Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Playing with transformers.
I do appreciate the suggestions. So far, the process I'm using has only cost a dollar. Any cheaper than that? Smile

They're actually turning out better than I thought they would. I've stopped doing the light sanding though. I keep taking a bit off the edges exposing metal. So I'm now just putting a light coat of WD40 on and then wiping off all the loose stuff with a rag. I'm getting rather excited to get this sucker put back together!
Korishan likes this post
-Mike G
Ones free time is easily available. But a ones dollar is hard to come by.

Mike, you don't happen to have a YT channel showing this progress, do you?
Proceed with caution. Knowledge is Power! Literally! Cool 
Knowledge is Power; Absolute Knowledge is Absolutely Shocking!
Certified 18650 Cell Reclamation Technician

Please come join in general chit-chat and randomness at Discord Chat (channels: general, 3d-printing, linux&coding, 18650, humor, ...)
(this chat is not directly affiliated with SecondLifeStorage; VALID email req'd)
I haven't recorded this particular project, but I do have a Youtube channel with various projects that I've felt like documenting:

If you're interested, I could maybe make a video.
-Mike G
Here's some info on transformers. I can make additions of there is a good suggestion...
@spinningmagnets, I've seen, and read, your article multiple times.

If I understand it well, is it possible to make an RSU welder for connecting copper strips directly to cells?
The power of lithium ion is in our hands!
We'll show them what we're made of!
Those experiments are forthcoming soon. I also want to experiment with using an RSU to spot-weld fuse-wire onto the cells. It holds the promise of making a solid connection with very little heat. 

Something I haven't added to the article yet (due to the lack of pics) is having a center tap on the secondary winding. Some of the industrial spot-welders had three screw-on taps on the front of the machine to access three different power levels. The user often controlled the heat at the weld by throttling the amount of time that the foot-switch was on. However, there is a cheap way to make the control of that type of system more stepped and discrete.

The secondary coil needs to be wound in the same direction as the primary, however, if you add a tap at the 1/3rd point, then you have three options for lengths of coil on the secondary.

I've heard it said that the higher voltage of the higher turn-counts plays a role, and...maybe it does. It is my understanding at this point, that it is the total watts that determines the amps that are provided. Having three coil lengths passing through the pulsing magnetic field provides three different amounts of copper mass to be acted upon.

Imagine you take the area of the secondary window, and calculate the diameter of the fattest wire that can give you 6 turns. Then, at the 2-turn point, you splice a wire where you then connect its other end to one of the electrode lugs. Then you continue to wind the other 4 turns. That provides three lugs that we'll call A, B, and C.

Connecting the two electrode cables to A and B provides 2 turns that pass through the magnetic field. Of course it's only 2-Volts, but it is the copper mass of three turns that interests me.

Connecting the two electrodes to B and C provides 4 turns, and connecting the electrodes to A and C provides 6 turns. Doing this can be a cheap, easy, and robust way to eliminate the need to add a variable control to the input on the primary coil.

You can do this with 1, 2, 3 turns...2, 4, 6 turns...3, 6, 9 turns...or finally 4, 8, 12 turns. Of course the more turns you fill the window with, the higher the volts, causing the amps to be lower. This means the lower turn counts will provide the highest possible amps. Welding cable has a high cross-section of copper with flexible multi-stranded wires, with a tough and thin insulation.

Once the actual amps that will perform well for a given job is determined, it is easy to calculate the smallest size of transformer that would work, however, it is likely that the common 800W small microwave type will be more than adequate, and they are easy to find for free once you start keeping your eye out for one. Plus if its free, there's no wasted money.


Another option for adjusting the amps (which will also not change the volts) is to take the largest E/I transformer coil you can find, and then grind away the two welds that hold the E/I sections together. Re-install the primary coil, and then fill the remaining window area with your chosen winding (fat wire is recommended, 2T or 4T).  

Mount the E section with the three legs pointing up, and replace the "I" section onto the top. If you plugged it in at this point, the I-section would vibrate wildly and fall off, so some type of clamp needs to be added to press it down onto the E-section. What this does is to provide you a way to adjust the size of the magnetic field. You loosen the E/I clamp, slide the I-section over, then re-clamp.

There are very old welding machines that adjusted the amps this way.
BlueSwordM likes this post
Wow, my last post on this thread was in January! Where has the time gone? Oh yeah, other projects.

Anyway, I guess I never told you all that I finally got the plates done. That was a while ago. But I've been stuck trying to decide whether to reuse the aluminum wire, or buy new copper wire. The aluminum has the advantage of being free, but I'd have a much lower power rating on the transformer when finished. New copper would give me a much higher maximum power, but I have to buy it. With limited funds, and multiple projects competing for money, it was actually something I had to think about for a while.

Finally decided on new copper. Not only that, but also a new tool to help me keep track of the number of turns. But first I had to get the old wire back off.

I picked 16 AWG which will give me a maximum power of 360W. Good thing I didn't go any bigger, as the coils just barely fit. Here's the counting machine I got with the spool adapted to it.

