nb. that with either PSoC, the delta-sigma component's internal clock was unsuitable and (wildly) inaccurate, and we had to use a discrete clock component to get both conversion accuracy and a sample rate that actually resembled what we had set, and that we had to choose a sample rate that yielded a desired clock rate that could achieved pretty precisely from a divider from the master clock, and that any ADC configuration selected in PSoC Creator besides Config #1 is ignored in favor of Config #1, and that changing your clock speed at runtime isn't accounted for even if you restart the conversion and the ADC itself, and that each of these points contradicts something that a Cypress app engineer told us during a face-to-face meeting. If we had known half of the problems we'd have with this uC before we'd designed our product we wouldn't have selected it.
|
|
#
?
Jul 7, 2012 19:05
|
|
|
Hey guys, quick question, sorry if this is the wrong spot for that. I have a little audio project I'm working on, and I'd like to replace a 20kOhm rotational pot (the volume control for the amp) with a slide pot. In general, 20kOhm pots seem to be kind of rare, and it's limiting my choices. So, my questions: 1) Can I use a different resistance pot in the circuit, say a 10kOhm one, or will that release the magic blue smoke somewhere? 2) Ideally, I'd like a pot like this: http://www.mouser.com/ProductDetail/ALPS/RS60112A600U/?qs=seHrhfPpLDybKKLMz8Fo1MxqgV5ySMh34%2f%252b3tbz3fSs%3d , but the slide length is 60mm and I'd rather have it be ~40mm. Does such a thing exist? I've checked digikey etc. Thanks
|
|
#
?
Jul 7, 2012 23:55
|
|
|
SupahDren posted:1) Can I use a different resistance pot in the circuit, say a 10kOhm one, or will that release the magic blue smoke somewhere? Chances are you'll be ok. A volume control is generally a voltage divider, and its the ratio that matters, not the exact resistance. You are reducing the input impedance of the amplifier a bit. If the amp is being driven by something with a highish output impedance, the pot will load it some and affect the overall amplitude. Chances are this is not a big deal, especially if you have a pre-amp before the volume control. That effect might be more if you have other passive components between your pre-amp and amp, however, such as tone controls. By the same token, the input impedance of the actual amp after your volume control loads the volume control, but since you're reducing the resistance that shouldn't be a problem. If you increased the resistance instead then you'd want to make sure the actual amp had higher input impedance. A 2X change is actually not that big in this context though.
|
|
#
?
Jul 8, 2012 01:26
|
|
|
It does matter if it's a linear or a logarithmic pot, though
|
|
#
?
Jul 8, 2012 03:42
|
|
|
Hey guys, I just wanted to say I got a new soldering iron tip and some flux. Apparently even Weller tips don't work very well after 30 years. Also flux. Flux is the best thing ever. It makes the desoldering pump actually work, and it lets you reflow tiny amounts of solder instead of making huge gobs. Use a flux. Flux flux flux.
|
|
#
?
Jul 8, 2012 21:28
|
|
|
Silver Alicorn posted:Hey guys, I just wanted to say I got a new soldering iron tip and some flux. Apparently even Weller tips don't work very well after 30 years. Also flux. Flux is the best thing ever. It makes the desoldering pump actually work, and it lets you reflow tiny amounts of solder instead of making huge gobs. Use a flux. Flux flux flux. Oh yeah, Kester rosin flux pens are the poo poo, and it's a shame that a lot of guides for people new at electronics don't mention them as often as possible. They are crucial for hand-soldering SMT components, and are a great help for through hole as well (since they help give you a much-larger effective heat transfer surface area, letting you reflow a solder ball more quickly and with less thermal stress). Since the solder wick I use has either no flux or lovely flux, applying it to that directly before using makes it work drat well too. That said, stay the hell away from weird foreign solder wick with lovely flux that leaves gunk everywhere.
|
|
#
?
