Playing with transformers.

Joined
May 25, 2017
Messages
548
I finished my first home made power supply, and have continued to play around with transformers, so I thought I'd start a new thread.

My first one turned out really good, and now I'm looking at doing more, including adding some safety features that I don't currently have.

image_pjjuhm.jpg


I found a much beefier cable, and even grounded the core.
I'm also looking at building a "current limiter" which is basically a high Wattage light bulb in series, so if I short something, the bulb coming on should be the worst that happens.

Another safety thing I want to make is an "isolation transformer" that is 1:1 ratio. The transformer I have is too small for that much wire, but check out this guy I found at the Habitat Store:

image_vcehuq.jpg


Compared to my first transformer, this guy is huge. Before I plugged it in though, I decided to give it a couple tests on the multimeter first. Everything was going well, until I found that there was continuity between the primary side and the secondary side. So, I started taking it apart to see what's inside.

image_kfpdcd.jpg


Inside, I found ordinary cardboard paper stuff wrapped around the coil, I figured I'd need to unwind the secondary coil to see what's going on, and I got out my favoritedismantling tools.

The first two plates were a bugger to get out, and got a little misshapen, but the rest came out no problem. Although the outside appears to be painted, it looks like moisture has infiltrated anyway. Lots of rusty dust and a bit of surface rust. I'll have to think on what kind of rust removal I want to do that won't be too invasive on the plates.

image_ekdbau.jpg


I unwound what I thought would be the secondary coil, and found that it is the only coil. ??? Is this one of those non-isolating transformers? My first transformer had a connection between the primary and secondary coils, and only had three wires total. I guess these must be a thing, since I've found 2 out of 2 now. So this will have to be completely rewound in order to get me the isolation I want.

The good news is the spool and wire are in perfect condition.

image_nxsdbj.jpg


Since the core is much bigger, I only need 228 Turns for the 122V primary, and probably the same to get 120-122V on the secondary. I mean, ideally I'd use a 1:1 ratio, but there's always losses.

Although the label says it's good for 1000VA, It seems the core is good for about 500W according to the math I've found online. Honestly, I can't think of any use for 500W, but it's good to know your limits.

I've also noticed that the 16 AWG wire is made of aluminum. I can't seem to find info on aluminum Amp ratings for DC. Copper wire was hard enough. For now, I'll guess that it will conduct as well as the next size smaller for copper. At 18 AWG, that'd be 3.2A. I may just buy all new copper wire. For the amount I'd need, that would get pricey though.
 
Yeah, as you found "auto transformers" are not isolated and the output can just be a tap off at some point in the primary winding.
Aluminium wire is not good & I'd dump it & buy some new copper stuff. it snaps easily & is a pain to solder.
To clean the laminations, maybe a quick sand + use a rust converter with phosphoric acid?
Get some of the proper transformer tape if you're going to use this for safety isolation.
 
To add to Redpacket, I would also say get some enamel spray. And every layer, spray a layer of enamel on the under tape, and then after winding, spray ontop of the wires, put another layer of tape, spray the tape, etc, etc.
You say the wire came off in perfect condition, but problem is, enamel gets brittle after a while, and can crack/flake and you not even know it. Adding the extra enamel will help add a layer of protection.
You could probably use nail polish, as well. But that's a bit more expensive, and I'm sure the wife/gf/daughter wouldn't like that much :p
 
Once I found out there was such a word as "autotransformer", searches suddenly started returning useful results. Like this video:

Unfortunately, the length of copper wire I'd need to replace the aluminum is outside my price range for now. I'll do my best with the aluminum wire. I've been able to do the primary winding ok, though I may have more windings than I need. I was able to keep the wire under a fair amount of tension, and it went on straight and uniform.

image_gmekip.jpg


The rust converter suggestion makes me wonder about black oxide as a coating. I've started cleaning off the plates, and there's definitely some kind of coating on there. You can see the shiny metal vs. the mat gray coating still on the rusty pieces:

image_osppko.jpg


Maybe it's paint? I've got lots of paint. I found a video that explains the purpose of laminated sheets vs. a solid piece of metal:

Also, I've been digging around looking for a better explanation of what goes into calculating the number of primary coil turns. You Tube in this case was useless. It's all just the ratio between the primary and secondary. I did find a few web sites and forums trying to explain it. I found it amusing that one fellow insisted that it is impossible to calculate the number of primary turns unless you already know the number. If that was true, we wouldn't have transformers at all, LOL. One of the better websites I found explaining the math was this one:Bright Hub Engineering

Anyway, I finally came across a few places where the following formula was mentioned: T/V = 1/4.44*0.0001*B*CA*f
Took me a while, but I was able to figure out what it means.

T/V = Turns per Volt
4.44 seems to be some kind of constant. Probably comes from yet more math.
0.0001 is a conversion factor between B and CA
B= Flux Density, measured in Webers / square meter.
CA = Core area, measured in square centimeters. This is where that conversion factor comes in.
f = frequency in Hz of the alternating current.

