|
My gut instinct is you should be ok, though what is your estimated flight time? I would highly recommend assembling all the subsystems and measuring their current draw while running them. It's a pain in the rear end but it's the only way to be confident. You'd have to do something insane to end up with a fireball, the real danger is running out of power too early and having no way to find your payload when it lands.
|
#
?
Sep 6, 2018 22:59
|
|
|
|
| # ? May 31, 2024 15:05 |
|
|
johnnyonetime posted:I've got a project I'm working on but I wanted to verify if my thinking is correct before I get too far down the rabbit hole. As with everything in aviation, redundancy is key. Yeah, it might be more annoying to have to have four packs and chargers, but AT LEAST put your telemetry stuff on its own power supply. That way, if the cameras go crazy and suck all your power you won't lose the whole package. Test on the ground before your flight, and test it inside a freezer to be doubly sure your cooler will work. edit: there's another page.
|
|
#
?
Sep 6, 2018 23:56
|
|
|
johnnyonetime posted:Will this be safe or will I end up with a high altitude fireball? Im not sure what this question is actually asking. Are you asking: 1) Will LiPo batteries and my electronics live in X hour transition from sea level to XXkft altitude and temp from XX°F to Y°F? Answer: I have no idea, but you should test it on the ground with your freezer and a vacuum pump + ice. 2) Do I have enough power to run this setup? No idea, use a power supply or something to measure the current draw of everything and figure it out. 3) Can I use a DC-DC converter instead of batteries? Generally, yes, though again try them out int he environment you want to try. I would also advise you to figure out redundant power. Also, I'd be careful with your connectors on your components. Dont just trust stuff to stay together, use tape, solder, etc.
|
|
#
?
Sep 7, 2018 03:12
|
|
|
I'd think the cameras might not be very happy if their previously clean stable battery power is replaced with relatively noisy buck converter output, have you actually tried that and left the cameras running for a while? I know CCD's can be pretty temperamental...
|
|
#
?
Sep 7, 2018 05:17
|
|
|
Also depending on how fast it comes down, plan for having water condense out on your electronics when they're cold and the air is warm and humid.
|
|
#
?
Sep 7, 2018 05:27
|
|
|
Mega reply time!quote:shovelbum According to my little DP50V5 power supply: trackuino = 12v, 150ma, 1.54watt 1.3ghz video transmitter + quadcopter camera = 12v, 400ma, 3watt Canon SD1000 = 3.7v, 100ma , 3watt Canon SD1000 = 3.7v, 100ma, 3watt I might be able to get by using 2 buck converters since the input voltage matches on both sets of devices quote:Splode Based on other flights I've seen online I am predicting a 3-5 hour flight time quote:babyeatingpsychopath That's probably a smart idea. I might go with 2 smaller capacity lipos and isolate the trackuino. That's a good idea testing it in my freezer! quote:CarForumPoster That's a good idea testing it in my freezer! As far as connectors go I'm trying to stick with Anderson Powerpoles and Dupont connectors everywhere. The powerpoles have these little plastic U shaped pieces that hold them together and I've got little plastic thingies for the Dupont that holds them together. quote:ate all the Oreos Didn't think about that. I do have an LC Filter from DCPAV that I can put in line with the cameras: https://www.readymaderc.com/products/details/l-c-filter-for-wireless-audio-video-systems I've ordered two cameras off eBay and I'm expecting them to arrive tomorrow. There are several "dummy battery" models on Thingiverse for Canon cameras and I planned on 3D printing two and hardwiring since I didn't think the built in batteries would last the entire flight. quote:Foxfire_ I've got big dessicant packs from work? I might put one or two in there for good measure. I'm doing this in Colorado and the relative humidity here is usually low but again, didn't think about that. Thanks for all the input! I've got some testing to do this weekend.
|
|
#
?
