Author Topic: Transformer Winding  (Read 2499 times)

0 Members and 1 Guest are viewing this topic.

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Transformer Winding
« on: July 15, 2019, 11:24:42 am »
I'm trying to wind my own transformer for a homemade PSU, and I'd like to ask a few questions first to avoid costly mistakes. Firstly, regarding turn ratio, lets say I wanted a 2:1 step down. How many turns should actually be made and why? You could do anything from actually only doing 2 and 1 turn to 200 and 100 turns and beyond. How do you calculate the "sweet spot" so to say? Secondly, would plugging a transformer with no load into mains cause a short? And lastly, would winding multiple layers be ok? Thanks in advance to anyone with info.
its your boi, dj ethanol
 

Offline TheHolyHorse

  • Regular Contributor
  • *
  • Posts: 179
  • Country: se
  • You don't need to be confused, just understand it.
Re: Transformer Winding
« Reply #1 on: July 15, 2019, 12:05:30 pm »
If you Google "winding your own transformer" you'll find all the formulas you need.  :-+
 

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Transformer Winding
« Reply #2 on: July 15, 2019, 12:09:06 pm »
I did and all I read is theory and no advice on practical design. I'd just like to make sure that I dont melt wire or trip my circuit breaker, which is why I made this post on a board frequented by enthusiasts and professionals.
its your boi, dj ethanol
 

Offline forrestc

  • Supporter
  • ****
  • Posts: 720
  • Country: us
Re: Transformer Winding
« Reply #3 on: July 15, 2019, 12:18:07 pm »
I'm trying to wind my own transformer for a homemade PSU, and I'd like to ask a few questions first to avoid costly mistakes. Firstly, regarding turn ratio, lets say I wanted a 2:1 step down. How many turns should actually be made and why? You could do anything from actually only doing 2 and 1 turn to 200 and 100 turns and beyond. How do you calculate the "sweet spot" so to say? Secondly, would plugging a transformer with no load into mains cause a short? And lastly, would winding multiple layers be ok? Thanks in advance to anyone with info.

I'm really not trying to dampen your enthusiasm, but the concept of someone winding an AC transformer as a hobby project for something which is going to be connected to the AC power line seems to be not all that safe.   As in 'could start a fire' or 'could cause you to be electrocuted' not that safe.

There are lots of things done with commercial, UL listed, transformers to keep the end users safe, such as inherent current limiting, proper insulation of the primary from the secondary, and so on.   If you get any of these wrong, the outcome is potentially deadly.

I would strongly encourage you to find a commercial transformer for your power supply project.   If at some point you want to experiment with winding your own transformers, then I'd probably start with transformers which aren't plugged into the AC line directly.
 
The following users thanked this post: codingwithethanol

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Transformer Winding
« Reply #4 on: July 15, 2019, 12:18:54 pm »
I'm trying to wind my own transformer for a homemade PSU, and I'd like to ask a few questions first to avoid costly mistakes. Firstly, regarding turn ratio, lets say I wanted a 2:1 step down. How many turns should actually be made and why?

This is akin to asking, "how long is a piece of string?" The number of turns is inversely proportional to frequency, magnetic flux density and core area, and directly proportional to peak voltage. For typical mains frequency transformers the maximum (or saturation) flux density is somewhere around 0.8 to 1.1 Tesla (8000 to 11000 Gauss), while ferrite for high frequency transformers saturates around 0.4T.

The equation I like to use regardless of transformer type (unambiguously valid without weird conversion factors) is:

N = (Vin * Ton) / (Bpk * Ae)

Where Vin is peak voltage (square or sine); Ton is the time that voltage is applied in microseconds; Bpk is the allowed flux density excursion for said voltage and time, in Tesla; Ae is minimum or effective core area (e.g. - the center pole in an E core, the cross-sectional area of a toroid, etc.) in square millimeters.

Ton is either 8333 or 10000us for 60Hz/50hz mains, respectively, while Vin is around 170V for US mains or 310-350V for most of the rest of the world. Typical silicon steel used at mains frequencies can tolerate up to 1.1T, but it is good practice to wind a few more turns to bring that down to 0.8T.

Secondly, would plugging a transformer with no load into mains cause a short? And lastly, would winding multiple layers be ok?

If there are sufficient primary turns such that the transformer doesn't saturate then, no, the transformer will draw very little current from the mains when the secondary is unloaded. The better the transformer the lower this no-load (or "magnetizing") current will be. Otherwise, primary current is directly proportional to secondary current (technically only limited by the heating of the windings!).

Ideally, the number of layers should be minimized to reduce losses from proximity effect and leakage inductance, but these effects mostly apply to high frequency transformers; you need not worry about them too much at mains frequencies.

All that said, it might be less work to start with an existing "split bobbin" transformer (in which primary and secondary are physically separated) with the right power and primary voltage rating, then unwind the secondary and count the number of turns vs. the voltage rating. Wind the new secondary using the same turns per volt ratio, as that will automatically incorporate the core area and flux density parameters.

