|
Zero VGS posted:One other question, you said that if the pack is at 315v, the solar charger might toss 320v at it, and the AC charger would see that. What would prevent the AC charger from then saying "OK, I'll charge at 325v" and the solar charger follows up with 330v, as they try to leap frog each other? If they worked the way you’re postulating, i.e. “add 5 V to whatever I just measured on the terminals”, each charger would leapfrog itself, even with no other charger connected. Cycle 0: charger measures 315 V at the terminals, outputs 320 V In the next cycle, the charger would measure 320 V, then decide to output 325 V. There is a single measurable voltage across any two points on a circuit at any given time. There is no way to separate the voltage into “battery voltage component”, “charger № 1 voltage component”, and “charger № 2 voltage component”. That is, they cannot be separated at a single point in time. What a charger can do is disconnect its output and see what the voltage on the battery terminals sags to when it’s not being charged (keeping in mind that it will decay for some time after disconnect). I have a 12 V car battery charger that works like this. I can hear the whine as it charges the battery, and maybe once a minute it lets up and checks the battery’s state. This is one way that certain smart chargers could get confused if there are multiple chargers connected. They try to measure the battery at rest but the other chargers are alternately charging and checking, too. But they’re not going to respond by applying infinite voltage because that violates their current limits and is not physically possible. What they might do is see data that doesn’t make sense and refuse to charge the battery. Platystemon fucked around with this message at 20:01 on Mar 13, 2017 |
#
?
Mar 13, 2017 19:55
|
|
|
|
| # ? Jun 3, 2024 04:10 |
|
|
DethMarine21 posted:I have one of these Dangerous Prototpyes ATX power supply breakout boards and it's pretty good. Can't say I've ever used it with breadboards though. They have a nice acylic case available for it too. Hmm, I do like the binding posts on that and it wouldn't be too hard to glue that and a bunch of breadboards to a backplate... thanks for the tip!
|
|
#
?
Mar 13, 2017 20:09
|
|
|
Zero VGS posted:
The solar charger shouldn't be outputting 350V. It should be setting the voltage for an appropriate charge current, just like the other charger. That said, it does still seem like an opportunity for catastrophic confusion (at least one charger will be seeing an elevated voltage and be unaware of other current flowing into the battery), so it sounds like the engineer you spoke to is making at least one assumption about the rest of the system. I don't know what that might be, but you are correct that there is only one V_batt, so generally, if you have multiple sources, you would choose between them *upstream* of charging circuitry. Maybe they're right and they know something you and I don't, but you need to be in on that secret before you wire anything up. Edit: VVVVVVV babyeatingpsychpath, this came up earlier. The goal is to power the system while charging, so disconnecting the series cells isn't an option. Combine that with the relatively high solar voltage, which is slightly more natural to use for series charging, and it's not crazy to want to do it that way. KnifeWrench fucked around with this message at 01:46 on Mar 14, 2017 |
|
#
?
Mar 13, 2017 21:14
|
|
|
Zero VGS posted:Thank you, that makes a ton more sense. I have an entire 75kwh Tesla battery pack, which at a Tesla Supercharger station will take 120kw without breaking a sweat. The AC charger I'm getting is 6.6kw and the solar panels will put out at maximum 5kw (this is if I construct some fold-out canopies for them to triple the surface area and I'm in direct sunlight. The charger engineer said others have repurposed the Tesla packs to upgrade other EV's like I have, and the packs are so efficient that they don't even need cooling while charging until you get up to around a 20kw rate. Just to be clear, the Tesla pack is a bunch of 24V packs you connected in series, right? Why don't you have a contactor assembly that switches the packs into parallel for charging and back to series for powering your motor? Series-charging packs is always fraught with difficulty, but if the Tesla packs are smart enough, it may not matter. Also, the solar charge controller should be MPPT, so it'll output its maximum current at all times based on solar input, and you can measure its current and have your AC charger make up the "missing" current in constant-current mode. One charger isn't going to piggyback voltages to infinity; solar cells perform pretty terribly once they leave their maximum power spot.
|
|
#
?
