How Power Transformers REALLY Work

We take a slightly more in-depth look at how power transformers work than the simple turns ratio equation relating input to output voltage. We discuss the inductance of the primary and observe the magnetizing current on an oscilloscope, as well as seeing what happens when the core goes into saturation. We also look at how the primary "knows" there is a load on the secondary and why current in a winding is reduced if the number of turns and hence voltage is increased. Finally we look at basic design compromises reacquired to decide how big or small a transformers is for any given load while keeping efficiency high and core and copper losses low.

[@princesswalt4010]

“I don’t have a wormhole at my disposal” are words I should use much, much more often!

[@ElectromagneticVideos]

Wouldnt the universe be more fun if we all had convenient wormholes for time and faster than light travel :)

[@davida1hiwaaynet]

Thanks for this great video! I have always appreciated how useful transformers are for the electrical hobbyist and technician. I have several 3kVA and a 5kVA transformers that I use for testing. They step 240V up to 480V for use of industrial VFD drives and other cool stuff.

[@ElectromagneticVideos]

So glad you liked it David! Yeah - what would we do without them! You have some big ones! I have never played with a VFD - have top try that sometime. By the way newark.com has a bunch of current transformers drastically marked down on clearance right now - just ordered a bunch - might be useful for your stuff if you wanted a permanent ammeter on you big transformers.

[@cocosloan3748]

The quality of the copper of old windings is much better than what they use these days ! TY for this very educational video 👍

[@ElectromagneticVideos]

Is the copper quality poorer? Maybe having been recycled and mixed with impurities? Or some other reason? Glad you like the video!

[@cocosloan3748]

@@ElectromagneticVideos Recycled and mixed with impurities is the correct answer 😋 Have a wonderful day 👍

[@ElectromagneticVideos]

@@cocosloan3748 Interesting. The best part of posting videos is learning stuff like that from comments like yours. Thanks! Have a great day too!

[@yafois988]

@@cocosloan3748 A commonly overlooked element!!

[@cocosloan3748]

@@yafois988 Yeah - and you know -less conductive metal like aluminum creates loses and heat - and the transformer can overheat easier.. Also cant deliver full power ...

[@hectorpascale1013]

Underrated Channel! Subscribed ;) @40:45: The "Isolation" Top/Bottom makes no sense IMO, when the distance left and right to the "E-I Core" (and between primary and secondary) is much shorter. Maybe the same Core was used as well for fabrication of different types with more windings?

[@ElectromagneticVideos]

Thanks! You know, I think your right - I'll bet the core was a standard core and they hey just filled in the unused part with epoxy or some such goop!

[@d46512]

Makes me want to try winding my own someday. Old desktop UPSes are a good source of EI cores. In your sacrificial specimen the HV transformer, why does it seem that the coil isolation distance on one side of the cutaway is greater than the same distance on the other side? I hope you consider this series on power electronics with such an excellent balance of theory and practice.

[@ElectromagneticVideos]

I think its just inexact manufactuing. It looks like they somehow positioned the coils so they didnt touch, and then poured some sort of insulatng goop (epoxy?) over everything to hold everythin in place and provide high voltage insulation and protection from moisture (a real issue with high voltage transformers causing insulation to break down and char causing a short). If you want to wind your own transformer, the best DIY book for that is the 100 year old book "The Boy Electrician" by Alfred P Morgan. Its an amazing book. Being pre - copyright times do a google search and you will find many digitized versions available as a free download!

[@BTW...]

