Learning electronics: Release the Magic Blue Smoke

PDP-1 · # ? Apr 24, 2014 02:03

Stabby McDamage posted:

Couldn't you take the oscilloscope, put it in AC mode (removing the DC offset), and poke it random places (including that power line) until you find the 100kHz source?

movax posted:

100kHz seems kind of low to me actually for a laptop rail supply, especially for recent models. I'd be expecting 500kHz or higher to let them use smaller inductors / parts to save on room, there are a lot of inexpensive controllers that will run from 500kHz to 2MHz (w/PLL) available now. I imagine there's an Intersil or Linear Tech 5V supply somewhere in that laptop that might be the culprit.

Anything else connected to the laptop other than the ADC board? Same behavior with/without an AC adapter attached?

ANIME AKBAR posted:

Most USB peripherals don't care if the 5V line is noisy as all hell. So it's not surprising if some manufacturers build lovely supplies for them.

Have they/you actually probed the 5V line to look for the switching noise? Tried with another laptop or PC? It could be a nearby piece of electronics too. At work we used to have a laptop with an aftermarket universal charger and it would consistently spew out EMI at several specific frequencies.
Have they actually tried putting a LPF on the analog inputs to kill the interference?

Thanks everyone for the comments and suggestions. We found the problem today which turned out to be an SMPS that was running another part of the experiment but was sitting close enough to the ADC box to magnetically couple its switching noise onto the lines we were trying to measure. We moved the SMPS further away and tightened up some of the wiring so we didn't have any large open loops to act as noise antennas and the 100kHz jitter went away.

We also found another issue where the signals from one channel appeared to be coupling a DC offset onto their neighboring channels. Right now we think this is because our sensors look like a 10mV voltage source in series with a 2MOhm Thevenin resistance, so they can't source/sink enough current to fully 'clear out' the capacitance on the ADC's input stage as it multiplexes from channel to channel. We're going to try sticking an instrumentation op-amp on the signal sources in hopes that the lower output impedance of the amplifier will be able to pull the ADC's input to the correct level before it makes a reading.

Debugging noise sources is such a goddamn pain in the rear end sometimes. :negative:

Adbot · # ? Jun 12, 2024 02:58

Cyril Sneer · # ? Apr 24, 2014 02:18

SandBox posted:

I'm certainly no expert, but I can ramble a bit about my dealings with hobby lcd screens and maybe it will be useful for you:

I started out wanting to do the same (simple) things you're describing, and I found that the screens with drivers are a lot easier to talk to with MCUs. The driver chips act as a small frame-buffer that is mapped in one way or another to the pixels on the screen. This makes it pretty easy to just feed pixel data into the memory as you go, and have the driver chip handle everything else. I would typically write some pixels to memory and then tell the driver chip "I'm done you can update the screen now". The driver chip will typically have some way of specifying which direction is "up" on the screen (i.e. the specific way memory should be mapped into the panel; top down left to right etc). Most driver chips will have a number of ways to interface to an MCU; 8bit parallel and 16bit parallel are fairly common. The vast majority of driver chips attached to large lcd panels will take 24bit parallel, which I found frustrating as PIC32 chips can only output 16bits from their parallel module. You could always attempt to bit-bang the 24bit parallel stuff but that would be irritating and slow. Some driver chips for small displays have SPI/I2C but they are slow as balls.

I guess with no lcd driver chip the MCU would have to handle all of the signal and timing aspects (i.e. the screen just zips away refreshing at whatever Hz and it's your job to program the MCU to inject the right pixel data at the right time to form a coherent image on the screen). Many hobby lcd projects I've seen that use driver-less screens use FPGAs to handle the timing and data processing required. This is probably way too annoying to implement in practice.

You've probably noticed that I haven�t mentioned any graphics stuff yet, it's all been single pixels placed into memory. The only way I found I could display graphical elements on my screens was to write my own routines in the MCU. My text routine would take a string as an input and construct a small array of pixel colour values and push it to a small portion of the driver chip frame buffer, one pixel at a time. HOW TEDIOUS. Drawing circles was even more irritating. Thankfully, there are example snippets of code out there so that you don�t need to bother with this stuff as much. (See: http://www.newhavendisplay.com/exprog_page.html) These example sites are fantastic as they show you how to initialise the display etc.

