Question : Profiling an unknown power inductor or power transformer

[BravoV]

I think we've seen too many threads asking identification for an unknown inductor or transformer that people salvaged or found. This makes me curious if there is a way, although limited, to get a certain usable specification out of it using non destructive method ?

Non destructive here means without the need of cracking the core which sometimes is fully epoxied or potted, or need to unwind the wires, all data gathered only through it's termination points and external physical inspections.

Assuming we have these pre-conditions before we start the probing :

Tools :
- A decent LCR meter like those popular handhelds such as MS5308, DE-500 or similar.
- An inductor saturation tester for finding the current saturation point, like this example -> HERE.
- A scope and/or sig-gen, maybe ? 

Inductor/transformer physical informations :
- Ability to measure or at least take an approximation of the wire diameter size from the termination points.
- Winding count and tapped count if any.
- Core's shape/size/volume.
- Winding topology if they're visible from outside like overlapped, stacked windings and etc ?
- Anything else that I missed regarding the needed info from the physical inspection on the inductor/transformer ?

Understand that there are limits on inductor specifications that we can gathered for example like the important factor isolation and creepage limit and etc.

Questions :

Aware the outcome definitely will not 100% accurate, is this doable ?
Can we somehow get "good enough" specifications to re-use the inductor/transformer later for one off project if needed someday ?

Of course the user also required to be aware of the risks involved like lacking of the isolation limitation info and etc. For example its not wise to use in a thousands volt circuit, or for welding (high current) using a small iron core transformer salvaged from a cheap a wall wart AC to DC converter, or a power ferrite inductor from those cheap car DC->USB cellphone converter/charger to be used in 1000 watt power switching circuit. 


PS :
Please see this question in the context of enthusiast/hobbyist perspective, understand that these activity probably will be useless in commercial/pro environment, since it will be much more economical to buy an off shelf ready product which fit in as required by the circuit. But for pack rat like me that love to salvage used inductors/transformers, I will be more happy if I can identify and use them even though not 100% suitable but still acceptable, compared to buying a new inductor even its cheap. Hoping I'm not alone here. 

[Jay_Diddy_B]

Hi,
With an LCR meter and method for determining saturation can you can reverse engineering a transformer.

Turns ratio(s)

Step 1

Locate all the windings using an ohm meter.

Lets say we found two windings.

Measure the inductance of each winding using the LCR meter.

Example:

Winding 1: 22uH

Winding 2: 2.44uH

Since L=N x N x Al (N = number of turns, Al = inductance factor)

Since the inductance ratio = 9:1

The turns ratio = 3:1

We also know that you can not have fractions of a turn, the turns are always integers.

So the transformer must have

3 turns : 1 turn, 6 Turns : 2 turns, 9 Turns to 3 turns etc...

We can get an idea of the number of turns using the inductor saturation tester.

Faraday's Law of Induction: V=N dPhi /dt

Phi = total flux = B x A (Where B = Flux density , A = cross section of the core)

V= N x A x dB/dt

For Ferrite we can assume that the saturation flux density is 0.35 Tesla. (Most power grade ferrites are similar)

The area is the area of the center limb of the core. You can measure this with calipers.

With the Inductor saturation tester you can measure how many volt microseconds it takes to saturate the core. We will measure the 22uH winding.

Let say it takes 200 Vus. 200E-6

Let say the area is 25mm² (5mm x 5mm) 25E-6m²

From V=NA dB/dt

N = Vt /A x dB  = 200E-6 / 25E-6 x 0.35 = 22.8 turns

But from the turns ratio measurement we know that this number is an integer and a multiple of 3.

So the number of turns is either 21:7 or 24:8 (You can not really tell)

If the construction of the transformer allows you add a winding, then you can measure ratio between the winding that you know and the unknown windings.

You now have the turns ratio, the saturation current, the primary and secondary inductance. You can now look at some datasheets from Pulse Electronics or Wurth to find a transformer that is similar to determine the power rating.

Remember in some circuits like Flyback power supplies, you need to identify the phasing, the starts and finishes of the winding)

This best way is to experiment with a transformer of known characteristics.

I hope that this helps.

Jay_Diddy_B

[bingo600]

@JDB

That was a neat explanation 

Even i could understand it

/Bingo

[BravoV]

Thank you so much Jay_Diddy_B ! 