Mar 13, 2017 23:46
|
|
|
babyeatingpsychopath posted:Just to be clear, the Tesla pack is a bunch of 24V packs you connected in series, right? Why don't you have a contactor assembly that switches the packs into parallel for charging and back to series for powering your motor? Series-charging packs is always fraught with difficulty, but if the Tesla packs are smart enough, it may not matter. If there's some kind of off-the-shelf contactor assembly that can switch 14 Tesla 24v packs between full series and full parallel safely, then by all means link me because that could be useful for a bunch of stuff.
|
|
#
?
Mar 14, 2017 02:34
|
|
|
Zero VGS posted:If there's some kind of off-the-shelf contactor assembly that can switch 14 Tesla 24v packs between full series and full parallel safely, then by all means link me because that could be useful for a bunch of stuff. What kind of currents are you looking at? The contactors don't have to be rated to switch full current since you're not going between series and parallel while the thing is running. You could even set up some manual switches if you were careful. It's the 14 packs that make this more challenging. edit: I know there are some Square-D contactors that let you set up individual poles to be NO or NC; you could get 2-pole contactors with one side set up NO and the other NC and use those. I'm pretty sure both sides are break-before-make, to avoid any interesting short circuits. babyeatingpsychopath fucked around with this message at 06:27 on Mar 14, 2017 |
|
#
?
Mar 14, 2017 06:20
|
|
|
Zero VGS posted:Thank you, that makes a ton more sense. I have an entire 75kwh Tesla battery pack, which at a Tesla Supercharger station will take 120kw without breaking a sweat. The AC charger I'm getting is 6.6kw and the solar panels will put out at maximum 5kw (this is if I construct some fold-out canopies for them to triple the surface area and I'm in direct sunlight. The charger engineer said others have repurposed the Tesla packs to upgrade other EV's like I have, and the packs are so efficient that they don't even need cooling while charging until you get up to around a 20kw rate. Battery charging capability is usually specified as a factor of the pack's total capacity per hour, in units abbreviated C (not to be confused with coulombs, or coulombs/sec which is amperes). For instance, a battery rated 1.0C charge rate can accept a current that will charge the battery's entire capacity in 1 hour. For a 5000mAh battery pack, that's 5 amps (+ a little for charge inefficiency); for your 75kWh pack, it's 75 kilowatts. You might know this already. A modern lithium cell can usually accept an average charge rate of 0.5-1.0C (charging from empty to full in 1-2 hours) without being damaged or requiring cooling. If you have a ton of batteries in a pack you may need some cooling where a single cell doesn't, because of the lack of airflow to the cells in the middle. Charging your 75kWh Tesla pack at 11.5kW is 0.15C, which is just above what's usually considered a trickle charge (0.1C or less). You could possibly even charge that pack in constant-voltage mode if that's all the current you're planning to put in -- but I wouldn't try it in my house.
|
|
#
?
Mar 14, 2017 07:29
|
|
|
What's the next level of connector up from plain Dupont for fairly small gauge wire, preferably available in relatively inexpensive bulk? I crimp Dupont currently, but it's fiddly and a little frustrating.
|
|
#
?
Mar 14, 2017 16:24
|
|
|
Stabby McDamage posted:What's the next level of connector up from plain Dupont for fairly small gauge wire, preferably available in relatively inexpensive bulk? I crimp Dupont currently, but it's fiddly and a little frustrating. Molex JST-XHP is pretty good, at least it has latches: https://www.amazon.com/Hilitchi-JST-XHP-housing-Connector-Adapter/dp/B015Y6JOUG/ It's just as fiddly and frustrating to crimp but once you're done that fucker stays in the housing and the connector stays connected. I also just got this box o' JST SW connectors that look cool but I haven't actually used them yet: https://www.amazon.com/Hilitchi-520Pcs-Housing-Terminals-Connector/dp/B0188DMF3A
|
|
#
?
Mar 14, 2017 16:35
|
|
|
I try using IDC connectors as much as I can, because I hate crimping thin gauge wire. IDC (TE AMP LATCH/ "FC" on Aliexpress) have the same pitch and footprint as the dupont connectors, don't need to be stripped and crimped, and have built in strain relief. They come in both latched and unlatched ("FC2") versions too.
|
|
#
?