I got to the stage where I could write lines of text, draw lines and squares and small sprites on a screen, but it was all pretty tedious and my code-base was huge. Capturing touch screen events was something I never got working properly.

Only recently I found a new chip that claimed it could handle all of the graphical element creation side of things: http://www.ftdichip.com/Products/ICs/FT800.html
I've tried it out and it works really well. The chip acts as an intermediary between the lcd screen and the MCU (I could be wrong but I believe you still need an lcd panel with a driver chip included). Interface from MCU to the FT800 chip is via SPI; the FT800 chip handles the entire interface to the display driver. Note that the FT800 chip can talk to any lcd driver chips in up to 24bit parallel. Laying down graphical elements is done with specially crafted SPI commands. There are also some very nice in-built buttons/switches/gauges that can be drawn on the screen by calling specific SPI commands. Another feature is that the FT800 chip can handle resistive touch screen event captures. I've used this chip in my projects and I got as far as displaying text and coloured boxes before getting distracted.

If you want to use the FT800 chip without having to fab your own board or buy a separate screen then I can recommend: http://www.mikroe.com/add-on-boards/display/connecteve/
You could just mount the thing, connect power and SPI and be done with it. The tricky part would then be how to craft the SPI messages to do exactly what you want to do in terms of displaying text/sprites etc.

Massive effort post but there you go, hopefully is of some use to you. Fire away with questions if I'm completely off the mark with my ramblings.

Yes, this is some nice insight, thanks!

Blotto Skorzany · # ? Apr 24, 2014 02:19

PDP-1 posted:

We also found another issue where the signals from one channel appeared to be coupling a DC offset onto their neighboring channels. Right now we think this is because our sensors look like a 10mV voltage source in series with a 2MOhm Thevenin resistance, so they can't source/sink enough current to fully 'clear out' the capacitance on the ADC's input stage as it multiplexes from channel to channel. We're going to try sticking an instrumentation op-amp on the signal sources in hopes that the lower output impedance of the amplifier will be able to pull the ADC's input to the correct level before it makes a reading.

Are you feeding your signal directly into the switch-cap input of the ADC without a buffer amp? Because that'll cause that.

ANIME AKBAR · # ? Apr 24, 2014 12:20

I'd really hate to think the NI DAQs don't at least have a decent buffer stage between the input and the ADC.

Shame Boy · Shame Boy Mar 2, 2010

So I got one of those neat ZIF sockets and want to build a reconfigurable programmer with it. Basically, I want to be able to map voltage, ground, 3 input and 1 output pin to any pin on the ZIF socket dynamically, using a microcontroller. I came up with this design for handling it:

The idea is you'd pull one of each set of *Select pins high, which would route things from the programmer (left) to the device (right). I only drew one input (top left), one output (bottom left) and 4 device pins, but it would just be this pattern for the rest of them.

Will this work? I've played around with MOSFET's before but never quite like this. Will the gate voltage on the MOSFET's mess with the data lines? Are there better ways of doing this?

EDIT: All the resistors in the picture are really high value pull-down resistors, if that wasn't clear.

Shame Boy fucked around with this message at 18:22 on Apr 24, 2014

asdf32 · # ? Apr 24, 2014 19:25

Parallel Paraplegic posted:

So I got one of those neat ZIF sockets and want to build a reconfigurable programmer with it. Basically, I want to be able to map voltage, ground, 3 input and 1 output pin to any pin on the ZIF socket dynamically, using a microcontroller. I came up with this design for handling it:

The idea is you'd pull one of each set of *Select pins high, which would route things from the programmer (left) to the device (right). I only drew one input (top left), one output (bottom left) and 4 device pins, but it would just be this pattern for the rest of them.

Will this work? I've played around with MOSFET's before but never quite like this. Will the gate voltage on the MOSFET's mess with the data lines? Are there better ways of doing this?

EDIT: All the resistors in the picture are really high value pull-down resistors, if that wasn't clear.

