Transformers??

In a nonresonant transformer, K < 1. Assuming the meter isn't broken, you're perhaps seeing the small-field nonlinearity of the core. Putting the windings in series-aiding increases the flux density in the core, maybe by enough to get you over the sticky spot at zero.

Cheers

Phil Hobbs

What frequency of test?

Not sure what the "student" is to draw from this; perhaps T1 is bifilar or heavily interleaved, while T2 is side-by-side / banked?

If T1 has more capacitance, and the test frequency is unlucky, it could account for the apparently high inductance (some reactance canceled with capacitance --> impedance is higher --> measured inductance is higher).

The important thing to remember about components is, they aren't. A resistor is only a resistor over a range. A capacitor is only a capacitor over a range. And so on. Lumped equivalent models or transmission lines give better agreement, but even those get complicated quite quickly, and don't account for certain physical phenomena very well.

So, perhaps T1 isn't a transformer in this range. :)

Tim

-- Seven Transistor Labs Electrical Engineering Consultation Website:

Check; in a LINEAR (we do not know if that is the case here), NON-resonant (we do know that is the case here) K

• BINGO!

• "test frequency" is not known via this "reliable" meter.

• Nonsense.

Isn't this the same meter that measured 20mH on your planar coil and in the circuit AND in the active test circuit you made up appears to be more like 11mH?

From your readings the 'scale' appears to be a bit wonky.

Not sure about how you derived all the numbers at the end of each line, but by definition the the coupling of the MUTUAL inductance is 1.

Given your four measurements you can indeed find Lp leakage, Mp mutual as seen from primary, Ls leakage[as seen from secondary], Ms [as seen from secondary], AND 1:a the turns ratio. yes, five 'unknowns' from four data points BECAUSE Mp and MS are related by turns ratio giving you an extra equation, so you really had five equations and five unknowns, thus you can get there from here.

Using the last line of data for T2, shows that T2 is wound 1:1, with what appears as a primary inside bobbin [low leakage] and a secondary outside bobbin [much higher leakage: Lp = 0.434mH, Ls = 3.094mH, Mutual is

In LTspice, T2 modeled with these parametrs absolutly recreatee your last line of data.

HOWEVER, T1 is NOT wound 1:1 and I could find no physically realizable way to reconcile that next to last data line. It appears the ratio was on the order of around 1:1.2 or as high as 1:1.25, but the leakage inductances just don't work out right.

So again I don't trust that meter. go back and put T1 in a 'proper' test circuit and measure the parameters with individual instruments you can trust.

Here is the series of equations to find all the parameters using your four measurements: four measurements: Lp+Mp=26.42mH Ls+Ms=26.43mH Lp+Mp+Ms/a + Ls + Ms +a*Mp=134.2mH Lp+Mp-Ms/a + Ls + Ms -a*Mp= 26.94uH the turns ratio is 1:a

note Mp=a^2*Ms therefore let Mp=M and Ms=a^2*M

now you again have four equations and 4 unknowns Lp+M=26.42mH Ls+a^2*M=26.43mH Lp+M+a*M + Ls + a^2*M +a*M = 134.2mH Lp+M-a*M + Ls + a^2*M - a*M = 26.94uH

you can't get there using that data. what's wrong? don't know, could be shorted turn, but be far more informative to measure T1 in an ACTUAL circuit over a decent bandwidth.

Also, I'll bet using your OWN circuit to measure, you will get repeated data! Unless you stored the trannies in the fridge and it took time to warm them up.

To avoid Aioe from getting upset about the number of lines in this posting, I'm going to post the 'test' circuits in another response.

Here is the test circuit showing all four configurations to demonstrate T2 as the data set in the last line:

Version 4 SHEET 1 27404 996 WIRE 624 -352 464 -352 WIRE 848 -352 704 -352 WIRE 1088 -352 912 -352 WIRE 1312 -352 1168 -352 WIRE 464 -304 464 -352 WIRE 848 -304 848 -352 WIRE 912 -304 912 -352 WIRE 1312 -304 1312 -352 WIRE 464 -176 464 -224 WIRE 848 -176 848 -224 WIRE 912 -176 912 -224 WIRE 1312 -176 1312 -224 WIRE 624 -16 464 -16 WIRE 848 -16 704 -16 WIRE 1088 -16 912 -16 WIRE 1312 -16 1168 -16 WIRE 464 32 464 -16 WIRE 848 32 848 -16 WIRE 912 32 912 -16 WIRE 1312 32 1312 -16 WIRE 464 160 464 112 WIRE 848 160 848 112 WIRE 912 160 912 112 WIRE 1312 160 1312 112 WIRE 624 384 464 384 WIRE 848 384 704 384 WIRE 1088 384 912 384 WIRE 1312 384 1168 384 WIRE 464 432 464 384 WIRE 848 432 848 384 WIRE 912 432 912 384 WIRE 464 560 464 512 WIRE 912 560 912 512 WIRE 848 592 848 512 WIRE 1312 592 1312 384 WIRE 1312 592 848 592 WIRE 624 784 464 784 WIRE 848 784 704 784 WIRE 1088 784 912 784 WIRE 1312 784 1168 784 WIRE 464 832 464 784 WIRE 848 832 848 784 WIRE 912 832 912 784 WIRE 464 960 464 912 WIRE 848 960 848 912 WIRE 912 960 912 912 WIRE 912 960 848 960 WIRE 1312 960 1312 784 FLAG 464 -176 0 FLAG 848 -176 0 FLAG 912 -176 0 FLAG 1312 -176 0 FLAG 1312 160 0 FLAG 848 160 0 FLAG 912 160 0 FLAG 464 160 0 FLAG 464 560 0 FLAG 912 560 0 FLAG 464 960 0 FLAG 1312 960 0 SYMBOL ind 608 -336 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Lp1 SYMATTR Value {Lp} SYMBOL ind 1072 -336 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Ls1 SYMATTR Value {Ls} SYMBOL ind2 832 -320 R0 WINDOW 0 -57 34 Left 2 WINDOW 3 -70 66 Left 2 SYMATTR InstName LMp1 SYMATTR Value {LMp} SYMATTR Type ind SYMBOL ind2 928 -320 M0 WINDOW 0 -52 36 Left 2 WINDOW 3 -67 70 Left 2 SYMATTR InstName LMs1 SYMATTR Value {LMs} SYMATTR Type ind SYMBOL voltage 464 -320 R0 WINDOW 3 24 96 Invisible 2 WINDOW 123 24 44 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value "" SYMATTR Value2 AC 1 SYMATTR InstName V1 SYMBOL res 1296 -320 R0 SYMATTR InstName R1 SYMATTR Value 10MEG SYMBOL ind 608 0 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Lp2 SYMATTR Value {Lp} SYMBOL ind 1072 0 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Ls2 SYMATTR Value {Ls} SYMBOL ind2 832 16 R0 WINDOW 0 -57 34 Left 2 WINDOW 3 -70 66 Left 2 SYMATTR InstName LMp2 SYMATTR Value {LMp} SYMATTR Type ind SYMBOL ind2 928 16 M0 WINDOW 0 -52 36 Left 2 WINDOW 3 -67 70 Left 2 SYMATTR InstName LMs2 SYMATTR Value {LMs} SYMATTR Type ind SYMBOL voltage 1312 16 R0 WINDOW 3 24 96 Invisible 2 WINDOW 123 24 44 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value "" SYMATTR Value2 AC 1 SYMATTR InstName V2 SYMBOL res 448 16 R0 SYMATTR InstName R2 SYMATTR Value 10MEG SYMBOL ind 608 400 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Lp3 SYMATTR Value {Lp} SYMBOL ind 1072 400 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Ls3 SYMATTR Value {Ls} SYMBOL ind2 832 416 R0 WINDOW 0 -57 34 Left 2 WINDOW 3 -70 66 Left 2 SYMATTR InstName LMp3 SYMATTR Value {LMp} SYMATTR Type ind SYMBOL ind2 928 416 M0 WINDOW 0 -52 36 Left 2 WINDOW 3 -67 70 Left 2 SYMATTR InstName LMs3 SYMATTR Value {LMs} SYMATTR Type ind SYMBOL voltage 464 416 R0 WINDOW 3 24 96 Invisible 2 WINDOW 123 24 44 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value "" SYMATTR Value2 AC 1 SYMATTR InstName V3 SYMBOL ind 608 800 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Lp4 SYMATTR Value {Lp} SYMBOL ind 1072 800 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 5 56 VBottom 2 SYMATTR InstName Ls4 SYMATTR Value {Ls} SYMBOL ind2 832 816 R0 WINDOW 0 -57 34 Left 2 WINDOW 3 -70 66 Left 2 SYMATTR InstName LMp4 SYMATTR Value {LMp} SYMATTR Type ind SYMBOL ind2 928 816 M0 WINDOW 0 -52 36 Left 2 WINDOW 3 -67 70 Left 2 SYMATTR InstName LMs4 SYMATTR Value {LMs} SYMATTR Type ind SYMBOL voltage 464 816 R0 WINDOW 3 24 96 Invisible 2 WINDOW 123 24 44 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value "" SYMATTR Value2 AC 1 SYMATTR InstName V4 TEXT 864 -424 Left 2 ;1:a TEXT 776 -392 Left 2 !k1 LMp1 LMs1 1 TEXT 1656 -344 Left 2 !.param a=1 M=24.716mH\n.param Lp=0.434mH LMp={M}\n.param Ls=3.094mH LMs={a*a*M} TEXT 776 -464 Left 2 !.ac dec 100 100 10kHz TEXT 864 -88 Left 2 ;1:a TEXT 776 -56 Left 2 !k2 LMp2 LMs2 1 TEXT 848 312 Left 2 ;1:a TEXT 776 344 Left 2 !k3 LMp3 LMs3 1 TEXT 864 712 Left 2 ;1:a TEXT 776 744 Left 2 !k4 LMp4 LMs4 1 TEXT 256 -272 Left 2 ;Primary\n25.15mH TEXT 248 64 Left 2 ;Secondary\n27.81mH TEXT 280 464 Left 2 ;Adding\n93.39mH TEXT 256 856 Left 2 ;Subtracting\n3.527mH TEXT 184 -392 Left 2 ;inductance = 1V/(I(Lp1)/(2*pi*freuency) TEXT 1232 -48 Left 2 ;inductance = 1V/(I(Ls2)/(2*pi*freuency) TEXT 208 344 Left 2 ;inductance = 1V/(I(Lp3)/(2*pi*freuency) TEXT 192 744 Left 2 ;inductance = 1V/(I(Lp4)/(2*pi*freuency)