I know this is basic, but for me, this is a really helpful knowledge, printed and stickied in front of my workbench now ! 

If the construction of the transformer allows you add a winding, then you can measure ratio between the winding that you know and the unknown windings.

Does this mean I can get an accurate turn counts on the unknown windings if I can managed to sneak in there my own winding with few turns ? Sorry, example of exact steps if you don't mind.

Remember in some circuits like Flyback power supplies, you need to identify the phasing, the starts and finishes of the winding)

Please explain further about this, I'm not sure I can have a strong grip in this matter.


Again, really appreciate the help given. 

[Jay_Diddy_B]

Hi,
I had a few minutes so I took a transformer that I have in my junk box. I added a winding of 5 turns of 30awg wire-wrapping wire like this:



I now have a winding with a known number of turns.

I now measured the inductance of all the windings:

5 Turn reference winding 3.83uH

Primary 233uH

Secondary 25.9uH

Aux 4.58uH

The Al value of the core can be calculated from the 5 Turn winding

Al = L/N²

= 3.88 /25 = 0.155uH per turn²

For the other windings we have:

Turns = SQRT (L / Al)

Primary Turns = SQRT (233/0.155) = 38.7 (has to be 39)

Secondary Turns = SQRT (25.9/0.155) = 13

Aux Turns = SQRT (4.58 / 0.155) = 5.4  (5 turns)


Turns ratio is Primary : Secondary : Aux = 3 : 1 : 0.4


The phasing is indicated on a transformer with 'spots' like this:



The spots indicate polarity. If the one spot is positive, the other spots will also be positive. This is important in a flyback supply. The transformer is arranged so that when the primary switch is closed the rectifier diode will be reverse biased.



Jay_Diddy_B

[Odysseus]

And don't forget leakage inductance or coupling coefficient, which is very important for simulating switching converters. It's found by first measuring inductance of one winding while the other winding is shorted.  Once L_short is found, then the coupling coefficient, K, is calculated as sqrt(1-L_short/L). Then the uncoupled portion of the winding, L_leak, is (1-k)*L while the coupled portion of the winding, L_mutual is k*L.

See http://en.wikipedia.org/wiki/Leakage_inductance for more details. It also gets more complicated for multiple winding transformers.

[The Electrician]

The responses so far have dealt with ferrite cored switcher transformers.

Several years ago there was a thread on another forum considering this problem with typical small silicon steel transformers:

http://forum.allaboutcircuits.com/showthread.php?t=38273

[BravoV]

Jay_Diddy_B, that is exactly I wanted, great & thank you Sir ! 

The phasing is indicated on a transformer with 'spots' like this:

..<snip>..

The spots indicate polarity. If the one spot is positive, the other spots will also be positive. This is important in a flyback supply. The transformer is arranged so that when the primary switch is closed the rectifier diode will be reverse biased.

Can we use sort of low voltage pulse/signal at the primary while observing the secondary result using scope for identifying the phase ?

And don't forget leakage inductance or coupling coefficient, which is very important for simulating switching converters. It's found by first measuring inductance of one winding while the other winding is shorted.  Once L_short is found, then the coupling coefficient, K, is calculated as sqrt(1-L_short/L). Then the uncoupled portion of the winding, L_leak, is (1-k)*L while the coupled portion of the winding, L_mutual is k*L.

See http://en.wikipedia.org/wiki/Leakage_inductance for more details. It also gets more complicated for multiple winding transformers.

Great to know, thank you ! 

The responses so far have dealt with ferrite cored switcher transformers.

Several years ago there was a thread on another forum considering this problem with typical small silicon steel transformers:

http://forum.allaboutcircuits.com/showthread.php?t=38273

Wow ... huge thread, bookmarked, that will take sometimes to read and digest, many thanks ! 

Yeah, I think with current trend that switching power is dominating the world, while iron core transformer is going to extinct I guess, I think most of the cases here on the salvaged inductors, majority will be ferrite cored type.

[Neilm]

Yeah, I think with current trend that switching power is dominating the world, while iron core transformer is going to extinct I guess, I think most of the cases here on the salvaged inductors, majority will be ferrite cored type.

Iron core transformers will still have their place, but not in small / home applications. They all have "small" written into the design spec and the only way to do that is go for high frequency and that means ferrite.