Perhaps if you drive the select pin with a higher voltage than the signals you're selecting. Vgs needs to remain >~3V to maintain the fet in an on state and the signal you're trying to pass will be on S. You also need to factor in the intrinsic diode and make sure you get it right (a single fet will only block in one direction).

A better option would be to use analog switch or analog mux chips which are designed for this purpose. Off the top of my head look at the NLASB3157 for an example of a 2:1 analog mux.

Shame Boy · Shame Boy Mar 2, 2010

asdf32 posted:

Perhaps if you drive the select pin with a higher voltage than the signals you're selecting. Vgs needs to remain >~3V to maintain the fet in an on state and the signal you're trying to pass will be on S. You also need to factor in the intrinsic diode and make sure you get it right (a single fet will only block in one direction).

A better option would be to use analog switch or analog mux chips which are designed for this purpose. Off the top of my head look at the NLASB3157 for an example of a 2:1 analog mux.

Ah, I was looking for something more along the lines of a solid state "switch" originally but the only thing I could come up with were solid state relays meant for switching things like mains power which seemed overkill for this (and overly expensive, like a buck each). I'll check those out, thanks!

EDIT: Actually, could I just use something like a 4051?

Shame Boy fucked around with this message at 19:35 on Apr 24, 2014

asdf32 · # ? Apr 24, 2014 20:07

Parallel Paraplegic posted:

Ah, I was looking for something more along the lines of a solid state "switch" originally but the only thing I could come up with were solid state relays meant for switching things like mains power which seemed overkill for this (and overly expensive, like a buck each). I'll check those out, thanks!

EDIT: Actually, could I just use something like a 4051?

Yep on a quick look that's an example of an analog mux like I was talking about.

Note I'm not sure you need analog here, I went that direction because that's what your fets would have given you.

Analog can be convienient when you have varying directions or varying voltages, which you might here.

Shame Boy · Shame Boy Mar 2, 2010

asdf32 posted:

Yep on a quick look that's an example of an analog mux like I was talking about.

Note I'm not sure you need analog here, I went that direction because that's what your fets would have given you.

Analog can be convienient when you have varying directions or varying voltages, which you might here.

Cool, thanks! I figured there was some kind of thing out there already that did exactly what I was trying to do but I couldn't figure out what words to cram into google to get what I wanted (mux! :argh:

)

Delta-Wye · # ? Apr 26, 2014 18:58

I assume that the 3 inputs and 1 output is JTAG? If you're just mapping JTAG around a connector, a digital mux would make sense over an analog mux.

Shame Boy · Shame Boy Mar 2, 2010

Delta-Wye posted:

I assume that the 3 inputs and 1 output is JTAG? If you're just mapping JTAG around a connector, a digital mux would make sense over an analog mux.

JTAG, +5V and ground, yeah. Oh and maybe two pins for a crystal or resonator.

asdf32 · # ? Apr 26, 2014 23:46

Delta-Wye posted:

I assume that the 3 inputs and 1 output is JTAG? If you're just mapping JTAG around a connector, a digital mux would make sense over an analog mux.

But don't overlook some of the benefits of analog. For example, an analog mux/switch driven with 5V can pass anything <5V, including 3.3, 1.8 etc. Also good ones can actually be better from a propagation delay perspective because they pass what comes in immediately, and don't wait until some threshold to start driving their output.

Also, because they don't actively drive they can simplify signal integrity (although they have some R and C, but often quite low). For example I try and put a source series resistors on every digital buffer output, but I'll skip that in the case of an analog mux/switch, which I see as a ~10-20 ohm series resistance - generally no big deal.

I have a current design that uses the NLASB3157 to mux JTAG signals from two sources (FGPA <-> flash/Header <-> flash). It works well for this purpose.

And again, they handle bidirectional signals (like I2C) perfectly.

JawnV6 · # ? Apr 29, 2014 17:28

How do those shake watches generally work? Is it piezoelectric energy harvesting or something more complex? I've kinda been lucky in that every time it's been suggested we had other reasoning for not wanting to use it, but something came up that seems like a fit from my relative ignorance.