Mar 14, 2017 17:17
|
|
|
Stabby McDamage posted:What's the next level of connector up from plain Dupont for fairly small gauge wire, preferably available in relatively inexpensive bulk? I crimp Dupont currently, but it's fiddly and a little frustrating. I know ate all the Oreos already said JST twice but there is a kind of connector commonly used in low power RC hobby stuff that is usually just called JST. I don't know if there is a series designation, but they are usually red and only 2pins. You can buy them pre-crimped as pigtails if that is easier for you. Like this: https://www.amazon.com/pairs-Plug-Connector-Battery-Female/dp/B0089RAXB2
|
|
#
?
Mar 14, 2017 17:45
|
|
|
Quick clarification if you're searching: Molex and JST are two different manufacturers. "Molex" sometimes gets tossed around as a style of connector, but it's a specific brand, and it can cause confusion to conflate the two.
|
|
#
?
Mar 14, 2017 17:58
|
|
|
KnifeWrench posted:Quick clarification if you're searching: Molex and JST are two different manufacturers. "Molex" sometimes gets tossed around as a style of connector, but it's a specific brand, and it can cause confusion to conflate the two. Yeah sorry I just know them as "Dupont-style," "Molex-style" and "that thing that comes on RGB LED tape" style 
|
|
#
?
Mar 14, 2017 18:30
|
|
|
Sagebrush posted:Charging your 75kWh Tesla pack at 11.5kW is 0.15C, which is just above what's usually considered a trickle charge (0.1C or less). You could possibly even charge that pack in constant-voltage mode if that's all the current you're planning to put in -- but I wouldn't try it in my house. Constant voltage meaning I can just tell both chargers to jam 348 volts no matter what, and put in whatever amperage they can supply? I assume chaerging gets very slow as it nears 100% due to the lack of potential difference? It seems like that's the "safest" option, to set both chargers for a fixed 348v, and let my BMS keep an eye on the temps/balance just in case, since I can only hit 0.15C in a very specific scenario (plugged in at pubilc charging station with all solar panels deployed with direct sunlight at optimum angle) and would be at trickle-charge rate the rest of the time. This is being used in an EV (with a sub-frame fiberglass battery box) and not a house, plus all the Tesla modules have per-cell fusing, so it seems pretty safe. Tesla fires are scrutinized and I believe have only happened from road debris impact, crash, or one time from Supercharging. With regards to the solar charger, the more I'm reading up on MPPT solar chargers, it seems like you use one with a PV array that has a voltage higher than the battery, and the MPPT just converts excess volts into amps to make the most efficient use of the solar panels... is there any reason I couldn't instead use a PV array with a voltage slightly lower than the battery, and use a boost regulator to lower the amps as needed to bring the output to a locked 348v? It is pretty much impossible to find an MPPT that goes above 48vdc, since RVs/Boats only have accessories that go up to 48vdc (to keep the voltage "touch safe") and homes use AC inverters.
|
|
#
?
Mar 14, 2017 18:36
|
|
|
Note: I am not an electrical engineer or an electrician so I would strongly suggest you get a professional's opinion on this before you try anything. I believe that if you have a constant-voltage supply of infinite current capacity, and you put 348v across the discharged battery, it will draw as much current as it can according to the battery pack's internal resistance. That would probably be thousands of amps and the pack would explode. In your case, you know that your charger can only provide 11.5kW max, which at 348v is ~33A. No matter the voltage differential, you aren't going to be putting more than that much current into the pack, and as the differential decreases, the current will drop (though not until the pack is close to fully charged, because of the very low internal resistance). If your pack can handle 33A input at any state-of-charge without cooling, I believe that theoretically you can just use a CV charger, with the current limit set by what the charger can physically produce. If you burn down your car, feel free to tell the fire department that a guy on the internet told you how to wire it up. e: btw you should get in the habit of saying "current" instead of "amperage" so that electricians don't make fun of you Sagebrush fucked around with this message at 20:58 on Mar 14, 2017 |
|
#
?