@@ElectromagneticVideos Yes, old poor quality design and manufacturing has the isolation distance vary so much in that old old HV Tx cut open. They used a varnish to impregnate the windings into the paper layers. Just like motors that have armature and field windings, the components are fully immersed in a varnish/lacquer bath and then baked. You notice on the HV windings it's broken down into several paper insulated layers, rather than a 100% copper winding mass. That helps with insulation resistance in the HV section. Modern HV transformers are wound differently, having primary and secondary windings on separate spools that are stacked inside the 'E' cores. This helps achieve segregation, and cheaper/more efficient production. Cores are smaller too. These days it's common to find 3 types of laminated construction types. The alternating stacked E/I core (like that old one shown), The E/ I core that has no interleaving (the I stack welded closed along 2 external edges, being cheaper and easier to produce), and The F core (same core shape of 2 voids for wiring) where the stamped laminations are of identical shape, stacked in an alternate interleaved pattern (being cheaper to produce press tooling). You may also find small lamination stacks inserted between primary and secondary windings, filling a section of the 2 core in the E/I construction. The torroid cores are the simplest, usually a continuous steel strip coiled up. Damn torroid transformers should be a lot cheaper than the old E/I core design, but they are not for some mysterious reason! These days it's common to find (token) thermal CB protection built into small consumer level transformers, but not located where it gets hottest under load - in the centre of the core where heat doesn't dissipate quickly. Decent high power audio transformers will have protection windings, not unlike your yellow feedback loop used here... to give an indication of overloading and magnetic saturation... rare but they are used sometimes. Then there are the ferrite cores... something else again. I've seen strange ultra thin fragile sheet torroid cores used in filter chokes that splinter into tiny sharp fragments when crushed.. the typical material used for them is ferrite that also fragments but not as nasty splintery. You want to take a look at supply infrastructure transformers one day, if you can. The type that stepdown the HV 22kV/2.2kV down to consumer LV 240V or 120V typically rated around 200kVA. Wound very differently and using copper busbar rather than wire. ;-) Biggest I've seen are Power Station Transformers rated at 350MVA at 550kV. All manner of protection systems employed with such monsters. As always, great video content from you. Keep up the good works, mate.

[@ElectromagneticVideos]

@@BTW... Fascinating! I have never heard of audio transformers with a saturation detection winding. Limit the power to keep out of saturation? Or used for feedback to drive even a bit more into saturation to try and linearize the system? Funny you mention toroid cores should be cheap - my only guess is the novelty. There also seems to be a mystique about them as if they worked on any different in principle from normal cores. Whenever I mention toroids I inevitably get comments/questions from the various "free energy" people. I suspect if the cores were called doughnut cores it wouldn't get that response. If ever get access to larger transformer I would certainly do a video on it. Sounds like you have worked with some giant transformers? Are you involved in power s systems? I usually deal with microwatts rather than megawatts.

[@TimPerfetto]

Thank you

[@ElectromagneticVideos]

You welcome!

[@h7qvi]

Feed the induced voltage through an RC integrator and display on y axis. Display current on x axis. You'll get the BH loop

[@ElectromagneticVideos]

You know, I was thinking of trying to display the BH curve but the video was getting (and ended up) way longer than I wanted. Might be really interesting to do that for a bunch of transformers and see if there are significant differences, or if modern ones are different from older ones

[@h7qvi]

@@ElectromagneticVideos if you have a digital cro, you might be able to do the integration on the cro

[@ElectromagneticVideos]

@@h7qvi I do have a digital storage scope but I'm not sure it has an integration feature (it may have). Unfortunately it has fan, not noisy but would have to experiment to see if its loud enough to be noticeable on video. I do like the idea of a simple RC integrator - would also be easy for anyone with a simple scope to duplicate if they wanted to try the experiment.

[@miltonthecat2240]

@@ElectromagneticVideos Plotting the BH curve sounds interesting. If you do revisit transformers in another video, and plot some BH curves, please include the toroidal core transformer. Those toroidal core are expensive, and they need special coil winding equipment, but they are pretty interesting. For example, I once was asked to build a variable AC power supply that could deliver several hundred watts at several thousand volts, to test another piece of equipment. After a little playing around, I discovered that a toroidal transformer that looked very similar to the one you have was quite happy operating at 500Hz instead of 50/60Hz; the core losses at 500Hz were still quite low. That allowed me to put 300V at 500Hz into the secondary winding without saturating the core, using a hefty bench AC power supply that we already had (not a cheap piece of equipment, btw), and step up to several thousand volts at the primary winding, now used as the output. (Yes, this thing was extremely dangerous! In hindsight, the HV transformer from a microwave might have sufficed for the testing.) Another suggestion if you revisit the transformers in another video - I've been told that a strong permanent magnet can destroy a transformer by saturating the core, but I have never investigated this. Apparently this is how a solar storm can destroy the power grid? Anyway, it would be interesting to see how a permanent magnet shifts the B-H curve. Another very interesting video, thank you. I wish I had understood transformers better much earlier in my career, it would have saved me a lot of grief.

