200A Power Supply has some Crazy Inductors
Вставка
- Опубліковано 1 вер 2022
- New User Get $30 User Coupons jlcpcb.com/CYT
Other Users Get SMT Coupons via Contacting JLCPCB Facebook
Examination of a 12V power supply with a 216A rating.
Subscribed to my 2nd channel? Watch all my 3 minute videos on Julian's Shorts: / @juliansshorts
Interested in my new garden workshop? Follow the entire build on Julian's Shednanigans:
/ @juliansshednanigans
Join me on Odysee/LBRY: lbry.tv/$/invite/@julian256:d
Octopus Energy referral link: share.octopus.energy/aqua-birch-918
(We each get £50)
#ElectronicsCreators - Наука та технологія
It's not an inductor, it's the transformer! And then there's the rectification diodes (actually there's probably MOSFET for active rectification) under the heatsink near the outputs.
Julian didn't mention output rectification!!!
I agree the last "inductor" is the power output transformer and the beefy "wires" to the heatsink that would have the rectifier under it are beefy to carry the very high output current.
We didn't get a good look at the 1 1/2 turn toroid near the rectifier heatsink, but I wouldn't be surprised if there is another winding underneath, making it a current transformer, to monitor output current.
The 3 metal terminals he talk about should most likely be the 3 terminals for the transformers secondary side, the middle one being the middle tap/negative terminal foe the 12v output?
@@tomgeorge3726 1 1/2 turn current transformer is not a good current monitor at DC. I think it is just a output filter. But it is posible they use magnetic sensors like in dc clamp meters.
That heatsink on the output has limited cooling for the transformer next to it where the thin laminated turns connect to the 3x busbars (center-tap ground), the heatsink is mainly to cool the Surface-mount Synchronous Rectifier's MOSFETs under the 2x thickest points of the heatsinks as the metal mounts protrude under the heatsink to pring the pressure onto the two halves of the rectification.
The transformer is stacked the way it is for a higher frequency to allow for such a small transformer, the plates are flat so as to distribute the surface over the inductive fields of the thinner wires so as to couple more current into the "skin" of the copper winding (The whole surface is the "skin" of the turn hence the flatness to enable better coupling), then the stacking is to parallel the turns as multiple thin transformers in one thicker package, thus reducing footprint and allowing for better coupling.
The design allows axial insulation to be relatively thick whilst having zero need for creeping insulation between input and output and thus allowing better cooling, especially as the primary input couples better to a heat-plate like flat of copper... Some of these inductors are known to have heatsinks on top of them (Especially in those 1ft by 1-inch by 4inch server power supplies at 2KW).
Inductors from top to bottom, the top tall inductor is likely for a 5v logic,
The next two for rectification synchronization stage of current-mode self-oscillating power supply designs,
The one with the "tube" (heat-shrink tubing) is for the main MOSFET switching current going through the transformer so the controller can change the duty-cycle (synchronous cycles to miss) on the transformer,
The tall purple inductor is part of the Power Factor Correction (Required by business laws in many countries it seems)
Basically the whole PSU is a resonant mode power-supply with phase-skipped regulation.
The best thing about this design is the low-noise output due to the low filtering requirement (narrow band-pass inductor, the one with 2x thick as turns in the top right-most corner) and the ability to put a lot of single-rail wattage in a relatively tight space (Good for servers where space is premium).
The downside is this tech is costly to design and make, yet some ATX 12V0 power supplies for hobbyist/home computing are starting to use this design for their 1KW+ supplies, especially as these can be tuned for 90%+ efficiencies (or calibrated via an onboard microcontroller as most server PSUs do)
The specialized windings in the inductors and Transformers are because you can't simply use very heavy wire to handle the 200 amp currents. The skin effect at switching frequencies is very high in large conductors and causes very high copper losses. So the windings are built of multiple strands of thinner wire.
I'm working on a very powerful flashlight, and the kind of current I'm working with has been pushing me into the territory of flat-wire inductors. There are videos of the manufacturing process on UA-cam. It's really fascinating.
I also saw a small SMPS transformer that used a flat copper strip as a winding. I measured it, and it was about equivalent to 3.5mm² wire.
We use some 1500 W 5 volt power supplies for electronics at work. So 300 amps continuous out of a unit about half the height of this one. Just two huge thumb-sized slabs as the output connectors and a deep sense of wonder from the user as to how the heck do they manage so much current handling in such a small box and very stable voltage and no noise either.