Mar 14, 2017 20:54
|
|
|
Knowing the power output doesn't necesarily mean you know its limit behavior. It might trip and poo poo down rather than actively current limiting. It certainly might but limit but it completely depends on the source. MPPT specifically means that it continuously optimizes to draw as much watts from the solar panels as possible (which have a non-linear and variable VI curve). As you say, that typically means stepping down, but the MPPT part could be done under any scenario. You could certainly use a boost converter in theory. If you series enough panels to get near 90V you'll be in "PFC" territory where you'd find a ton of options, ICs, app notes and demo boards for 90 to ~370V boost converters. Google PFC boost.
|
|
#
?
Mar 14, 2017 22:53
|
|
|
Zero VGS posted:Constant voltage meaning I can just tell both chargers to jam 348 volts no matter what, and put in whatever amperage they can supply? I assume chaerging gets very slow as it nears 100% due to the lack of potential difference? It seems like that's the "safest" option, to set both chargers for a fixed 348v, and let my BMS keep an eye on the temps/balance just in case, since I can only hit 0.15C in a very specific scenario (plugged in at pubilc charging station with all solar panels deployed with direct sunlight at optimum angle) and would be at trickle-charge rate the rest of the time. You need to figure out how to actually charge a single Tesla pack. Lithium technology is not as simple as "voltage in = battery charge." Unless there's some intelligence in the pack itself that makes it that easy, you are going to have to do some work. Lead-acid batteries can take so much abuse that, if you were using those, I'd say just go with whatever. Even NiCd is about that easy apart from an every-six-months-or-so cell balancing charge. But you can permanently ruin a lithium battery by overcharging or overdischarging it exactly once. It's also absolutely possible to turn that battery into a large fire-producing device that cannot be extinguished by water by treating it like a lead-acid or NiCd. So, how do you charge a single Tesla battery pack? Does the pack have internal battery management circuitry to prevent over-rate charging? over-voltage charging? Undervoltage conditions? Overcurrent discharge? Overtemp? If the answer to all five of these questions is "yes," then treat the thing like a Lead-acid pack, put 350V across it all the time, and let your battery chargers figure everything else out. If the answer to ANY of the questions is "no," then you need to do homework.
|
|
#
?
Mar 14, 2017 23:18
|
|
|
babyeatingpsychopath posted:So, how do you charge a single Tesla battery pack? Does the pack have internal battery management circuitry to prevent over-rate charging? over-voltage charging? Undervoltage conditions? Overcurrent discharge? Overtemp? If the answer to all five of these questions is "yes," then treat the thing like a Lead-acid pack, put 350V across it all the time, and let your battery chargers figure everything else out. If the answer to ANY of the questions is "no," then you need to do homework. I have that covered, the Tesla 24v packs are wired as 6 series, and something like 100 parallel. Each pack also has a 7-pin JST cable to hook into it's own slave BMS (the slave BMS won't work unless in a Tesla talking to the main system, which no one has hacked yet). Since I have 14 of the 24v modules, that's 84 in series altogether, so I got an 84-series battery management system on the way: http://www.orionbms.com/ The Orion BMS measures pack state of charge, balances out all series cells, has thermistor probes to check pack temps, ammeter to check draw from load, and it can communicate with contactors to disable either charger if they exceed the parameters I program into it (has a nice Windows app to set parameters, which you then flash over CAN bus adapter). It actually directly commands the AC charger over CAN bus to tell it exactly how to behave, but there's nothing like that to control a solar charger so that's why I'm on my own there. asdf32 posted:You could certainly use a boost converter in theory. If you series enough panels to get near 90V you'll be in "PFC" territory where you'd find a ton of options, ICs, app notes and demo boards for 90 to ~370V boost converters. Google PFC boost. It seems like PFC only applies to AC-DC conversions? I think I just need to find a very high wattage DC boost converter, but I can't seem to find anything near 5kw. Sagebrush posted:Note: I am not an electrical engineer or an electrician so I would strongly suggest you get a professional's opinion on this before you try anything. I appreciate that y'all reiterate this every couple pages. Yes, I will get multiple licensed professionals overseeing this when I put everything together; right now I'm still shopping and designing the overall system.
|
|
#
?