Opioid · # ? Apr 29, 2014 17:53

JawnV6 posted:

How do those shake watches generally work? Is it piezoelectric energy harvesting or something more complex? I've kinda been lucky in that every time it's been suggested we had other reasoning for not wanting to use it, but something came up that seems like a fit from my relative ignorance.

You mean an automatic watch? There is zero electricity involved in those. The movement of the wearer's arm and hand spin a circular weight inside the watch that then stores this kinetic energy in a spring system to release it in a controlled fashion for the timekeeping. You can also give the spring energy by winding it like a normal mechanical watch

Shame Boy · Shame Boy Mar 2, 2010

I've seen ones that have a magnet on a track around the outer perimeter of the watch and coils along the track so that as the magnet passes them it generates little pulses of electricity. Walking causes the magnet to move around on the track and generates enough power to run a crappy LCD and a crystal.

EDIT: I can't find the one I'm talking about, but I found this one that just has a teeny tiny generator in it: http://en.wikipedia.org/wiki/Automatic_quartz

poeticoddity · # ? May 1, 2014 04:57

Quick question about power supplies:

Is there something I should be looking at to supply 5V to 2 700mA LEDs and some sort of control hardware (tbd) that's cheaper or better than this which has been properly safety tested and would be suitable for making a few decorative lamps?

Cyril Sneer · # ? May 1, 2014 23:30

Arrgh so I'm still a bit overwhelmed by my lcd touchscreen options.

For example, I found this:

http://www.crystalfontz.com/product/CFAF320240F035TTS

which is a kit that includes the lcd, touchscreen, and driver chip with a breakout board. Its $86.00 The lcd+touchscreen only is $53.

Similar modules on digikey range from about $35 - $75. Is there substantial difference in quality between manufacturers or something? \

Are screens generally interchangeable with driver chips? Like, could I take this module:

http://www.digikey.ca/product-detail/en/NHD-3.5-320240MF-ATXL%23-1/NHD-3.5-320240MF-ATXL%23-1-ND/2165878

and drive it from say an SSD2119 (ie. are all 320 x 240 screens and drivers the same)?

edit: Or, like, there's this driver: http://www.newhavendisplay.com/app_notes/HX8238.pdf which I could use with this screen: http://www.alibaba.com/product-detail/3-5-tft-lcd-touchscreen-display_800710318.html which is only $15! But...like...do I need to find drivers or something that I can compile to my uC? (I don't want to have to write display code myself)

Cyril Sneer fucked around with this message at 23:39 on May 1, 2014

asdf32 · # ? May 2, 2014 01:31

poeticoddity posted:

Quick question about power supplies:

Is there something I should be looking at to supply 5V to 2 700mA LEDs and some sort of control hardware (tbd) that's cheaper or better than this which has been properly safety tested and would be suitable for making a few decorative lamps?

That gets the job done. What are you controlling the LED current with?

poeticoddity · # ? May 2, 2014 01:49

asdf32 posted:

That gets the job done. What are you controlling the LED current with?

Usually with these LEDs I just put a transistor and a resistor in the circuit and PWM it to adjust brightness. I'd been mulling over using a 555 timer, but I'm actually thinking this would be a good project to jump from full Ardunios to ATTinys. I'm open to suggestions, though.

priznat · # ? May 2, 2014 02:30

Cyril Sneer posted:

Arrgh so I'm still a bit overwhelmed by my l But...like...do I need to find drivers or something that I can compile to my uC? (I don't want to have to write display code myself)

I was gonna say, depending on what you're developing for have a look at driver support before deciding on a panel. I got saddled with an LCD the board designer just chose for a Zynq based project and it had no Linux drivers whatsoever. Simply by choosing another display it would have been a trivial process but instead I had to spend days building my own drivers for it.

I'm sure if it is for a Pi or a beagle board or something there are a few options. Or if it's for something like Linaro check out their driver folders for ideas.

Stabby McDamage · # ? May 2, 2014 03:42

poeticoddity posted:

Usually with these LEDs I just put a transistor and a resistor in the circuit and PWM it to adjust brightness. I'd been mulling over using a 555 timer, but I'm actually thinking this would be a good project to jump from full Ardunios to ATTinys. I'm open to suggestions, though.

