I see there one big problem: You did all the switching and sensing on the GND side. That means as it works (the current sensing, current limitting, load switching,...), it will make voltage differences between these grounds. But this GND is common with the USB communication, connected inside of the TV. That means the "fire stick supply output" GND is shorted out with the "power input" GND, once it gets connected together into one TV. You should just have made all the current sensing and regulation on the positive side (so using PMOS devices,...), keeping the one single common GND.
I agree. I am surprised it worked at all when he used it on his actual TV. My only explanation for that is that his TV's port already has soft current limiting, meaning he could have gotten away with just connecting the capacitors directly to the USB port power.
This is exactly why I wrote down "not tested for longer time" on the disadvantage side. Now longer testing would not have revealed this problem because my TV works with my solution. But if I would have other TVs to test I think I would have found this problem sooner. It seems like I ignored that the USB GND and HDMI GND are apparently most of the time connected together. In theory that should mean that charging the supercaps up and adding the FireTV stick should work just fine. But afterwards the lower MOSFET would get shorted which means we remove the current limiter. Since there are still the supercaps in parallel to the Fire Stick that should not be such a big deal breaker. But yes. I think this is a problem in my circuit which I didn't find with lab bench power supply testing and with my TV testing. This is all theory in my head right now by the way. So the problem can be bigger or smaller. Switching the positive votlage side with P-Channel MOSFETs would solve the problem. That would require a bit of a redesign but the main components should stay the same. Sorry about that.
@@greatscottlab Hey I'm sorry but I didn't exactly get what the above commenter explained. Could you perhaps elaborate in a yt short or a community post or even here.
I'd added 2 LEDs to show is in charging and when ready, also a protection diode to get polarity protection and * feedback current from supercaps to tv. *Edit: avoid
It still amazes me that we have devices like this today that are so small and are basically full fledged mini computers that can stream 4K video while using only around 200ma of power at 5V!
Hardware decoders are amazing, you can do the same video reproduction task with either gobs of processing power, or a small fraction of it with the correct hardware. I think even plain MPEG 2 video at those resolutions would put a strain on a processor with that little power consumption if they didn't have the hardware decoders, then again back at your point, it's still really impressive how said decoder can do such a feat with so little power.
Whether or not it was 4k isn't the point. I'm guessing the poster is around my age. I'm 53 and remember being in computer classes where a GB was theoretical. No one believed at the time something could be so large. I remember being hot shit and paying $400 for a single 4MB SIMM of RAM. The fact technology has progressed this far is the point he/she was making. Shit, I remember a time when there was no internet.
Important! Pls add two similiar high ohmic resistors in parallell to supercaps!! This will help them with staying balanced! Otherwise you run risk with age that one blows up due to over voltage. Capacity may vary as they age = they will come more and more unbalanced! I approve of protype development process. I do the same.
@@Kirillissimus never heard of these. Max I saw was a 3v supercap. I believe 5v is just 2 in series hidden within. But who knows maybe they really produced 5v supercaps
Not a great method. Bottom balancing on ELDCs can cause major issues because they can actually flip polarity when at 0 volts. It is imperative to never drain them down fully. Doing this can do internal and external damage.
@@EVLS10 Not sure what you mean with bottom balancing as parallell resistors will always try to maintain 1/2 x Uin over each capacitor at all times. Complete discharge is unavoidable due to current circuit design. I do agree that parallell resistors is the bottom tier solution and cannot keep up in very dynamic situations. This design is usage is pretty chill though, where the caps will at 99.99% of time just stay hovering at 5V without much going in or out giving much time for the resistors to do its job. In the case of rapid charge or discharge; parallell resistors are much much better than nothing, as it will correct errors with enough time, as opposed to not being corrected at all. During a full rapid discharge, with unbalanced cap size there is the possibility that it will force the lower capacity cap into reverse voltage, with discharge resistors it will atleast try to minimize the time spent in error state. I can't forsee that as worse than having the cap forced into reverse voltage for a longer duration so; I'm not sure what the fuzz is about but I'm intrested to know. Adding parallell diodes would limit the reverse voltage situation, clamping it to one diode forward drop. (But if one has gotten this deep into patching, one might aswell get a proper solution like SAB mosfet, op-amp balancer or maybe a well thought out zenerdiode)
@@Paxmax it means that your method will intentionally or not bring both cells to 0 volts after a long time of being disconnected. You are providing a constant drainage path which is a no no with eldcs
Nice to know supercaps have been improved a lot. They didn't use to have any significant energy capacity. They were just for maintaining a voltage on something like static RAM, etc.
Might be better if the output voltage cutoff had some hysteresis, so it gets to a decent level, but will then allow it to drop to the minimum working level, otherwise it may cut off prematurely as the capacitor discharges. Could also use a boost converter so that the voltage does not fade with capacitor discharging. Finally, it appears that the switching and any drop is in the 0V side, not the +V, this would be bad if the USB ground is not isolated from the HDMI ground
The main ic is a monolithic power MP2617H. worht checking the datasheet, really cool ic, it does everything, charges the lithium cell, regulates the maximum input current and provides higher output current (from the battery) than the input current.
Cool idea. Though I have couple of ideas to consider. Since you are charging the 2 capacitors to a combined higher voltage that they can tolerate by themselves, I would put in some protection for too high "cell" voltage. Supercapacitors (at least in the range of 100s of Fs) can be easily be damaged with too high voltage. Also the op. Amp used to turn the load on does not have any feedback to prevent oscilation when reference voltage is near the compared voltage. Just a small amount of feedback to the compared voltage would made the operation to work a bit like a schmitt trigger.
This is not regarding any video in particular but all in a whole, THANK YOU for your content I love watching them even though I sadly cant afford to do like 99% of any of it I find them very entertaining and knowledgeable I really love your style of explaining how it works or why it works that way the most Thank You again
I think I'd rather power with an external supply over this solution tbh. In the past I've used a relay to turn off the device when I've required a synced power down.
exactly my thoughts, you could tap from the mains inside the tv to connect a seperate 5v powersupply that offers enough amps, add a usb port to the tv and connect a relay in between.
I didn't watch the video closely, but it seems he made a contraption instead of simply buying a USB power supply? They're only $5-$10 each, for one that does 2 amps. I don't see the point of making something like this when they're cheap and available. I did not pay attention to the video, though.
He used Super Capacitors and not a Li-ION Battery. It's job is to only deliver required short current spikes while being powered by current limited input and not to be used as UPS I think the Buy Version is capable of that stuff.
Interesting design, I thing you should probably add a voltage balancing circuit of some kind for the series super-cap, dispersion in capacity may result in premature deterioration.
or at very least a 100k-1M parallel drain resistor on each so the higher cap will drain faster to some safe equilibrium while not wasting a tonne of power.
Kudos for beating a major manufacturer with better design! For me this is way above my capabilities to make, I would just hook up a phone charger to the 220v TV cable and use it to supply voltage to the Fire Stick.
Excellent rendition. You say the circuit is simple, but there is a lot going on. You truly have big brain, Scott! Thanks for sharing! I really enjoy your videos. They are great supplemental material for my EE degree. Take care you wizard!
If cost is a thing I would simply use an old usb phone charger, they are >500mw. What would be handy, like a sort of PC PSU, offering different voltages, for for use outside the pc. Near my tv I have a router, nvidia shield, DAB+ radio unit. amplifier, raspberry pi. Most are 5v, some 12v. Every device has its own power supply. Why not have 1 power supply, supplying power to all devices. This is probably more efficient too (less watts). And as a bonus, being able to turn on/off per device.
I also thought using a lipo was not a great idea. Trickle charging like this would be good for NiMH cells too. I do like the super cap solution too. Ok further watching and this is awesome. I need to make a few of these.
I've seen some high end charge management chips for Li-Ion charge the battery to only 80% capacity - e.g. 4volts instead of 4.2volts, this doubles or triples the life span of the cell, other ways of doing this too is to allow the battery to fall to 80% from fully charged, doing a smaller charge back up to 4.20volts, meaning the cell is never heard at fully charged.
Nice solution! I was wondering if it would be sensible to limit the input current draw to 450mA, so that the TV's usb port isn't being driven so hard/close to its max mA?
@ True, but as GS mentioned, not all are capable of exceeding their ratings - in this case USB1.1 base PD. It’s worth noting that it’s fairly common to have 500~700mA (nominal) PTC fuses on USB (1.1-2) ports. Being a fuse, the PTC does not _actively_ limit the current (at least not very rapidly ;) ). Instead, after a period of time - say 500-2500ms (depending on several factors) - it’s resistance will become extremely high, cutting power. As such, as you mentioned, it’s not at all uncommon to see ports that will handle 500-1000mA spikes without a hitch, but completely fail when they are asked to deliver that same current continuously for even a second or two. It’s also important to note that the recovery time for PTC fuses and very dramatically, from milliseconds to hours. Both factors combine to make the device featured in the video necessary, in the absence of an external power source.
Personally, this is one of those problems that the DIY is less of a value proposition than just trying to find an outlet to plug the stick into. Even if its cheaper than the commercial product, so is a USB charger I already had. But admittedly I'm WAY lazier than you :P Great video overall though!
@@JeanRobLast Ironically this is the exact reason I gave up on running my Chromecast off the TV port (my TV has a 1 amp port which should be enough but still caused it issues from time to time). Specifically, the Chromecast supports HDMI CEC, which lets me turn the entire set up on by just telling Google Assistant to boot up my Chromecast or just directly streaming to it. The Chromecast then automatically turns the TV on, and can even turn it off afterwards. Presumably the Fire Stick can do this too. The result of that is you can ignore the TV remote entirely and everything works off of a single control. If you use the TV turning on to turn on the Chromecast/Fire Stick you still then need to switch over to whatever is controlling the other device.
@@clancywiggum3198 yes the fire tv does that too, also it doesn't need to boot every time and it can update when not in use. Power consumption when asleep is negligible
@@clancywiggum3198 Good to know, i never explored that possiblity
2 ปีที่แล้ว +2
I may have done an overkill then. I opened up the tv and the sick, put a dc plug inside after the fuse, added the included charger inside in a corner, added a heatsink to the stick and changed its wifi antenna for a broken laptop one, the hdmi cable goes out from a small carved hole and plugs to the tv, so it has better wifi, can run colder and I have only one cable visible from the tv.
Hmm, now this makes me want to try a similar solution for my 3D printer. I've got a raspberry pi hooked up directly to the printer's PSU, so I have to make sure to shut down the pi before I turn off the power (if just for good practice's sake). Would love to read the state of the PSU and shut down the pi automatically, but obviously that would mean it'd have to be on for just slightly longer than the rest of the system.
@@jerrb7991 Yeah, I've been thinking about that too, same with just buying a smart switch or something, but those draw power while they're off (of course, so you can remotely turn them on again, but which I don't necessarily need) and I'm not confident enough to work in a relay on 220V mains. So it's not that your idea doesn't work (or even might be better than mine in general purpose), but I've kinda backed myself into a wall with my requirements ¯\_(ツ)_/¯ Still, appreciate the idea, thanks 👍
@@LordHonkInc For what it's worth the idle power draw of Zigbee stuff tends to be really low (my IKEA Zigbee unit uses less than 0.1 watts for instance)
I literally went to setup my mums stick on Sunday, and got the "low power" message haha. Great alternative. Is that power adapter advertised as having a li-ion battery, do customers know that it's not going to last?
Yea this would make the battery solution better. Maybe a supercap+battery hybrid would fix this issue. Using USB C would be the way to go to have data and power all travelling through the same port. You could even forget about HDMI this way, and USB C wouldnt have the top current limit
I could see a potential issue with the circuit. You are using low side switching on FireStick GND, and this works fine on bench power supply. But when connected to TV, the GND of usb output is tied to GND of hdmi port, and thus the low side fet switch is bypassed. Now the GND FireTVStick goes straight to GND USB.
@@greatscottlab I would like to see actual voltage and current measurements when connected to a real TV. My suspicion is the TV's USB port is soft limiting the current and your circuitry is being bypassed.
While I don't really have a use for a Fire Stick or similar, I think Thist is a very interesting idea for averaging current draw in many applications - I can imagine it would be a great way to power remotely installed solar powered devices that have an occasional need to draw extra power (to operate a motor or servo etc). I'll be bookmarking this :)
Excellent solution Scott. There are a number of circuits that supercapacitors are better but manufacturers always jump to lithium ion batteries which increases prices and decrease life.
I just ran into my own case of battery VS capacitor. The raid card I was trying to use required a battery backed ram cache to enable RAID 5. While searching for one I realized that they also made flash backed ram cache modules. These use a capacitor to dump the ram to flash when the power cuts out. The capacitor is rated for 80 seconds of use, while the identically sized battery is rated for 3 days. However the battery is only rated for 3 days for the first 3 years, and if the power is out for more than 3 days I'm stuck. With the FBRC I can last indefinitely without power and the capacitor can last longer than 3 year.
In my line of work as an AV Integrator I have to deal with these blasted firesticks. I have seen newer TVs come with 1A rated USB ports that is great so it can be powered from the TV. It's kinda hard to mount a TV and have a USB wallwart sticking out three to four inches from the electrical outlet, doesn't let the TV lay flush and screws up stuff. Never know what equipment the customer has until I show up. Would be nice if the HDMI 2.1 spec could use that +5V pin with around 800mA to 1A of current so stuff like the firestick and chrome cast can be powered from the HDMI port itself.
I'd love a PCB for this. Something I could put in a 3D Printed case and mount on my TV. Maybe there's a way to use this for multiple USB Devices? I have a PS1 Classic running off USB now too I'd also love to power via the TV safely.
I remember there was an app that would allow you to create the pcb from the schema. It also had an option to optimize the size of the board and to create single or double layered pcbs as well.
I wonder if it would be better using only parallel capacitors at a lower voltage and a buck converter for the output more akin to the way power-banks work. No balancing issues and the charge voltage is not so close to the USB input voltage.
Great Scott, Great Edutainment! I’m curious about when you build your beta circuit in free air. Do you prefer to build this way because it allows you to understand the circuit better? Do you prefer free air vs breadboard? Do you prefer free air because of convenience, easy vs hard, fast vs time consuming? Some videos of free air circuit construction can be more entertaining than the actual circuit application information, personally I prefer deeper material like yours vs the “Sensory Satisfaction Art”, but a balance is a good thing. I ask you the free air construction question because if there is real value to it, vs breadboarding, I might be willing to give it a try. Thanks 🙂
"never above at any time" at 03:52 while it shows more than 800mA ;-) btw at 09:18 the IRL540 seems to be connected in strange way: drain to the ground?
Fantastic! I didn't know about that thing! 😃 I was waiting to buy 2 new smart TVs... Now I'm getting 2 of those! 😂 Really great work as usual, dude! Stay safe and creative there! 🖖😊
It might be overkill, but I'd love to see a version of this circuit that uses a buck converter to charge the supercap (so you aren't burning half the energy when you initially charge the caps) and a boost converter to make the 5V output (so you can drain the caps lower than 4.5V and use more of their energy storage capacity). You'd be able to start the Fire TV before the capacitor is fully charged with that sort of design.
Is there no problem switching the GND of the Fire TV instead of the +5V line? In my opinien this will not work, because the HDMI GND will short the switching mosfet for the FireTV.
since you have that ambient light (the ws2812b led strip on the back) setup hooked to your tv, you probably have a raspberry pi as well, in your case why not use power from the pi's supply, it would atleast be of 2-3a
@@kentahirono but thats when you have a pc, its tv so you need to use a pi, also ya Hyperion can also use wled as a ambient light, but all of this dosent change the fact that there are ton of ws2812b leds so there must be a beefy power supply
I have build a circuit like this one. what you really looking for is a HOT SWAP CONTROLLER. you can put 2 of them and chain the Pgood and Enable pins together to separate the super caps. you might also just be able to make a simple soft start circuit using a mosfet and some timing caps.
Perhaps combine the two approaches, allowing for a quick first-time boot up with the caps, and then later, even faster boot ups by using the batteries just temporarily to cover for any charge the caps may have bled while the thing was powered off, using the batteries a lot less than if they were the only thing there?
I would've thought a simple transistor constant current source would be good enough, rather than using an op amp and voltage reference? Or did you try this and find some problem? And I would've just used a resistor divider on the super caps, fed into a comparator rather than an op amp, then you just need a pull-up resistor rather than a gate drive and pull-down resistor.
Or plug it into a 1amp wall adaptor, that also means its powered when the TV is in standby (if your TV doesn't power USB whilst I'm standby) and you can also turn on the TV by issuing voice commands
For me, some of the components aren't sold anywhere in my country. Even there are some shop that sell it, it can't be verry cheaper. I also cant find it on my country online shopping website. I cant use Amazon or eBay for some reason. But your diy version is realy interesting.
How few components are required to build a negative resistance? Not sure is cheap parts exist but a simple 550mA current limiter with no voltage drop below 500mA placed in front of a few big super caps to charge from the supply.
Great vid. The DIY vs Buy decider for me is whether this is a general circuit that can power a generic device with a USB source or if it is purpose built for the firestick.
I'm not much of an electronics nerd but I was wondering. How hard would it be to scale something like this up to act like a UPS for SBC's? Like could you power a RasPi long enough, given enough supercaps, to shut it down properly in case of a power outage? I guess that would mean that the first part of the circuit isn't really needed because current limiting off a cellphone charger isn't *THAT* essential (given that it should have its own limiting) so I guess the bigger issue is providing enough current through the second part of the circuit.
You could simply buy Y Type cable which is 2x usb standard port input, each can be plugged to one usb pory in tv and microusb output . I have this kind of cable and therr are no more notification about undervoltage. It's the simpliest and the cheapest solution.
This seems like a considerably better alternative that would've been considered at a design discussion meeting during the planning phase of the commercial product. Such a design is so much better that I really wonder why it was not chosen.
Very cool.. was just thinking of a problem with external hard drives for recording video.. I guess you can also run into problems with over current. You could just bypass the data lines and be able to run an HDD with this too. I'm just not sure if you need to connect the data lines after power is supplied..
If I had an "unsmart" tv and needed a fire stick, I would just power with a phone charger laying around. As the USB shuts down at 500+ mA, there is always a risk of frying the motherboard with super capacitors' high initial current spike. However, the buffer circuit is quite good at maintaining it under 500mA and offers all necessary features as an energy reservoir. anyone needing this kind of technology should definitely give it a try.
If you plug your outlet is close enough you can actually swap the outlet out for one with a USB port that'll power the device. I have one and I believe both ports output 3.6 amps. Which would be enough for the firestick. Also the outlets are about 20 bucks. If you're good with changing an outlet and faceplate it would be simple task.
