Finding turns ratio of an unknown transformer(small)

I tried simulating the circuit using PSpice.But it showed no oscillating output.
The psipce library(eval.lib) had Q2N6059 power transistor which is rated 100V and 12A.Hence I decided to take input ac voltage source accordingly.
Rather than simulating the entire circuit ,I simulated only the RHS of the circuit.That I assumed that the voltage across 6.8u cap remains almost constant and hence I replaced it with a voltage source of 100V.
Also the lib had a zener of 4.7V.Hence I used it instead of 6.6V.

Also I agrued with myself that the circuit must oscillate without connecting the secondary.As when no load is connected ,it supplies only to the LED ,hence the current drawn is quite less and thus secondary coil effect can be ignored.I may be incorrect on this aspect.I took inductor values L1 and L2(as in previously posted schematic) both as 100u.
But still I tried simulating the circuit with the following assumptions.But I got constant output.
Here is the psipce code:
;
.lib eval.lib
r1 1 4 1Meg
r2 1 2 150k
r3 5 0 1.5
r4 4 7 470k
r5 7 8 15
;
c1 4 7 1n
c2 1 2 3.3n
c3 5 6 47u
;
d1 3 2 D1N4002
d2 6 8 D1N4002
;
x1 3 4 5 Q2N6059
;
dzener 6 4 D1N750
;
l1 1 9 100u
rl1 9 3 0.6
l2 5 8 100u
k1 l1 l2 1
;
v1 1 0 100
.tran 1u 10u
.probe
.end

btw for transient analysis upto 10s too the output remains constant.
Also i assumed primary coil resistance as 0.6Ohms for L1(which I measured with multimeter)

thanks Mr.Nigel .I will go through blocking oscillator(I guess it was used in joule theif) and get back.

Something still looks funky in the schematic near the base drive. Should that point between the two diodes be grounded?

@Mr.RB I rechecked the that point between two diodes(1n4007 and zener anode) but it is not grounded.

savveg. How do you have L1 and L2 polarized in your sim? The end of L1 connected to Q1-C, needs to be in phase with the the end of L2 connected to Q1-E.

If you know the cross sectional area of the wire and you can measure the DC resistance of the windings you can find the wire length using R=pL/A where R is the resistance measured, p = resistivity of copper (eg. ohm-cm units), A is the cross sectional area and L is the lenght. You then might be able to estimate the number of turns.

When measuring the DC resistance use the Kelvin method. Use a resistor in series with a power supply to set the current and measure the voltage across the resistor and the voltage across the winding. Get the current flowing through the circuit and the voltage across the winding and computer R. You won't have to worry about meter effects on the measurement.

You could try a safe current and have enough voltage resolution.

Use a wire guage table and a dial caliper to get an idea of the diameter of the wire and get from the wire guage table or computer the cross-sectional area (Area of a circle with a wire diameter d).

savvej said:

@Mr.RB I rechecked the that point between two diodes(1n4007 and zener anode) but it is not grounded.

Click to expand...

Something doesn't seem right. There is an emitter resistor that is typically used to reduce Q1 base biasing if current through Q1 is too high, but your entire base drive circuit is floating and not connected to the other side of that resistor in any way, making it useless and redundant.

I have a feeling some part of that base drive circuit should be connected to "ground" ie to the bottom of that emitter resistor.

ChrisP58 said:

savveg. How do you have L1 and L2 polarized in your sim? The end of L1 connected to Q1-C, needs to be in phase with the the end of L2 connected to Q1-E.

Click to expand...

I changed it accordingly but still no good.Is the secondary o/p coil(with LED -R load) necessary for the circuit to oscillate?

savvej said:

I changed it accordingly but still no good.Is the secondary o/p coil(with LED -R load) necessary for the circuit to oscillate?

Click to expand...

Quite possibly, switchmode PSU's commonly don't work without a load.

However, as you've no way of accurately simulating the transformer (or even finding out it's characteristics) it's not surprising it won't work.

Simulators aren't real life, and they often don't work as real life does.

The attached LT spice circuit follows the schematic in post 19. Some components have been changed to match those included in the standard LT distribution. Inductance values are just what I found would make it start.

It does oscillate, but does not reach a point of regulation. It stops running at about 3.5mS. Probing around shows the nominal DC point at the base of Q1 goes to low to keep it running. Using a longer sim time shows that it will restart briefly after a while.

Thnx for your effort ChrisP but the compoenents don't appear when I open it in LTspice.Maybe some problem with the libraries.Could you post a snapshot of the simulation on the forum?

I downloaded the new version of LTspice(released just yesterday ) .I guess you have used it coz it opens properly in the new one.I am analysing it.I simulated it for 100m but the oscillations did'nt restart.I am trying for larger times.Let me know what you are getting too..

*I simulated for 10 sec.Except initial osc. it did'nt continue.