|
tonberrytoby posted:I had a similar problem, it turned out the intermittent fault was at a sensor. It displayed the wrong heat for long enough to shut the system down from the bios, but not long enough to get logged by software. This one is an oh poo poo shutdown. Totally instant even with the protection thing in the BIOS disabled. It swears blind that there's some vague horrific voltage problem when I have it enabled. None of the onboard voltage sensors reflect this. But the light tap to the case causing it makes me think PSU. it's barely warm when it does it too so it's not like it's overloading horribly. I've considered connecting a multimeter even my CRO to rails to see what's going on but it's probably too quick to record or even perceive. The previous power supply used to shut down under load too but I'm pretty sure it was under-rated because the previous graphics card used to draw a lot more too and drag the +12v rail down hard. It lives on in the computer in the loungeroom playing videos etc without issue. Should I dig in and see what I can find? That is after taking precautions. I'm no stranger to that either because I usually strip dead PSUs for useful components before getting rid of them.
|
#
?
Jul 18, 2013 03:33
|
|
|
|
| # ? Jun 4, 2024 06:22 |
|
|
PSUs are a thing where if you do it wrong, you've fried your machine. By all means, open it up, it sounds like there may be a lovely solder joint or screw bouncing around in there or something, but if you're not totally sure what the problem, a new power supply is like, forty bucks.
|
|
#
?
Jul 18, 2013 05:56
|
|
|
That kind of instant shutdown could also be caused by some kind of intermittent short. But unless it's really expensive, just buy a new one.
|
|
#
?
Jul 18, 2013 11:03
|
|
|
ante posted:a new power supply is like, forty bucks. A thousand times this. Just but a new one from a big box store, and if it doesn't fix it you can easily return it.
|
|
#
?
Jul 18, 2013 17:55
|
|
|
ante posted:PSUs are a thing where if you do it wrong, you've fried your machine. By all means, open it up, it sounds like there may be a lovely solder joint or screw bouncing around in there or something, but if you're not totally sure what the problem, a new power supply is like, forty bucks. I've never killed one from taking a look. If it's nothing that a visual inspection or a bit of continuity checking can't find it goes on the heap. What magical place do you live in where a PSU no fashioned from tin cans, spit, cereal box and fencing wire costs $40? I think I paid like $70 or $80 for this one and while the design seems solid it's still not a brand name one like Corsair. Corla Plankun posted:A thousand times this. Just but a new one from a big box store, and if it doesn't fix it you can easily return it. There are none. Online or nothing I'm afraid.
|
|
#
?
Jul 18, 2013 22:17
|
|
|
How does load capacitance play into selecting a crystal to drive an AVR uC? I'm looking at newark and I see ranges from like 7 to 30pF. I guess first of all I'm not really sure how the load capacitance plays into selection, so I have no idea how to work out which crystal to spec out. Is this somewhere in the datasheet for my AVR or should I be able to math this out myself?
|
|
#
?
Jul 19, 2013 03:43
|
|
|
Yeah, the AVR data sheet should have a section on crystal selection with charts showing crystal speed and recommended capacitance.
|
|
#
?
Jul 19, 2013 04:18
|
|
|
For reference, I'm talking about the crystal's OWN load capacitance, not the two capacitors that sit on the AVR's xtal lines. The reference layout has two capacitors whose values are listed in the chart, but the crystal itself has this mysterious "load capacitance" that I don't know how to factor in. If I need to at all, I mean. Like for example, this search comes up with a few crystals, but the have load capcitances ranging between 15pf and 50pf so I don't know if I need to worry about matching that with something or not.
|
|
#
?
Jul 19, 2013 04:48
|
|
|
Martytoof posted:For reference, I'm talking about the crystal's OWN load capacitance, not the two capacitors that sit on the AVR's xtal lines. The reference layout has two capacitors whose values are listed in the chart, but the crystal itself has this mysterious "load capacitance" that I don't know how to factor in. If I need to at all, I mean. Good question---I remember being confused by this the first time I had to select my own drat crystals. The load capacitance doesn't actually refer to the equivalent capacitance of any part of the crystal. The load capacitance refers to the external capacitance required to maintain a stable oscillation at the specified frequency in a parallel resonant crystal. The load capacitance is a factor of the values of your two external caps, as well as stray capacitance on the board and in the microcontroller's oscillator circuit. This is a fantastic resource: http://www.foxonline.com/pdfs/xtaldesignnotes.pdf Fox Electronics has a bunch of other good app notes on crystal theory, if anyone is is curious.
|
|
#
?