Using ATtinys is easy. Just install a Tiny core (either Arduino-tiny or the MIT one), get a cheap USBasp programmer, make a simple breakout cable (Gnd, Vcc, MOSI, MISO, SCK, Reset) by chopping up a 10- or 6-pin ribbon, and pick an ATTINY85 (8 pin) or ATTINY84 (14 pin).

poeticoddity · # ? May 2, 2014 03:57

Stabby McDamage posted:

Using ATtinys is easy. Just install a Tiny core (either Arduino-tiny or the MIT one), get a cheap USBasp programmer, make a simple breakout cable (Gnd, Vcc, MOSI, MISO, SCK, Reset) by chopping up a 10- or 6-pin ribbon, and pick an ATTINY85 (8 pin) or ATTINY84 (14 pin).

I should have clarified: I have a programmer already and I get the gist of it, I've just never actually had a project that didn't warrant using an Arduino, so I've never gotten around to playing around in just AVRC.

Stabby McDamage · # ? May 2, 2014 04:19

poeticoddity posted:

I should have clarified: I have a programmer already and I get the gist of it, I've just never actually had a project that didn't warrant using an Arduino, so I've never gotten around to playing around in just AVRC.

Ugh, why'd you want to use plain AVR C? I use Arduino language for everything, including (and especially) bare chips. You can just bust out raw register access when you need to, but keep it high level when you don't. I don't see any reason to give up the Arduino niceties as a baseline. (Note, that doesn't mean you have to use their 'IDE'...)

poeticoddity · # ? May 2, 2014 05:13

Stabby McDamage posted:

Ugh, why'd you want to use plain AVR C? I use Arduino language for everything, including (and especially) bare chips. You can just bust out raw register access when you need to, but keep it high level when you don't. I don't see any reason to give up the Arduino niceties as a baseline. (Note, that doesn't mean you have to use their 'IDE'...)

I'm trying to familiarize myself with it so I don't have to use the bootloader and so I can write slightly faster code for lab equipment. Going from Arduino to AVR C for a project this simple strikes me as a good way to benchmark how much more time I should expect to accomplish the same task with slightly faster code. Plus, I've wanted to know how for a while.

FSMC · # ? May 2, 2014 11:36

Cyril Sneer posted:

Arrgh so I'm still a bit overwhelmed by my lcd touchscreen options.

For example, I found this:

http://www.crystalfontz.com/product/CFAF320240F035TTS

which is a kit that includes the lcd, touchscreen, and driver chip with a breakout board. Its $86.00 The lcd+touchscreen only is $53.

Similar modules on digikey range from about $35 - $75. Is there substantial difference in quality between manufacturers or something? \

Are screens generally interchangeable with driver chips? Like, could I take this module:

http://www.digikey.ca/product-detail/en/NHD-3.5-320240MF-ATXL%23-1/NHD-3.5-320240MF-ATXL%23-1-ND/2165878

and drive it from say an SSD2119 (ie. are all 320 x 240 screens and drivers the same)?

edit: Or, like, there's this driver: http://www.newhavendisplay.com/app_notes/HX8238.pdf which I could use with this screen: http://www.alibaba.com/product-detail/3-5-tft-lcd-touchscreen-display_800710318.html which is only $15! But...like...do I need to find drivers or something that I can compile to my uC? (I don't want to have to write display code myself)

What uC are you using?

There are generally two levels of code. First is the driver to interface with the lcd driver chip, that's usually basic like initialisation code, and display this color at x, etc. Then you have the useful code for drawing lines, GUI etc. Basically you should be able to write the low level drivers from the spec sheet, but it helps if someone else has already done it. There are a few libraries for the later, you might need to put in some glue code.

At 320x240 most displays will have a display driver integrated into the display so you can't really mix and match.

I would recommend doing some prototyping with the itead stuff, they will have schematics and drivers, GUI's etc. Then you can start your own design from scratch or just copies theirs. The cost of their displays are still cheaper than most of the ones you looked at http://imall.iteadstudio.com/display/im120906008.html

Stabby McDamage · # ? May 2, 2014 14:43

poeticoddity posted:

I'm trying to familiarize myself with it so I don't have to use the bootloader and so I can write slightly faster code for lab equipment. Going from Arduino to AVR C for a project this simple strikes me as a good way to benchmark how much more time I should expect to accomplish the same task with slightly faster code. Plus, I've wanted to know how for a while.

