Finding specifications of a transformer

[DaJMasta]

I bought a lot of a bunch of random parts a while back and ended up with quite a few (200?) small transformers.  I did some looking around and I've been totally unable to find any specifications, which doesn't seem so uncommon for transformers, so how would I best go about characterizing them enough to use them in a design?  I've got some equipment, so I'm confident I have something that could get at least ballpark figures, but I'm not terribly familiar with magnetics and don't really know what the primary stats would be, let alone the best ways to determine them.

The transformer in question is marked 188209-101 Rev 5 and is made (or repackaged and sold) by Communication Associates with the part number 1487000, I've tried contacting them for any information, but haven't heard anything.  It's a smallish transformer that's probably too big for any kind of signal coupling, and the one bit of packaging I have describes it as a "switcher", so I assume it's designed for use in lowish current SMPS.  There are leads on either side, two isolated coils on each, and they've got to have different numbers of turns, measuring 0.5 ohm, 0.3 ohm on one side and then 0.3 ohm, 0.2 ohm on the other with what looks like about the same gauge wire.

Should I start by sacrificing one and just counting the turns?  Is injecting a signal a better approach (input to output voltage ratio to figure turns?)  When I inject a signal, should I use additional diode protection?  I assume the frequency is 10s of kHz to a MHz or so if it's used in a switcher with a physical size like this, so is the best way to figure out what it's been optimized for to just sweep the input frequency and look for the highest output amplitude?  Is there anything to look for in terms of voltage isolation if it turns out to be designed for higher voltage use?

Should be an interesting exploration of transformers for me, in any case, but if there are more pertinent details or better ways to find things, I'm all ears.  I've got a sig gen, a scope, a good multimeter, an LCR meter, etc - probably enough equipment to get the information that I'm looking for.

[Jwillis]

If you have a signal generator you can find the ratio of the winding's .First identify which pegs have continuity and mark them. The higher resistance is usually the  primary. For laminates test frequencies should be between 50 Hz and 500 Hz .Ferrites are a little more involved because it depends on the permeability of the core.But you can test those between 500Hz to as high as 30kHz.
Laminates are geneally used at 50 - 60 Hz and ferrites are used at higher frequencies for switching power supplies.If the wires are very fine then the transformer is likely a signal transformer. 
Find out with a DMM what the voltage is on your signal generator.

For example mine is  6.5 volts AC at full amplitude of 20Vpp

Set wave to a sine wave at connect your generator to a primary side of transformer and adjust the frequency until you get the best output voltage at the secondary side.

Say the reading is 0.4 volts AC at the secondary side and your generator outputs 6.5 volts ac on the primary side . So the ratio is 6.5 : 0.4     Divide 1 by 0.4  that gives you 2.5 as a multiplier for the other part of the ratio  so your ratio becomes 16.25:1 This will give you an idea of the voltage output on the secondary based on what goes in the primary.

For example 120 volts divided by 16.25 equals around 7 volts on the secondary side.

You can download a free program called Xformer Designer .Handy little program. To get the core area measure the width and length at the center of the bobbin. plug that into XD .Plug in a voltage for the primary of say 120 volts ac (this may be different where ever you live.).Plug in the best frequency you got for ferrites or 50 - 60 Hz for laminates.Use 1 flux for most laminates and  1.5 -4 for ferrites.
Hit calculate and this program gives you a close idea of how many turns for that coil and the Maximum VA of the core.Generally you want to half the VA .
To find the Amps at the secondary ,plug in the secondary voltage for example from above 7 volts. Hit calculate .Notice the VA of 12.05 VA remains the same but the turns change.
The current I in amps is equal to the apparent power S in volt-amps, divided by the voltage V in volts:

I(A) = S(VA) / V(V)      I(A)= 12.05 / 7   I(A) = 1.7 amps  MAXIMUM   Divide that by 2

Although this gives a close idea of an unknown transformer it does work fairly well.I've done this lots and haven't encountered any problems.

Also its interesting to see how transformers of different types respond to varying frequencies .Ferrites don't work well at low frequencies and laminates don't work well at high frequencies.

[tautech]

Obviously a SMPS transformer from the low resistance windings and the ferrite core material.
By virtue of no two windings having the same resistance it won't be used in a center tapped configuration.
My guess (presuming all windings are same wire gauge) is highest value 0.5 ohm is primary and its mate 0.3 is feedback for the controller IC run supply.
The secondaries will be a dual voltage windings more than likely.