Jul 8, 2012 21:46
|
|
|
Silver Alicorn posted:Hey guys, I just wanted to say I got a new soldering iron tip and some flux. Apparently even Weller tips don't work very well after 30 years. Also flux. Flux is the best thing ever. It makes the desoldering pump actually work, and it lets you reflow tiny amounts of solder instead of making huge gobs. Use a flux. Flux flux flux. Wash your boards (DI water preferably) after you're done going crazy with the flux 
|
|
#
?
Jul 9, 2012 17:30
|
|
|
That's only good if you use water soluble flux. Denatured alcohol is my choice.
|
|
#
?
Jul 9, 2012 18:03
|
|
|
This is a real real real long shot, but I think this is the closest thread to anything on SA where someone might know: Does anyone here have a Tektronix TDS 420 digitizing oscilloscope? It's giving me an Attn/Acq POST failure which is fairly common for this scope, though I'm not looking forward to swapping out all the caps on-board. My question is regarding the A03 CPU board serial port. I'm trying to get a console readout of the scope's POST, and I should be able to by just hooking up a serial port header, then connecting a null modem, then connecting that to my computer. I didn't have a serial header handy so I "borrowed" the serial header from the option card in slot 5. Unfortunately when I boot the scope I still don't see any console messages coming across the serial line. I'm not sure why this would be, unless the RS232 header I borrowed uses some other nonstandard pinout maybe? Not really sure where to go from here. Like I said, I know I need to start replacing caps, but I was hoping to read the console out before I tore this scope to bits. Thanks! (the scope actually seems to work fine, other than channel 2 which is incredibly off, even when measuring the reference post)
|
|
#
?
Jul 9, 2012 20:10
|
|
|
Have you tried posting on the official Tektronix forum? Lots of people over there have that scope and can help. Have fun replacing those caps!
|
|
#
?
Jul 9, 2012 21:51
|
|
|
Oh balls, I didn't even know this existed. Thanks a ton!
|
|
#
?
Jul 9, 2012 21:53
|
|
|
Slanderer posted:Oh yeah, Kester rosin flux pens are the poo poo, and it's a shame that a lot of guides for people new at electronics don't mention them as often as possible. They are crucial for hand-soldering SMT components, and are a great help for through hole as well (since they help give you a much-larger effective heat transfer surface area, letting you reflow a solder ball more quickly and with less thermal stress). Any recommendations on brands or specific products for flux? I consider myself decent at through hole soldering, but I have no idea how to get started with SMT stuff. A quick Google search left me with two types: a liquid-y pen type and a Vaseline-esque paste. Which would you recommend and why, if you don't mind?
|
|
#
?
Jul 10, 2012 03:18
|
|
|
I use the liquid flux pen. It's convenient. I have the gel stuff, but I've never actually used it because I don't want to be bothered to scoop it up and put it down on a board when I can just take a second and dab with the flux pen. This is the one I use.
Krenzo fucked around with this message at 03:47 on Jul 10, 2012 |
|
#
?
Jul 10, 2012 03:40
|
|
|
I'm looking for a differential line driver/receiver pair for transmitting a clock signal over twinax/shielded twisted pair (differential Z0=100ohms). The link is pretty short; like 1-2 meters, but immunity to interference is an issue. There are tons of LVDS parts out there, but I need some more specialized performance: 1. Operating frequency of at least 150MHz (300Mbaud) 2. Low part-to-part skew (like less than 1ns for the driver and receiver combined) 3. Preferably the receiver should be specced with a wide input common mode voltage range, since this link is going to be operating in strong transient E fields and will probably have substantial common mode interference. 4. I'd like something with a somewhat higher margin between driver strength and receiver sensitivity (most LVDS drivers seem to be 350mV/100mV for ~10dB, I'm looking for something with 20dB) I'm having trouble finding all these in one part. If I look at simple differential line drivers, I can satisfy all the requirements except 2. If I look at clock buffer parts, I can satisfy 1-2 but not 3-4 (most clock buffers are meant for driving short PCB traces, not actual cable, so they're not as robust to interference). Any recommendation? This is my first time making a high bandwidth digital link myself. edit: Actually I've found one part that looks like a good candidate for the receiver, maybe the transmitter as well: http://www.ti.com/lit/ds/symlink/sn65cml100.pdf However it only specifies operation for driving 50 or 25ohm single ended lines, but that means it should be fine with 100ohm differential lines, right? Also it has CML outputs, and therefore needs a terminator bias voltage (Vtt) on the outputs, and the datasheet doesn't make it clear what that output needs to be. Does it depend on the type of logic is being driven? Like if I use one of these to driver a LVDS input, does Vtt need to be the input common mode voltage of the LVDS (VCC/2)? ANIME AKBAR fucked around with this message at 13:27 on Jul 11, 2012 |
|
#
?
