Cheap pulse transformer as a ni-crhome wire heater...

[mook Jonhon]

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated the
test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at 200KHz
and 50Vin. The secondary is attached to a 4 ohm Ni-chrome wire that I
need to drive with 1A RMS for a few seconds at a time at low duty
cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this kind
of load and high pulse current but I need galvanic isolation low cost
and short non-custom design. It is still reasonablly efficient for
what i need.


My question is, I am able to apply almost double the 50V-us I measured
in the initial test. with 100Vin I can apply a 2uS pulse which should
be 50V * 2us or 100V-us, and not see the primary current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will saturate
as a lower v-us product. Open circuit on the secondarty would have
it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can be
applied to the primary without saturating? I thought the peek flux
density was only determined by the Volt seconds applied to the primary,
number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?

 

[server]

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated the
test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at 200KHz
and 50Vin. The secondary is attached to a 4 ohm Ni-chrome wire that I
need to drive with 1A RMS for a few seconds at a time at low duty
cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this kind
of load and high pulse current but I need galvanic isolation low cost
and short non-custom design. It is still reasonablly efficient for
what i need.


My question is, I am able to apply almost double the 50V-us I measured
in the initial test. with 100Vin I can apply a 2uS pulse which should
be 50V * 2us or 100V-us, and not see the primary current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will saturate
as a lower v-us product. Open circuit on the secondarty would have
it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can be
applied to the primary without saturating? I thought the peek flux
density was only determined by the Volt seconds applied to the primary,
number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?

Given an ideal transformer (no DCRs, no leakage inductance) with a
shorted secondary, you can\'t saturate it because you can\'t apply any
voltage across the primary; it\'s a dead short.

To see what\'s happening, you can Spice an ideal transformer and then
add the copper resistances and leakage inductance. The volt-second
saturation happens at the leads of the internal, inaccessable ideal
transformer.

Your internal ideal transformer probably still saturates at 50 uv-s.
But if the secondary is nearly shorted, the voltage across the ideal
internal primary is not what you apply at the terminals; there\'s
voltage drop in the copper. It just looks like the saturation v-us has
gone up.





--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[Tim Williams]

Easy: primary voltage is lost to DCR and leakage. As these drop more
voltage, flux in the core drops.

At some point, you\'ll be dropping more across the transformer than across
the load, giving terrible efficiency. Then you absolutely need a bigger and
better transformer.


It\'s not clear from the numbers given, just how much you\'re dropping, or if
this is on the right slope to be correct.

There\'s also the difference between single ended and full-wave saturation:
it\'s not clear which one you tested.

Leakage can be dealt with by going to a lower frequency -- of course, you
need more flux for that, too: in other words, a bigger transformer.

Lunch, free, etc.

Tim

--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Design
Website: https://www.seventransistorlabs.com/

\"mook Jonhon\" <mook@mook.net> wrote in message
news:qcPfH.143499$RY8.38545@fx48.iad...

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated the
test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at 200KHz
and 50Vin. The secondary is attached to a 4 ohm Ni-chrome wire that I
need to drive with 1A RMS for a few seconds at a time at low duty
cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this kind
of load and high pulse current but I need galvanic isolation low cost
and short non-custom design. It is still reasonablly efficient for
what i need.


My question is, I am able to apply almost double the 50V-us I measured
in the initial test. with 100Vin I can apply a 2uS pulse which should
be 50V * 2us or 100V-us, and not see the primary current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will saturate
as a lower v-us product. Open circuit on the secondarty would have
it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can be
applied to the primary without saturating? I thought the peek flux
density was only determined by the Volt seconds applied to the primary,
number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?

 

[mook Jonhon]

jlarkin@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated
the test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at
200KHz and 50Vin. The secondary is attached to a 4 ohm Ni-chrome
wire that I need to drive with 1A RMS for a few seconds at a time
at low duty cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this
kind of load and high pulse current but I need galvanic isolation
low cost and short non-custom design. It is still reasonablly
efficient for what i need.