• 1) Those numbers are *calculated* from measured values: M=(L aiding-L opposing)/4, then K=M/(sqrt(L1*L2). 2) you have your so-called definition all wrong. Assume one has a "perfect" RTTY bifilar-wound toroid (what i had for T1 was cheaply wound), L1=22mH, L2=22mH, L series aiding=88mH, L series opposing=0. That would give one an M=22mH and then K=1. You DO remember RTTY, do you not? Famous for using 22/88mH inductors like this. Maybe way before our time?

• NOPE! *NO* "bobbin". It is a toroid, pri on "left" half,sec on "right" half. NO "of course", but the intention does strongly exist,and the number of turns are at least rather close to each other. If what you mean by Lp is primary inductance, and Ls is secondary inductance,may i suggest you go back and READ what i measured? BTW, here you contradicted yourself (above: "MUTUAL inductance is 1").

• *WRONG*! T1 is absolutely and exactly wound 1:1; it is BIFILAR and one can SEE the wires on the toroid. As Mr Hobbs indicated, that saturation is likely (and on further thought) resonance may also be possible. Saturation,at least partial, kills the relationship pri-sec.

• 1) that is from T2..NOT T1. 2) Where did this Mp, Ms and "a" come from? Trying to derive "a" for T1 using data for T2 is not too swift...especially one knows a-priori that a==1 in the bifilar-wound toroid.

• Why are errors "ALWAYS" attributed to shorted turns?

• You are nuts! 1) I used the factory built AADE meter as indicated, and these toroids have been sitting at room temperature for months (and still are).

RobertMacy wrote: For T2, (from numbers used) which is NOT bifilar wound, one cannot assume that "a" is exactly one, altho it should be close.

Not assumed, as I said there are five equations and five unknowns, so the ratio is indeed 1:1, I did NOT assume, the solution showed it.

I sort of doubt that the AADE meter is going to saturate an inductor of reasonable size. I was talking about the nonlinearity that iirc some cores exhibit at low field. It's sort of like static friction--once you get the domain walls moving, they move more easily, i.e. the permeability goes up when you get away from zero amp-turns. But there are folks round here who know a lot more transformer nitty-gritty than I do.

Cheers

Phil Hobbs

Telephone line loading coils!

Huh? I got stuck at "...your so-called definition all wrong." I'll look at Lptotal related to Lstotal and k=to your equation, I NEVER use that model. I ALWAYS break the model down into more basic components.

Of course you looked at the LTspice schematic to better understand the definitions? where Lp and Ls are the LEAKAGE for the primary and for the secondary and the mutual inductance is just that, the mutual inductance between primary and secondary.

