Xicon small transformers inductance

[rankot]

I am puzzled with this for a long time, mainly because there is no trustworthy data - what is the inductance of small Xicon audio transformers sold by Mouser? I have came across this table somewhere on the web:

42TL022 1.5K Ohms .56 Henry
42TL021 4K Ohms 1.5 Henry
42TL018 7K Ohms 2.6 Henry
42TL019 10K Ohms 3.75 Henry
42TL025 17K Ohms 6.4 Henry
42TL017 20K Ohms 7.5 Henry

But I have one 42TL021-R and when I measure it with my GM328A transistor tester / LCR meter, I get completely different values:
6.11H / 162.8 Ohm primary (670mH / 82 Ohm Ct)
146mH / 67 Ohm secondary (inductance below measurement range / 33 Ohm Ct).

Does anyone here have any of those transformers and accurate inductance meter, so he can confirm any of those values?

[Rob Strand]

6.11H / 162.8 Ohm primary (670mH / 82 Ohm Ct)
146mH / 67 Ohm secondary (inductance below measurement range / 33 Ohm Ct).

web post = 42TL021 4K Ohms 1.5 Henry

From the specs:
42TL021-RC, Rp=148, Rs = 64, Np = 1320, Ns = 440

So the DC resistances Rp and Rs are close.

From the specs we see Np/Ns = 3.  If we take your secondary inductance measurement Ls = 146mH and the turns ratio we would predict a primary inductance of Lp_est = 146 * 3*3 = 1.31H.  So the estimate isn't far off the 1.5H value from the web.

We know the turns ratio must be 2:1 on the primary, we don't have to rely on the spec.  If we look at your 6.11H primary measurement we would predict a centre-tap measurement of 6.11H/4 = 1.53H. That does not agree with the measured value of 670mH.  We could also predict the end-to-end inductance of the primary of 4 * 670mH = 2.7H.   

The main point is all the measurements are inconsistent with the turns ratios.  When I see that I suspect something is wrong.

Some theories:
- The high DC resistances are affecting the inductance measurement.   While this can occur on cheaper meters i wouldn't expect a problem on the primary (as the inductive reactance XL is much larger than the winding resistance).  The secondary could have a small error.   That's the opposite of what we see.

- The inductance depends on the drive level and drive level on the meter is affecting the results.

At this point I'd probably set-up at 1kHz sine-wave generator to produce 1Vrms; measured with your DMM.
You could then use a circuit like figure 2, then measure the output voltage and calculate predict the inductance. Maybe use R = 47k to 100k.  (There's many ways to do this)
http://physics.indiana.edu/~urheim/p109s13/p109_manual_2013S/13_Low_high_passFilter.pdf

You can check the set-up by putting a known capacitance (say 22n to 100n) in there in see you if can get the correct results.

EDIT: The calculations are,

XL  = XC  = R / sqrt ( (Vi/Vo) ^ 2 - 1)
L  =  XL/ (2 * pi * f_test)
C = 1/(2 * pi * f_test * XC)

You want Vo to be small but not so small that the meter doesnt' have many digits in the measurement.
Feel free to play with R and the drive level.

[Rob Strand]

You might want to check out the graph on page 7, showing how inductance can vary with drive level.  Fairly normal for un-gapped iron cores:
http://www.tsf-radio.org/forum/im/257734ransformer_measurements_and_modeling.pdf

[R.G.]

Finding THE inductance of an ungapped E-I core is a lot like finding the hfe of a bipolar transistor. There ain't one.

What there is varies. Ferromagnetic materials - like transformer iron - have an incremental permeability (that's how "inductability" is measured) that varies with the amount of flux in the iron. The incremental permeability is the slope of the B-H curve, and for linear irons, this is a slanted S curve. The slope varies. Incremental permeability also varies with the magnetic history of the iron. If it's been fed a bunch of DC, it comes to rest at a different point in the B-H curve than if it's been fed a big, then slowly decreasing AC sine wave that gently leaves it at B=0, H=0, or just congealed by cooling slowly from its Curie temperature.

And it varies with construction. The permeability of good, linear(ish) tranformer iron is 10,000 to 15,000 times better than the permeabilty of free space. That means that if the magnetic flux has to flow through 1/10,000 as much air gap as it does iron path, then the effect of the air and the iron are of about equal influence on the measured inductance. In fact, this is how you make a fixed-value inductor - you use the magnetic material to channel the M-field through a specific part of air-filled three-space, and make the empty three-space be what determines the inductance, not the flakey iron. And given that the air gap being tiny is important, how well-squeezed-together the E-I laminations are makes a big difference in the measured inductance.

What's a transformer designer to do??
Well, the smart ones specify the frequency response of their transformers, not the inductances. And they specify that only as a minimum, telling the buyer you'll get at least this much, but maybe more. They then design the trannie so that the shoddiest, sorriest, worst example that the manufacturer will let slip through will be that minimally good. The best ones may be way better.

So if you want a specific value of inductance, don't use transformer windings to do it. If you just want a biggish inductor that's probably enough, they are a GREAT value.

[rankot]

I'd like to build my own inductor, but 1.5H is way too big w/o core, so I wanted to use those transformers.

[Rob Strand]

I'd like to build my own inductor, but 1.5H is way too big w/o core, so I wanted to use those transformers

The issue you are seeing is the inductance is varying with different drive levels.     The equipment you are testing the inductors with is measuring the inductance at different drive levels so you are seeing different (or inconsistent) inductance measurements.    In the real circuit you might find the drive level is fairly low.  If you measure the inductance at similar drive levels you will get a measurement which is closer to what the circuit will see.