Sep 7, 2018 16:18
|
|
|
johnnyonetime posted:I've got big dessicant packs from work? I might put one or two in there for good measure. I'm doing this in Colorado and the relative humidity here is usually low but again, didn't think about that. IMO you're better off protecting against the effects of condensation than trying to prevent it. Conformal coating exposed CCAs is my suggestion. Any of your consumer electronics, if still enclosed, are probably low risk. That said, leave them in the freezer for 2 hours and then walk straight outside on a hot day. Does it work fine? Then youre probably fine.
|
|
#
?
Sep 7, 2018 20:37
|
|
|
Yeah I don't think dessicant packs alone would be fast enough, and as the thing comes down moist air would rapidly rush in and condense out before they could remove the moisture. Maybe if the box is sealed airtight except for two designated air pipes with check valves - one that allows air to vent directly to the outside for use on the way up and another that routes incoming air through a bunch of silica gel. The silica gel would be cold on the way down too, so the moisture might just condense out on that even if it doesn't absorb it fast enough. Or you could just spray everything with conformal coating but where's the fun in that when you could over-engineer it instead 
Shame Boy fucked around with this message at 21:43 on Sep 7, 2018 |
|
#
?
Sep 7, 2018 21:41
|
|
|
I read a post mortem of a HAB recently, and conformal coating over everything was a last minute "why not" thing that ended up saving it
|
|
#
?
Sep 7, 2018 22:39
|
|
|
Let's say I want to run a Raspberry Pi Zero from a 12 sealed lead acid battery. What would be the most energy efficient way to do it? I don't really have a reason or anything. I've just been wondering about it. All I can think of is a linear regulator or a buck converter. The former would probably burn a fair bit just with the regulator. The latter I really have no idea. Not sure how clean the supply would be either. I ran an esp8266 off a <5v > 5v buck converter once using a AA that couldn't power my mouse any more, but that was just to see what would happen. It worked. But that could have been due to whatever filtering the USB power suply on the esp8266 does. So, yes it's just a theoretical question, which I might try just for the sake of it.
|
|
#
?
Sep 9, 2018 01:49
|
|
|
They already sell off the shelf solutions for what you are trying to do. It's called a Battery Elimination Circuit, and they will take 5-25vdc and switch it down to 5-9vdc, without too much noise or power loss. it's used in the RC model world to run radio gear off a primary battery that has too high of a voltage.
|
|
#
?
Sep 9, 2018 01:59
|
|
|
You'd use a buck regulator if you want any efficiency whatsoever. But as the post above mentioned, it'll be easier and cheaper to buy a finished product than build your own.
|
|
#
?
Sep 9, 2018 02:15
|
|
|
This probably isn't *the* most power efficient way to do it, but I bet one of those alligator clip to 12V socket adapters + a good quality 12V USB adapter is reasonably close. It also has the added benefit of being a great emergency preparation to have on hand, so if you want to try it out it's not exclusively useful for one project.
|
|
#
?
Sep 9, 2018 02:54
|
|
|
MRC48B posted:They already sell off the shelf solutions for what you are trying to do. I believe a "Battery Elimination Circuit" is just RC people speak for a buck converter. If the output of the buck converter is too noisy (I bet it's not), you can slap a big cap on there to smooth it, but it should already have a cap for that exact reason, so I think it's fine. +1 on conformal coating it. Just buy a can and hose it all down.
|
|
#
?
Sep 9, 2018 04:49
|
|
|
Interesting. I have a couple of adjustable buck modules that can raise and lower voltage kicking around. I'm a big fan of putting big caps on inputs anyway so that's no big deal. I certainly do it with SBCs because most power supplies suck. Like I said though, absolutely no idea why I'd do it. Maybe some odd headless thing if I put a small solar panel on it to maintain the charge. Just see if it's viable. I dunno. I've always wanted to set up a mesh net doing something like that with espXXXX uCs, but have absolutely no reason to do so. Can't think of anything for them to talk about.
|
|
#
?
Sep 9, 2018 06:24
|
|
|
Main power supply -> buck converter -> LDO linear regulator is something I remember learning in school for clean but more efficient power. You set the buck output to something just above the minimum input voltage of the LDO, then you don’t need a massive low pass filter cap and get relatively efficient conversion.
|
|
#
?