Jul 19, 2013 05:46
|
|
|
Brilliant, bedtime reading material. Thanks for this  edit: Also apparently I have no idea what the difference between a crystal and oscillator are  But I'm about to find out! some kinda jackal fucked around with this message at 05:51 on Jul 19, 2013 |
|
#
?
Jul 19, 2013 05:49
|
|
|
General_Failure posted:What magical place do you live in where a PSU no fashioned from tin cans, spit, cereal box and fencing wire costs $40? PcCaseGear: Corsair 350W - $45AU Thermaltake 500W - $45AU Corsair 450W - $49AU Aerocool 450W - $49AU (Never used this brand personally but I've heard good things about it.) Newegg: Antec 350W - $30US Coolmax 500W - $35US (lovely sounding brand name but seem to be pretty reliable - they tend to overstate the maximum output of their PSUs though.) Seasonic 300W - $38US Cooler Master 400W - $39US Antec 450W - $40US Seasonic 350W - $40US Thermaltake 430W - $40US Corsair 430W - $46US Pick one, add shipping.
|
|
#
?
Jul 19, 2013 11:22
|
|
|
My string of 240 ws2811 LEDs arrived and I got it up and running easily with the FastSPI library. What I've noticed though is that the lights seem to have very bad dynamic range. Each RGB channel is specified by an integer from 0 - 255, and I've found that anything below 13 and the lights are off; incrementing from 12 to 13 they suddenly to turn on dimly; from 13 to about 100 they increase in brightness, then above 100 or so I can't see any difference up to 255. For fast fading its not super noticeable but it really kills the ability to do more slower/subtle effects. Google doesn't seem to find any similar comments on this. Any suggestions? Maybe I should bring this up in the arduino forums.
|
|
#
?
Jul 19, 2013 16:27
|
|
|
Martytoof posted:Brilliant, bedtime reading material. Thanks for this Generally a crystal is just that: a crystal. It's "hammered" by some external circuitry (i.e. like that found in a MCU) and should resonate at a specific frequency, which is amplified and generates a clock inside said MCU. There are different modes / cuts of crystals but that's extra reading if you want to know more. An oscillator (to me, anyway) generally implies that its output signal is a logic-level square-wave clock signal. It may need its own external crystal or have one built in, but from a system perspective, an oscillator provides a single-ended or differential clock signal that I can feed to a device. Temperature compensation can be applied here as well.
|
|
#
?
Jul 19, 2013 16:38
|
|
|
Cyril Sneer posted:Any suggestions? Maybe I should bring this up in the arduino forums. I'm not an expert on LED dimming but I think this is typical of LEDs due to the non linear response of our eyes, if you move to a log ramp instead of a linear ramp the dimming should look more linear. Unfortunately this means the resolution of the PWM is pretty low, on PICs they usually have 10-bit PWM or more which makes it slightly less of a problem but short of making your own PWM generator with a log response/external high precision linear PWM I suspect there's not much to do...
|
|
#
?
Jul 19, 2013 17:14
|
|
|
Not sure if I should classify this as electronics, or programming or whatever, but since I guess this is kindof both... anyone have any experience with the TI Sensortag? http://www.ti.com/ww/en/wireless_connectivity/sensortag/index.shtml?DCMP=sensortag&HQS=sensortag-bn http://makezine.com/2013/04/18/teardown-of-the-ti-sensortag/ It seems pretty great, considering it's just 25$ I have some places around the house that I'd like to add sensors, mainly for detecting simple stuff like if the mailbox has been opened and stuff like that, and it seems like this thing would be pretty perfect for it, with some modifications. Since it's only 25$, I went ahead and ordered 10 of them. If I get bored, I can probably go ahead and re-sell them for more than 25$ here in Sweden anyways, so might as well go for broke... Using the mailbox as a simple example. Planning to rewire one of the push-buttons of the sensortag to the opening of the mailbox, with a closed circuit when the box is closed, and open circuit when opened. Then I'll just throw together some software on my computer for reading from the Sensortag, and it should "just work", if I understand this correctly. Buffis fucked around with this message at 19:49 on Jul 19, 2013 |
|
#
?
Jul 19, 2013 19:47
|
|
|
I'd never thought of putting my phone into a rocket before but now I'm consumed by the idea.
|
|
#
?