Jul 11, 2012 12:58
|
|
|
ANIME AKBAR posted:I'm looking for a differential line driver/receiver pair for transmitting a clock signal over twinax/shielded twisted pair (differential Z0=100ohms). The link is pretty short; like 1-2 meters, but immunity to interference is an issue. There are tons of LVDS parts out there, but I need some more specialized performance: Have you looked at Pericom? Something like the PI90LV047/048 pair? We use that here for some synchronization signals. We also utilize BLVDS xcvrs like the DS92LV0xx series from TI for our high-speed I/O box. quote:edit: Actually I've found one part that looks like a good candidate for the receiver, maybe the transmitter as well: http://www.ti.com/lit/ds/symlink/sn65cml100.pdf Also, is fiber an option? Light gives no fucks about noise.
|
|
#
?
Jul 11, 2012 16:38
|
|
|
ANIME AKBAR posted:I'm looking for a differential line driver/receiver pair for transmitting a clock signal over twinax/shielded twisted pair (differential Z0=100ohms). The link is pretty short; like 1-2 meters, but immunity to interference is an issue. There are tons of LVDS parts out there, but I need some more specialized performance: Why do you need < 1 ns of part to part skew? Data capture? You said you're putting a few meters of cable in line, so that would push the total time of flight to something > 1 ns. FYI, that TI chip specs part to part skew for two chips on the same PCB (i.e. at the same voltage and temperature). This assumption won't hold true for two different boards with different power supply variations. I'd also try to AC couple, if possible, as that will reduce the effects of DC common mode noise. You might want to take a look at Micrel or IDT - they both make decent clock/data buffers.
|
|
#
?
Jul 11, 2012 17:29
|
|
|
SnoPuppy posted:You might want to take a look at Micrel or IDT - they both make decent clock/data buffers. I use Micrel clock buffers. They work great. I can send a 600 MHz clock signal from one of their buffers over USB 3 cable just fine. ANIME AKBAR posted:However it only specifies operation for driving 50 or 25ohm single ended lines, but that means it should be fine with 100ohm differential lines, right? Right ANIME AKBAR posted:Also it has CML outputs, and therefore needs a terminator bias voltage (Vtt) on the outputs, and the datasheet doesn't make it clear what that output needs to be. Does it depend on the type of logic is being driven? Like if I use one of these to driver a LVDS input, does Vtt need to be the input common mode voltage of the LVDS (VCC/2)? It depends on what standard the receiver uses. Take a look at this.
|
|
#
?
Jul 11, 2012 20:13
|
|
|
movax posted:Have you looked at Pericom? Something like the PI90LV047/048 pair? We use that here for some synchronization signals. We also utilize BLVDS xcvrs like the DS92LV0xx series from TI for our high-speed I/O box. quote:Looks more like some kind of translator/repeater? I'm not entirely sure that's what you want to be using to transmit a clock signal. quote:Also, is fiber an option? Light gives no fucks about noise. SnoPuppy posted:Why do you need < 1 ns of part to part skew? Data capture? quote:You said you're putting a few meters of cable in line, so that would push the total time of flight to something > 1 ns. quote:FYI, that TI chip specs part to part skew for two chips on the same PCB (i.e. at the same voltage and temperature). This assumption won't hold true for two different boards with different power supply variations. quote:I'd also try to AC couple, if possible, as that will reduce the effects of DC common mode noise. quote:You might want to take a look at Micrel or IDT - they both make decent clock/data buffers. Krenzo posted:I use Micrel clock buffers. They work great. I can send a 600 MHz clock signal from one of their buffers over USB 3 cable just fine. quote:It depends on what standard the receiver uses. Take a look at this. I also found that there's a LVDS version of that TI repeater, with the same specs otherwise. I might use the CML version though anyways because it can have higher output swing. Thanks for the replies so far! ANIME AKBAR fucked around with this message at 00:54 on Jul 12, 2012 |
|
#
?