The other way to look at it is if someone has specified a  42TL021-R in a circuit then it doesn't matter what inductance is written on the schematic, what the actual value is, or what crazy behavior it has.   The designer chose that so it should be OK (putting the wide tolerance aside).

If you want to make a large value inductor which varies less with current you would probably need to use a high mu Ferrite core.  To avoid it being too large and not having to wind too many turns it would have to be gapless ( cores have an effective small gap).    It will have a wide tolerance.    You will find "linear" ferrites particular the ferrites suited to filter applications will give workable solution and will have a region where the inductance doesn't vary too much with current.     The ultimate size depends on the maximum current (and resistance)

(If you want to decrease tolerances you need to get a gapped core.  It will need a lot more turns.)

EDIT:
Look here, p23, for an example
https://www.americanradiohistory.com/Archive-Electronics-Today/70s/Electronics-Today-1975-01.pdf

[rankot]

XL  = XC  = R / sqrt ( (Vi/Vo) ^ 2 - 1)
L  =  XL/ (2 * pi * f_test)
C = 1/(2 * pi * f_test * XC)

You want Vo to be small but not so small that the meter doesnt' have many digits in the measurement.
Feel free to play with R and the drive level.

I tried to measure small 600:600 Ohm isolation transformers I have at hand.

With 1kHz signal and R=100k (using figure 2 from above mentioned article - High pass RL filter), I get this result:
Vi = 0.896V (dropped from 1V when nothing connected)
Vo = 0.010V
X_L = 1116 Ohm

It shall be 600 Ohm, if I get this well, but is almost double?

I tried to measure with 100n cap instead of inductor and the measurements are pretty correct (I get 130n for 100n cap).

Measuring Xicon 42TL gave 660 Ohm: 15.5k.

So it seems that measurements are correct, now I must sort little TFs I have.

[Rob Strand]

With 1kHz signal and R=100k (using figure 2 from above mentioned article - High pass RL filter), I get this result:
Vi = 0.896V (dropped from 1V when nothing connected)
Vo = 0.010V
XL = 1116 Ohm

It shall be 600 Ohm, if I get this well, but is almost double?

The XL value does not need to be 600ohms.   For example, if you measured at lower frequency XL will turn out lower and if you measured at a higher frequency XL with turn out higher.     XL is roughly proportional to frequency.   If you have a small 600ohm transformer rated down to 300Hz  you might see XL somewhere around 300ohm (half of 600ohms).  For example if we scale your XL(1kHz) =1116 ohms down to 300Hz, XL(300Hz) ~ 1116 * 300 / 1000 =335 ohms.  So it doesn't look bad.   If XL at 300Hz was larger than 335ohm I still wouldn't be too worried.

Vo = 0.010V

As a check you could decrease the 100k until you get say 0.05 to 0.1V and see if you get similar XL values.  Some DMM's, especially the true RMS ones can have larger errors at voltages less that 1/10th the full scale voltage.

Measuring Xicon 42TL gave 660 Ohm: 15.5k.

So it seems that measurements are correct, now I must sort little TFs I have.

Your measurements work out at L = 2.5H.

FYI, I took all the web measurements and the turns from the 42TLxxx datasheet
then computed the average AL factor for the core.  There's no info on how the
web measurements were made.

             Web                            |    Spec    Est
                                                |     N       AL
                                                |               (nH/t^2)
42TL022 1.5K Ohms 0.56 Henry   |   930    647
42TL021 4K Ohms 1.5 Henry   |   1320     861
42TL018 7K Ohms 2.6 Henry   |   1720  879
42TL019 10K Ohms 3.75 Henry   |   2000  938
42TL025 17K Ohms 6.4 Henry   |   2800  816
42TL017 20K Ohms 7.5 Henry   |   2840  930

All
   AL_av =  845 nH/t^2

Removing 42TL022

   AL_av = 885 nH/t^2

That predicts about 1.54H (L= AL * N^2) for the 42TL021.

For a gapless core, not bad compared to the 2.5H you measured.
Anyway your measurements are probably a better estimate for that transformer.

BTW, the method I gave to measure the inductance is OK to get ball-park estimates.
There's a number of areas where small errors can creep in.   The DMM accuracy would
be a contributor.

Vi = 0.896V (dropped from 1V when nothing connected)

Another source of error is loading.  Like if your DMM (or CRO) has a 1M input impedance
you need to factor in the effect of loading.

The fact the inductance varies with drive level makes life difficult coming up with a single inductance value.
If you are interested, you could check at a few drive levels.  If your meter isn't accurate down to low AC levels it might be hard to know if the meter is causing the change or the inductance drive level!

[Esppse]

I've been using this chart for a while.

http://guitarnuts2.proboards.com/thread/7477/gibson-varitone?page=2

They discontinued a lot of these, I was hoping to get more to screw around with. BTW dumb question, are the Xicon inductors polarized like a guitar pickup? Or does it not matter when building a varitone/midrange cut control?

[Rob Strand]

I've been using this chart for a while.

Very cool thanks.  I'll have a look at the details of that later.

In that thread someone quoted:

- 42TL017: Primary 10.40H-10.96H
- 42TL021: Primary 2.140H-2.191H-2.354H

So maybe the big spread sheet has some drive level issues as well.

BTW dumb question, are the Xicon inductors polarized like a guitar pickup?

Used as an inductor they don't have a polarity.     
Used as a transformer the input and outputs will have a a relative polarity.
If the inductor is placed in an *external* magnetic field it will have polarity.