Mar 15, 2017 00:35
|
|
|
rawrr posted:I try using IDC connectors as much as I can, because I hate crimping thin gauge wire. IDC (TE AMP LATCH/ "FC" on Aliexpress) have the same pitch and footprint as the dupont connectors, don't need to be stripped and crimped, and have built in strain relief. They come in both latched and unlatched ("FC2") versions too. That sounds absolutely perfect, but could you link to a specific example? TE AMP is a whole giant line of like 200,000 different connnectors, and I didn't find much conclusive looking for FC or FC2 on ebay, aliexpress, or digikey.
|
|
#
?
Mar 15, 2017 01:05
|
|
|
https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20170314172221&SearchText=fc+idc
|
|
#
?
Mar 15, 2017 02:23
|
|
|
Zero VGS posted:I appreciate that y'all reiterate this every couple pages. Yes, I will get multiple licensed professionals overseeing this when I put everything together; right now I'm still shopping and designing the overall system. I just keep reiterating it because "how can I build a homebrew charging system for my 350-volt 75-kilowatt battery pack that can push probably 2000 amps no problem and uses technology that is known to occasionally catastrophically explode" is about the most dangerous electronics question we've had in this thread so far. I'd have the same hesitation if you were asking about how to make your own computer-controlled gas heating system, or how to cast your own brake rotors for your car, or something like that. You're working with stuff where an error could kill you in several different ways.
|
|
#
?
Mar 15, 2017 03:42
|
|
|
Sagebrush posted:about the most dangerous electronics question we've had in this thread so far. I don't know, I'm personally a fan of the guy who wanted to build a thing for shocking his junk. I mean sure it wouldn't be able to kill more people than just him but it gets points for creativity 
|
|
#
?
Mar 15, 2017 04:00
|
|
|
Zero VGS posted:Constant voltage meaning I can just tell both chargers to jam 348 volts no matter what, and put in whatever amperage they can supply? I assume chaerging gets very slow as it nears 100% due to the lack of potential difference? It seems like that's the "safest" option, to set both chargers for a fixed 348v, and let my BMS keep an eye on the temps/balance just in case, since I can only hit 0.15C in a very specific scenario (plugged in at pubilc charging station with all solar panels deployed with direct sunlight at optimum angle) and would be at trickle-charge rate the rest of the time. No. Whenever you connect 2 voltage sources together in parallel, they fight. One wins, takes all the current, and the other looses, and effectively acts as another load on the circuit. If you're unlucky your load is enough to make either source alone fail in which case the first voltage source fails, then the other one fails. This is because voltage sources have tolerance. For example, just because you have it set to 348, the output it could be, 347.5 - 348.4v with 1 figure of rounding. (which is still better than 1% accuracy here. It wouldn't surprise me to see a tolerance of 5% here, which would be +/- 12.4v, or a potential difference of 24v between the 2) So when one source attempts to regulate to a higher voltage, it will dump current into the circuit to attempt to bring the voltage up. The lower one will attempt to sink current (badly, because voltage sources are pretty much never designed for this) and basically won't supply any at all. There are devices designed to be paralleled. They use feedback and/or external control to enable the sharing of current.
|
|
#
?