Sep 9, 2018 07:34
|
|
|
meatpotato posted:Main power supply -> buck converter -> LDO linear regulator is something I remember learning in school for clean but more efficient power. You set the buck output to something just above the minimum input voltage of the LDO, then you don’t need a massive low pass filter cap and get relatively efficient conversion. Making a lot of assumptions here. Let's take my Pi Zero. It takes a 5v input. AFAIK the only thing which utilises that besides the 5v GPIO pins is the MicroUSB port. I believe the rest is 3v3. So perhaps a Pi Zero could run happily off a 5v output from a buck converter. USB devices should have their own filter cap / regulator(s) / whatever unless they are pretty garbage, like some of the crap I own. I guess what I'm saying is the scenario which you mentioned for good efficiency and stability already exists. I mean I think I have some low dropout 5v and 3v3 linear regulators handy, but they might not be needed. Not that I can actually think of a use for my idea. But that's not really the point. I'd just like to chuck a small solar panel on the battery and use it for ???
|
|
#
?
Sep 9, 2018 09:46
|
|
|
Remember that a lot of the raspberry pi variants draw quite a lot of current, so a usb power supply might not actually be up to the task
|
|
#
?
Sep 9, 2018 12:11
|
|
|
General_Failure posted:Interesting. I have a couple of adjustable buck modules that can raise and lower voltage kicking around. Most buck converters lose efficiency at an appreciable rate below a certain output load (usually about 20-30% full load in this power range) which actually becomes so bad at a point that for circuits with low power modes you put a LDO parallel to the buck and swap which is enabled when you enter low power mode, but this is usually only a useful trick for things that have sleep modes <1mA, which I am not aware of the raspberry pi having, but which would definitely apply to the esp8266/variants in some type of solar powered mesh network where you'd want them to be asleep most of the time. For the pi zero on a SLA, from the schematics of the pi you can see that it has a 47u cap on the power input connector so you shouldn't worry too much about ripple upsetting things unless your buck converters are particularly egregious. If you don't trust them, consider something like these buck converters which can also conveniently be dropped in to circuits previously designed for use with fixed output linear regulators, but are a little overkill for this application. There are many options for small dc/dc converters so if you want to be particularly efficient you need to figure out what your maximum current consumption is going to be (pi + anything plugged into it / other 5v peripherals) and choose one based on that.
|
|
#
?
Sep 9, 2018 20:04
|
|
|
Splode posted:Remember that a lot of the raspberry pi variants draw quite a lot of current, so a usb power supply might not actually be up to the task According to this guy's testing, they're all sub 1A. The worst is 980mA for a 3B+ under 400% load. (aka, 100% load on all 4 cores). It's not testing the graphics processor, so there are probably situations that could bring up the total to over 1A. Heavy onboard network access could plausibly add another watt or so as well. But really, just make sure you have a quality USB supply. 2.1A units are everywhere. Use a car one, it'll take 12v native. On the BEC side, there are ones that go up to very high currents, because they expect you to need to power servos off of them as well. If you were trying to also power external devices like external networking or drives you could make the power demands arbitrarily high. Use a powered hub after the pi help spread the load around. meatpotato posted:Main power supply -> buck converter -> LDO linear regulator is something I remember learning in school for clean but more efficient power. You set the buck output to something just above the minimum input voltage of the LDO, then you don’t need a massive low pass filter cap and get relatively efficient conversion. Even if you gave an RPI extra clean 5v, it'd immediately be dumped into its onboard switching regulator. Which probably powers the processor directly. It's a good trick in general. One thing to note is that you often can't just use cheapie LDOs if you really care about noise. The important spec is Power Supply Rejection Ratio (PSRR). It's basically how much noise a regulator lets though. Cheap regulators will be kinda meh, and it gets worse as frequency increases. As switching regulators prefer higher frequencies, these quickly run into conflict. You also just can't throw big honking caps on everything to eliminate high frequency noise. They also have ripple specs. The general trend is that capacitance decreases as frequency increases, and the effect is more pronounced as the capacitor gets physically larger.
|
|
#
?