Ideally you should use enamel coated solid "magnet" wire for the winding(s) because less winding area will be wasted on insulation (and it will invariably have a higher temperature rating than "hookup" wire with a thermoplastic insulation), but you can use common stranded hookup wire in a pinch and/or for prototyping.

 
The following users thanked this post: codingwithethanol

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Transformer Winding
« Reply #5 on: July 15, 2019, 12:53:07 pm »
@forrestc
Yeah the thing is that inductors and transformers and radio circuits have always struck me as black magic, and I used to get discouraged trying to learn about them having no real way of understanding what was going on. You can measure volts and amps with a multimeter, and its all very intuitive learning to probe around in a circuit; but the only tool you get for visualizing magnetic fields as an amateur is "oh hey look iron filings YAYYY ok clean up time". So conquering this topic has long been a goal of mine, buying a transformer just wont cut it. I gotta make it myself and see it work.

@MagicSmoker
I ordered a crapshoot bag of ferrites from electronic goldmine and was lucky enough to get a c core with a serial number on it. Here's the specs:

Material
ferrite

Surface Treatment (Non-Core)
phosphate coated

Special Features
electrical characteristic,initial permeability,5000 percent; effective permeability,3800 percent; saturation flux density,3600 gauss at 2 oersteds; coercive force,0.08 oersted

(The website which I found this on is pretty sketchy, but whatever)

I calculated the cross-sectional area of the ferrite as 36mm2

I have a roll of 30 gauge magnet wire, as well as some other random short lengths taken from scrapped stuff.

According to the formula you gave:

N = (170V * 8333us) / (3600Gs * 36mm2) ~= 10.93

What does N represent exactly? I only know it to mean the turns ratio, but I would guess here it means minimum turns on either the primary or secondary. Also given that interpretation, the result seems far too low, is there an error in calculation?


its your boi, dj ethanol
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Transformer Winding
« Reply #6 on: July 15, 2019, 01:29:39 pm »
...
@MagicSmoker
I ordered a crapshoot bag of ferrites from electronic goldmine and was lucky enough to get a c core with a serial number on it.

You can't use ferrite at mains frequencies. Well, you can, but it will take a bazillion turns.

N = (170V * 8333us) / (3600Gs * 36mm2) ~= 10.93

What does N represent exactly? I only know it to mean the turns ratio, but I would guess here it means minimum turns on either the primary or secondary. Also given that interpretation, the result seems far too low, is there an error in calculation?

N is number of turns, yes, and the flux density is supposed to be in Tesla, not Gauss. Ignoring the fact that ferrite is supremely unsuitable for this task, the correct answer would be 109,306 turns. In this case, both the lower saturation density of ferrite (0.36T vs. 1.1T for Si steel) and the small core area of 36mm^2 are working against you.

 
The following users thanked this post: codingwithethanol

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Transformer Winding
« Reply #7 on: July 15, 2019, 01:52:55 pm »
Thank you for your answer, you may have just saved me quite a bit of anguish! I wasnt aware that the frequency was such an issue, I presumed the permeability was the only limiting factor.
its your boi, dj ethanol
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: codingwithethanol

Offline ArthurDent

  • Super Contributor
  • ***
  • Posts: 1193
  • Country: us
Re: Transformer Winding
« Reply #9 on: July 15, 2019, 04:05:46 pm »
As others have pointed out you have taken a view that a transformer is a two-step process of turns wound on primary to turns wound on secondary when it is actually a three-step process. The input voltage on the primary winding is first ‘converted’ and stored as a magnetic field in the core. Next, as this magnetic field collapses, the magnetic energy is ‘converted’ or induces a voltage into the secondary winding where the turns ratio of primary to secondary determines the output voltage. For a fairly small transformer the primary turns/volt needed to not saturate the core could be in the 4 to 10 turns/volt range and the D.C. resistance of the winding could be around 50 ohms. So the number of primary turns needed for 120vac could be between 480 and 1200.

While it may be easier to visualize a transformer in terms of voltage ratios, what you are doing is converting one form of energy (power) to another, then converting the intermediate power back to the first form on a separate winding, or electrical to magnetic then back to electrical. An oversimplified model of hydropower is converting the energy of falling water to rotational energy to electrical energy in a generator. For hydropower to work you have to have enough input power to convert to enough rotational torque (intermediate power) to turn the generator to produce enough electrical power for the load (electrical grid).

Decades ago when there were plenty of scrapped electronic devices that used tubes, old transformers were plentiful but they never had the output voltages I needed. If I had a device I needed to power I would check what the approximate wattage of the device was then try to find a transformer in my saved pile that had about the same wattage rating. I could then carefully disassemble it, unwind the secondary, remembering to count the turns so I would know the turns/volts necessary for my new winding, and wind a new secondary. Almost all transformers back then had windings to power vacuum tube filaments that were either 6.3 or 5 volts so it was easy to calculate the turns needed by counting the number of turns on those low voltage windings.