My question is, I am able to apply almost double the 50V-us I
measured in the initial test. with 100Vin I can apply a 2uS pulse
which should be 50V * 2us or 100V-us, and not see the primary
current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of
spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will
saturate as a lower v-us product. Open circuit on the
secondarty would have it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can
be applied to the primary without saturating? I thought the peek
flux density was only determined by the Volt seconds applied to the
primary, number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?


Given an ideal transformer (no DCRs, no leakage inductance) with a
shorted secondary, you can\'t saturate it because you can\'t apply any
voltage across the primary; it\'s a dead short.

To see what\'s happening, you can Spice an ideal transformer and then
add the copper resistances and leakage inductance. The volt-second
saturation happens at the leads of the internal, inaccessable ideal
transformer.

Your internal ideal transformer probably still saturates at 50 uv-s.
But if the secondary is nearly shorted, the voltage across the ideal
internal primary is not what you apply at the terminals; there\'s
voltage drop in the copper. It just looks like the saturation v-us has
gone up.

Thats what I was thinking. The current in vs current out ratio is
held even at the high V-us numbers so I think the \"internal ideal
transformer\" is not saturated but being redced in voltage by the high
DCR of the windings and the high primary current.


I was thinking it was something more \"magical\" that that.

You know what they say about transformers. Theres more than meets the
eye.

 

[legg]

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated the
test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at 200KHz
and 50Vin. The secondary is attached to a 4 ohm Ni-chrome wire that I
need to drive with 1A RMS for a few seconds at a time at low duty
cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this kind
of load and high pulse current but I need galvanic isolation low cost
and short non-custom design. It is still reasonablly efficient for
what i need.


My question is, I am able to apply almost double the 50V-us I measured
in the initial test. with 100Vin I can apply a 2uS pulse which should
be 50V * 2us or 100V-us, and not see the primary current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will saturate
as a lower v-us product. Open circuit on the secondarty would have
it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can be
applied to the primary without saturating? I thought the peek flux
density was only determined by the Volt seconds applied to the primary,
number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?

Vt = BNA

V=volts
t=seconds
B=teslas
N=turns
A=cross-sectional area of core in meters^2

B is limited by saturation, so deltaBmax = Vt / NA

If the core flux is saturated in opposite polarity with
each pulse, then deltsB max = Bsat x 2.

Left to its own devices, an open circuit winding will
only reset flux to a remenance value, under the influence
of stored magnetization energy. So you don\'t get full Bsat
swing in single ended pulse repetition.

Bsat reduces with temperature in most pulse grade core
materials.

RL

 

[server]

On Fri, 09 Oct 2020 09:24:45 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

jlarkin@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated
the test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at
200KHz and 50Vin. The secondary is attached to a 4 ohm Ni-chrome
wire that I need to drive with 1A RMS for a few seconds at a time
at low duty cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this
kind of load and high pulse current but I need galvanic isolation
low cost and short non-custom design. It is still reasonablly
efficient for what i need.


My question is, I am able to apply almost double the 50V-us I
measured in the initial test. with 100Vin I can apply a 2uS pulse
which should be 50V * 2us or 100V-us, and not see the primary
current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of
spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will
saturate as a lower v-us product. Open circuit on the
secondarty would have it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can
be applied to the primary without saturating? I thought the peek
flux density was only determined by the Volt seconds applied to the
primary, number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?


Given an ideal transformer (no DCRs, no leakage inductance) with a
shorted secondary, you can\'t saturate it because you can\'t apply any
voltage across the primary; it\'s a dead short.

To see what\'s happening, you can Spice an ideal transformer and then
add the copper resistances and leakage inductance. The volt-second
saturation happens at the leads of the internal, inaccessable ideal
transformer.

Your internal ideal transformer probably still saturates at 50 uv-s.
But if the secondary is nearly shorted, the voltage across the ideal
internal primary is not what you apply at the terminals; there\'s
voltage drop in the copper. It just looks like the saturation v-us has
gone up.