That's cool, and I'm not trying to argue on the internet, but I just want to note that you don't need to use the Arduino bootloader. I don't use it ever -- even on bare ATmega328P's. The bootloader chiefly lets you program over serial/USB, but if you use a USBasp, you program directly, so you can take bare chips straight from the box and flash arbitrary Arduino code onto them, no bootloader needed. (For ATTINY's, the MIT core doesn't even have a bootloader, so the "flash bootloader" option only sets the fuses, which makes for an easy clockrate selector.)

Moist von Lipwig · # ? May 4, 2014 20:40

My router blew a cap and I'd rather just replace the cap vs the whole drat thing but I wanna make sure I order the right one here. I'm looking on Digikey and there's Aluminum-Polymer Caps, Silicon Caps, blah blah etc

The outside of the current cap says 16v 680�F LZ 105c 1114T HERMEI but I'm not sure which variety of cap to get with these specs, any help?

poeticoddity · # ? May 4, 2014 21:47

LARD LORD posted:

My router blew a cap and I'd rather just replace the cap vs the whole drat thing but I wanna make sure I order the right one here. I'm looking on Digikey and there's Aluminum-Polymer Caps, Silicon Caps, blah blah etc

The outside of the current cap says 16v 680�F LZ 105c 1114T HERMEI but I'm not sure which variety of cap to get with these specs, any help?

If it's 680�F, it's most likely an electrolytic cap. Take a look at the blown cap in the router and if it's a cylinder, it's most likely electrolytic, which are just listed as "aluminum" rather than "aluminum-polymer" on digikey.

Moist von Lipwig · # ? May 5, 2014 00:09

poeticoddity posted:

If it's 680�F, it's most likely an electrolytic cap. Take a look at the blown cap in the router and if it's a cylinder, it's most likely electrolytic, which are just listed as "aluminum" rather than "aluminum-polymer" on digikey.

Yup, definitely aluminum then, thanks for the help :tipshat:

asdf32 · # ? May 5, 2014 00:27

LARD LORD posted:

My router blew a cap and I'd rather just replace the cap vs the whole drat thing but I wanna make sure I order the right one here. I'm looking on Digikey and there's Aluminum-Polymer Caps, Silicon Caps, blah blah etc

The outside of the current cap says 16v 680�F LZ 105c 1114T HERMEI but I'm not sure which variety of cap to get with these specs, any help?

You might be surprised if you google for the actual part. My Delta bench grinder blew a cap and rather than take a chance (I don't know anything about AC motors) I just bought the actual replacement. Decent brands have partslists and you can find the parts online.

sixide · sixide Oct 25, 2004

http://www.taiwantrade.com.tw/EP/hermei/products-detail/en_US/742292/Low_IMPEDANCE/

Looking at the LZ series catalog you should be able to find a direct replacement from a more reputable brand (that's available in the US).

Moist von Lipwig · # ? May 5, 2014 02:04

asdf32 posted:

You might be surprised if you google for the actual part. My Delta bench grinder blew a cap and rather than take a chance (I don't know anything about AC motors) I just bought the actual replacement. Decent brands have partslists and you can find the parts online.

Already tried it and only got a few results indicating the exact same issue. I'm really bad with electrical math so I didn't want to be subbing voltages or capacitances and blowing something up :ohdear:

sixide posted:

http://www.taiwantrade.com.tw/EP/hermei/products-detail/en_US/742292/Low_IMPEDANCE/

Looking at the LZ series catalog you should be able to find a direct replacement from a more reputable brand (that's available in the US).

Thanks very much!