[@BTW...]

19:48 - I'd take care using power resistors of that design. They are known to go off like fireworks under sustained operation at or above rated wattage. The ends blast out, rather than just crack and burn due to the enclosed aluminium case. The higher the rated wattage the higher the explosive hazard.

[@ElectromagneticVideos]

Good point! Lucky I have never experienced such an event!

[@yafois988]

ELI the ICE man!

[@ElectromagneticVideos]

In case anyone is wondering, ELI the ICE man = [E] leads current [I] in an inductor [L] (that's the ELI part) and current [I] leads voltage [E] in a capacitor [C] (that's the ICE part)

[@a3b36a04]

I've noticed a slight circular discoloration on the lamination around the windings at 35:52. Does it actually trace the magnetic field strength inside the transformer or is it due to uneven (over)heating?

[@ElectromagneticVideos]

What always amazes me is when someone like yourself notices something in a video that I never saw! Your right about the discoloration. I just looked at the actual core (I still have it) and I think its actually some sort of glue or lacquer used to stick the lamination together that is what is actually discolored. And it does seem to follow the filed as you suggested. I think your right about heating since it is in the center of the core legs. So with the max fields being there and being away from the edges where heat could leak out a bit (and perhaps even leak into the fairly sparsely filled winding areas) those center core areas probably do get hottest. How interesting! Thanks for pointing that out!

[@retrozmachine1189]

The crest being chomped off the peak of the mains sine is all too common these days due to large numbers of non-PF corrected switching power supplies and other non-linear loads. I see a similar thing on the mains here.

[@ElectromagneticVideos]

Oh - of course! I guess its the large current dray at the peak of the peak of the voltage flattening it out. That is so incredible that there is so power drawn by non-PF corrected switching power supplies that it can do this to the main voltage. Sure shows the need for better input circuitry on power supplies. Thanks for telling us the cause!

[@retrozmachine1189]

@@ElectromagneticVideos It's a problem of our own making really. Some countries require PFC but only in larger loads which makes sense because of the cost of implementing it in small devices. The catch is the wattage threshold allows a lot to fly under the radar. Wall plugpacks, phone chargers, LCD TVs, LED ceiling lighting, even inverter fridges all fall into what most countries consider to be low wattage. The fridge might only have a typical cooling cycle draw of 40 watts. Worse yet the PFC required may not even have to take the harmonic draw into consideration and only apply to the purely capacitive/inductive component of larger loads such as inverter air conditioners.

[@BTW...]

17:31 - Your 'load feedback' loop (yellow meter cable) is an example of why you should NEVER inadvertently create short circuit loop windings when bolt mounting torroid transformers inside equipment cases. It doesn't end well.

[@ElectromagneticVideos]

On my various videos there have been comments with stories of seeing the damage from that sort of thing. The toroids I bought had mounting hardware with them - probably easiest way for the manufacturer to prevent complaints from those types of issues.

[@BjornV78]

32:26 I have a lot of power adapters from the 90's with various voltages and current outputs that still have a real transformer inside, and until recently I didn't know that the lamination was to limit Eddy currents, I always thought this was to prevent the metal plates from vibrating (Which is the result of the Eddy currents). But I also did not know that these plates are painted so that they are isolated from each other, but what I do not understand is that in some transformers these plates are welded together on the sides. Does this weld not affect the Eddy currents, or am I missing something?

[@ElectromagneticVideos]

You know, I never thought about the welds- very insightful question! The trick is that they are all on the outside of the core and none are on the inside, so they do not create a large loop for a lot of current to circulate. A tiny circular current would flow within the thickness of the weld, but that would be relatively small. Same would apply to the bolts that are often used in place of welds to hold the core together - again only on the outside of core so no loop is created. Thanks for that question!