Flux Density seems to be somewhere around 1 to 1.3 W/sq m. This has to do with the specific alloy and how it's put together. In the case of laminated steel sheets, people seem to figure the value at 1.2 W/sq m.
The frequency is of course, 60Hz where I live, so that's pretty easy.
CA is the thing that will change from core to core. But that's a measurement that is easy to make.

Before I learned all this, I'd been using the formula T/V = 42 / CA to get the Turns per Volt, without knowing where the number 42 came from. This new formula tells me that some assumptions were made, as to the Flux Density and frequency. I did some algebra and came up with a new number. Since T/V and CA were the only values likely to change from transformer to transformer, I got T/V = 31.28 / CA. If this is true, then it puts my T/V at 1.39, so I should only need 170 turns total rather than 228. It also means I'd need significantly less wire. 265 feet instead of 354 feet. That's a rough estimate though.

I'm also trying to figure out if I can estimate the length of wire based on the resistance. I found a calculator website that actually works with aluminum, and so far, it seems to work. I get 1Ohm on my primary coil, and I've guessed that the length is about 3877 cm. According to the calculator, the length and diameter of the aluminum wire should get me about 0.9 Ohm. I'm hoping that if there were a short in the windings, it would show up as a much lower resistance reading. My thinking there could be wrong though.

I'llhave a current limiter between the transformer and the outlet, so if there is a short, the worst that will happen is the light bulb will turn on.
 
Raw aluminium has a conductivity of 65% IACS.

Also, unless the aluminium is copper/nickel/tin plated, I wouldn't even try to solder it. It is extremely hard to do, requires preheating, aggressive fluxes.

Too time consuming and hard to solder aluminium wire.
 
Nice finds :)

I've come across the 42 as the magic number, too. ...... Hmmm, now that I think about it, it 'is' the answer to Life, the Universe, and Everything

The videos were of interest, but a little boring in the deliver, for me, anyways ;)

I like the formula you came across. It allows for a bit more precision on the windings, which is good. Can help mitigate losses.
For your laminations, you can use clear fingernail polish, or any kind of thinly applied lacquer. You could even use lacquer you'd get in the paint department. Might be cheaper overall as you can get a pint, and it'd last a looooong time doing many many laminations.
To prepare your laminations, you could use some muriatic acid to super clean them. This will remove any and all rust as well as any grease that may be present. The other added benefit is that it slightly etches the metal, which makes applying the lacquer a little easier to stick to.
You could either spray the lacquer on with a spritzer style sprayer, or you could just dip the sheet into the can (if using the can from the paint store) and hang them to dry. Any excess will drip off and should leave you with a fairly even coating.

Yes, if there was a short in your wires, then the resistance would be calculated as being lower than expected, as it's taking a short cut.
 
Re the laminations, they are usually painted or coated with something insulating - you actually want them (mostly) insulated from each other to stop or reduce eddy currents.
If they were perfectly bright, clean metal, no paint & done up hard with bolts or welded, etc the eddy currents would see the core as one solid block & more power would be dissipated in the core.
The shape of them (thin sheets, oriented "longways") is also helping this.

The same principle happens in some of the (typically yellow) ring cores you find in switching power supplies, they use iron particles & the molding compound, etc keeps them apart.

Eddy currents are basically from some of the core acting like mini sized "shorted windings", ie metal in a "circle" in the magnetic field.
Yes any shorting of windings will quickly lead to fireworks, high current flows, etc. Think of how DIY spot welders from microwave transformers work....
 
Yeah, I found the aluminum impossible to get solder to stick to. I ended up making some tiny aluminum crimps, and covered the joint with shrink wrap. So far, I've only had to do one. I'm also not sure what kind of ends to put on the wires. The original covers have sockets on both sides, I'm thinking because it's reversible. I'd like to have the mains side be permanently installed, rather than have a "suicide plug". I did find some insulated connectors that would fit perfectly in the existing holes. I painted one green in case I decide to use them.

image_xyyoyu.jpg


I've been doing a bit of body work on the side covers. They've got chipped paint, rust and bent spots I'd like to fix. I found some gloss white enamel appliance paint from 20 years ago that's still good. It will be perfect for repainting those guys. Since the label is now inaccurate, I've removed it just to make sure someone in the future won't accidentally pick it up thinking it's still a reversible autotransformer and get a bad surprise.

image_zwusfv.jpg


I've watched a few of Stan's videos. He's a bit odd, but what I like is they're fairly straightforward, and I don't have to skip through long fancyintros that ultimately make the rest of the video a letdown. I've been wading through a lot of that lately, LOL.

I dug around in the basement today, and didn't find any lacquer, but I did find urethane floor sealer. Maybe that will do as a laminant for the plates. I also found some acetone I can use to get them nice and clean before coating. I got quite a few sheets cleaned up the last couple days, but it's pretty tedious. Still, it's pretty satisfying getting the rust off and ending up with shiny metal.

image_iqquzv.jpg


image_iiskcd.jpg
 
The covers are done, more or less. Turned out better than I expected.

image_eolmwm.jpg


The ancient appliance paint took longer do dry, but is otherwise just fine. The insulated connectors fit perfectly in the existing holes. I think it looks pretty sharp. :)
since I'm only making this guy to play with low power stuff, I'm not going to bother making the mains side both 120V and 240V. Just 120V will do, and will save lots of room and wire for the primary.

For the secondary, I think I'll make a center tap since I have the holes and connectors. I'm going to make it 60V for now, as I've got a future project in the back of my head that might need that Voltage, and this will save me time and effort later if that's the case.

I've been continuing to read and watch videos about isolation transformers, and apparently, a truly isolated transformer shouldn't have a ground on the mains side? I'm thinking that I may as well add the ground though, as the neutral is already tied to ground both at the house service panel, and at the neighborhood transformer. Also, I spent all that time painting one of the connectors green, I've just got to use it, eh?

Still cleaning the core plates. Still tedious.
I've got about 74 pieces de-rusted, but there's still 120 left to do. Maybe I need to start listening to podcasts or something while I work, LOL. I've got enough done though to start cleaning and coating them. I'm thinking that the 3 ends of eachE and one edge of each I should be free of coating so they can make contact during reassembly.
 
Looking good! :cool:

You could do several taps, and just have a jumper inside the box to change your output voltages. That way you wouldn't need to redo the whole thing when you change projects ;)

For the grounding, it doesn't matter if the ground/neutral are bonded in the main panel. There's a reason why ground and neutral isn't bonded in the sub-panels. Might want to look up "Ground Differential" to get an idea of how it works.
But basically (and I'll admit I'm a little hazzy on the details here) is that there is slightly less resistance on the ground wire because it goes straight to ground/earth sooner. It should never carry load under normal operation conditions for this reason. This could cause the breakers to see more current flow and pop sooner.
Remember, on 110/120, Neutral carries amps back to the source while a device is in use.

I'm not sure why they wouldn't ground an isolation transformer (which most transformers are anyways). All the transformers I've seen have been grounded in some way. Either ground is directly bolted to the unit, or through grounding of the metal chassis it's bolted to.

Maybe others can chime in and clarify this issue.
 
Finally got all the plates cleaned down to bare metal!

I spent a while looking for ways to get black oxide on the plate, and I think this videosums up the details pretty well:

I've already started degreasing with acetone. It's what I normally use on metal that I'm about to paint.
Not sure if etching is needed, as I've just run them all on a scotch-brite wheel, but it's easy enough, and I've got the distilled vinegar.
I've got the salt, but I'll have to go get some hydrogen peroxide. Unless I can find some around the house...
And it turns out I've already got a little electric burner, so I won't be doing this in the kitchen, LOL.

I think I will use deionized water for boiling. I've got pretty hard water, and deionized is cheap, even for me. :)
 
Wow, interesting. Learned something new today :D I didn't know 'rust bluing' was a thing.
 
Redpacket said:
Yeah, as you found "auto transformers" are not isolated and the output can just be a tap off at some point in the primary winding.
Aluminium wire is not good & I'd dump it & buy some new copper stuff. it snaps easily & is a pain to solder.
To clean the laminations, maybe a quick sand + use a rust converter with phosphoric acid?
Get some of the proper transformer tape if you're going to use this for safety isolation.

Remember that isolation between the plates must be maintained! No electrical connection between them !!
 
Hi folks! It's been a long while, but I haven't given up yet. I've been re-rusting my plates, and though it's not as easy as YouTube suggested, it's still going well.

I got all the plates cleaned with acetone, etched in white vinegar, and the rusting process has been kick-started with hydrogen peroxide and salt.

image_isssvj.jpg


It's not as uniform as I had hoped, maybe because I didn't get all the coating off like I thought I did. But they're still rusting quite nicely, and boiling them is definitely converting the orange rust into black rust.

image_rxbbyj.jpg


I've found that now the rusting process has begun, I only need to let them set out for a few hours, and the bare spots start to rust on its own, and can be re-boiled. Here's a few that are part way through the blackening process:

image_lgwzid.jpg


At this point, it's taking longer because I don't have the time to sit and watch the "soup" boil more than once every couple days. But more than half the plates are just about done.
 
I must have missed it, but why are you going to all the trouble?
Are you making something else out of the trans.?
 
When I took the transformer apart, the plates were all rusty. What I should have done was to lightly sand the rust down and then maybe boil the plates and oil them so as to keep the insulative coating intact. But I didn't know any better, so I cleaned them all down to shiny bare metal, and then discovered I'd have to re-insulate them somehow. I chose black oxide, as I was concerned with lacquer or urethane being all sticky and making reassembly difficult.

So I ended up making a lot more work for myself than I could have. But there is light at the end of the tunnel! My first few plates are finished, and I think they turned out really good:

image_iphush.jpg


I gave them a very light sanding and a thin coat of WD40. I need to test one to see if it's less conductive than bare metal, but everyone says it will be, so I'm not too worried. Only 183 left to go!
 
After trying numerous different products, I found this at a NAMES model expo.
It is the best that I have used.


Trying to get a picture to come up. :huh:
 
You have to click the image after uploading
 
Back
Top