Made in Japan and very clean units, impressive.
Hi, whats the power supply model number?
@@LucasRubini I don't remember exactly the model but Cosel 1500 Watt modular units in slim form factor. Probably PBA1500 based on pictures I found with a quick search.
Was hoping for a little demonstation, like the 200+ amps vaporising a wire or something!
That big inductor looks to be the main switching transformer - the high current secondary windings are joined to the plates to get the 200+ amps safely to the equally large rectifiers under that heatsink.
Could the big toroid be for a passive PFC? He should give it to "DiodeGoneWild" for reverse engineering with schematic.
@@raymundhofmann7661 I was thinking same lol
If the short-circuit protection isn't onerous, you might have a nice spot welder there too. 😁
The large inductor after the mains filter is likely for the active power factor correction. The small taller transformer might be the auxiliary power supply, two of the smaller ones might be gate drive transformers one for each half bridge. The third one with the black strap on it might be for primary current measurement. The small flat bar inductor is likely to be in series to the output transformer. A possible controller might be a UCC2895 phase shift PWM controller.
The design of this is reminiscent of the 1500+ watt server PSUs that we have seen over the last 15-20 years or so in the server industry. As others have said the big inductor looking thing attached to the heatsink appears to be the switching transformer and that heatsink with rectifiers is probably why they are able to get over 200Amps out of the unit in a resonably small space (smaller units do exist with larger power output in some servers). Still an impressive feat of design and electrical engineering to build something like this.
THAT is a serious power supply... I'll take two :)
Would love to see this loaded up and voltage checked at various loads to see how stable it is. The output of the unit could be (for test purposes) fed into a hefty (3000w?) pure sine wave inverter. Feed the 240v output of that into your car. Then a bit of working out to see how efficient the whole setup is.
Input has common mode choke, then bridge rectifier and inrush limiting, then active PFC to make the bus voltage across the big capacitors around 400VDC. Then 4 MOSFET transistors in a bridge, with the one small transformer with the black loop doing current sensing, feeding the planar transformer via the other flat wound inductor to control rise time. Primary windings in parallel so as to couple well with the centre tapped secondary winding, likely 2oz copper on a double sided PCB, with each side forming one turn, and the centre tap having lots of vias to connect. The other 2 small transformers are there to drive the 2 half bridge mosfet transistors, probably 4 SMD devices in parallel for each half, to do active rectification, as otherwise at 200A a schottky diode would dissipate close to 600W. Tall yellow transformer a auxillary power supply, driving both the primary side controller, the fan and the active rectifiers, as they would otherwise need a more complex transformer, and easier to have a dedicated separate power supply than try to fit the extra winding on, plus allows a wider voltage adjust range.
Then output side your classical LC filter, with likely a freewheel diode to keep the inductor from saturating, and a lot of capacitors in parallel to handle the riple current, though I would say those capacitors are inadequate, 200A means they would ideally have 20A ripple, while most of that type of solid electrolyte are rated for 5A for rated life. Some will have higher ripple anyway due to trace impedance, so they will not make it long term at 200A.
Input relay shorts the NTC units, as otherwise on 110VAC they would melt themselves off the board, and the 100A at 115VAC is because the PFC can barely handle the current at lower mains. Would say this is actually a 50A supply, with the peak current before shutdown being 200A, and run it accordingly, as running at 200A will likely lead to it definitely failing in a little over a year. Derate and it will probably work for a good while, though the fan will most likely fail first.
There has to be some output rectification after the main transformer. I guess that's what under the heatsink.
The metal strips in the output transformer are the secondary coil and the heatsink is for the rectifying diode which is hidden under it. I have seen a similar transformer in a psu for Motorola BTS but it outputs 27v.
That is the electric version of an air start unit - shit tonnes of air to wind up something bigger at the end of the pipe (in your case a miner box). In the air unit case, probably a boeing 747 ;)
To put it in perspective, 216A -- if it could be sustained across a 1 ohm load -- would require a resistor with a rating > 46kW (!) Put another way, a 1 milliohm connection resistance would heat up with nearly 5 watts. I^2R gets out of hand quickly!
And yet it only pulls just over 2kW out the wall which is barely anything if you own an EV, and have read the data plates on the side of CCS chargers. Power is strange.