Jul 19, 2013 20:10
|
|
|
Cyril Sneer posted:My string of 240 ws2811 LEDs arrived and I got it up and running easily with the FastSPI library. What I've noticed though is that the lights seem to have very bad dynamic range. Each RGB channel is specified by an integer from 0 - 255, and I've found that anything below 13 and the lights are off; incrementing from 12 to 13 they suddenly to turn on dimly; from 13 to about 100 they increase in brightness, then above 100 or so I can't see any difference up to 255. For fast fading its not super noticeable but it really kills the ability to do more slower/subtle effects. Google doesn't seem to find any similar comments on this. This guy did a pretty nifty colored pixel grid display using WS2801 LED strings, and near the end of the writeup he describes having to use a kind of gamma correction curve to give the human-perceived output colors/brightness scale more linearly with input 0-255 values. Sounds like you may be running into the same issue.
|
|
#
?
Jul 20, 2013 01:06
|
|
|
I'm pretty sure the latest version of fastspi has some colorspace corrections (such as log-based curves) built in to it. You have to specifically use them, but you might want to open up some of the header files and have a look around in lieu of complete documentation. That was how I discovered the built-in HSV/RGB conversion stuff as well as some fun and nifty rainbow-generation routines.
|
|
#
?
Jul 20, 2013 03:37
|
|
|
JawnV6 posted:I'd never thought of putting my phone into a rocket before but now I'm consumed by the idea. My dad used to make these things he called CompuROC's back in the late 90's early 2000's that were based on PIC processors which did everything that thing's doing (minus bluetooth) without a giant phone contributing nothing but dead weight and flash memory. He event built a tiny carefully-tuned centrifuge to test them  My favorite part was that he refused to use a compiler or an IDE because it was "too complicated" so he wrote all the software for it in assembly using notepad  It was fairly complex software as well, since it had self-test modes and the entire thing was triggered and stopped by these "separation pins" he built and modeled after actual rocket separation detectors. My dad's a terrible nerd is what I'm trying to say. Shame Boy fucked around with this message at 11:24 on Jul 20, 2013 |
|
#
?
Jul 20, 2013 11:20
|
|
|
I'd love to see some pictures of those, if you have any. Sounds super cool.
|
|
#
?
Jul 20, 2013 22:32
|
|
|
I am going to wire up a 3.5mm stereo jack as an aux input for my car. Unfortunately I was in a rush at RadioShack and got a 5 connector model, which as I understand it is useful if you want a different source until you plug in. This should give you an idea of the diagram: | \/------------------5 | 4 (on side of 5) | | | 3 (on side of 2) | /\________________2 | |______________________1 Does this look correct in terms of the connections for a stereo jack: Ignore 3 and 4 1 Ground 5 Right channel 2 Left channel Hmm it is not displaying the spacing properly. (5) is for the middle part of the plug. (2) is for the tip. Tyro fucked around with this message at 22:59 on Jul 20, 2013 |
|
#
?
Jul 20, 2013 22:41
|
|
|
I want to set up a dynamo on my bike like this article describes. But instead of sending a current straight to the phone, I'd like to send it to a rechargeable battery, and then to various electronics (like lights, a phone charger, etc); that way I have a reliable current powering the lights whether I am pedalling or not. What type of light battery pack would work for this purpose?
|
|
|
#
?
Jul 21, 2013 03:24
|
|
|
I can't help you with the battery, but I did that with a little DC motor last year at Burning Man and it was pretty sweet.
|
|
#
?
Jul 21, 2013 03:42
|
|
|
A hobby radio-control grade lithium polymer pack is the ideal mix of size, weight and energy density. I power the various lights and electronics on my biycle  with the Zippy 20C packs from HobbyKing and they work great.However, the thing about lithium packs is that they have to be charged carefully so that you don't blow them up. If you used regular heavy old nicads or nimh cells, you could essentially just hook them up to the dynamo with a diode and go. With lithium cells you need to regulate the charge with something like this unit from Adafruit: http://www.adafruit.com/products/280 That unit will only charge single cells (for safety reasons), so you're limited to the nominal 3.7v provided by a li-poly cell. If you're planning to drive a phone charger or run a 5v microcontroller you'll need a boost converter to bring the voltage up to the proper level: https://www.sparkfun.com/products/10255 As an aside, in my case, I actually use 3-in-series (11.1v nominal) li-poly packs rather than the single-cell type. The ~12v source lets me run things like electroluminescent wire inverters at higher efficiency; to power the central microcontroller I have a little 5v buck converter, and the head and taillamp LEDs are all running on their own constant-current driver boards that accept the 12v in with about 80% efficiency. I don't have any generators or dynamos and just charge the packs with a regular balancing charger. e: the charger that you linked to is pretty poorly constructed (a perfboard just sitting out in the open on the wheel in San Francisco weather?) and will be inefficient at high speeds. The linear regulator is an especially poor choice if he's seriously producing 30 volts at the generator going downhill but needs 5v for the phone. If you want to build something like that, at least put it in a good waterproof case please, and there are much better options (buck converter) for bringing the voltage down more safely and efficiently. Sagebrush fucked around with this message at 04:20 on Jul 21, 2013 |
|
#
?