For the wire size needed, I could check the current rating of the secondary, if the transformer had that info listed, or check an article in a Popular Mechanics reference book I had on winding transformers. I bought an AWG gauge to check the wire size then I could use a wire table to see the circular mils of that wire size and use a conversion factor to convert from circular mils needed to amps delivered. If I remember correctly it was about 750 cir mils per amp so a #18 wire that has 1624 cir mils, divided by 750, would be good for about 2 amps. Double checking the current rating on the label on some transformers and the measured wire size against the cir mils calculations generally agreed.

Here is a link to a short video that shows pretty much how to wind a transformer but near the end where they are putting the bolts through the 4 corner holes they do not use an insulating washer so this can create a partial shorted turn, wasted power, and heating of the bolts.
 
Bottom line is that, in the interest of safety, unless you really know what you’re doing it would be wise to buy the transformer you need and not try to wind one yourself.

 
 
The following users thanked this post: codingwithethanol

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Transformer Winding
« Reply #10 on: July 15, 2019, 06:35:41 pm »
...The input voltage on the primary winding is first ‘converted’ and stored as a magnetic field in the core. Next, as this magnetic field collapses, the magnetic energy is ‘converted’ or induces a voltage into the secondary winding where the turns ratio of primary to secondary determines the output voltage.
...

This is a common misconception: the energy delivered to the secondary from the primary in a true transformer is not actually stored in the core, as the flux from the secondary amp*turns exactly cancels out the flux from the primary amp*turns. Some energy does get stored in the core as a result of magnetizing it, but this energy is automatically returned to the supply every cycle and doesn't cause a net power draw, anyway, (except for various losses) because the magnetizing current is phase shifted from the supply voltage by 90 degrees (it is purely reactive, in other words).

One prominent exception is the flyback transformer - energy from the primary is stored in the core before transferring to the secondary - because the primary and secondary don't conduct at the same time. As a result, flyback transformers have to be significantly larger to transfer a given amount of power than their conventional counterparts. EDIT - and they need an air gap, which does most of the actual energy storage, oddly enough.

Another exception is the transformer in an LLC converter, which both stores energy and acts as a true transformer, depending on conditions at any given time, but this - and other resonant mode topologies - is a bit beyond what the OP was asking about...  :P

« Last Edit: July 15, 2019, 07:57:02 pm by MagicSmoker »
 
The following users thanked this post: soldar, codingwithethanol

Online tautech

  • Super Contributor
  • ***
  • Posts: 29811
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Transformer Winding
« Reply #11 on: July 15, 2019, 08:36:48 pm »
...........................

Both mention https://ludens.cl/Electron/trafos/trafos.html
This ^^^
OP, to get a good idea of what's involved study it.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 
The following users thanked this post: codingwithethanol

Offline bob91343

  • Super Contributor
  • ***
  • Posts: 2675
  • Country: us
Re: Transformer Winding
« Reply #12 on: July 15, 2019, 08:57:10 pm »
In order to design a transformer one needs to figure the flux density first.  Looking in catalogs of core materials one sees that a particular material saturates at some specified density.  So you pick a density considerably lower than that.  Reason: if you approach saturation, the inductance drops and the input current climbs rapidly.

Once you decide on a density, allowing for normal variations in materials and input voltage, you can select a primary number of turns for a given cross section area of core.  That will insure that you can apply a given voltage at a given frequency without excessive primary current.  If you choose a small core, you will need more turns of wire, which takes up more space.  If you use finer wire you can't handle as much current.  So the cross section area of the core has a major influence on the amount of power you can transform.  In addition, a particular core provides a window area to accommodate the windings.  Again, too small a core gives insufficient room for enough wire.

An optimum design allocates about half the window to the primary and half to the secondary.

So we pick a core, decide how many primary turns we need to keep the flux density low enough and still not take up more than half the window.  So that determines wire size.  If that results in 100 turns, then the secondary can be the same (1:1 transformer) or more or less depending on the ratio we want.  Fewer secondary turns allows for larger wire size, etc.

In general, a conservative design runs perhaps 500 circular mils of wire per Ampere.  That keeps the copper losses low and avoids a hot transformer.

There are other factors one needs to take into consideration, such as insulation.  On an iron core for typical power line transformers, it's not important to make a neat winding other than making it space efficient.

I could ramble on.  I have designed and built many transformers over many years and am aware of the pitfalls and tradeoffs.  I don't intend for this message to be taken too literally as a design guide but more as an introduction to the subject, allowing the reader to decide how to proceed.  The literature is out there, and has been out there for many years.  Yes it's theoretical, but it has to be for a safe and efficient design.

In other words, it's not for beginners.
 
The following users thanked this post: codingwithethanol


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf