A Buck Converter Too Far - DC Power Transmission Project Fail
Vložit
- čas přidán 11. 09. 2024
- I found a neat-looking buck converter on eBay with a 60V maximum input voltage and a 30V maximum output voltage. The trimmer potentiometer let me take the output voltage up to 40V with predictable results.
You can't blame your actions on the manufacturer. They clearly state that the maximum output voltage is 30v and yet you ignored that and preferred to use the controller chip manufacturers specs of 36v max instead, which you then also decided to ignore as the pot had more adjustment left. Go and stand in the naughty corner! No playing with your boost / buck converters for a week!
It's rather silly design if device can kill itself from the (meant to be used) settings. It should be limited to max it can handle so that you can't set it any higher or then at least protect itself. If this was designed for James Bond then I agree that self-destruction option is always a nice thing.
ShadowMancer LOL! I would agree with you if this was a standard domestic product like a television or a toaster but this product is clearly designed for use by electronics hobbyists and the like so should not need the same idiot proofing! If this were a radio or CRT television and the user had taken the cover off, then adjusted a trimmer pot to its far end, causing a serious malfunction no manufacturer would accept liability - and rightly so. Just because this buck converter wasn't in a case doesn't mean that it should be designed not to be adjusted beyond safe values, after all, it's not intended for use by the ignorant masses.
My pov is that it would add 10cent to the price to limit the max of the pot and not worth it since most people who buy this has to set it manually anyway. {LIKE IF YOU AGREE DISLIKE IF YOU DISAGREE}
It's a mark of poor circuit design if you don't have matched components.
Doesn't really matter who a product is intended for, built in safeguards is just good engineering practice.
imagine my shock when turning the potentiometer beyond the chips power rating lead to catastrophic failure
Best line... "Well if it blows up... it blows up"
Who else was yelling the following?
JULIAN!!
Get that 300w bench power supply with the 1602 LCD!
it takes 60V in (and out!)
I don't know if you do this Julian, but I treat every coloured, ribbed, ridged, one-side is transparent, *whatever* as positive (or negative in your case). That way, I never have to think, "hmm, what logic should I use this time to remember". I just automatically know something different/special about strand X over strand Y, strand X is positive.
Really helps when you're trying to re-wire your stereo at 3 am and you're drunk!
No, I do it differently every time, get it wrong and blow things up :)
if I go to make something that has a run of wire and have several intermediate steps and think I might forgot what conductor of the figure 8 wire I used as negative or positive, I lay out my run, select which will be which and then put a small bit of red heatshrink over the + end at the source, then go to the destination and do the same. Saves expensive mistakes :)
Great video, as always! I wouldn't have continued to turn the pot, but I'm happy that you did.. Not because its broken now, but who else would have done it? =D You and Clive perhaps ;)
Really love the finished result on the banana leads! Such a small and simple thing, but I've seen way more ugly ones than good ones and yours is just beautiful! =D I hate you for showing the spider though! ;) haha
So you took the 30V max output buck converter up to "36v which was the datasheet max, the pot still hadnt reached its end limit, so I carried on going, got to about 40v" -- and it blew up, what a shock. You trusted the pot, which probably wasnt a good idea. What was the point of the voltmeter then? Sigh. You should try building a custom solution instead of using 3 poorly suited off the shelf converters in series
DC-DC converters mean high-frequency square waves, which means you're probably going to need a circuit board. Such things are not trivial - you have to either spend a lot of time making your own, or wait a few days turnaround on revisions from a supplier. For something as common as a DC-DC converter, it's faster and cheaper just to buy one as a module.
Agreed, board design for switch mode regulators can make the difference between the thing being stable or not. It takes even professional designers several revisions to get a new one right. It ain't something you can just knock up.
LT3800 specs say "Telecom 48v applications" and this does not mean Power Over Ethernet. DC Distribution power is 48v in Telecom nodes (central office buildings and expansion nodes - eg DMS, remote frames, regen's, etc..).
I think the problem is in bad ferrite coil. Judging by color, it is actually not ferrite ring at all, but rather noise suppressing coil, which has low efficiency for any power conversion. It is ment to be only for removing noise after some buck or boost converters, not as active coil.
Word of advice, never trust those data sheet( if any) of those cheap disposable Chinelectronics, esp those that's been poorly translated. When they say up to 60V, discount it by 50%, so 15V is perhaps doable. 10A supply means only 6A is you are lucky.
They may also be using components that are not original or have failed QA tests (underperforming).
Was it in fact the chip that burned?? I was hoping you would do a autopsy on the board.
If your going to be using "unsuitable" connectors outdoors even short term I would recommend at the very least buying a tube of dielectric silicon grease for ignition wires/automotive lamps to extend the life/keep resistance down. This should be plastic/rubber safe, but even smearing petroleum jelly or general grease would be better than nothing...
Hi Julian, I haven't read any comments below. But I would suggest that you use AC , not DC to carry power over long distances. As it is done currently in utility power lines. Surprisingly you will loose lots of power in just a few feet when using DC. Love your show, Cheers mate.
I'm curious How does it do 10Amps with no heat sinks I have some droks that get pretty hot with large sinks on doing 8amps.
This is a high speed switch mode power supply. It doesn't regulate. When a transistor is on, there is lots of current, but no voltage, thus, no wattage. When the transistors turn off, there is full voltage across them, but no current, thus no wattage. The output of a pulse width modulator is the supply voltage times the percentage of duty cycle. if it switches a perfect square wave, 24 volts becomes 12 volts. If a 10 amp load is placed on the output, the 24 volt source only has to supply 5 amps. Theoretically, 24 volts X 5 amps=120 watts, 12 volts X 10 amps=120 watts. Realistically, you would put in about 125 watts, to get 120 out, because nothing is 100% efficient, but these types of converters can exceed 97% efficiency under ideal conditions. The whole circuit probably dissipates 5 watts of heat.
New modern FET's are astonishly efficient! Nevertheless gluing heat sinks is a good idea.
Definitely more bang for your buck
I doubt it would have worked very good even if it survived.
That big boost converters output voltage will probably drop a lot under load if the input voltage is lower than around 18V from what i read and experienced myself.
This probably has also something to do with the input current, which should definetly be less than 10A.
So with all that conversion it will probably only work good for light loads and the smaller loss of the wire at the higher voltage would be lower than the loss from all that conversion but it might still be interesting to see if that is true and where the limits are.
Not to be redundant or spoil your fun, but... why not just use a COTS 12VDC-120/240VAC inverter and outdoor-rated power cord to a conventional AC-powered bench supply or thousands of other common appliances?
Are such inverters too inefficient or produce too choppy a sine wave? Surely there are better efficiency/cleaner waveform inverters on the market fairly dirt cheap and well packaged for outdoor operation. I'd think there should be tons of them for auto/marine applications.
Sorry, I was thinking you'd be supplying power from a solar-charged 12v battery set.
That would be wayy too boring.
And its more efficient to use boost and buck converters
After all, Julian does have about 500 of the things
Did you even measure it? I think you blew up the volt display, NOT the DC2DC converter.
That is a design fail. Looks like they had to develop the adjustable part themselves (the datasheet shows no "typical application" for an adjustable version). I don't know how exactly they designed their circuit, but if they took the app notes and just replaced the two feedback resistors (basically a voltage divider) with a pot, you can set that pot to a position where you make the chip regulate to a higher voltage than the current sense inputs can take (40V max according to datasheet). That's not the fault of the chip. It's the designer's fault of, first, using a chip that's not generally intended for adjustable operation (otherwise you'd see an example of that in the datasheet) and secondly skimping on a single resistor in series with the pot, that could have prevented this. And obviously noone ever bothered to test the design and turn the pot from one end to the other at 60V input.
The gate charge thing is also exceeded above around 40V, most likely even below 30V already (but I don't think that's the reason why it fried the chip).
So you have 12V boosted up to 60V to bring it down again to 30V where you would prefer 40V?
Why don't you just boost up your 12V to your preferable 40V directly?
Well, in 2019 the same exactly product supplied with 54v dc still has the same issues. The tpca 8050 mosfet just exploded.
If I had found your clip before buying the board...
Anyway, you are doing a great job!
At about timestamp 8:50, this type of brushing can cause damaging static charge. Over the years, I have learned that an atomizer (mist sprayer) of tap (not distilled) water sprayed in the air above (not directly above) the PCB's and a bit on the brush, kills this high static charge.
I do believe that the linear regulator built into the control chip died because of input overvoltage, and it probably powered most (if not all) of the circuitry inside that lil chip
The sign of a practiced solderer, one who can hold the object and feed in solder with the same hand.
At about 9:40, I like to nearly fill' the spade tube (cylinder) with solder, then when I am ready to insert the tinned wire into it, I reheat the tube 'till the solder will accept the tinned wire applied to it with a bit of pressure. A bit of practice, and you will know just how much to fill the tube initially. Many electricians frown on a "no-crimp" method, but I use it as you do here--never had any problems--but I have had some crimps pull out!
48 VDC for telecom equipment predates POE. Telephone exchanges use 48 VDC because historically they used lead-acid batteries (making them immune to black-outs. I believe they were originally up 56 V a long time ago, but later standardized down to 48 V but this is conjecture on my part.) You can actually get PC power supplies with 48 VDC inputs for installations in telecom sites. (at least you could in 1997 which is when I was involved in that industry)
Hello Julian, interesting videos. Can you tell me if there is a specific name for the pen device you use to turn the trimmer screw with?
Julian. Do you think it is possible to increase a step up converters efficency by just replacing the diode and the inductor to something beefyer? My 5v Step up converters (1A red ones on ebay) inductor and diode are only things getting hot, so could this be the solution?
It did not blow because of you setting output voltage to 40v, while listing did say it would be unstable, but you had no load..
so, let's see...
Fets are 41nC, so looking at 3800 datasheet, max safe input voltage should be
10A seems a lot for something without a heatsink. Maybe it doesn't need one because of its high efficiency? Guess we'll never find out now!
Sorry for the fail Julian. I had my very first boost converter do the very same thing. Lasted all of 5 minutes, the made a fizzle sound and let off a pungent smoke. So been shy about what boost converter to buy since.
Been thinking about this idea further, and perhaps a concept similar to MPPT could work well for efficiency, where the transmit end adjusts the boost voltage depending on demand and conditions to achieve a better efficiency, much like how the boost-buck converter adjusts the boost stage to the best value for the buck stage. I realise boost buck do all their magic with one micro-controller, but you could achieve something similar by using an NRF for communication between the two ends. That way you don't always have to boost up to the full 60 (or whatever the maximum happens to be).
Interesting ideas
@2:25 I think the the opposite way. The greater the difference between input and output, the less work converter has to do. If the difference is low then duty cycle will be higher hence more work.
all these converters could vendor from a fuse, I don't no if you want a slow blow fuse, or a faster fuse, you are the expert, maybe you will do a video on that. and where on the converter to put a fuse
9:35 those plugs are supposed to be pressed, not soldered.
What about pressed then soldered?
@@mashedpotatoes5323 You're probably making a joke, but you should not solder wires in a crimped connection. Neither before nor after the crimp. In consumer applications it probably doesn't matter much other than needing to be repair more often. But there's a reason why aerospace industry requires crimp-only connections, because adding solder to the mix makes the connections mechanically fail more frequently.
An analogy to what Julian did is to put R rated tires on a car and then try to go 150 MPH. Its just safer to try things with electronics than cars.
I love your DC-DC series!! Keep it up!
What's the max of the meter? Couple of my self power meters are max at 30 volts.
Hi Bill, in the video he mentioned the power supply for the meter was tapped off the input side of the boost converter ie. 12 battery. The sense terminal is connected to the output side.
This self-powered meter works from 5v to 120v.
The module should never be operated above 20V output!
From the data sheet, page9:
In LT3800 converter applications with output voltages in the 9V to 20V range, back-feeding VCC and VBOOST from
the converter output is trivial, accomplished by connecting diodes from the output to these supply pins.
Deriving these supplies from output voltages greater than 20V ***will require additional regulation to reduce the feedback voltage***.
Exactly that circuitry to limit Vcc to 20V is missing, Vcc is fed from a simple diode out of Vout.!
Probably the chip will run off specs until 30V, but 20V is the regular limit for Vcc. Period!
The best is to immediately replace the 50K pot with a 20K one.
so why do you use a step-up and step - down converter when you could use only step - down converter
I'm wondering if the instability above 15V is because it needs a different compensation network at that voltage.
Most bucks I've used were only ever designed to run at one voltage.
Hey I have that same converter that little skinny one it's junk I tried using 55v in to reduce down and it smoked it.
There is recently some better buck down units that accept up to 70v but you pay a bit more for then. There is a few for 14 bucks but are like 180w. Not bad.
Newest buck I've seen lately is a blue one that's 800w 70v to 58v max looks similar to the 1500w red ones.
i have one of these but can't seem to get the voltage to change at all.. under load or not.. feeding in 20v and it's spitting out 20v, regardless of how mcuh i turn the pot or which direction :-/ can't figure out if i'm doing something wrong or if this thing is just useless :-( trying to get down to 12-15v for a cigarette socket adapter of a usb-c source :-/
48V is also every major phone switch ever in the us...which is why PoE uses generally 48V nominal.
The exact reason 48V is used is because its the highest voltage classified as "extra low voltage" so most things are engineered not to exceed it (at least nominally) for safety. actually i think there's law governing it such that if you want to exceed 48V nominal in a commercial device you have to get extra compliance tests done which you might aswell avoid if you don't absolutely need high voltage.
Reminds me of long ago when I was attending the South East Asia Signal School in December, 1970. We were assigned to climb short poles (indoors) about four feet above the floor, and work in a terminal that was housed at the top. There were several so many of us could work at one time, and behind a one way glass sat an instructor who watched us work. Should anyone touch the wrong terminal, he would send a ringing current through the terminal to light us up a bit. Many of my fellow students actually gaffed out and fell to the floor from those little shocks of about 120 volts AC. As you said, the 48 Volts was there all the time, but you wouldn't really feel that, but see sparks should you cross those wires. Strange thing about that school, the US Students had to work to pass the classes, however the Korean students who worked beside us would pass no matter their ability. I asked one of the instructors about this, he explained that they once had failed a Korean and later learned he was put to death for dishonoring his unit and his people. Different strokes for different folks I guess.
i would put the output of the boost in the garden, directly into the buck-boost that you want for your PS.
I wonder, did you put a new pair of chips on this thing later and kept on the specs of the board?
a VERRY important part is also the diode... if the max. voltage of that is (during normal operation, so not even look at the absolute max.)overloaded, the whole circuit will die in most cases.
About the input: since you are using a boost-converter, the voltage there will not be stable enough to be under the limits. Although the longer wires will do some protection you should build in a filter and protection against the peek-voltages on the input as well or keep much lower.
Just get a cheap and small scope (or a nice second-hand) and you will surely see why they will give you problems, especially if not in its loades specs.
Did you get a good working solution in the end?
Used to have a saying that could well apply here, I have seen it applied to nearly everything from a fatal shooting, to car accidents, to well, just about anything else including stubbing one's toe, "shit happens!" See now doesn't that make very thing all right once again. Thanks for the informative videos, I always look forward to your great shows, and have learned way more then I ever Imagined I could for a You Tube Channel.
British sheds in summer time = The Spidularity. The point of time in the universe when everything is spiders
did you check the total gate charge of the mosfets on that converter? I used that same LT part on a high power application. if you use high power logic fets, you need to add a precision 5v regulator to supply Vcc so that the internal regulator doesn't fail which is probably what happened here. Silly Chinese engineers
It's not just an OVP circuit has kicked in that requires the input voltage to be turned off and on is it? (Don't think it will be, but one can hope).
Unfortunately I suspect that this is simply the wrong chip for the job and some fool has blindly copied the "typical application" schematic. The Vcc pin is absolute max of 24V. The typical application shows it powered through a diode from a fixed 12V output, but a quick skim of the datasheet appears to read that if Vout < 9V, you'd need a boost circuit to power the chip and if Vout > 20V you'd need a step down regulator to power the chip. Therefore having a pot that simply adjusts the output between 1.25V and 30V wouldn't work on this chip without some complex circuitry.
Your spade lugs to banana jack cables are beautiful, too bad they'll be corroded to hell by Christmas if you leave them outside like everything else. ;)
Why do you keep buying different DC to DC converters? Could you address that in your next video?
I may stop buying so many now!
I bought few converters myself,I like learning and figureing out different use for them.
No no no Julian, don't listen to mdesm2005... please keep reviewing more DC-DC converters, for me they are most favourite vids from you. I bought tons of dc-dc converters from e-bay just because of your reviews, I especially like to fiddle with MingHe ones with lcd, they are fun to operate, exploring their possibilities etc. :)
what part of "why" don't you people understand. I even follow my question by specifically asking for a response.
I would think that going from your battery to an A/C transmission line and then DC rectification into your buck/boost would make more sense and eliminate some components, no?
RIP ... a moment of silence for the departed :'(
Hi Julian, what is the special tool called that you turn the 25turn pots with ? I like it cause it seems handy :)
I like the way you do the maths on your circuits Julian, but I dont think you give enough leeway for the "China Constant", I like to consider a device twice the size that a Chinese Manufacturer specs set, and they may not blow up as much, as often or as easily as they tend to, while the Chinese gear does have a place in our lives, it is a bit like the cars in the Hollywood Movies that tend to blow up every time they they have an accident!
Good vid, always learnin stuff on ur channel, Thanks!
It might have been nice to see what voltage was making it through the long "HV" transmission wire to the now blown DC-DC...
the converter is protcted by chip, not by fuse, so you should have enough chips in backhand :-)
Yes. Chips will gallantly sacrifice themselves to protect your fuse!
If you order another could you sneek a zener in between the terminals of the pot, to limit the max output voltage?
I did the something like that with a different one.The capacitors blew on that one. Designing a power supply with capacitors (or other parts) that don't have enough head room is just bad design.
Also, what about a low voltage cut out to protect the battery?
How are you Mr. Ilett?
Can b this module use for charging a 12v battery from 19v adapter?
According to my magic Chinglish decode ring:
'more better' = 'best'.
Output 1.25-15V best, the performance will not be stable when output more than 15V
What a shame. Nice bit of sunshine too! Best wishes.
What boards do you recommend that allow you to exceed all the warnings and still get away with it?
If the whole idea is to avoid power loss, using I^2•R wouldn't it be easier to just drop the current?
If you use your capacitor bank as your main power source then you could trickle charge it from the lead acid out back. Using pwm or something to limit the current along the wire to minimal levels would also limit the power loss.
Firecul42
because cutting your current by 60% reduces power loss in the wire by 84%, compared to doing nothing. still not sure if the losses in the wire are high enough to warrant the effort though.
edit
reducing the current is exactly what he's trying to do. I can't work out the math in my head while on the phone fit your method. it might work because he does not use the line continually.
Cory Marshall Sorry I guess I didn't explain very well. I know the theory behind what he is doing and I just wanted to suggest an alternative.
Instead of transferring all the power all the time it's in use, I suggest a constant trickle to top off the capacitor bank.
If he uses the caps for his supply then they could fulfil his requirement for the majority of cases while being efficient by having a low current and therefore power loss in the supplying wires.
Hi Julian...great video! I need a small buck converter that will buck down 54 volts at 2.1 amps down to 42 volts where the amps can be as high as 5 amps. Can you recommend a small buck converter that can do this please?!
Is the surface (part of the battery) not metal ? Because you put the convertor only partly on the paper sticker.
(I'm a newbie!) Why not just set the boost converter so that it delivers 37v at the end of the long wire run and plug these directly into the D3806?
Exactly, DC looses more voltage over long cable runs than AC does. this is one of the reasons why mains supply is always AC. You simple can't pump DC over long distances.
DC doesn't lose more they both will lose the same for identical average currents, the reason AC is used for transmission is just because its much easier to make super efficient transformers to convert voltages up and down than to make a dc/dc converter.
I think some of the 600w boost converters seem to have the same problem.
Hello there,love your vids...
What you might try is 5/6/7 x 12 volt batteries in series or whatever to eliminate the need for voltage reconversion at the receiving end. i.e boost voltage (synchronous boost?) to charging voltage for 5/6/7 batteries in series and after that leave them on float.You would be able to get/pull 12/24/36.. etc volts across any 1/2/3.. etc batteries.
Hmm, that's not such a crazy idea :)
You know Julian electronics run on smoke. If you let the smoke out it doesn't work anymore :D
Hi Julian... love the videos, but please don't put your experiments on, or near your lead acid batteries. A spark can ignite the hydrogen gas in them...I saw it happen to a co-worker once...not a good experience.
so someone makes a board and state 30 volts max 15 recommended, you push it to 40 and it dies and you state it`s a fail, i disagree, you knew its limits and chose to go over them, at 30 volts it might outlast everything else in the curcuit, there is no way to know now.
My mppt controller needs to be given power through my car alternator. Problem with that is too many amps and not enough volts. I guess we are working on the same thing but my solar panels will be there for summer but winter time I think I'll be ? dependent on the boost converter, running off the car alternator. And then into the mppt controller.
I found a review of this unit by HKJ.
"At 60V I reached 2.5A and the module died. This prevent me from doing more test!"
"This is not a 60V converter, I would never rate it above 50V."
lygte-info.dk/review/Converter%20DC-DC%2010A%206.5-60V%20to%201.25-30V%20UK.html
Nice video Julian, you should buy a nice soldering station.
He should buy a PSU first
It's no fun to buy a PSU when you can build one. And have some magic escape every now and then. :-)
I had a chuckle on the "special purpose" comment.
You might want to do a CZcams video on some of these mods (try the 8V reg first, it's easy for
Oops, forgot the IRF3205's are not LL, ...IRFZ44's then, but at only a few amps at most!
Bet if you replace the LT3800 it would work again, if you can manage it.
If it were me, I'd find the resistance of the pot at 36V, and install a parallel resistor to prevent it going over 36V. Something the chinese manufacturer should have done themselves. Seems a common fail with these cheap units.
Well the IC itself is often most of the cost of the entire board and can sometimes be hard to obtain outside of the manufacturer or specified distributors such as Digikey and they tend to charge a modest amount for shipping alone, unlike the Chinese sellers who offer free shipping on everything, so you're almost always better off just grabbing another board instead.
I wish they would use fuse of some type,to protect our eqiptment
Dang your about Mid Evil on that poor battery. Boxes are only few bucks.
hello sir julian... can you do for me how to explain the dc to dc ZK-4KX buck boost converter. Thank you
That efficiency suckd if you can only get half of 60 out lol
What type of cable did you use for those nice connections (with the banana plugs)? I'd like to buy a spool and it would be nice if you could help.
Rapid Electronics 0.75mm tri-rated. Part code 01-1830.
Julian Ilett
Thank you very much!
Aren't those self power voltage meters limited to 30v? If so it could have been that meter dying not the board.
This self-powered meter is definitely 120V
Julian Ilett ah. A self destructing unit. Cool.
Well it would have suited your requirement, but it won't now.
seems all his synchronous bucks are fated to die prematurely.
+iceberg789 Yes, that's true. There's another one coming. How should I destroy it?
ah..... idk, pl find one of your unique ways. =]
Could you set the boost converter to 30 vdc to match the buck converter on your bench?
I think the whole idea is he wants to use a high voltage transmission line setup to reduce voltage drop over those long wires ( and also for a bit of fun ). He's got a massive roll of wire, just run two cable runs in parallel would also work :)
Why don't you try a DPS5015 DC-DC buck converter. It takes 60v input and outputs 0-50v regulated with a nice color display. They also seem to be very efficient. I did some testing here - czcams.com/video/ddbyVrk4P2Q/video.html . Jump to about half way where I do some testing of efficiently.
That's got to be the ultimate buck converter - I'll see if I can get one. Cheers.
you must have a room full of buck and or boost boards
I solder a 14G wire to each brass TTpot screw, finger adjust now, no driver!
why bother looking at the data sheet for the chip and the board if you were going to keep turning the output up to 11? why are you surprised that a product built as cheaply as possible doesn't have protection against stupidity. of course setting the output 11% higher than rated isn't going to end well. I'm surprised as normally you are more precise.
Same things happend to my LTC3780 boost buck regulator. Current limit set too low and fried up the mosfet
I have come to the conclusion it is not worth it, my solar input 200W system October 2016 is dismal, it is OK for experimental use only as it would never pay for itself.
hello Julian, do you plan to play again with arduino and modules one of this day ?
Yes, to make a hygrometer with OLED display.
+julian ilett I guess you made a todo list huh :)
Chinese amps or international ones?
+Julien Ilett can you do an episode on arduino motor drivers
:-) Keeping your wires which have been snipped off. Me too!
Dingo spider stole my baybay
The product is for hobbyist not a consumer product. A well designed fool-proof consumer product of the same functionality would cost at least three times if not more.
This made me smile. Good Video though
Mike bkk