Jul 12, 2012 00:51
|
|
|
ANIME AKBAR posted:Those pericom devices seem to have too much max skew from part to part (up to 1.5ns). Yeah, that's because they're designed for a max signal of 200 MHz. ANIME AKBAR posted:What parts? Are they actually specified for cable driving? SY58032U and SY58011U. Both parts have a skew of 20ps and 15ps. You have to get the parts rated for GHz clock speeds to get such small skew. They say nothing about driving cables, but as long as your cable is differential, I don't see why your differential signal would have problems with it. I certainly don't have a problem. Oh, and the second part I listed is driving air in my system, not a cable. 
Krenzo fucked around with this message at 02:25 on Jul 12, 2012 |
|
#
?
Jul 12, 2012 02:22
|
|
|
Krenzo posted:SY58032U and SY58011U. Both parts have a skew of 20ps and 15ps. You have to get the parts rated for GHz clock speeds to get such small skew. quote:They say nothing about driving cables, but as long as your cable is differential, I don't see why your differential signal would have problems with it. I certainly don't have a problem. Oh, and the second part I listed is driving air in my system, not a cable. Also it occurred to me that I'll need a high speed comparator to create the clock initially (from the sine wave output of my DDS), so I'll probably just have one of those be my line driver (no point in having an extra stage I think...).
|
|
#
?
Jul 12, 2012 13:08
|
|
|
I'm thinking of discretizing an elliptic filter (implementing in software). Is a matched z-transform my best bet?
|
|
#
?
Jul 13, 2012 13:13
|
|
|
On another note, some people might be interested in some poo poo that's been happening in power electronics the last few years. People are starting to catch on that silicon devices are reaching their theoretical limitations, and have started looking at new materials, including GaN. And it turns out that GaN is so superior to Si for making power transistors in every way that it's only a matter of time before they get replaced in most applications. The only limit is really people's willingness to adapt to new technology (which is pretty easy for these things, since they're still FETs in the end). EPC is currently at the front of GaN FETs for power applications. I've been using their stuff for a while, and I'll probably never use Si again when I can help it. Some group made a 50MHz boost converter with ~90% efficiency out of GaN devices. This is some pretty ridiculous poo poo. With this kind of performance you could increase the power density of most power supplies by tenfold at least.
|
|
#
?
Jul 13, 2012 13:19
|
|
|
Otto Skorzeny posted:I'm thinking of discretizing an elliptic filter (implementing in software). Is a matched z-transform my best bet? What do you plan on using the filter for? http://www.mathworks.com/help/toolbox/control/ug/bs08hih.html has a pretty good rundown of various discretization methods.
|
|
#
?
Jul 13, 2012 18:25
|
|
|
 ELECTRICAL PROBLEMI've got this friend Rocky at my apartment complex. He's this WWII vet with an old television. It's got a HD/LD converter connected with a wallwort, and a pair of rabbit ears running into the converter. Today, when I touched the rabbit ears to try and adjust them, I received a shock. It was the "finger touching a hot pan" sort of shock, not one of these "oh poo poo I'm going to die call 911 turn it off turn it off turn it off" kind of shocks. However I had my Fluke volt-alert, and when I held it next to the whole antenna arrangement it sang from about 1/4th of an inch or closer. It looks like the shield is either energized or is holding a charge. I'm still wondering what the hell caused it. I did notice that the set is really old and has a two-prong plug. Could the set have the frame grounded to one of the prongs? I've heard of older appliances that have done this, and supposedly those things can be dangerous. I'm wondering if the TV has some kind of ground connection or is trying to ground via the antenna? I'm typically much better at solving electrical problems on much larger equipment - I usually don't deal with little appliances and the like.
|
|
#
?