Mar 15, 2017 04:25
|
|
|
Zero VGS posted:Constant voltage meaning I can just tell both chargers to jam 348 volts no matter what, and put in whatever amperage they can supply? I assume chaerging gets very slow as it nears 100% due to the lack of potential difference? It seems like that's the "safest" option, to set both chargers for a fixed 348v, and let my BMS keep an eye on the temps/balance just in case, since I can only hit 0.15C in a very specific scenario (plugged in at pubilc charging station with all solar panels deployed with direct sunlight at optimum angle) and would be at trickle-charge rate the rest of the time. Double posting for emphasis that the answer is no, you can't parallel like that. What constant voltage means is that the charger attempts to keep the voltage constant. Imagine for a moment a 0-1v rechargeable battery. If you connect a 1v voltage source when it's fully charged no current will flow. If it was discharged to .9v you'd have current proportional to whatever its internal resistance is. If it's discharged to .1v it'd draw much more current. Makes since right? Now, what if that battery can only take so much current. If it was discharged enough the battery would draw enough current to damage itself. In the case you'd need to current limit the current, this is done by reducing the voltage, and since the voltage is no longer constant, it isn't constant voltage. Since you typically limit the current to some constant maximum value, until the voltage rises high enough that its the peak for the battery, and then stays there as a constant, this is known as CCCV charging (named for the Constant Current Constant Voltage mode change of course.) If you never discharge enough to over current your battery on whatever your charging voltage is you can just use a constant voltage, and skip the smarts that make constant current possible. Your battery chemistry also needs to be able to take a constant voltage when it's fully charged (lead acid can, lithium ion can't. dealing with this problem is one of the things that the protection circuitry is for)
|
|
#
?
Mar 15, 2017 04:44
|
|
|
Zero VGS posted:
Power factor correction applies to AC but the basic boost topology is after a rectifier meaning it's DC to DC. The PFC part is irrelevant but will provide the most common example of boost converters in this general voltage/power territory. Many or most would run off DC as-is, others would require minor modifications (like removing the rectifier).
|
|
#
?
Mar 15, 2017 05:14
|
|
|
hey I'm trying to save a bit of money by making my own defibrillator, any tips?
|
|
#
?
Mar 15, 2017 08:43
|
|
|
Cumslut1895 posted:hey I'm trying to save a bit of money by making my own defibrillator, any tips? All you need is a second-hand Tesla battery, boost converter and solar charger. Job's a good'un Here's the build video: https://www.youtube.com/watch?v=55dTFEStZVc
|
|
#
?
Mar 15, 2017 13:03
|
|
|
Is this the best place to asked about DIY speaker builds/repairs? I picked up a pair of towers speakers for cheap at an estate sale and I think the tweeters are blown: no continuity on a multimeter and no response from a AA battery tapped to the terminals. The more I learn about speakers the more questions I'm coming up with. I'd like to get these working for cheap, knowing that they probably won't ever sound great since I'm not an audiophile. They're tower speakers with one tweeter, one mid and two woofers each. There's no identifying information on the cabinet whatsoever. The tweeters are Motorola but don't state an impedance. After googling the number on the back I found someone claiming they were KSN 1036, and the data sheet I found for the KSN 1036A has a picture which looks correct. Response range is 3 kHz to 40 kHz and sensitivity is 90 dB. The hole is 2.385" and the tweeter outer diameter is 3.75". All the tweeters I've looked at on Amazon are either too big or too small to fit. I can make the hole bigger if needed though. The mids say D12EC65-01F which leads me to Pioneer NM-1201 but I can't find any specs on them other than the 8 ohm on the back. The woofers say C30EU80-51F along with 8 ohm, Taiwan, R. O. C. and 81 8142. I can't find anything on these at all. There doesn't appear to be any real kind of crossover except for a capacitor on the mid's positive terminal. The cap says CHIUS / 10 MFD / 50 WV / NP-820 and appears to be unpolarized (one lead on each end instead of two on the same end, and no arrow printed on the outside). The speakers appear to all be wired in parallel: the positive terminal on the back goes to both woofers positives, and then 2 more wires go from one woofer's positive to the tweeter and mid positives. Same with the negatives. What replacement tweeters should I get and how do I wire these up to get speakers that work with an 8 ohm amp? I'm not afraid to solder some resistors and capacitors if it helps the quality a lot but I'd prefer a cheap and simple fix.
|
|
#
?
Mar 15, 2017 23:43
|
|
|
Cpt.Wacky posted:Is this the best place to asked about DIY speaker builds/repairs? There's a car audio thread in AI if you want cross post there too. edit: oh, didn't know pioneer made home audio
|
|
#
?