Please correct me, if i am wrong. As the li ion battery will never get completely discharge while plugged in. So, the assumption of 500-3000 being " less" in this case is a bit too much. So for the most life of the Commercial product, Li ion battery will be full and as it will not get discharged even to half, its cycle completion will almost never happen. So, you can use that device even longer than 3000 days.
Hi GreatScott, what is your opinion on the Raspberry Pi Pico? Have you already given it a try? Are there any chances that you will include one in your future videos? And if so, would you rather recommend using Arduino or Raspberry Pi Pico for Electronics projects? Thank you!
I’d be curious if this could be adapted for use inside a high current draw guitar pedal(900mA) so it can be daisy chained w other pedals w/o causing excessive noise. Yes there are already market solutions to this issue but there’s always a new way to skin a cat
Mine was doing what you described when using the ethernet adapter they sell, apparently it doesn't supply enough amps for both the eth and fire stick when updating
Hi ,Great video as always 👍.One of the ICs is MP2617 a switch mode battery charger and the other IC is also from Monolithic power systems but not sure what the PN is ?
Hello GreatScott! You say that you've created 0.1 V reference throught the help of TL431. Isn't TL431 designed for Vref around 2.495 V? Does driving the Vref so low create problems for the TL431 and blow it up? Furthermore, isn't problematic taking voltage off of Vref of the TL431? Shouldn't be the voltage taken off of Cathode when needed instead of Vref? I'm asking all of this because I am buidling something with TL431 right now... I'm confused a bit. Schematics state that the voltage needed is taken from Cathode pin and also state that Vref must not fall below 2.480 V (I think it was 2.480 the minimum value for the available Vref voltage, maybe I am wrong but I am sure that Vref was not stated that it could go as low as 0.1 V)... Otherwise, great channel, great videos.
I personally think that your DIY version is better considering that have super capacitors instead of battery, much better solution. For example, i use DIY power bank made by myself with super capacitors to start up my car instead of regular car battery. Much better, much clean and more reliable, and the life span is much bigger than regular acidic battery, as well as it is smaller and "environmental friendly" compared to Lead-Acid regular car battery. Great video and great channel thumbs up
A modified version of this circuit which can handle a 24v input would be super useful for 3d printing guys (like me) who run Octoprint. 🐙 As you can use a simple momentary button to shutdown a RPI safely, if the circuit could take a 24v input and when the input power is switched off, provide a momentary connection between GPIO3 and GND, the RPI could shut down safely when you turn off the printers main power switch with the super caps providing enough power to allow the RPI to complete its shutdown procedure. Great video.
The adapter he's talking about for sale on Amazon has some reviews that say it doesn't work after some use. It may work for a bit before failing and causing the Fire TV to restart over and over in a loop
A cheap li-ion battery bank would produce 5v at 1 amp if chosen for the purpose. The battery terminals could also be connected to a TP4056 set to provide 500ma towards the battery and thus, via its inverter, about 400ma contribution to the 5v load. The lithium ion battery will sit at 4.2v except for short periods when the load is demanding more than the TP4056 is set to provide so the number of full charge cycles is not going to be approached.
Holding a lithium battery at maximal charge also decreases lifespan. A significant increase in lifespan can be achieved by holding the cell at 3.7v with a custom charging circuit but at that point your complexity is even higher than the supercap circuit and it still will last far less.
I see there one big problem: You did all the switching and sensing on the GND side. That means as it works (the current sensing, current limitting, load switching,...), it will make voltage differences between these grounds.
But this GND is common with the USB communication, connected inside of the TV. That means the "fire stick supply output" GND is shorted out with the "power input" GND, once it gets connected together into one TV.
You should just have made all the current sensing and regulation on the positive side (so using PMOS devices,...), keeping the one single common GND.
I agree. I am surprised it worked at all when he used it on his actual TV. My only explanation for that is that his TV's port already has soft current limiting, meaning he could have gotten away with just connecting the capacitors directly to the USB port power.
Agreed switching the ground is a recipe for disaster.
Yes... wow...
This is exactly why I wrote down "not tested for longer time" on the disadvantage side. Now longer testing would not have revealed this problem because my TV works with my solution. But if I would have other TVs to test I think I would have found this problem sooner. It seems like I ignored that the USB GND and HDMI GND are apparently most of the time connected together. In theory that should mean that charging the supercaps up and adding the FireTV stick should work just fine. But afterwards the lower MOSFET would get shorted which means we remove the current limiter. Since there are still the supercaps in parallel to the Fire Stick that should not be such a big deal breaker. But yes. I think this is a problem in my circuit which I didn't find with lab bench power supply testing and with my TV testing. This is all theory in my head right now by the way. So the problem can be bigger or smaller. Switching the positive votlage side with P-Channel MOSFETs would solve the problem. That would require a bit of a redesign but the main components should stay the same. Sorry about that.
@@greatscottlab Hey I'm sorry but I didn't exactly get what the above commenter explained. Could you perhaps elaborate in a yt short or a community post or even here.
I'd added 2 LEDs to show is in charging and when ready, also a protection diode to get polarity protection and * feedback current from supercaps to tv.
*Edit: avoid
Good ideas for improvement :-)
Is there a schematic out there of these improvements?
@@jonmayer Not yet.
@@greatscottlab And capacitors needs some big parallel resistors for long term balancing.
Go for it
It still amazes me that we have devices like this today that are so small and are basically full fledged mini computers that can stream 4K video while using only around 200ma of power at 5V!
Hardware decoders are amazing, you can do the same video reproduction task with either gobs of processing power, or a small fraction of it with the correct hardware. I think even plain MPEG 2 video at those resolutions would put a strain on a processor with that little power consumption if they didn't have the hardware decoders, then again back at your point, it's still really impressive how said decoder can do such a feat with so little power.
@@Kalvinjj iirc rpi3 can decode 1080p24 hevc 10bit with SW decoding and consume about 5V1A. But it has a general purpose low budget chip.
Was he streaming it at 4k?
Whether or not it was 4k isn't the point. I'm guessing the poster is around my age. I'm 53 and remember being in computer classes where a GB was theoretical. No one believed at the time something could be so large. I remember being hot shit and paying $400 for a single 4MB SIMM of RAM. The fact technology has progressed this far is the point he/she was making. Shit, I remember a time when there was no internet.
@@llspragulus im interested in it, whether its the point or not, since there was a lot of difference in power consumption between 4k and 720p.