Sep 9, 2018 23:42
|
|
|
Aurium posted:According to this guy's testing, they're all sub 1A. The worst is 980mA for a 3B+ under 400% load. (aka, 100% load on all 4 cores). quote:
quote:
I said Raspberry Pi Zero for a reason. Admittedly I didn't state it. Because they are pretty stripped and don't have WiFi, ethernet or an internal USB hub. Again, not sure what I'd use it for but it'd be interesting to see what is needed to keep it running. Maybe put an occasionally on OLED and an RTC on it. I have both. eInk display would be interesting, but I don't have any of those. quote:
quote:
quote:
Yeah, I know it doesn't work that way. However I have found that the RPi boards can have absolutely brutal current spikes which can cause sag and corruption. A decent cap helps a bit with that. I actually have a pi daughterboard I made which uses two core automotive wiring to supply power via GPIO, which I had connected to a 5v, 5A supply. Yes, it was an act of anger. Eventually after going through a lot of power supplies and cables I finally found a pair that could provide enough current all the time. I've been meaning to get one of the voltage dropping bucks with a fixed 5v output and USB plug. The only fixed 5v output one I have raises the voltage. The adjustable ones go both ways.
|
|
#
?
Sep 10, 2018 05:11
|
|
|
General_Failure posted:I've seen that side before at some point. I find the results a little odd but I guess their gear is way better than mine. I doesn't surprise me too much really. They used a USB connector as a power jack in the first place, so they had a power budget. The other thing is that the Broadcom SoC the first RPi used (the BCM2835) was drastically under used, and so it didn't use much power either. You see, the BCM2835 was originally supposed to be a media box controller. As such it needed a bunch of internal processing to deal with video encoding and decoding. They chose to do it with what's basically a multicore DSP coprocessor, much much more powerful and general purpose than what we typically think of when we talk about video accelerators, GPUs and hardware media codecs. This piece of the chip, called VideoCore is what's actually what's in charge of the chip, what controls the boot process, etc. It's not a setup that most chip manufacturers use. The VideoCore IP has been used by Broadcom only. It's behind a bunch of NDA's. It doesn't seem to have much actual benefit, and it's hard to use code designed for more typical ARM setups. So most often it runs a simple accelerator program, and it's treated as though it's much simpler than it is. It's not considered to have done particularly well in the market, and there's speculation this is why the RPi foundation got such a sweetheart deal. The ARM core that most of the code you actually care about on the RPi was originally put there to handle a small amount of data marshaling, and to be a pretty minimal linux computer to run a GUI, network, etc. Broadcom never expected it to do much, so it's tiny and low power. So basically, the original RPi is mostly runs on the auxiliary computer with the main one mostly turned off. The later BCM SoCs ones just stuck larger and larger ARM cores on it (it's what was actually being used after all), and left the VideoCore half there largely as a backwards compatibility fossil. quote:I've got heaps of the things. Maybe I should prod at them to work out their current overhead. quote:Yeah, I know it doesn't work that way. However I have found that the RPi boards can have absolutely brutal current spikes which can cause sag and corruption. A decent cap helps a bit with that. I actually have a pi daughterboard I made which uses two core automotive wiring to supply power via GPIO, which I had connected to a 5v, 5A supply. Yes, it was an act of anger. Eventually after going through a lot of power supplies and cables I finally found a pair that could provide enough current all the time. A decent supply with a good cap is pretty much all the RPi needs. You're totally right about the current spikes, sag and corruption. I was more going off on a tangent on there's low noise, and then there's low noise.
|
|
#
?
Sep 10, 2018 22:28
|
|
|
I got my babbys first PCB from dirty pcbs, with the $15 DHL it would've been super fast except for the hurricane. I only made one glaring error that will require dremeling! They even converted my lazy little tabs to proper mouse bites
|
|
#
?
Sep 18, 2018 17:22
|
|
|
shovelbum posted:I got my babbys first PCB from dirty pcbs, with the $15 DHL it would've been super fast except for the hurricane. I only made one glaring error that will require dremeling! They even converted my lazy little tabs to proper mouse bites  Do the dremeling outside and wear some sorta particulate filter mask because you really don't want to inhale FR4 dust if you can help it.
|
|
#
?