Mar 17, 2017 02:01
|
|
|
There used to be a speaker building thread in DIY that was really awesome, but is now archived. You could start another and there might be some interest. You might want to post in the home audio thread or the audio questions thread.
|
|
#
?
Mar 17, 2017 17:34
|
|
|
Man I have to check this thread more often...asdf32 posted:Akbar do you have experience with SIC or Gan personally? Most of my work revolves around using EPC eGaN FETs. I'm quite sure this is what TI's LMG5200 has inside it. The LMG3410 is definitely something else, possibly Transphorm devices. The cascode devices are more user friendly since they can handle higher gate voltage overdrive. But if you can make do with <200V devices then the EPC FETs are pretty much unbeatable right now. But getting their best performance requires good layout (at least 4 layers, <8mil trace/space).
|
|
#
?
Mar 18, 2017 12:13
|
|
|
ANIME AKBAR posted:Man I have to check this thread more often... The LMG3410 has a series silicon enhancement device for normally off behavior but that device is turned on just once, and then depletion GaNs only are switched at high frequency. So it's an improvement over the Transphorm hybrid devices because the silicon fets themselves don't need to switch. The TI solution is obviously interesting because it takes care of the gate drive. What are you using for gate drive for the EPC devices? And what do you mean about the layers specifically? For properly routing the gate drive or for dealing with high frequency in general? I'm being asked to design a high voltage (~400V out) low power (~1A) inverter eventually that needs to be very small so I'll be taking another look at this stuff sometime. Right now, even in a higher current variant with larger fets the isolated gate drive takes up about as much board area as the fets themselves. So while the EPC devices are extremely small there are diminishing returns using them unless the gate drive solution can also be sized down. The TI devices are very attractive there though there would be risk in designing around this new special chip. Cumslut1895 posted:hey I'm trying to save a bit of money by making my own defibrillator, any tips? My feeling is that the guy isn't doing anything illegal and you have to ask stupid questions to get started somewhere. asdf32 fucked around with this message at 14:57 on Mar 18, 2017 |
|
#
?
Mar 18, 2017 14:49
|
|
|
asdf32 posted:My feeling is that the guy isn't doing anything illegal and you have to ask stupid questions to get started somewhere. Pretty much, I mean I'm building the world's first 100% electric RV, and practicality dictates that it can charge from an EV plug or rooftop solar panel array... let's just say there's not exactly a playbook for this kind of thing. I've built electric bikes from the ground up, it's the same thing here except all the numbers are bigger. Bigger size, bigger weight, bigger voltages. I do plan to live to enjoy the fruits of my labor, I'm taking the safety very seriously. Remember that in the late 90's and early 2000's, you couldn't walk into a Tesla dealer and buy an EV, the only option was doing an electric conversion yourself, and there are thousands of success stories on EVAlbum.com But now with plenty of competent OEM options, people don't need to do conversions anymore, they can buy stuff ready-made better and cheaper. Except for RVs... I can make it work but I have to start with a blank slate and design for the highest efficiencies I can because there's the extra challenge of powering the climate control and fighting the aerodynamic drag. I did just order a 6.6kw UHF charger that is completely potted, it only came out in the last couple months. 93% efficient AC-DC conversion at 100% load, what a time to be alive.
|
|
#
?
Mar 19, 2017 02:31
|
|
|
taqueso posted:There used to be a speaker building thread in DIY that was really awesome, but is now archived. You could start another and there might be some interest. After getting into electronics and trying to find the areas I like best, I wanted to do something with speakers but found it a pretty vertical slope after learning how speakers work and building the bare bones simplest stuff. Anything I looked at online seemed caked in audiophile stuff, when all I want is to learn more by having a project to make something probably not that great but probably good enough to begin with. Does that thread help with that? Or is it more for dedicated, knowledgeable stuff?
|
|
#
?