Mid air circuits should be considered art
I think so too ;-)
and a fire hazard
@@heartlessalice5801 A discharge hazard at least.
I wouldn't want to be hit by those capacitors
Here in Brazil some people call them spider circuits
@@TheNumbuh121 in english it is called "dead bug"
Important! Pls add two similiar high ohmic resistors in parallell to supercaps!! This will help them with staying balanced! Otherwise you run risk with age that one blows up due to over voltage. Capacity may vary as they age = they will come more and more unbalanced!
I approve of protype development process. I do the same.
Or just use a single one. I believe there are some 5V rated ones although they somewhat rare.
@@Kirillissimus never heard of these. Max I saw was a 3v supercap.
I believe 5v is just 2 in series hidden within.
But who knows maybe they really produced 5v supercaps
Not a great method. Bottom balancing on ELDCs can cause major issues because they can actually flip polarity when at 0 volts. It is imperative to never drain them down fully. Doing this can do internal and external damage.
@@EVLS10 Not sure what you mean with bottom balancing as parallell resistors will always try to maintain 1/2 x Uin over each capacitor at all times. Complete discharge is unavoidable due to current circuit design. I do agree that parallell resistors is the bottom tier solution and cannot keep up in very dynamic situations. This design is usage is pretty chill though, where the caps will at 99.99% of time just stay hovering at 5V without much going in or out giving much time for the resistors to do its job. In the case of rapid charge or discharge; parallell resistors are much much better than nothing, as it will correct errors with enough time, as opposed to not being corrected at all. During a full rapid discharge, with unbalanced cap size there is the possibility that it will force the lower capacity cap into reverse voltage, with discharge resistors it will atleast try to minimize the time spent in error state. I can't forsee that as worse than having the cap forced into reverse voltage for a longer duration so; I'm not sure what the fuzz is about but I'm intrested to know. Adding parallell diodes would limit the reverse voltage situation, clamping it to one diode forward drop. (But if one has gotten this deep into patching, one might aswell get a proper solution like SAB mosfet, op-amp balancer or maybe a well thought out zenerdiode)
@@Paxmax it means that your method will intentionally or not bring both cells to 0 volts after a long time of being disconnected. You are providing a constant drainage path which is a no no with eldcs
This is opening the barrier to new invation in DIY. Love Great Scott. Miss the old intro tho.
I second getting back to old intro.
Nice to know supercaps have been improved a lot. They didn't use to have any significant energy capacity. They were just for maintaining a voltage on something like static RAM, etc.
Might be better if the output voltage cutoff had some hysteresis, so it gets to a decent level, but will then allow it to drop to the minimum working level, otherwise it may cut off prematurely as the capacitor discharges.
Could also use a boost converter so that the voltage does not fade with capacitor discharging.
Finally, it appears that the switching and any drop is in the 0V side, not the +V, this would be bad if the USB ground is not isolated from the HDMI ground
Good idea. As you’re probably aware, it’s trivial to make this tweak to the circuit presented.
Charging time will increase 🌚
The main ic is a monolithic power MP2617H.
worht checking the datasheet, really cool ic, it does everything, charges the lithium cell, regulates the maximum input current and provides higher output current (from the battery) than the input current.
would be a pretty decent choice for a powerbank
Thanks for the feedback. I will definitely have a look :-)
how do find that IC ?))))
Cool idea. Though I have couple of ideas to consider. Since you are charging the 2 capacitors to a combined higher voltage that they can tolerate by themselves, I would put in some protection for too high "cell" voltage. Supercapacitors (at least in the range of 100s of Fs) can be easily be damaged with too high voltage. Also the op. Amp used to turn the load on does not have any feedback to prevent oscilation when reference voltage is near the compared voltage. Just a small amount of feedback to the compared voltage would made the operation to work a bit like a schmitt trigger.
If you want to be safe you can parallel them and use a cheap booster to get 5v and a Cc step down
This is not regarding any video in particular but all in a whole, THANK YOU for your content I love watching them even though I sadly cant afford to do like 99% of any of it I find them very entertaining and knowledgeable I really love your style of explaining how it works or why it works that way the most Thank You again
Great stuff. I like how you take three fundamental circuits and come up with simple solution.
Thanks ;-)
DIY won definitely. Good idea in general and very interesting video. Thank you and good luck.
Thank you!
I think I'd rather power with an external supply over this solution tbh. In the past I've used a relay to turn off the device when I've required a synced power down.
exactly my thoughts, you could tap from the mains inside the tv to connect a seperate 5v powersupply that offers enough amps, add a usb port to the tv and connect a relay in between.
My tv has both 500ma and 1A ports. Although using a separate power supply is cheaper still. Good video
I didn't watch the video closely, but it seems he made a contraption instead of simply buying a USB power supply? They're only $5-$10 each, for one that does 2 amps. I don't see the point of making something like this when they're cheap and available. I did not pay attention to the video, though.
Amazing work! That circuit is genius and the possibilities to use is infinite 👍
Thanks👍
Could this also work as a rudimentary UPS to make the connected device survive short power downs?
He used Super Capacitors and not a Li-ION Battery. It's job is to only deliver required short current spikes while being powered by current limited input and not to be used as UPS
I think the Buy Version is capable of that stuff.
@@pauljackson2126 He did show however that you can power the TV stick for up to two minutes in this configuration, longer with larger/more caps
@@resneptacle then again larger caps means longer charging time
@@resneptacle actually, if you want to keep it only on for like short period, then I guess you can use higher power caps to make a poor man's UPS
Sure, just depends on the down time you're looking to accommodate.
Interesting design, I thing you should probably add a voltage balancing circuit of some kind for the series super-cap, dispersion in capacity may result in premature deterioration.
or at very least a 100k-1M parallel drain resistor on each so the higher cap will drain faster to some safe equilibrium while not wasting a tonne of power.
Kudos for beating a major manufacturer with better design! For me this is way above my capabilities to make, I would just hook up a phone charger to the 220v TV cable and use it to supply voltage to the Fire Stick.
Excellent rendition. You say the circuit is simple, but there is a lot going on. You truly have big brain, Scott!
Thanks for sharing! I really enjoy your videos. They are great supplemental material for my EE degree.
Take care you wizard!
If cost is a thing I would simply use an old usb phone charger, they are >500mw.