Sep 18, 2018 17:25
|
|
|
Also I think I brought them up before but what's y'all's opinion on ALLPCB? I've used them for a bunch of boards lately and they seem to produce good results, are very cheap and come with free express DHL shipping. I'm not trying to advertise for them or anything but I'm curious as to why I don't really hear anyone else use them at all, is there a horrible secret I don't know about  The only thing I can think of is their system isn't very user-friendly, you have to know how to export your gerbers the exact right way and there's no validation after you upload them (though I hosed it up once and they sent me an email the next day explaining what I did wrong) so it's probably not great for beginner use...
|
|
#
?
Sep 18, 2018 17:29
|
|
|
Use pcbshopper.com Most of these sites are nearly identical, and even use the same fabs on the backend. I haven't noticed any serious quality issues from the 5 or 6 different places I've tried. For a bit, allpcbs had $5 for ten pcbs, including shipping, and eight days to your door. I think it's more normal now.
|
|
#
?
Sep 18, 2018 17:39
|
|
|
ante posted:Use pcbshopper.com Yeah that's where I found ALLPCB to begin with ante posted:For a bit, allpcbs had $5 for ten pcbs, including shipping, and eight days to your door. I think it's more normal now. Yeah I remember that, the price is back up to like $20 for [five copies of] a 100mm by 100mm board now but that's fine. It's also only ever been 5 days to my door, idk if that's a region thing though.
|
|
#
?
Sep 18, 2018 17:53
|
|
|
It looks like price and shipping ends up around the same for a lot of these once you adjust qty and board size, it's great how fast stuff can get to my house these days
|
|
#
?
Sep 18, 2018 18:08
|
|
|
Are there any of these prototype board houses that do small run board stuffing for not crazy money?
carticket fucked around with this message at 19:11 on Sep 18, 2018 |
|
#
?
Sep 18, 2018 19:07
|
|
|
Check out Seeed Fusion. Maybe more expensive but better for a few reasons is Macrofab, based in Texas. They have a rad quoting interface, too
|
|
#
?
Sep 18, 2018 19:14
|
|
|
Seeed Fusion looks like a reasonable start. I still need to go through a prototyping phase, but I'd love to have an integrated solution. It's not practical for me to solder or try to learn toaster reflow on some of the parts, though. A house that can assemble it with BGAs also gives more options.
|
|
#
?
Sep 18, 2018 19:22
|
|
|
I have a stupid theoretical question and I'm hoping some of the EE heavy folks here can help out. Can the relative difference in the mass of an electron and a hole be used to make any human-perceptible difference in a solid state device? I'm thinking "no," either because the scales are ridiculous and throttling every pathway full of holes still won't be a fraction of as much mass as the chip packaging, OR I'm forgetting something about how they're balanced and even in a pathological case there's a local balance.
|
|
#
?
Sep 19, 2018 00:22
|
|
|
JawnV6 posted:I have a stupid theoretical question and I'm hoping some of the EE heavy folks here can help out. What do you mean by human perceptible difference? If you accept kind of abstract things as being human-perceptible, physicists over time have done careful measurements and have created tables of effective masses of electrons and holes in semiconductors. Also, mobility is usually inversely proportional to effective mass, and we can measure charge carrier mobility in many ways (hall measurement, resistance measurement of a semi-conductor with known charge carrier density). These are maybe cop-out answers. If you mean like whether the difference in effective mass of electrons and electron-holes could be measured using the gravitational force, I'm not sure. I'm also not sure if it is legal to use the concept of effective mass with gravity. When the concept of effective mass is invoked, it is usually in contexts outside of gravity. When talking about electrons and holes in semi-conductors, gravity is really weak compared to the electromagnetic force and is usually ignored, and this is why transistors work just as well when right side up versus when they are upside down. Maybe you could also ask this in the physics thread in SAL. There might be someone there who understands condensed matter physics (the newer name for solid state physics) really well and could better answer your question: https://forums.somethingawful.com/showthread.php?threadid=3781321&userid=0&perpage=40&pagenumber=1 silence_kit fucked around with this message at 03:01 on Sep 19, 2018 |
|
#
?