Mar 19, 2017 12:32
|
|
|
asdf32 posted:What are you using for gate drive for the EPC devices? quote:And what do you mean about the layers specifically? For properly routing the gate drive or for dealing with high frequency in general? quote:I'm being asked to design a high voltage (~400V out) low power (~1A) inverter eventually that needs to be very small so I'll be taking another look at this stuff sometime. Right now, even in a higher current variant with larger fets the isolated gate drive takes up about as much board area as the fets themselves. So while the EPC devices are extremely small there are diminishing returns using them unless the gate drive solution can also be sized down. The TI devices are very attractive there though there would be risk in designing around this new special chip. Rolling your own isolated gate drive is definitely a pain in the rear end, especially after considering you'll need to adjust the dead times to within a few nanoseconds to get optimal results. At one point I was toying with the idea of taking on the littlebox challenge and I was planning on using SI8610BC isolators with UCC27611 gate drivers for EPC2010C FETs in a four level inverter. I think I was able to get all of the power stage, including isolators and gate drivers, in a 2"x1.5" area. Never got past the drawing board though. ANIME AKBAR fucked around with this message at 13:09 on Mar 19, 2017 |
|
#
?
Mar 19, 2017 13:04
|
|
|
Zero VGS posted:Pretty much, I mean I'm building the world's first 100% electric RV, and practicality dictates that it can charge from an EV plug or rooftop solar panel array... let's just say there's not exactly a playbook for this kind of thing. Have you worked out the math on how many batteries you will need to move this anywhere (including the energy needed to move the batteries)? There's a reason that EVs are tiny, and I'm pretty sure that reason is that distance-per-charge falls off dramatically as weight increases.
|
|
#
?
Mar 19, 2017 15:36
|
|
|
ANIME AKBAR posted:For hard switched stuff I'm using the LM5113, but that only works up to 100V. I'm nearly 100% certain the gate drive inside the LMG5200 is a LM5113 chip. Yeah got it. 2x1.5 is still huge compared to the devices themselves and in the context of that challenge. I'm using SILabs too, their integrated isolated gate drivers like Si826X. I don't think that exists in a form tailored for GaN yet. And what are you using for control out of curiosity? Digital or analog control IC's? I'll be doing all the control in the FPGA which is work but means I know I can deal with dead time etc (though with finite timing resolution). Yeah I'm having a hard time seeing how EPC fits into my current requirements but I'm still staring at the Transphorm stuff which are basically drop in replacements for regular devices up to 600V. We have to do a couple PFC designs and they have some interesting "Totem Pole" PFC demo boards.
|
|
#
?
Mar 19, 2017 16:07
|
|
|
Corla Plankun posted:Have you worked out the math on how many batteries you will need to move this anywhere (including the energy needed to move the batteries)? There's a reason that EVs are tiny, and I'm pretty sure that reason is that distance-per-charge falls off dramatically as weight increases. I'm using the battery pack out of a Tesla 75D, which has an EPA range of 256 miles and a curb weight of 4600+ pounds. My RV is basically a Ford Ranger with a U-Haul style box as a cabin, and will actually have a lighter curb weight of about 3900 pounds. The problem is that aerodynamics matter more then weight, and this has a lot of frontal area and bad drag, but keeping driving speeds low is a cheap way to mitigate that.
|
|
#
?
Mar 19, 2017 16:08
|
|
|
Fat Turkey posted:After getting into electronics and trying to find the areas I like best, I wanted to do something with speakers but found it a pretty vertical slope after learning how speakers work and building the bare bones simplest stuff. Anything I looked at online seemed caked in audiophile stuff, when all I want is to learn more by having a project to make something probably not that great but probably good enough to begin with. The car audio thread might be helpful since they sometimes build special speaker enclosures for their boom cars. This site is the best I've found so far at explaining the concepts and process of building your own stuff instead of following someone else's recipe for a specific design: https://www.diyaudioandvideo.com I think I'm going to write off my repair project. It's too hard to find an appropriate tweeter and the foam gaskets on the woofers are falling apart the more I handle them. I may keep the cabinets around for an eventual complete rebuild someday though.
|
|
#
?
Mar 19, 2017 17:07
|
|
|
|
| # ? Jun 3, 2024 04:10 |
|
|
I believe you can 're-foam' the woofers but it's just something I've read about in passing.
|
|
#
?
Mar 19, 2017 18:01
|
|