You could spot weld batteries with that.
probably this belongs in the hands of DiodeGoneWild.
Nice board, but the white gunk makes it nearly impossible to service.
I am guessing that the large tall inductor is part of the PFC circuit, and that is why the caps in parallel are 450V the mains is rectified with a small cap usually approx 1uF then a boost circuit gives around 400V to the large 450V caps, the the switch mode PSU runs from this elevated voltage. The relay is part of the inrush circuit.
The devices under the heatsink I am guessing are Mosfets as part of a synchronous rectifier.
I'd love it if you could measure the power factor 😃
Interesting
KissAnalog has been doing a series on the design of these types of power supplies. It might be helpful to watch his videos for some insight on how these components work together.
Thanks, I will :)
PSUs in general or these extreme high power ones?
@@alexstone691 High powered smps. All the parts of the circuits. PFC circuits in them etc. He's an engineer that has been designing them for a long time. Currently he's working on the design of a 300watt model for a Class D audio amplifier, but they're for all kinds of applications ofc. Check out Eddies channel.
Squeaktastic new Avatar !!...cheers.
@@andymouse Cheese!! but this one is an imposter, and not me
I use the low tech approach and use Ant powder :-) high power ant miners may bring in the coins but leave the shed overrun with the little buggers;-)
It´s the output transformer.
Julian, would have loved to see it plugged in and some voltage reading taken through the unit. Maybe get more of a clue how those inductors are working.
Julian! 😄
I understand nothing but am intrigued
So what is the controller, the mosfets, diodes, other transformers etc. etc. ?????
3:45 The unusually tall inductor is probably two toroidal cores stacked together...
Also the creepage and clearance limits increase as you go higher, for CE most will limit the equipment to 2000m or 3000m i can't remember the correct altitude. but there are not many places above the normal limit.
But why?
@@Poebbelmann Air pressure! As the pressure drops it is easier for electrons to jump the gap, less getting in their way. So to maintain the same level of safety gaps need to increase. As I said before there are not many places above that height so you just put it in the specification as a do not exceed limit.
@@TheEmbeddedHobbyist I don't think so. As air pressure drops, there are less particles to ionize and therefore the gap could get smaller to obtain the same level of isolation.
@@Poebbelmann if that was so why would all the safety standards increase the gap requirements? Also the voltage required to break down the gap of a spark plug would reduce, where in fact the voltage has to increase to jump with the higher pressure. Have a look for a copy of the standards required by the Low voltage directive (LVD), EN61010 might have it in. Been a few years since I needed to know and the old grey cells are wearing out fast ;-)
@@TheEmbeddedHobbyist Yep, you are right. There is the "Paschen-Law" describing this penomenon. Air needs to have the largest isolation distances at pressures of about 1 mBar. At lower and larger pressures you will need less air gap. My assumption was partly wrong because I knew, that under high vacuum you would need much less isolation distance than under normal conditions. This for itself is right but there is no linear correlation.
smps with PFC...The more expensive the SMPS, the more difficult it will be to repair..😅 i like cheaper smps ..but i want used expensive 1...😂
very sceptical that you could safely pull 200a across those three posts - at that current you are normally looking at M10 ring terminals and very thick cable
I guess it would depend on the length of time. Also, I would think that unit would just burst into flames and smoke if you tried to draw that current for any length of time even with heavy gauge connectors. Like 3 minutes. I am guessing. 200+ Amps is a lot of current at any (car/household) voltage, AC or DC. I have an older and slightly smaller (footprint) power supply for model car battery charging. 120V in, 12V/5V out (US). The 12V out is 10.5 Amps. I can run my 1000W power (car type) amp up to a certain level and it exceeds the 10.5 amps (really loud but cuts out automatically when not enough current). I paralleled the power supply with a 12 V rechargeable gel cell and the amplifier was happy. (I'm doing this on the bench, 20 years ago). My Ammeter only went to 10A, so I don't know what it was drawing when I was shaking my neighbor's wall hangings. (The speakers were 4 Ohm). It wasn't 200 Amps. The power supply was basically replaced by the battery but it kept the battery charged. (Yes, I disconnected the two when the amp was off! 😳) Total draw below 20 Amps. Had to be. And that is a lot. Could it be one of those Chinese mistranslations where they really meant "20 Amps"? It would make sense but Julian would know that. Can I get 400 Amps out of my US 120V? I thought 10.5 was a lot to expect from something about this size from a (120V) wall outlet. I think Julian is "teasting" us. (Just coined it; "Teasing AND Testing". I think we are being "Roasted"! 😂)
I wouldn't want to push it to the full 200A. It was bought (by the previous owner) to run two 800W miners. So it has been used at 133A.
twisting electrons around
Does it meet it's 200A rating???? Good video.