Jul 21, 2013 03:49
|
|
|
Ephphatha posted:You didn't specify how much power you actually needed but these should suffice for a basic PC. drat. thanks for that. Newegg isn't an option though unless they've recently started shipping here. I've seen a lot of thermaltake PSUs but had no idea on their reliability. Might repurpose this PSU for my Apple IIgs. Did one with a uATX supply for my IIe years ago. I bought another for the IIgs but ants built a nest in it in the shed so I'm not convinced of its viability.
|
|
#
?
Jul 21, 2013 23:08
|
|
Sagebrush posted:A hobby radio-control grade lithium polymer pack is the ideal mix of size, weight and energy density. I power the various lights and electronics on my biycle Awesome, thanks.
|
|
|
#
?
Jul 22, 2013 00:33
|
|
|
I ordered a Corsair 450W PSU. Thanks again. Now I'm really curious because I used a really comprehensive PSU calculator http://www.extreme.outervision.com/PSUEngine to work out what this system needs. It thinks 258W recommended, which I believe because it fits with my rough mental arithmetic. So when the new PSU arrives I'm going to go hunting within the old one to see what caused such a strange failure.
|
|
#
?
Jul 22, 2013 07:27
|
|
|
Alright, building the rectifier/regulator for my mixer PSU tonight. No more dilly-dallying. Remind me the brief points on how to do a bipolar linear. Transformer center tap goes to PCB ground, I float the diode ring and then run each of those through +/- filtering and regulation? Edit: Say I wanna use 7812/7912's for my regs. I need to run a voltage divider off the output to ground, putting the 7812's ground pin at the middle, right? I calculated that a 390/100 ohm divider will give me ~3.0 volts at the junction if it gets 15 volts from the regulator out. Jonny 290 fucked around with this message at 20:30 on Jul 22, 2013 |
|
#
?
Jul 22, 2013 19:45
|
|
|
Jonny 290 posted:Alright, building the rectifier/regulator for my mixer PSU tonight. No more dilly-dallying. Yeah if you want to use a single bridge rectifier, you ground the center tap of the supply. I don't understand the regulator thing though. If you're using 7812/7912 and you want +/- 12v out, you don't need any extra resistors or dividers at all. You just tie the ground leg to ground, the input to your filter cap, and the output is at +/- 12v. Those are fixed voltage regulators, so they will have 12v between output and ground pin. Technically any regulator can be made 'variable' by adding a resistor from out to 'gnd' leg, which always have a fixed voltage (like 12v in your 7812, or 1.25v in a normal variable reg), so fixed voltage across fixed resistance gives you a current, then a resistor from gnd leg of regulator to circuit ground will have a specific voltage drop. So the final output voltage will be reference voltage of regulator plus voltage drop on ground leg resistor. Again this is only if you want higher voltage than the reference or fixed voltage of your regulators.
|
|
#
?
Jul 23, 2013 02:22
|
|
|
I'm assuming he wants an extra reference voltage for something... 490r on 12v will dissipate over a quarter watt all by itself (V2/R), though, you might consider somewhat higher values and buffer it with a follower if you're worried about loading the divider. Also bypass it to reduce noise. Alternately, there is no law against having an extra 3V regulator hanging off the 12V.
|
|
#
?
Jul 23, 2013 03:09
|
|
|
IIRC he wants a 15V supply and presumably only has 12V parts available.  This is how you want to hook up the transformer for a single winding with center tap. Putting 3V at the ground reference will increase the output voltage to about 15V but I'd still prefer to get 7x15s.
|
|
#
?
Jul 23, 2013 09:12
|
|
|
I.....I found a 7915 in my junkbox last night. And of course I don't have a 7815. God hates me. (Yeah, I need +/- 15V, and a 5V low current logic rail).
|
|
#
?