Thats what I was thinking. The current in vs current out ratio is
held even at the high V-us numbers so I think the \"internal ideal
transformer\" is not saturated but being redced in voltage by the high
DCR of the windings and the high primary current.


I was thinking it was something more \"magical\" that that.

You know what they say about transformers. Theres more than meets the
eye.

There\'s magnetization and hysteresis and eddy currents and capacitance
and skin effect and even prop delay. It just gets worse.

I want to make many channels of 150 volt isolated
exponentially-decaying fast-rise pulses into 50-ohm coax. There is
some secret, probably exploding device a few km away on the other end.
I need an exotic pulse transformer, fast rise and lots of volt-secs.
Turns out that a cheap Coilcraft dual-winding inductor is perfect. The
way to find that out was to get one and test it.

Coilcraft is great. Heather answered my email on a Sunday and sent 10
samples on Monday.



--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[mook Jonhon]

jlarkin@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 09:24:45 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

jlarkin@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net
wrote:
https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back
to >> > zero (saturation). I reversed the primary windings and
repeated >> > the test in the opposite direction and got the same
reaponse.
It saturated at ~50v-us so 20% margin between datasheet rating
and >> > actual. good job TDK.


I checked this against a spice model (simetrix) and got very
similar >> > results EP5- 3c96 with 42 turns on the primary (Got teh
same >> > inductance so I called it close enough). TDKis probable
using N45 >> > material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at
200KHz and 50Vin. The secondary is attached to a 4 ohm
Ni-chrome >> > wire that I need to drive with 1A RMS for a few
seconds at a time >> > at low duty cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for
this >> > kind of load and high pulse current but I need galvanic
isolation >> > low cost and short non-custom design. It is still
reasonablly >> > efficient for what i need.


My question is, I am able to apply almost double the 50V-us I
measured in the initial test. with 100Vin I can apply a 2uS
pulse >> > which should be 50V * 2us or 100V-us, and not see the
primary >> > current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of
spiking.

If course I need to limit the time to 10s of mS when testing to
not >> > burn the transformer up sice the load is experienceing much
higher >> > currents than 1A at this point.

If Increase the resistance on the secondary, the primary will
saturate as a lower v-us product. Open circuit on the
secondarty would have it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that
can >> > be applied to the primary without saturating? I thought
the peek >> > flux density was only determined by the Volt seconds
applied to the >> > primary, number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux
generated by >> > the primary voltage?


Given an ideal transformer (no DCRs, no leakage inductance) with a
shorted secondary, you can\'t saturate it because you can\'t apply
any >> voltage across the primary; it\'s a dead short.

To see what\'s happening, you can Spice an ideal transformer and
then >> add the copper resistances and leakage inductance. The
volt-second >> saturation happens at the leads of the internal,
inaccessable ideal >> transformer.

Your internal ideal transformer probably still saturates at 50
uv-s. >> But if the secondary is nearly shorted, the voltage across
the ideal >> internal primary is not what you apply at the terminals;
there\'s >> voltage drop in the copper. It just looks like the
saturation v-us has >> gone up.



Thats what I was thinking. The current in vs current out ratio is
held even at the high V-us numbers so I think the \"internal ideal
transformer\" is not saturated but being redced in voltage by the
high DCR of the windings and the high primary current.


I was thinking it was something more \"magical\" that that.

You know what they say about transformers. Theres more than meets
the eye.

There\'s magnetization and hysteresis and eddy currents and capacitance
and skin effect and even prop delay. It just gets worse.

I want to make many channels of 150 volt isolated
exponentially-decaying fast-rise pulses into 50-ohm coax. There is
some secret, probably exploding device a few km away on the other end.
I need an exotic pulse transformer, fast rise and lots of volt-secs.
Turns out that a cheap Coilcraft dual-winding inductor is perfect. The
way to find that out was to get one and test it.

Coilcraft is great. Heather answered my email on a Sunday and sent 10
samples on Monday.