Again, huh? as stated above, the standard use of the terms are shown in the accompanying LTspice schematic. Lp is NOT the whole primary inductance, just the primry leakage. AND, the data is equivalent to the model values. Now the question is why are they different? I don't know, perhaps the secondary has a thick isolation layer between the core and the winding. Or there's an error in the data. Can't tell unless you actually make some actual measurements. Like you did with that planar coil where the ACTUAL value was around 11mH, verified in your osc. circuit, too, NOT the 20mH as indicated on your meter.

Does not matter if the transformer was made on a coat hangar wire. The model is based upon the data and the data makes little sense.

As I said earlier, isn't that the same meter that measured 20mH for the inductor that actually measured 11mH in your test circuit?

I thought you said your meter operates near 1kHz. Resonance is unlikely. saturation more likely as volt-sec could be violated. All the better to make your OWN test circuit to SEE what's going on. you can get frequency, you can see if the volt-sec of the core is exceeded - a lot more trustworty information.

I thought I used the data from T1 in those equations

quoting the next to last line: "T1 | 26.42 | 26.43 | 134.2 | 26.96u | 33.54 | 1.27 |" you're right I mis-typed 26.94u should be 26.96u But even with that small change there is no physically realizable solution to model that transformer.

I'll avoid humour in the future. {well, maybe more like 'attempts' at humour]

Then you tell me, why are your readings all over the place? Non repeatable readings SHOULD be a major wake up call that something is not right.

Why shorted turns? because the way I test and work with transformers, [actual circuitry] I already KNOW that there is NO DC bias, KNOW that there is NO saturation, KNOW how close to resonance I am, and KNOW that the coercivity is being exceeded enough to at least turn on the core. so the only thing really left from getting wonky data comes from measurement error or shorted turns. True I jumped to the conclusion based upon my experience.

I forgot, the way you're doing it has far more unknowns.

PS: the characterisics in the data set for T1 are supportable *if* you are saturating your core.

I didn't think it would either. But coercivity...single coil not quite turn on the core. double coils [adding] turn it on makes the inductance look higher than 4X. And bifilar wound means the double coils [anti-phase] yield almost zero inductance. Yeah, that makes sense.

It's just that bifilar is rarely used on a power core which must sustain AC mains voltages and you can live with higher coercivity cores [they're always ON]. Hmmm, not likely a common mode choke as in an RFI/EMI filter either.

I keep coming back to putting that trannie in a circuit so you can 'see' what's going on.

Really? Did not know that. Thanks.

• I said no such thing; Do not know how the AADE makes its inductance measurements.

• That is also my guess.

• Did i or did i not call my AADE "reliable"? The "repeatability" of the measurements shw how (un)reliable it is.

• Not for power; it is from the RTTY daze..

Didn't soak in until now. duh! I used to design a lot of that equipment too.

looks like a VERY well made, bifilar wound common mode choke. From memory the old stuff used to have +/-20V compliance to drive a maximum of 20 mA in those old current loops. Why high coercivity was allowed in the core still eludes me. High coercivity core would really weaken inductance value as the net field would almost always center around zero.

Again to be certain, put into an 'active' circuit. You can use the soundcard as transmit AND receive then sort out what's going on over some band with a 'free' and accurate instrument.

Thanks for posting. I now have a reason to compare several models. The first is the one I use the most: primary leakage in series with mutual, and secondary is the mutual and secondary leakage. Like to compare to the model using Lptotal and Lstotal and some coupling factor, k where k is that mutual/sqrt(Lstotal*Lptotal). I've always wondered how well that works. Then compare after adding the material's parallel resistance representing core losses. I think I checked once and found they're identical, but will check again now.

Finally realized a major source of confusion. Your model consists of two inductors with a 'magical' coupling coefficient that yields each winding's leakage inductance AND the windings ratio! My model consists of turns ratio, primary leakage, secondary leakage, and a mutual inductance with k=1. I use this model to gain control of what's going on over a wide spectrum. Each of the components in my model I understand their origin, etc. where as with your 'simpler' model, not sure how to do. [yes, I'm dense.]

So...I just ran comparisons of those two modeling techniques and found that at low frequencies they are NOT identical! amplitude is within numerical accuracy the same, but the phase shift is different. only slightly,, but different. And is worst where the windings subtract.

I then used the two types of models to compare an actual implementation of a 600 ohm 1:1 telephone transformer by Stancor [that little cube] and found that within numerical accuracy the two models yielded identical answers.

Interesting, again thanks for posting.

I used one to replace the audio output/modulation transformer in a CB radio, back 40+ years ago. I wound a secondary to drive a speaker, in receive mode.