What would be handy, like a sort of PC PSU, offering different voltages, for for use outside the pc. Near my tv I have a router, nvidia shield, DAB+ radio unit. amplifier, raspberry pi. Most are 5v, some 12v. Every device has its own power supply. Why not have 1 power supply, supplying power to all devices. This is probably more efficient too (less watts). And as a bonus, being able to turn on/off per device.
I also thought using a lipo was not a great idea. Trickle charging like this would be good for NiMH cells too. I do like the super cap solution too.
Ok further watching and this is awesome. I need to make a few of these.
I've seen some high end charge management chips for Li-Ion charge the battery to only 80% capacity - e.g. 4volts instead of 4.2volts, this doubles or triples the life span of the cell, other ways of doing this too is to allow the battery to fall to 80% from fully charged, doing a smaller charge back up to 4.20volts, meaning the cell is never heard at fully charged.
Explications are so clear, really good job. I learn so much from your channel! Thank You!
Your mid-air soldering is art
Nice solution! I was wondering if it would be sensible to limit the input current draw to 450mA, so that the TV's usb port isn't being driven so hard/close to its max mA?
Usb ports are rated for continuos .5A output.
The 500mA are already a low rating. Most of these ports can deliver 1A, some even 2A.
@ True, but as GS mentioned, not all are capable of exceeding their ratings - in this case USB1.1 base PD.
It’s worth noting that it’s fairly common to have 500~700mA (nominal) PTC fuses on USB (1.1-2) ports. Being a fuse, the PTC does not _actively_ limit the current (at least not very rapidly ;) ). Instead, after a period of time - say 500-2500ms (depending on several factors) - it’s resistance will become extremely high, cutting power. As such, as you mentioned, it’s not at all uncommon to see ports that will handle 500-1000mA spikes without a hitch, but completely fail when they are asked to deliver that same current continuously for even a second or two. It’s also important to note that the recovery time for PTC fuses and very dramatically, from milliseconds to hours. Both factors combine to make the device featured in the video necessary, in the absence of an external power source.
Personally, this is one of those problems that the DIY is less of a value proposition than just trying to find an outlet to plug the stick into.
Even if its cheaper than the commercial product, so is a USB charger I already had.
But admittedly I'm WAY lazier than you :P
Great video overall though!
Having your firestick/google TV stick turning on and off with the TV is priceless ...
@@JeanRobLast Ironically this is the exact reason I gave up on running my Chromecast off the TV port (my TV has a 1 amp port which should be enough but still caused it issues from time to time). Specifically, the Chromecast supports HDMI CEC, which lets me turn the entire set up on by just telling Google Assistant to boot up my Chromecast or just directly streaming to it. The Chromecast then automatically turns the TV on, and can even turn it off afterwards. Presumably the Fire Stick can do this too. The result of that is you can ignore the TV remote entirely and everything works off of a single control. If you use the TV turning on to turn on the Chromecast/Fire Stick you still then need to switch over to whatever is controlling the other device.
@@clancywiggum3198 yes the fire tv does that too, also it doesn't need to boot every time and it can update when not in use. Power consumption when asleep is negligible
@@clancywiggum3198 Good to know, i never explored that possiblity
I may have done an overkill then. I opened up the tv and the sick, put a dc plug inside after the fuse, added the included charger inside in a corner, added a heatsink to the stick and changed its wifi antenna for a broken laptop one, the hdmi cable goes out from a small carved hole and plugs to the tv, so it has better wifi, can run colder and I have only one cable visible from the tv.
Hmm, now this makes me want to try a similar solution for my 3D printer. I've got a raspberry pi hooked up directly to the printer's PSU, so I have to make sure to shut down the pi before I turn off the power (if just for good practice's sake). Would love to read the state of the PSU and shut down the pi automatically, but obviously that would mean it'd have to be on for just slightly longer than the rest of the system.
you could also made a circuit that first shut down the raspberry, then shut down itself and the power supply with a relay
@@jerrb7991 Yeah, I've been thinking about that too, same with just buying a smart switch or something, but those draw power while they're off (of course, so you can remotely turn them on again, but which I don't necessarily need) and I'm not confident enough to work in a relay on 220V mains. So it's not that your idea doesn't work (or even might be better than mine in general purpose), but I've kinda backed myself into a wall with my requirements ¯\_(ツ)_/¯
Still, appreciate the idea, thanks 👍
@@LordHonkInc For what it's worth the idle power draw of Zigbee stuff tends to be really low (my IKEA Zigbee unit uses less than 0.1 watts for instance)
The pi has an overlay to get a pin High or Low when it shutdown. Look for gpio-poweroff
I quit my engineering college just to learn from youtube its was cheap and reliable
Scott is a wizard! He can solder components in mid air!
Your perfboard skills are incredible. Could you please make a video where you give us little tips and tricks?
I literally went to setup my mums stick on Sunday, and got the "low power" message haha.
Great alternative. Is that power adapter advertised as having a li-ion battery, do customers know that it's not going to last?
7:21 That's gotta be the biggest mid air spaghetti I've seen in quite a while lmao
It is art ;-)
A lot of TVs like Samsung will power off the USB ports after turning the TV off. But still, some TVs have USB ports powered even in standby mode.
Yup, same problem with my TV.
Yea this would make the battery solution better. Maybe a supercap+battery hybrid would fix this issue. Using USB C would be the way to go to have data and power all travelling through the same port. You could even forget about HDMI this way, and USB C wouldnt have the top current limit
I could see a potential issue with the circuit. You are using low side switching on FireStick GND, and this works fine on bench power supply. But when connected to TV, the GND of usb output is tied to GND of hdmi port, and thus the low side fet switch is bypassed. Now the GND FireTVStick goes straight to GND USB.
I tested it all with two real TVs and never experienced such problems.
@@greatscottlab I would like to see actual voltage and current measurements when connected to a real TV. My suspicion is the TV's USB port is soft limiting the current and your circuitry is being bypassed.
While I don't really have a use for a Fire Stick or similar, I think Thist is a very interesting idea for averaging current draw in many applications - I can imagine it would be a great way to power remotely installed solar powered devices that have an occasional need to draw extra power (to operate a motor or servo etc). I'll be bookmarking this :)
What about taping into the 5v power rail of the TVs power supply?
I would love to add something to the USB port instead of opening up the whole TV.
Excellent solution Scott. There are a number of circuits that supercapacitors are better but manufacturers always jump to lithium ion batteries which increases prices and decrease life.
