I've found (in my limited experience) that relative to the open inductance the leakeage inductance is pretty similar when shorting one or all other windings. Unless the coupling to a winding for some reason is extremely poor, any shorted winding should short the others.
How big a difference are you getting?
Does it matter if you're off by a few percent%
How big a difference are you getting?
Does it matter if you're off by a few percent%
No, that will give the stray inductance of primary to secondary1 in parallel to the stray inductance of secondary1 to secondary2. Leave the other secondary open circuit, Then if you need to measure the stray inductance between the 2 secondaries separately by leaving the primary open circuit. In an actual circuit the coupling between secondaries will have some effect, you may need to calculate the effect but at least the values will have been measured
When I measured the secondaries with the primary and other secondary
shorted then I got approx 4,4uH for one and 4,9uH for the other.
When the other secondary was open then I got 15uH and 17uH.
I measured this at 1MHz. This is not 100% true inductance since the
phase never reached 90deg, the highest I got was 80(@1MHz).
I think the need for shorting or opening the secondary B during
stray inductance measurement of sec. A depends on whether the secondary B
will be switched on or off at the time when the rectifying diode on
sec. A is closing but I have not looked into that more deeply at the moment...
I forgot to mention, that the transformer itself is 2x115v|2x25v 225VA toroid
from Noratel.
shorted then I got approx 4,4uH for one and 4,9uH for the other.
When the other secondary was open then I got 15uH and 17uH.
I measured this at 1MHz. This is not 100% true inductance since the
phase never reached 90deg, the highest I got was 80(@1MHz).
I think the need for shorting or opening the secondary B during
stray inductance measurement of sec. A depends on whether the secondary B
will be switched on or off at the time when the rectifying diode on
sec. A is closing but I have not looked into that more deeply at the moment...
I forgot to mention, that the transformer itself is 2x115v|2x25v 225VA toroid
from Noratel.
Last edited:
A transformer with multiple windings does not have a stray inductance, it has lots of stray inductances. You need to determine exactly what you mean to measure, which is likely to depend on exactly why you want to measure it.
At higher frequencies you are also measuring the effect of stray capacitances too. There will always be resistance as well, from both copper and iron losses.
At higher frequencies you are also measuring the effect of stray capacitances too. There will always be resistance as well, from both copper and iron losses.
Agreed, you need to be specific in what you are looking for.
If your question is "how do I measure impedance in a three winding transformer", there is a very specific procedure and equations in ANSI/IEEE C57. Most of us are familiar with a two winding test procedure, but the 3 winding is more complex. For those of us who perform short circuit studies on 100MVA three winding xfmrs, this is necessary to understand. I can provide you with the specifics if you are serious, but we are only talking about a little 225VA unit.
Don't count on a highly inductive result either; that's reserved for the big boys. A little xfmr like this will actually be mostly resistive.
If you are only interested in the inductance number, you will still have to go through the impedance method. Then again, I may not understand exactly what you are looking for.
If your question is "how do I measure impedance in a three winding transformer", there is a very specific procedure and equations in ANSI/IEEE C57. Most of us are familiar with a two winding test procedure, but the 3 winding is more complex. For those of us who perform short circuit studies on 100MVA three winding xfmrs, this is necessary to understand. I can provide you with the specifics if you are serious, but we are only talking about a little 225VA unit.
Don't count on a highly inductive result either; that's reserved for the big boys. A little xfmr like this will actually be mostly resistive.
If you are only interested in the inductance number, you will still have to go through the impedance method. Then again, I may not understand exactly what you are looking for.
The reason for this was quite simple - to calculate snubber resistors
for the rectifying diodes.
Now after I ran few simulations it appears that it really doesn't matter
that much if the stray inductance is 15nH or 5nH, with the inter-winding
capacitance of ca 65pF the ringing is damped in 2-3 periods in both of
these situations using the same (510ohm) resistor value...
for the rectifying diodes.
Now after I ran few simulations it appears that it really doesn't matter
that much if the stray inductance is 15nH or 5nH, with the inter-winding
capacitance of ca 65pF the ringing is damped in 2-3 periods in both of
these situations using the same (510ohm) resistor value...
It will give you the coupling ratio "K".
K= SQRT(1-LLeakage/LMagnetizing)
More germaine for the folks who design switchers.
I find the following model from Basso to work well and faster than using "K"
.subckt TFR_8K_8 1 2 3 4 Params: Ratio=0.0316
LLeak1 1 20 9e-3
RSeries1 20 30 46
RPri 0 2 1e6
C1 1 3 900e-12
C2 1 2 500e-12
E1 5 4 Value={V(30,2)*Ratio}
G1 30 2 Value={I(VM)*Ratio}
RS_Sec 6 3 1e-6
VM 5 6
.ends
Where Ratio = SQRT Zprimary/Zsecondary
You should read Morgan Jones' article on snubbers in Linear Audio Vol 5. Revealing!
jan
Personally I have found Jim Hagerman's article on snubbers completely useless. You have a bunch of DIY'ers running around mad trying to measure or calculate transformer parameters to determine what value R and C are appropriate. Just throw a scope on it and determine optimal values by experiment. You can get amazing results with a very simple process. No math needed.
Yes, snubbers (in this application) don't have to be exact. All you need to do is add some capacitance (to slow the transient) and resistance (to damp the resonance). In many cases the transformer winding will have enough resistance anyway. Much more important is to keep circuit loops small to reduce induction.
I got some valuable insights from Hagerman's article, most especially that oscillatory ringing occurs because there's an underdamped LC resonant circuit which is stimulated by dI/dt pulses. So you need to snub the LC circuit; you don't need to snub the diode(s). A point that Morgan Jones vigirously reiterated in Linear Audio v5, just last month. Eighteen years after Hagerman.
But I don't bother to measure the transformer L&C, nor do I bother to calculate the snubber component values; I just connect the transformer to my special purpose test jig (link), and adjust the snubber trimpot for optimum damping. It only takes ~ 90 seconds to dial in a snubber that damps out the ringing completely, at the damping factor that I happen to prefer: Zeta=0.8. Now I've got a 2C-1R snubber that optimally damps this particular transformer, for any and all board layouts & rectifier choices.
My method is spelled out in this thread:
http://www.diyaudio.com/forums/tubes-valves/207017-rectifier-injecting-noise-into-heater-supply.html
Start with the smallest C that affects the ring, go with the largest R that snubs it out. A little trial and error and it's done in 20 minutes.
http://www.diyaudio.com/forums/tubes-valves/207017-rectifier-injecting-noise-into-heater-supply.html
Start with the smallest C that affects the ring, go with the largest R that snubs it out. A little trial and error and it's done in 20 minutes.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.