Jul 14, 2012 00:47
|
|
|
It sounds like the whole system is ungrounded. Is there any way to run your voltmeter between the neutral prong and a good earth nearby? I'm guessing the guys in the wiring thread have a much batter grasp of this than I, though.
|
|
#
?
Jul 14, 2012 01:32
|
|
|
sixide posted:What do you plan on using the filter for? So it turns out that I'll be able to get the results I want in terms of noise attenuation and and settling time from a simpler filter with a minor tweak (although I've got a fair bit of testing to do yet to make sure my approach didn't screw up in some stupid way), but that link is wonderful. Thank you for posting it!
|
|
#
?
Jul 14, 2012 01:46
|
|
|
Three-Phase posted:
After having to look into it recently, it seems that some old appliances hooked up the case in strange and unpredictable ways. Honestly, though, I'm still not sure on the point of grounding the case/frame of appliances with newer 3-prong plugs, since I've also read a ground current from a faulty appliance often won't trip a breaker, so you're hosed anyway!
|
|
#
?
Jul 14, 2012 04:42
|
|
|
Slanderer posted:Honestly, though, I'm still not sure on the point of grounding the case/frame of appliances with newer 3-prong plugs, since I've also read a ground current from a faulty appliance often won't trip a breaker, so you're hosed anyway! Grounding doesn't trip a breaker, it's to make sure there aren't potential differences between different grounds, or ground and a chassis or other metal part. Because neutral carries a lot of current (same current as hot), the voltage difference V=IR between neutral and ground can be more than you'd safely like, so you can't use neutral as a ground reference. (And it gets miswired from time to time.) Normally, there should be little current returned via ground as opposed to neutral. If a power supply is isolated vs. power and floating, it can get charged up well above "actual" ground (meaning whatever else a human might be in contact with); the metal is basically a capacitor. Properly grounded the charge gets returned to ground, but it shouldn't be a lot of current. In cases where there's a danger ground might end up carrying a lot of current, you turn off hot and neutral with a GCFI, which interrupts the power driving the current. A breaker that disconnects a ground would be dangerous.
|
|
#
?
Jul 14, 2012 05:09
|
|
|
I read about the old american AM radios recently, article mentioned that a lot of the time the two prong lead had one end connected to chassis ground. So you could try switching the plug around (and then put on a polarized plug). My understanding of the GFCI is that it compares the current in the L and N wires, and a difference of more than 25mA trips it since it's assumed to be going through your arm. I think it'll generally trip any time current runs through the PE wire too if it's a TN system. Since it's a TV I might be concerned about HV leakage too, if the antenna somehow got charged to 30kV that would pack a hell of a punch, but it doesn't sound like you were shocked by that kind of voltage. longview fucked around with this message at 10:24 on Jul 14, 2012 |
|
#
?
Jul 14, 2012 10:21
|
|
|
longview posted:I read about the old american AM radios recently, article mentioned that a lot of the time the two prong lead had one end connected to chassis ground. So you could try switching the plug around (and then put on a polarized plug). GFCI is just a coil around both leads (typically called hot and neutral) leading to a comparator, outputting to a relay. The trip on GFCI must be between 3-6mA for UL listing. Must not trip below 3mA, must trip at 7mA. GFPE (ground fault protection of equipment) can be up to 300mA leakage. GFCIs don't check the ground connection or monitor its current in any way on any of the GFCIs I've ever cracked open. The point of a ground is to provide a low-impedance current path back to source to cause a circuit protective element to open the circuit. If the chassis is connected back to ground in the panel and a hot hits the chassis, it should pop the GFCI, because the low-impedance path lets more than 6mA flow. If there's no GFCI, then hopefully all the wiring is big enough to draw 15-20A for long enough that the breaker on the circuit pops. On two-prong plugs, one prong should be wider; this is the neutral. This prong/wire should be attached directly to the chassis. If unpolarized, double insulation or equivalent is required. That said, a lot of homeowners replace plugs wrong and end up energizing the case. The appliance doesn't care; it's got no reference, and AC power is fine. The problem is when you touch the thing and can provide some kind of path to ground; you get shocked. Those voltalert things go CRAZY for HF. On some 2.4V/200mA LED drivers at 30kHz, mine will beep from a few inches away. I know 30kHz isn't HF, but the detector is usually an inductor tuned for 60Hz, so that's pretty high frequency, considering.
|
|
#
?