The counter machine is supposed to let you do winding faster with the little handle on the side, but 16 AWG wire is really stiff, so I wound up (no pun intended) just turning the spool by hand, and using the counter to keep track. It's actually very accurate.

167 turns later, the first coil is done.

I used the original insulator between the coils, and did another 167 turns on the second coil. It also has a center tap at turn 84, because the box it goes in has three holes in one side, so may as well. Gotta remember though that the center tap will be 60V, and at 3A, that means it's only good to 180W, as opposed to the 360W using the end taps.

I did a few maths before and after I started. Knowing the circumference of the spool, I estimated that I'd use about 124 feet of wire. That includes about 10% extra for good luck. Not knowing the circumference of the finished primary coil, I took my best guess that 167 feet of wire would be enough. Adding more extra, I figured a 300 foot spool would cover it. Seems everyone sells copper wire by the pound instead of the foot. So I got 2.5 lbs. of wire estimated at 315 feet. It all worked out, as i have a fair amount left over.

Later, I did more calcs and figured that at 124 feet, my primary coil should have about 0.5 Ohms from end to end. That is what I got on my multi-meter. Measuring the resistance of the secondary coil, I got 0.6 Ohms. That roughly translated back to around 150 feet. Or about 120% of the length of the primary. So I over estimated how much wire I'd need by quite a lot, but better to have more than enough than less.

I'm still curious about how many Watts the core could handle. The original label said 1000 VA. But then it was originally an auto-transformer able to plug into 110V or 220V. I've got separate coils for 120V so I don't know if that makes a difference. I recently came across a website that gives a few ways to estimate core Watts, so I'll be looking into that. Mostly out of curiosity.

I'm not sure I believe that 1000 VA label though. That aluminum wire was around 18 AWG. At 220V that would be 4.5A. Seems like a lot to be pushing though that little wire.
-Mike G
Well, I'm no further trying to predict how much power an unknown transformer core can handle. This transformer was rated at 1000W. One of the tricks I picked up was to square the surface area of the tongue cross section. In this case 22.44 sqcm x 22.44 sqcm = 503W. Half of the original rating. The other trick I found was to weigh the transformer. They are supposedly able to handle about 40W/kg. Mine weighed in at 4 kg, so 160W.

Moving along though, I put the plates back on the spool, and wired in a 12A braker, a mains Watt meter, and a 250W lightbulb for the load.

I also checked beforehand the various continuities and resistances to make sure nothing weird was happening. Everything was in spec.

So here are the results:
It was loud, but didn't trip the breaker.
121V in vs. 119V out, without load.
Under load, the input was 119V, 3.58A, 426W.
Output was 115.5V, 2A, 231W.

With Watts in / Watts out * 100 to measure efficiency, I've got a whopping 54.2% Impressive, no? No.

Here's what I think is wrong. In my enthusiasm to coat the plates with oxide, I think I made the coating too thick. It's certainly effective. The bare plates are almost no resistance, and the coated plates are thousands of Ohms. But not all the plates fit back in. I think I'm going to have to thin down the coating on all the plates until they do. Not enough iron mass in the core means the flux magnetic thing is wrong and it won't absorb enough energy from the mains current. I think that's why there's about 2.5A being used even without load.

Time to take it apart and do some sanding...
Korishan likes this post
-Mike G
The "loud" part is probably the laminations chattering ie physically shaking from the cycling 60Hz magnetic field. This is one of the reasons manufacturers vacuum dip in varnish. It's likely also ab indication towards the next points....
The ~200W loss in the transformer is much more serious. You'll need to treat it as faulty until resolved.... I suspect you have a shorted winding somewhere. If you run the transformer for a longer time (eg a few minutes, not unattended!) you'll probably find it's getting very hot - that 200W of energy is going somewhere! It might burn windings &/or melt the plastic former...
If you have a large number of the original laminations left out eg > maybe 15-20%? then the core might be saturating as suggested above. this will also cause high current flow. The sanding process might have reduced the total steel some as well.
If you have an oscilloscope, the later issue makes the output voltage waveform look "squared off" & the input current "peaky" ie when the core saturates, the input current spikes up.
Transformers under load are usually up in the >90-95% or so efficient.
Was the original number of turns on the primary the same as now?
Running off solar, DIY & electronics fan :-)
I have an oscilloscope, but it doesn't read more that 50V. I have a smaller transformer that could lower the Voltage, but I wonder if that would change the wave form?

I do have quite a few plates leftover, which makes me think it's a lack of core mass issue. The resistances in all the wiring is right where it should be if there's no shorts.

I was too dumb to think to count the original windings. Also, in it's first life, it was an auto-transformer. Not sure if that makes a difference.

3.5A is higher than I intended for the wire to handle. I used 16 AWG which I figured would be ok up to 3A. I think before I turn it on again, I want to thin the coating enough to fit most or all the plates back in.
-Mike G

Forum Jump:

Users browsing this thread: 1 Guest(s)