Sep 19, 2018 02:46
|
|
|
silence_kit posted:If you mean like whether the difference in effective mass of electrons and electron-holes could be measured using the gravitational force, I'm not sure. I'm also not sure if it is legal to use the concept of effective mass with gravity. When the concept of effective mass is invoked, it is usually in contexts outside of gravity (when talking about electrons and holes in semi-conductors, gravity is really weak compared to the electromagnetic force and is usually ignored--this is why transistors work just as well when right side up versus when they are upside down). I was thinking about a circuit whose center of gravity would shift due to a circuit state. Like a DRAM module that would be heavier with 1's instead of 0's. But there's a nagging feeling that I'm misunderstanding something, the effective in effective mass is probably it. silence_kit posted:condensed matter physics (the newer name for solid state physics)  dangit Thanks!
|
|
#
?
Sep 19, 2018 02:59
|
|
|
I feel like someone wrote a big article on this. I'm going to see if I can find this. E: http://www.ellipsix.net/blog/2009/04/how-much-does-data-weigh.html related to ferromagnetic storage, though. carticket fucked around with this message at 03:02 on Sep 19, 2018 |
|
#
?
Sep 19, 2018 02:59
|
|
|
JawnV6 posted:
I think this is a re-branding that is only well-known in physics--maybe it is to be inclusive to physicists who study liquids and other kinds of complicated matter (to be contrasted with the relatively simpler dilute gases and the high energy particles which are only found in billion-dollar instruments)? Wikipedia tells me that this re-branding started in the 1960's. I think every engineer calls it 'solid-state physics' though. JawnV6 posted:Oh right... this is one of my weaker areas, I was making the jump that effective mass was affected by gravity and that's proooobably not right. I"m not saying that it is impossible--I'm just saying I don't know if effective mass works like that. Maybe it does actually work like that, and some physicist has created some kind of solid state version of the Millikan oil drop experiment. Like when you turn an nMOS transistor on, is it actually very very slightly lighter since you have kind of converted 'heavier' holes in the semiconductor bulk into 'lighter' electrons at the semiconductor-insulator interface? I don't know. If so, it is probably is really difficult to measure though, since electrons are so light compared to protons and neutrons, and like you said, most of a transistor/circuit/chip/whatever by weight is just mechanical support and packaging.
|
|
#
?
Sep 19, 2018 03:27
|
|
|
Anyone know anything about audio amplifiers? I have a very old Yamaha amp (SR-50) I use for my PC sound and sometimes one channel will drop off, but will come back if I crank the volume up and then back down. I don't know much about electronics but can use an iron so I'd be happy to change out capacitors or something if that sounds like a possible suspect. I'm hoping the issue is strange enough to have a small number of likely sources I can check. I bought this thing back in 1990 or so, so I've had plenty of mileage out of it, but it's also too nice to just replace if there's some easy fix.
|
|
#
?
Sep 19, 2018 03:38
|
|
|
istr reading somewhere that all the electrons moving through every manmade electrical circuit on Earth have a total mass of like 3kg. I don't think that helps in your calculation at all, it's just a neat factoid
|
|
#
?
Sep 19, 2018 03:39
|
|
|
Welp, all the left-half boards got their thermal reliefs, and all the right-half boards got nothing. Of COURSE I did a right-half board first. Way to lazily copy-paste in KiCad and not check that those populated, me. Oh well, I have more left-half boards than I do of the expensive components anyway.
|
|
#
?
Sep 19, 2018 04:17
|
|
|
|
| # ? May 31, 2024 15:05 |
|
|
ALLPCB is fine, their marketing is a little scummy though. They like to spamvertize message boards and have like 20 different storefronts that all map to the same underlying company to try to slurp up traffic. Used them once since they do aluminum core boards in prototype quantity.
|
|
#
?
Sep 19, 2018 04:46
|
|