Jul 15, 2012 22:09
|
|
|
ANIME AKBAR posted:On another note, some people might be interested in some poo poo that's been happening in power electronics the last few years. People are starting to catch on that silicon devices are reaching their theoretical limitations, and have started looking at new materials, including GaN. And it turns out that GaN is so superior to Si for making power transistors in every way that it's only a matter of time before they get replaced in most applications. The only limit is really people's willingness to adapt to new technology (which is pretty easy for these things, since they're still FETs in the end). We used SiC JFETs for a high-efficiency converter project, they also performed quite well. Had no idea GaN was popular like this though; I wonder what the cost is like for those. Might be fun to order to play with.
|
|
#
?
Jul 16, 2012 05:56
|
|
|
movax posted:We used SiC JFETs for a high-efficiency converter project, they also performed quite well. Had no idea GaN was popular like this though; I wonder what the cost is like for those. Might be fun to order to play with. Cost wise they're more expensive than MOSFETs (like $3 each) but the point is they can do things that silicon simply can't. SiC will have the advantage over GaN in higher voltage/power applications for a while, but the theoretical yield of GaN is higher than SiC, and SiC is notoriously expensive to fabricate. I think SiC will always have a high temperature advantage though.
|
|
#
?
Jul 16, 2012 12:48
|
|
|
The advantages of SiC in some higher voltage high temp applications are why a lot of companies are pursuing it. There are more than a few future processes that are counting on C-Si swaps; it isn't the hardest thing to do but getting it into process as a single step, or better as a codopant, is certainly a thing being researched.
|
|
#
?
Jul 17, 2012 01:38
|
|
|
Well only a few extreme applications are going to really benefit from 600C operating conditions. And as far as I know there aren't any commercial devices rated for above 200C. My university does a lot of work on SiC (for ICs, not discrete devices), and from what I understand it's a brutally difficult process to control properly. Device yields are awful, which is one of the main reasons that SiC devices are preposterously expensive.
|
|
#
?
Jul 17, 2012 06:22
|
|
|
Think solid state power devices. There are a lot of interesting things one can do at those temps that couldn't be done before.
|
|
#
?
Jul 17, 2012 10:36
|
|
|
So I recently snagged a giant LCD TV from the side of the road (Polaroid TDX-03211c). Predictably, it doesn't power on. I looked around on the internet and this model tends to blow capacitors on the power supply. Unfortunately, I can't identify any caps that are obviously bad. None are bulging or leaking. I've pulled out the two caps that are most often cited as bad and measured them on my capacitor meter. Both caps are 10uF 450v. One cap registers as 9.98 uF on the meter, while the other registers as 8.90 uF. Does that mean the second cap is bad? Or can caps be off their rated values by a bit? Is there anything else I should look out for? I have a digital multimeter and a capacitance meter, but I'm afraid that's it. polyfractal fucked around with this message at 00:21 on Jul 18, 2012 |
|
#
?
Jul 18, 2012 00:08
|
|
|
Electrolytic caps have really wide tolerances on capacitance...
|
|
#
?
Jul 18, 2012 00:19
|
|
|
|
| # ? May 9, 2024 11:59 |
|
|
Slanderer posted:Electrolytic caps have really wide tolerances on capacitance... Hmm, well that sucks. Any suggestions on what to do next? I've checked the fuses and they are fine, none of the caps are obviously bulging...and if they can have wide tolerances than I guess I don't have any way to test them. Am I just poo poo out of luck?
|
|
#
?
Jul 18, 2012 00:50
|
|