[@BjornV78]

@@ElectromagneticVideos that makes sense, because this weld is done by laser and not very deep. Thank you for the quick reply. Grtz from Belgium

[@ElectromagneticVideos]

@@BjornV78 Your right - they are shallow - when I took apart one transformer to use its core for the various magnetism videos, I was surprised how little grinding I needed to do to get though the weld. Greeting from Canada!

[@Sparky-ww5re]

Cool demo. Sounds like you are an electronics instructor at a local trade school lol. When I took a class on AC/DC industrial electronics, one of my classmates whose dad was a mechanic, made the remark "transformers are like electrical gears!" At first I didn't understand his point, till I realized his point. A transformer's power input is equal to the power output (minus losses) but when you step the voltage up, the current is stepped down in proportion, and vice versa depending on the turns ratio. A soldering gun is a good example of a step down transformer with an extreme turns ratio, essentially a very low volt, very high current output, but the wattage is roughly the same , in this case the soldering tip, such as 260 watts, in the case of my Weller 550D 200/260 watt dual heat soldering gun.

[@ElectromagneticVideos]

Close! I'm an EE/CE and I taught university courses years ago. Actually loved teaching but would have gotten tired teaching the same stuff all the time. Funnily enough when i was a student the electrical machines course which included transformers was required and seemed like it would be uninteresting. How wrong I was - one of the most memorable courses I ever took, partly because of the great instructors and partly because of the lab which had large 600V AC and DC machines of every type and you got to do real stuff! So this has been great fun to look at that type of stuff again! I never though of the gears analogy - brilliant! I will remember that one! That mechanic was a smart guy! The soldering gun is such a great example that I never thought of! An interesting short video would be to measure the output current and voltage. If you don't have any objections on me stealing your great example idea, I will do a video. Let me know if you OK with that! So are you an industrial electrician? One of my neighbors is and has told me stories about things like giant bus bars vaporizing at nearby heavy industry plant. I generally only get to deal with Amps and milli-Amps so I always find that sort of stuff fascinating!

[@Sparky-ww5re]

@@ElectromagneticVideos yes go ahead and measure the output of a soldering gun. I actually tried doing that with my 400 amp clamp meter, and I couldn't get a peak reading. Initially more than 400 amps, but as the tip heated up went down to around 250 amps roughly, before I risked melting the plastic jaws on my meter. Perhaps a 600 amp clamp?

[@Sparky-ww5re]

@@ElectromagneticVideos I actually work on residential, and occasionally light commercial, and therefore familiar with 120/240 single phase, 120/208Y and 277/480Y 3ø, as well as 120/240 3ø delta with a high leg (208V to neutral, in the U.S. on panelboards that's typically the B phase, in older installations before 1975 it's typically the C phase, and required to be marked orange or other effective means) Never worked on 347/600Y because that is uncommon here in the states, but apparently is very common in Canada, or at least certain parts, though I've never been to Canada. When I was an apprentice one of the masters I was under was working with a group of electricians on a 120/240 delta system several years ago in an old restaurant near Memphis Tennessee, and while doing a panel upgrade the electricians were rushing, it was a late night on a weekend, one of them did not verify voltages before decommissioning the old equipment, the high leg was not identified or the phase tape wore off, and because the NEC requires the high leg to be on the B phase, assumed that was how it was originally set up. The high leg was originally on the A phase, and upon energizing the new panel, several thousands of dollars of 120 volt equipment smoked up when 208 volts went through. Lessons learned here, do not trust colors, do not trust other people, trust your meter.

[@ElectromagneticVideos]

@@Sparky-ww5re My soldering guns are all the smaller ones - 140W I think - so the currents should be less. One of the cheapo hardware store clamp on meters i have has a 1000A scale setting so that should be OK. Interesting how much the temperature increase lowered the current - I guess a small change in resistivity is relatively big when the driving voltage is 1V or so.

[@ElectromagneticVideos]

​@@Sparky-ww5re That's really cool! I have never found our how we ended up with 347/600 since we normally just copy whatever the US does. As far as I know, we are the only place in the world that fluorescent ballasts and LED drivers for 347V in - used in most commercial buildings of any size. What a lessen! Yikes! Trust no-one! On a much smaller scale I have a buddy who renovates houses on the side, usually places that the previous owner made a mess of. Anyway he was quite confused and asked me if I might have any idea what was going on with kitchen light that continually burned out. It was of course somehow wired to 240V although how that mistake could have been made and not caught by the previous owner I'll never know. He eventually got an electrician to fix that and inspect/correct the place including other deficiencies such as added outlets wired with speaker wire or coax cable! I'm sure you have seen lots of crazy stuff like that!