[Jay_Diddy_B]

Jay_Diddy_B, that is exactly I wanted, great & thank you Sir ! 

The phasing is indicated on a transformer with 'spots' like this:

..<snip>..

The spots indicate polarity. If the one spot is positive, the other spots will also be positive. This is important in a flyback supply. The transformer is arranged so that when the primary switch is closed the rectifier diode will be reverse biased.

Can we use sort of low voltage pulse/signal at the primary while observing the secondary result using scope for identifying the phase ?

Yes, you can use a pulse / function generator and a scope to identify the phasing.

You can also use the inductor saturation tester as the signal generator, and measure the other windings with a scope.

A third method is to use the LCR meter. Put two winding series, if the inductance goes up they are connected in phase, if the inductance goes down they are out of phase. Think of positive and negative  turns.

If I take the example that I measured, which had a turns ratio of 3:1

Primary: 238uH

Secondary: 22uH

In series the inductance will either be 238uH x ((3+1)² / 3²) = 423uH (in phase)

or

238uH  x ((3-1)² / 3²) = 105uH (out of phase)

Jay_Diddy_B

[M0BSW]

Hi,
With an LCR meter and method for determining saturation can you can reverse engineering a transformer.

Turns ratio(s)

Step 1

Locate all the windings using an ohm meter.

Lets say we found two windings.

Measure the inductance of each winding using the LCR meter.

Example:

Winding 1: 22uH

Winding 2: 2.44uH

Since L=N x N x Al (N = number of turns, Al = inductance factor)

Since the inductance ratio = 9:1

The turns ratio = 3:1

We also know that you can not have fractions of a turn, the turns are always integers.

So the transformer must have

3 turns : 1 turn, 6 Turns : 2 turns, 9 Turns to 3 turns etc...

We can get an idea of the number of turns using the inductor saturation tester.

Faraday's Law of Induction: V=N dPhi /dt

Phi = total flux = B x A (Where B = Flux density , A = cross section of the core)

V= N x A x dB/dt

For Ferrite we can assume that the saturation flux density is 0.35 Tesla. (Most power grade ferrites are similar)

The area is the area of the center limb of the core. You can measure this with calipers.

With the Inductor saturation tester you can measure how many volt microseconds it takes to saturate the core. We will measure the 22uH winding.

Let say it takes 200 Vus. 200E-6

Let say the area is 25mm² (5mm x 5mm) 25E-6m²

From V=NA dB/dt

N = Vt /A x dB  = 200E-6 / 25E-6 x 0.35 = 22.8 turns

But from the turns ratio measurement we know that this number is an integer and a multiple of 3.

So the number of turns is either 21:7 or 24:8 (You can not really tell)

If the construction of the transformer allows you add a winding, then you can measure ratio between the winding that you know and the unknown windings.

You now have the turns ratio, the saturation current, the primary and secondary inductance. You can now look at some datasheets from Pulse Electronics or Wurth to find a transformer that is similar to determine the power rating.

Remember in some circuits like Flyback power supplies, you need to identify the phasing, the starts and finishes of the winding)

This best way is to experiment with a transformer of known characteristics.

I hope that this helps.

Jay_Diddy_B

I'm worried , I even understood that explanation

[BravoV]

Yeah, I think with current trend that switching power is dominating the world, while iron core transformer is going to extinct I guess, I think most of the cases here on the salvaged inductors, majority will be ferrite cored type.

Iron core transformers will still have their place, but not in small / home applications. They all have "small" written into the design spec and the only way to do that is go for high frequency and that means ferrite.

Agree, while most iron cores for home appliances, 99.99% are easily identified since their primary as mostly for mains while secondary can be easily spotted by looking at the load circuit or simply measure the output.

[BravoV]

A third method is to use the LCR meter. Put two winding series, if the inductance goes up they are connected in phase, if the inductance goes down they are out of phase. Think of positive and negative  turns.

Simple enough, currently I only have the el-cheapo <$30 L and C meter, still I managed to identify one grabbed from my junk box, great. 

I guess its time to look for a decent LCR meter.

[BravoV]

About current saturation point, as in the saturation tester, if the "knee" bend is easily distinguishable, then its easy, but what if there is no sharp bend and it shows only curve like slope ?

How can we safely decide or take a smart guess at which saturation point is the one rated for the tested winding ?

Example like these below.