I just ran into my own case of battery VS capacitor. The raid card I was trying to use required a battery backed ram cache to enable RAID 5. While searching for one I realized that they also made flash backed ram cache modules. These use a capacitor to dump the ram to flash when the power cuts out. The capacitor is rated for 80 seconds of use, while the identically sized battery is rated for 3 days. However the battery is only rated for 3 days for the first 3 years, and if the power is out for more than 3 days I'm stuck. With the FBRC I can last indefinitely without power and the capacitor can last longer than 3 year.
In my line of work as an AV Integrator I have to deal with these blasted firesticks. I have seen newer TVs come with 1A rated USB ports that is great so it can be powered from the TV. It's kinda hard to mount a TV and have a USB wallwart sticking out three to four inches from the electrical outlet, doesn't let the TV lay flush and screws up stuff. Never know what equipment the customer has until I show up. Would be nice if the HDMI 2.1 spec could use that +5V pin with around 800mA to 1A of current so stuff like the firestick and chrome cast can be powered from the HDMI port itself.
I'd love a PCB for this. Something I could put in a 3D Printed case and mount on my TV. Maybe there's a way to use this for multiple USB Devices? I have a PS1 Classic running off USB now too I'd also love to power via the TV safely.
I remember there was an app that would allow you to create the pcb from the schema. It also had an option to optimize the size of the board and to create single or double layered pcbs as well.
I love your circuit diagrams and drawings
I wonder if it would be better using only parallel capacitors at a lower voltage and a buck converter for the output more akin to the way power-banks work. No balancing issues and the charge voltage is not so close to the USB input voltage.
The source limiter explanation actually helped me with my uni course, thanks
Great Scott, Great Edutainment!
I’m curious about when you build your beta circuit in free air. Do you prefer to build this way because it allows you to understand the circuit better? Do you prefer free air vs breadboard? Do you prefer free air because of convenience, easy vs hard, fast vs time consuming?
Some videos of free air circuit construction can be more entertaining than the actual circuit application information, personally I prefer deeper material like yours vs the “Sensory Satisfaction Art”, but a balance is a good thing.
I ask you the free air construction question because if there is real value to it, vs breadboarding, I might be willing to give it a try.
Thanks 🙂
i love the DIY (comic sans) or (impact) Buy (arial wordart?) font choice
also, that circuit is ART man
"never above at any time" at 03:52 while it shows more than 800mA ;-)
btw at 09:18 the IRL540 seems to be connected in strange way: drain to the ground?
One step closer to removing battery storage from our devices. Bravo!
Fantastic! I didn't know about that thing! 😃
I was waiting to buy 2 new smart TVs... Now I'm getting 2 of those! 😂
Really great work as usual, dude!
Stay safe and creative there! 🖖😊
You dont need his crap to use the fire tv stick. Just plug it in to the wall
@@JackTheAwesomeKnot I didn't even say anything about it. Take care of your own life.
It might be overkill, but I'd love to see a version of this circuit that uses a buck converter to charge the supercap (so you aren't burning half the energy when you initially charge the caps) and a boost converter to make the 5V output (so you can drain the caps lower than 4.5V and use more of their energy storage capacity).
You'd be able to start the Fire TV before the capacitor is fully charged with that sort of design.
Is there no problem switching the GND of the Fire TV instead of the +5V line? In my opinien this will not work, because the HDMI GND will short the switching mosfet for the FireTV.
thats that what that circuit is for. Now I know! All of the videos are cool!
since you have that ambient light (the ws2812b led strip on the back) setup hooked to your tv, you probably have a raspberry pi as well, in your case why not use power from the pi's supply, it would atleast be of 2-3a
You can use 2812b strips with an Arduino nano too, just let a SW like prismatik do the work.
@@kentahirono but thats when you have a pc, its tv so you need to use a pi, also ya Hyperion can also use wled as a ambient light, but all of this dosent change the fact that there are ton of ws2812b leds so there must be a beefy power supply
I have build a circuit like this one. what you really looking for is a HOT SWAP CONTROLLER. you can put 2 of them and chain the Pgood and Enable pins together to separate the super caps. you might also just be able to make a simple soft start circuit using a mosfet and some timing caps.
Perhaps combine the two approaches, allowing for a quick first-time boot up with the caps, and then later, even faster boot ups by using the batteries just temporarily to cover for any charge the caps may have bled while the thing was powered off, using the batteries a lot less than if they were the only thing there?
I would've thought a simple transistor constant current source would be good enough, rather than using an op amp and voltage reference? Or did you try this and find some problem? And I would've just used a resistor divider on the super caps, fed into a comparator rather than an op amp, then you just need a pull-up resistor rather than a gate drive and pull-down resistor.
This one is actually great, Scott!
Or plug it into a 1amp wall adaptor, that also means its powered when the TV is in standby (if your TV doesn't power USB whilst I'm standby) and you can also turn on the TV by issuing voice commands
Laughed way too much at the "This monstrosity" part with the sound effects :D (7:16)
Hello nice video, in the AMAZON product the first IC is an MP2617B from MPS that is Power Path Management/limiter for single Cell Li+ Battery.
I really like the way you show the resulted mid air soldered components🤣
For me, some of the components aren't sold anywhere in my country. Even there are some shop that sell it, it can't be verry cheaper. I also cant find it on my country online shopping website. I cant use Amazon or eBay for some reason. But your diy version is realy interesting.
You can build it easily? While in the background is a room of electronic building equipment?
Finally.... PSIM is sponsoring. I was thinking that only newbies like me uses PSIM.
Your free air soldered version can even be used as a smoke machine!
that is a great project. and soldering is one of the things i like.
Great video. Your videos inspire me to do more DIY. Your overview and detailed explanations are fantastic. Keep up the awesome work! Thank you!
How few components are required to build a negative resistance?
Not sure is cheap parts exist but a simple 550mA current limiter with no voltage drop below 500mA placed in front of a few big super caps to charge from the supply.
Great vid. The DIY vs Buy decider for me is whether this is a general circuit that can power a generic device with a USB source or if it is purpose built for the firestick.
I'm not much of an electronics nerd but I was wondering. How hard would it be to scale something like this up to act like a UPS for SBC's? Like could you power a RasPi long enough, given enough supercaps, to shut it down properly in case of a power outage? I guess that would mean that the first part of the circuit isn't really needed because current limiting off a cellphone charger isn't *THAT* essential (given that it should have its own limiting) so I guess the bigger issue is providing enough current through the second part of the circuit.
So basically a powerbank would also do?
You could simply buy Y Type cable which is 2x usb standard port input, each can be plugged to one usb pory in tv and microusb output . I have this kind of cable and therr are no more notification about undervoltage. It's the simpliest and the cheapest solution.
It also doesn't work if your TV has only one port or if both ports share a single 500ma power rail.
This seems like a considerably better alternative that would've been considered at a design discussion meeting during the planning phase of the commercial product. Such a design is so much better that I really wonder why it was not chosen.
DIY is always a winner if you try hard enough.
Very cool.. was just thinking of a problem with external hard drives for recording video.. I guess you can also run into problems with over current. You could just bypass the data lines and be able to run an HDD with this too. I'm just not sure if you need to connect the data lines after power is supplied..
Great video but just curious why not use a breadboard instead of soldering mid air?
This was quicker to do for testing
Hi. I'm pretty sure the first IC is an MP2617
3A Switching Charger with NVDC Power Path Management For Single Cell Li+ Battery
If I had an "unsmart" tv and needed a fire stick, I would just power with a phone charger laying around. As the USB shuts down at 500+ mA, there is always a risk of frying the motherboard with super capacitors' high initial current spike. However, the buffer circuit is quite good at maintaining it under 500mA and offers all necessary features as an energy reservoir. anyone needing this kind of technology should definitely give it a try.
if I might say... fantastic ideal and all, but, i'll not be easier to just use 2 usb ports from the tv in parallel to increase the current?
I love you Scott with heart...
If you plug your outlet is close enough you can actually swap the outlet out for one with a USB port that'll power the device. I have one and I believe both ports output 3.6 amps. Which would be enough for the firestick. Also the outlets are about 20 bucks. If you're good with changing an outlet and faceplate it would be simple task.
Please correct me, if i am wrong.
As the li ion battery will never get completely discharge while plugged in. So, the assumption of 500-3000 being " less" in this case is a bit too much.
So for the most life of the Commercial product, Li ion battery will be full and as it will not get discharged even to half, its cycle completion will almost never happen. So, you can use that device even longer than 3000 days.
DIY WINS FOR ME ANYWAY : ) cool little circuit you made there. plus i never knew that about those TV sticks. SO THANKS FOR THAT BIT OF INFO ; )
You're a genuis! I love your videos! 😍
Hi GreatScott, what is your opinion on the Raspberry Pi Pico? Have you already given it a try? Are there any chances that you will include one in your future videos? And if so, would you rather recommend using Arduino or Raspberry Pi Pico for Electronics projects? Thank you!
I’d be curious if this could be adapted for use inside a high current draw guitar pedal(900mA) so it can be daisy chained w other pedals w/o causing excessive noise. Yes there are already market solutions to this issue but there’s always a new way to skin a cat
Nice, thank you! Supercapacitor will replace batteries on small power applications.
Definitely possible :-)
Good build, but how would you make it simpler ? What can you give up in this build and what would be the trade-off(s) ?
A simple, 8 pin, microcontroller with a few peripheral components would do the job. The SW would be trivial to write.
Mine was doing what you described when using the ethernet adapter they sell, apparently it doesn't supply enough amps for both the eth and fire stick when updating
Hi ,Great video as always 👍.One of the ICs is MP2617 a switch mode battery charger and the other IC is also from Monolithic power systems but not sure what the PN is ?
Impressive manipulation of Voltage and Current Sr. Great Scoot.ill try your Circuit on my other Devices
Why does just hooking up a couple of caps in parallel not suffice? Seems to me a bit like a rectifier-filter problem. Did you try that, what happened?
Hello GreatScott! You say that you've created 0.1 V reference throught the help of TL431. Isn't TL431 designed for Vref around 2.495 V? Does driving the Vref so low create problems for the TL431 and blow it up? Furthermore, isn't problematic taking voltage off of Vref of the TL431? Shouldn't be the voltage taken off of Cathode when needed instead of Vref? I'm asking all of this because I am buidling something with TL431 right now... I'm confused a bit. Schematics state that the voltage needed is taken from Cathode pin and also state that Vref must not fall below 2.480 V (I think it was 2.480 the minimum value for the available Vref voltage, maybe I am wrong but I am sure that Vref was not stated that it could go as low as 0.1 V)... Otherwise, great channel, great videos.
you can use a resistor divider to get a lower voltage ref
Did i miss how we prevent the supercaps going out of balance exceeding the voltage limit on one of them?
I'd like to hear more about that OTII ARC power supply you are using
I think using an LC filter would have been much easier. Can you try it? I can't try because i don't have a firetv.
I don't think that will work this way.
I personally think that your DIY version is better considering that have super capacitors instead of battery, much better solution.
For example, i use DIY power bank made by myself with super capacitors to start up my car instead of regular car battery. Much better, much clean and more reliable, and the life span is much bigger than regular acidic battery, as well as it is smaller and "environmental friendly" compared to Lead-Acid regular car battery.
Great video and great channel thumbs up
A modified version of this circuit which can handle a 24v input would be super useful for 3d printing guys (like me) who run Octoprint. 🐙
As you can use a simple momentary button to shutdown a RPI safely, if the circuit could take a 24v input and when the input power is switched off, provide a momentary connection between GPIO3 and GND, the RPI could shut down safely when you turn off the printers main power switch with the super caps providing enough power to allow the RPI to complete its shutdown procedure. Great video.
The adapter he's talking about for sale on Amazon has some reviews that say it doesn't work after some use. It may work for a bit before failing and causing the Fire TV to restart over and over in a loop
A cheap li-ion battery bank would produce 5v at 1 amp if chosen for the purpose. The battery terminals could also be connected to a TP4056 set to provide 500ma towards the battery and thus, via its inverter, about 400ma contribution to the 5v load. The lithium ion battery will sit at 4.2v except for short periods when the load is demanding more than the TP4056 is set to provide so the number of full charge cycles is not going to be approached.
Holding a lithium battery at maximal charge also decreases lifespan. A significant increase in lifespan can be achieved by holding the cell at 3.7v with a custom charging circuit but at that point your complexity is even higher than the supercap circuit and it still will last far less.
I believe the bigger chip is a MPS MP2617 switching battery charger.