[@danblankenship5744]

Very nice, thanks for the demonstration and tutorial. It answered so many questions but it left me with a couple more questions. In the beginning, you had the two primary windings in parallel and the secondary windings were disconnected. You demonstrated how the current waveform developed peaks as the voltage exceeded the 100VAC point. You explained that this was due to the core material starting to saturate. The next demonstration showed what happened when you connected a secondary winding to an incandescent lamp. The current through the primary jumped up significantly, but after you changed the current sense resistor and adjusted the volts per division scale on the oscilloscope it appeared as if the spikes were gone. The input voltage remained the same so why was the core capable of handling the magnetic field without saturation? It seems that the same voltage and increased current would create more magnetic fields making the core go into full saturation. does it have to do with the secondary winding creating a reverse EMF of something that cancels some of the fields absorbed by the core? My second question is what would have happened if you would have connected the two primary windings in series (like it was for a 240VAC) and kept the primary voltage at only 120AC. I believe that these current spikes would have disappeared. Am I right in thinking that the voltage would need to exceed 200VAC before these current spikes would return? I would like to see videos on core saturation and how designers can determine the best type, size, and materials to fit the job.

[@ElectromagneticVideos]

Very good points - I guess I wasn't as clear as I should have been: Those horrible current spikes we saw are actually tiny in terms of current - about 0.01 Amps peak if I remember correctly. The light bulb is 60W. TO make things simple we will say it draws 0.5Arms (average). So the peak of its sine wave is about 1.4 times 0.5 = 0.7A. Our horrible 0.01Amp peak is still there and adds to that -so combined the max current would be 0.7 + 0.01= 0.71 if the sine wave peak and horrible peak was at the same time (they are not). But where ever it is its a tiny 0.01Amp increase might me just noticeable if we zoomed in on the scope. I should have done that. The effect of the current on the secondary would have no effect on the peak value of the B field in the core and the saturation peak current into the primary should remain unchanged. BUT - because of resistance in the primary wires, AC source supplying the system, the effective voltage as seen by the primary drops just a little when the load is supplied. So the magnetic fields in the core drops just a bit (as does the magnetization current) so we would get just a bit a less peak value of the saturation current and core would be just slightly less saturated. Your 100% right about if the primary windings were put in series the peaks would go. The primary would now be configured to 240V , so feeding it 120V would be like feeding the single 120V primary only 60V, and that was about the voltage where we saw a really nice sine wave current going in. By the way there is nothing preventing you from using a transformer at a lower than designed voltage. The limiting factor is the current rating of the windings, so if you had a 240V transformer rated at 1A primary = 240W, run it at 120V and you are still limited to 1A so the max power is now 120W. The core loss would be considerably less than at 240V, so if you wanted to minimize that - maybe for a load that only rarely draws power (maybe a doorbell) you could minimize the completely wasteful on-all-the-time core losses that way. As fare as details how designers determine the best core for an application - that's outside of my area of experience. Might be fun to research sometime for a video, but I wont promise anything soon! Glad you liked the video!

[@danblankenship5744]

​@@ElectromagneticVideos, Thanks, that makes much more sense to me now. I like that you took the time to show us those complicated formulas for the B and H fields, and brought them down to show how one is controlled by the voltage and the other by the current. That's a new concept for me, my mind's eye just combined them into one magnetic force. I'm glad you chose a topic that so many glaze over, and these tutorials need to be on Y/T for prosperity. So much happens when you decide to coil a simple wire.