The data label said 240v 16a but can an IEC plug/socket handle that because a UK BS 1363 plug top can't?
That's just probably there to indicate that you need a fair amount of inrush capability to run these so you should have it on a 16 amp circuit. Or it's just some piece of text some chinese person copypasted from somewhere because of reasons.
Anyways the unit will probably not draw 16 amps continuous, but if it delivers the 2600 W of DC power it will well exceed the 10 amp rating of a C13/14.
Not something a cheap overseas manufacturer would bother to care about of course, it'll probably last longer than the unit at full blast anyways.
Nope, theyre only meant to be 10A continuous - though whether it would survive in reality is a different matter, but standard says nope so I would go with nope
Oh I get it, it’s a PSU for a mining rig. Couldn’t work out what it was for until you waved the power leads around at the end.
The other clue is the Bitmain branding:)
@@JulianIlett be careful, there's a bot that's pretending to be you (at least, so ig)
how much for that psu ?
There's one on Ali for $400
@@JulianIlett not bad
@@JulianIlett I have over 10 spare ones (original Bitmain), never used. Will deliver in NL only though. Mine are I think 1600 to 1800 Watt. You do can parallel these. I wil sell a few, but keep a few as well since I plug my solar panels to them! Yes: DC IN, DC OUT. And I fed the DC 12 V in a 2000 Watt inverter to create 230 Volt. This way I can go off grid when needed! @Julian: did you read this?
I don't quite buy the claims of the power supply. A test would be highly interesting!
Haha, if you had seen (Bitcoin / Altcoin) miners at work, you would believe it! These PSU's really deliver that much power.
Hmm they do not seem to want anyone repairing it looking at the gloop all over the PCB 😞
old one now i'm confused SQUEAK!...cheers.
More want to keep the bang down when it fails.
it might be for mechanically securing the parts
Snpsu design = black art
SMPSU or just SMPS
Not so much if you break down the process AC - DC -AC - DC. Low frequency high voltage AC, to high voltage DC. Then high frequency high voltage pulsed DC (AC), to low voltage DC output. all the rest is mainly to do with PFC and OVP, OCP and temp.
Place your bet for the Switch Frequency: 250 kHz, 500 kHz, 1 MHz, 2 MHz, more...?
Would the ferrite still have permeability at 2 MHz or above ?
Did you say what you wanted this supply for?
Not yet :)
It would be interesting to see if it’s push pull and you can put 12v into it and get 240v output or not.
that'd be cool
I don't think that is how switching power supplies work :-)
????????? Hmmmmm , what !!!!!!!!!!!, don't think so.👎🤔🤔🤔🤔🤔🤔
@@brucepickess8097 push pull with 6 fet rectification is by far the most efficient way to transfer that amount of power. Given the number of fets on the ac side and the controller, it looks like 2 fets plus 4 and a pfc controller chip that gets you 384vdc and then a 2 fet soft switching resonant isolated dc/dc.
If I’m right it’s a push pull by definition and should work in reverse.
@@benbaselet2026 depends on the type. Push/pull ac/dc are the most efficient design. See also a ford f150 lightening. It’s a push pull isolated power supply very similar to what was shown here.
3:49 PCF inductors
See the other comments to know what is every inductor doing in this circuit
4:14 the two looking the same are auxiliary PSU for internal needs (controllers, etc...) and the other one is a current sensing transformer
5:15 you forgot to mention the PFC boost, which brings the rectified mains voltage up to 380, regulated, volta DC
6:17 it's another auxiliary trafo
6:22 it's probably the resonance inductor of the LLC stage
6:48 it's to mitigate the skin effect
mmm...IGBT
Is that for your crypto miner rig lol
Hmm?? so going big time in mining??
yeah I bought 2 more miners. they're fan-heater style and a bit quieter than other miners. will be used for office heating during winter.
@@JulianIlett every electric heater that isn't a heat pump should be one