EDIT: Oh look the exact same fried component

Moist von Lipwig fucked around with this message at 02:08 on May 5, 2014

Sagebrush · # ? May 5, 2014 03:28

Yeah, the nice thing about failures in electronic boards is that physics dictates that they will all tend to fail in a similar way given similar inputs. If one person puts reverse power into their router and it blows up a certain voltage regulator and a capacitor, you can probably expect that putting reverse power into your router will also blow up the same voltage regulator and capacitor. Common problems, common fixes.

poeticoddity · # ? May 5, 2014 04:06

LARD LORD posted:

I'm really bad with electrical math so I didn't want to be subbing voltages or capacitances and blowing something up

IIRC, in most circumstances the voltage rating for a capacitor isn't important as long as it's high enough to meet or exceed spec. Too low and you'll damage it, but too high and it just doesn't charge to full voltage, which is fine.

MRC48B · MRC48B Apr 2, 2012

Voltage ratings for caps are like a "Do not exceed" limit. Anything equivalent or higher and you will be fine.

movax · movax Aug 30, 2008

Depending on the voltage coefficient of the capacitor, putting a higher voltage cap in may actually result in better performance (closer to rated capacitance) at operating voltage. Certain types can "lose" 30% or so as they approach their rated voltage. Seen it the most with ceramics, there's a decent AN from Vishay with data on it.

ekuNNN · # ? May 5, 2014 15:13

Hey guys, I was wondering if it would cause problems if I switch between two similar power supplies while things are running. For example if I have a raspberry pi running on 5v from a battery and I want to switch to a backup power source that's the same voltage, can I just flip a relay or something?

Captain Cool · # ? May 5, 2014 23:36

What do I have to do to charge a lithium battery while drawing power from it?

I want to hook up a small solar cell, a lithium battery, a microcontroller, and a servo or motor. I don't know a lot about how a lithium charge controller actually works, but I think that the power draw from the motor will drop the battery voltage enough to confuse the charge controller. Can I run a wire to interrupt the microcontroller when the charge controller is sensing voltage? Or should I run a signal the other way, to disable the charger while running the motor? Or am I on the wrong track here?

ekuNNN posted:

Hey guys, I was wondering if it would cause problems if I switch between two similar power supplies while things are running. For example if I have a raspberry pi running on 5v from a battery and I want to switch to a backup power source that's the same voltage, can I just flip a relay or something?

I searched for power mux and found this discussion: http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=115485

Valdara · # ? May 5, 2014 23:44

I'm a physics teacher, and we are covering circuits. I had my students fooling with 9V batteries, holiday lights, and multimeters to look at series v parallel (they were astounded at finding the current the same everywhere in a series circuit, and that you can take out one light and still have the other going in parallel). One of my multimeters is now acting crazy. No, it's not the fuse, I promise. I checked it, changed it, tried it in a different multimeter, and it is 100% fine. The < 200mA setting works just fine with no issues. The problem comes when I try to measure on the 10ADC setting. It either reads 0 or some absurdly small value. I've opened up a few of them, and there is no discernible difference between the insides of the broken one and a not-broken one. I tried it on a few different circuits a few different times. I'm guessing the students just started sticking leads places because Hey! Toys with pointy bits! and accidentally hit both battery terminals at the same time. Is it now just a <200 mA and DC voltage meter only, or is there a way to get the 10ADC readings back?

Adbot · # ? Jun 12, 2024 02:58

SybilVimes · # ? May 5, 2014 23:51

Valdara posted:

I'm a physics teacher, and we are covering circuits. I had my students fooling with 9V batteries, holiday lights, and multimeters to look at series v parallel (they were astounded at finding the current the same everywhere in a series circuit, and that you can take out one light and still have the other going in parallel). One of my multimeters is now acting crazy. No, it's not the fuse, I promise. I checked it, changed it, tried it in a different multimeter, and it is 100% fine. The < 200mA setting works just fine with no issues. The problem comes when I try to measure on the 10ADC setting. It either reads 0 or some absurdly small value. I've opened up a few of them, and there is no discernible difference between the insides of the broken one and a not-broken one. I tried it on a few different circuits a few different times. I'm guessing the students just started sticking leads places because Hey! Toys with pointy bits! and accidentally hit both battery terminals at the same time. Is it now just a <200 mA and DC voltage meter only, or is there a way to get the 10ADC readings back?

10A will use a separate current sense resistor than the rest, it might have died - DMMs often have the high current input separate or separately fused for this reason.