However I'd still sacrifice one to check if it closely fits the transformer spec on P13 in the attached pdf.


A configuration like Fig 30 on P13 in the attached pdf is probably a good fit for this transformer.

[Jwillis]

For a SMPS transformer you can find a secondary winding for the high side of the power supply but not always.It depends what the transformer is configured for.Could be a current sense or a feed back .I'm not sure . I still have much to learn .As you probably know the higher the frequency the fewer winding's there will be.I've cracked open a few and found only a  single and even half winding's for a pair of taps,even twisted pairs for four taps .I have a couple with one big winding along with a few others on the secondary side.Probably for high current .Pulled those from a couple 2000w UPS that didn't work.

[DaJMasta]

Finally got back to this after a while and made all sorts of measurements, first DCR with my good meter in 4 wire mode, then L with the LCR meter at several frequencies, then I went through with my sig gen set to 1Vpp on a 50 ohm load and swept through the frequency until I could see it was sweeping down, and it looks like it's shaping up to be a little different than is was first guessed.  Then I sacrificed one to count the turns and look for any other oddities.

The pictured side's red and green conductors I dubbed the primaries for my results, since they had the lowest DCR, but I'm not totally convinced they are, considering that they turned out to be the same number of turns.  The same gauge wire is used in every coil and the E core is definitely a ferrite material.

For the "primaries" I measured a 0.1995 ohms and 104.9uH @100kHz on the red and 0.190 ohms and 95.97 uH on the green, both were 22 turns.  Starting with the red on the left as pin 1 and working around it like a DIP, that means the "secondary" between pins 8 and 6 measured 0.2155 ohms and 154.90 uH and counted 28 turns.  The other "secondary" on pins 5 and 7 measured 0.469 ohms and a whopping 963.2uH @100kHz and counted 70 turns.  The two "primaries" with 22 turns are wrapped outermost, with the 28 turn "secondary" under an layer of tape and the 70 turn coil in the center.

Using that naming scheme, I swept through frequencies, from 100Hz to 4MHz, and found that, surprisingly, they seemed to be pretty high frequency devices.  With the "fixed" output of the sig gen I measured the input on the primary tap being energized and the outputs of both secondaries with my scope using 10M impedance probes and recorded the voltage of all three points.  The graphs are attached, but I found the peak output voltage to be at about 1.6MHz from both primary taps, but with the changing input voltage thanks to the sig gen supplying the device, I'm unsure if this point would actually be the peak of current transfer.  If I take the same data and plot "efficiency" of the secondary voltage divided by the input voltage, the peak is much higher, between 3.25MHz and 3.75MHz.  Assuming the sig gen is doing a good job seeing the 1Vpp on its output terminal, I also worked out the impedance of the "primary" (voltage on the device times 50 ohms) and its peak seems to be just around 1.5MHz.  I don't know how valid or useful the efficiency and impedance calculations are, but conceptually I think they make sense.


Looking at the turns, could this actually be a transformer for a dual supply with the things I assumed are primaries as the secondary side and tying the middle two pins for a 0V point?  That would make it a step down transformer, probably from the 70 turn side with a feedback tap as the 28 turn coil?  Does that really make sense as a design?  Could also explain the color coding on the output wires, perhaps.

[DaJMasta]

Well I think I've got myself looking in a better direction, after counting the turns the disassembled transformer and given the frequency estimates initially, I figured my method of testing was likely flawed, so I decided I'd put a load on the output of the supposed secondaries and try measuring again.  Since the sig gen output is limited in power and because it drives a very low impedance load at some frequencies, I went with a fairly light load: a 1k resistor on each winding.  Then I went through the same methodology as the last time and came up with some more reasonable numbers: an operating frequency range between about 300kHz and 700kHz where the voltages peak on the output and where the ratio of input to output approximately matches the ratio of turns (generally, the 22:28 coupling seems to work more efficiently than the 22:70 coupling, and the primary winding on pins 1 and 3 is slightly more efficient than the one on 2 and 4).

[tautech]

Turns ratio is what I'd take notice of.
Say this is working with switched rectified mains of ~170V for US or ~320V for EU/Asia/Au, well you do the maths.

There's enough info in the switching controller pdf I attached earlier to work out if this transformer suits a CT application but I think based on a total of 44 vs 70 turns I'd suspect 70 turns was the primary and 2 x 22 for a + and - secondary rail.
My 2c.