[@ElectromagneticVideos]

@@danblankenship5744 I'm thrilled to hear that!!!!!! Its so hard to know how much or how little to put on KZbin or whether tying to just portray the gist of it with minimal details is understandable. "So much happens when you decide to coil a simple wire" - so true! Those formulas were such eye openers for me too - they are 2 of "Maxwell's Equations" which are 4 equations and are as fundamental as Newtons and Einsteins equations. But sadly Maxwell didn't have as good a publicity agent as the other two. Here is the really amazing part - mathematically put both of those equations together and you get a moving wave = light, radio waves etc and its speed falls right out of the equations! Thats a future video!

[@localverse]

@@ElectromagneticVideos 👍wow, would love to see that video: 'The rise of light!' Remember to mention it was Oliver Heavyside who redid Maxwell's original equations into the neatly brief versions we're now familiar with!

[@ElectromagneticVideos]

@@localverse 'The rise of light!' - thats a great title actually! Yeah - and I think Heaviside was also instrumental in developing Vector calculus for that very purpose. Another name that many have never heard of.

[@waelsadek81]

Isn't the bigger the volume of the core the more flux it can hold before saturation ?

[@ElectromagneticVideos]

You have touched on a very nuanced point: if the volume is bigger it will hold more flux, but the saturation is the result of the flux density, not total flux in the core. A power transformer does not depend on the amount of magnetic energy stored in the core (if that is where your question is leading to) other than to determine the inductive (magnetization) current. So core size is more a practical matter of how big things need to be to hold the windings and as a very general rule, the greater the volume/weight of the transformer (and also the core) the more power it can handle. The core is essentially sized to accommodate the windings. One type of transformer where the energy is stored in the core to be used in the power transfer process is a flyback transformer, where when the current to the primary is abrupt stopped, that energy is stored in the core is released over (a very short time) to the secondary as the magnetic field drops. Hope that helps!

[@waelsadek81]

@@ElectromagneticVideos Shouldn't a larger core allow more magnetic domains to align, resulting in a higher saturation point than a smaller core ?

[@ElectromagneticVideos]

@@waelsadek81 The saturation level depends on the material - look here en.wikipedia.org/wiki/Saturation_(magnetic) On the first graph, the saturation point of each material is somewhere around the "elbow" region where the line changes from vertical to horizontal. If the magnetic loop length of the core is longer, you need more Amp-turns to get to that saturation level, but that that saturation level (and curve) follows the same shape and the saturation level (in Tesla) is unchanged. Note that the Tesla is flux density. A core with greater cross sectional area gives you more flux - but the the field strength - flux density is still limited by the saturation level.

[@waelsadek81]

​@@ElectromagneticVideos Just to make sure that I understand correctly. The size of the transformer (in addition to the wire size) also corresponds to the voltage induced in the secondary ? The reason why I'm asking this is because the higher the power of the transformer the larger the core it usually has.

[@ElectromagneticVideos]

@@waelsadek81 Its the turns ratio from primary to secondary that determines the secondary voltage. So how do you determine the number of turns in the primary? You need enough turns such that the magnetic field in the core stays a bit below saturation. The current in the primary that magnetizes the core adjusts itself to create enough field that the voltage induced in the primary essentially equals the input voltage. A big part of the core size is to ensure there is enough space for the windings - which need thick wire to carry the current for high power applications.

[@wcvp]

So is 1 Tesla being just under where saturation starts consistent across most transformers?

[@h7qvi]

Is for the usual silicon steel laminations. Ferrites are usually 0.3 to 0.35T

[@ElectromagneticVideos]

@@h7qvi Very good point! I did read some modern alloys do somewhat better than 1T - do you happen to know - I always just think of 1T as an easy approximate number to remember.

[@h7qvi]

@@ElectromagneticVideos i haven't looked at materials data sheets for a long time. I find the square loop materials for magnetic amplifiers interesting. A use for core saturation is dc current transducers and magnetic compasses

[@ElectromagneticVideos]

@@h7qvi Same here! I just did come across something about new magnetic materials when searching the video and if wouldn't surprise me if the manged to raise the saturation levels a bit. I do have some vintage magnetic core memory. Might be interesting sometime to see if I can measure the BH curve on them with tiny currents and voltages.

[@maeanderdev]

​@@ElectromagneticVideos One somewhat new magnetic material is metallic glass, aka. amorphous metal. You may like to take a look at that, it's permeability is enormous