Jul 23, 2013 15:43
|
|
|
PDP-1 posted:This guy did a pretty nifty colored pixel grid display using WS2801 LED strings, and near the end of the writeup he describes having to use a kind of gamma correction curve to give the human-perceived output colors/brightness scale more linearly with input 0-255 values. Sounds like you may be running into the same issue. Bad Munki posted:I'm pretty sure the latest version of fastspi has some colorspace corrections (such as log-based curves) built in to it. You have to specifically use them, but you might want to open up some of the header files and have a look around in lieu of complete documentation. That was how I discovered the built-in HSV/RGB conversion stuff as well as some fun and nifty rainbow-generation routines. Great suggestions, lots to dig into here. I have more of an asthetic question about my LED panel now. I'm debating between the "box" pixel approach used in the PixelBrite panel, or the simple half ping pong ball diffusion technique. The box pixels are nice because they improve the pixel contrast and they diffuse the light nicely; but on the other hand an all-black (or possibly white) panel I don't think would look that great in my space. The ping pong ball approach will have more light cross-coupling and probably won't diffuse as well, but I feel like the panel itself would look much nicer as an art piece on the wall. Any thoughts?
|
|
#
?
Jul 23, 2013 16:25
|
|
|
I'm thinking of putting together an automation system for my tortoise table's lights, heat mats, etc, running off a Raspberry Pi or something like that. Given that all these plug into the mains, what's the simplest way of switching them on and off? Is there some sort of power extension strip that has the the ability to be controlled over USB or some sort of other interface? I'm not sure I want to start taking the things apart and wiring up the switches individually (and my girlfriend might not let me). I guess I wouldn't be averse to trying to build one if such a thing didn't exist.
|
|
#
?
Jul 25, 2013 15:46
|
|
|
chippy posted:I'm thinking of putting together an automation system for my tortoise table's lights, heat mats, etc, running off a Raspberry Pi or something like that. Given that all these plug into the mains, what's the simplest way of switching them on and off? Is there some sort of power extension strip that has the the ability to be controlled over USB or some sort of other interface? I'm not sure I want to start taking the things apart and wiring up the switches individually (and my girlfriend might not let me). PowerSwitch Tail, though they appear to be out of stock on the manufacturer's site.
|
|
#
?
Jul 25, 2013 16:00
|
|
|
Oh that looks good, apart from one snag, which I stupidly forgot to mention - I'm in the UK. I'll see if I can find a UK equivalent though, thanks.
|
|
#
?
Jul 25, 2013 16:01
|
|
|
I didn't see a more suitable thread for this so here goes: That's the mains filter from my Candy washing machine. You can't see it because it's melted but thats "N" on the right. I replaced the filter and the lead and everything seems to be in order (I don't see any smoke and my sheets seem to be getting cleaner - or at least wetter and spinnier). I'm just curious to know what the hell happened? Nothing else seems to be damaged. tl;dr: How/why does the "N" bit (I'm assuming "neutral") on a mains filter melt like that?
|
|
#
?
Jul 25, 2013 17:19
|
|
|
chippy posted:I'm thinking of putting together an automation system for my tortoise table's lights, heat mats, etc, running off a Raspberry Pi or something like that. Given that all these plug into the mains, what's the simplest way of switching them on and off? Is there some sort of power extension strip that has the the ability to be controlled over USB or some sort of other interface? I'm not sure I want to start taking the things apart and wiring up the switches individually (and my girlfriend might not let me). I'm going to be building basically this relay box soon for various purposes. Going to use it for Christmas lights, probably some sort of Halloween decoration, etc. Here's the relay board I bought, and the rest of it comes from the hardware store. I'm not planning on putting an Arduino in my relay box, but bringing the relay control connectors out to the side of the box for use with Arduino, etc. Sainsmart also makes relay boards with 2, 4, and 16 mains rated relays depending on your needs. EDIT: The relays are available on amazon.co.uk, and I'm assuming that mere mortals are allowed to buy electrical outlets, wall boxes, etc in the UK. I know some countries get a little finnicky about people who aren't licensed electricians doing their own electrical work. n0tqu1tesane fucked around with this message at 18:05 on Jul 25, 2013 |
|
#
?
Jul 25, 2013 18:01
|
|
|
Jerry Cotton posted:I didn't see a more suitable thread for this so here goes:
|
|
#
?
Jul 25, 2013 18:01
|
|
|
|
| # ? Jun 4, 2024 06:22 |
|
|
Jerry Cotton posted:tl;dr: How/why does the "N" bit (I'm assuming "neutral") on a mains filter melt like that? Similar to what tonberrytoby said, the Neutral line on a three-phase AC line is common between multiple phases and will therefore be a return path for all current. http://en.wikipedia.org/wiki/Ground_and_neutral#Three-phase_circuits
|
|
#
?
Jul 25, 2013 18:27
|
|