This is my experience as well with coilcraft. They also have a
duplicate of the TDK at 1/2 the price.

 

[mook Jonhon]

legg wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated
the test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at
200KHz and 50Vin. The secondary is attached to a 4 ohm Ni-chrome
wire that I need to drive with 1A RMS for a few seconds at a time
at low duty cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this
kind of load and high pulse current but I need galvanic isolation
low cost and short non-custom design. It is still reasonablly
efficient for what i need.


My question is, I am able to apply almost double the 50V-us I
measured in the initial test. with 100Vin I can apply a 2uS pulse
which should be 50V * 2us or 100V-us, and not see the primary
current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of
spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will
saturate as a lower v-us product. Open circuit on the
secondarty would have it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can
be applied to the primary without saturating? I thought the peek
flux density was only determined by the Volt seconds applied to the
primary, number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?


Vt = BNA

V=volts
t=seconds
B=teslas
N=turns
A=cross-sectional area of core in meters^2

B is limited by saturation, so deltaBmax = Vt / NA

If the core flux is saturated in opposite polarity with
each pulse, then deltsB max = Bsat x 2.

Left to its own devices, an open circuit winding will
only reset flux to a remenance value, under the influence
of stored magnetization energy. So you don\'t get full Bsat
swing in single ended pulse repetition.

Bsat reduces with temperature in most pulse grade core
materials.

RL

Thanks, That what the gray beards taught me long ago. working for 20
years but never ran across or even thought to look at intentionally
pushing a transformer to the Bsat limts before with various loads.

never too old to learn.

 

[Bill Sloman]

On Saturday, October 10, 2020 at 2:00:27 AM UTC+11, jla...@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 09:24:45 GMT, \"mook Jonhon\" <mo...@mook.net> wrote:

jla...@highlandsniptechnology.com wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mo...@mook.net> wrote:

<snip>

You know what they say about transformers. There\'s more than meets the
eye.

It\'s not the eyes that are failing here, but the fundamental grasp of what is going on.

There\'s magnetization and hysteresis and eddy currents and capacitance
and skin effect and even prop delay. It just gets worse.

If you refuse to think carefully about the basics of what\'s going on in the core, the elaborations are even more difficult to keep track of.

I want to make many channels of 150 volt isolated
exponentially-decaying fast-rise pulses into 50-ohm coax. There is
some secret, probably exploding device a few km away on the other end.
I need an exotic pulse transformer, fast rise and lots of volt-secs.

It\'s called a transmission line transformer. If you wind a transformer with twisted pair, the twisted pair acts as a transmission line - not a particularly perfect transmission line, since minature coax is a lot more nearly perfect - but quite good enough for lots of applications.

Turns out that a cheap Coilcraft dual-winding inductor is perfect. The
way to find that out was to get one and test it.

There \'s a large market for transformers that can carry a fast switching drive across an isolation barrier. They all seem to be wound with twisted pair.

Coilcraft is great. Heather answered my email on a Sunday and sent 10
samples on Monday.

He could have wound something for himself in ten minutes, if he\'d had a reel of the right wire and a few ferrite cores and formers handy.

He seems to have been frightened by a transformer at an early age and has steered clear of them ever since.

--
Bill Sloman, Sydney

 

[server]

legg <legg@nospam.magma.ca> wrote:

On Fri, 09 Oct 2020 01:40:38 GMT, \"mook Jonhon\" <mook@mook.net> wrote:

https://www.tdk-electronics.tdk.com/inf/85/ds/b82804a.pdf

The 670uH 1.5:1 version.

E-dt is rated for 40V-us.

I ran a test where I applied a step through a 50 ohm resistor and
watched to see how long it took for the voltage to nosedive back to
zero (saturation). I reversed the primary windings and repeated the
test in the opposite direction and got the same reaponse.

It saturated at ~50v-us so 20% margin between datasheet rating and
actual. good job TDK.


I checked this against a spice model (simetrix) and got very similar
results EP5- 3c96 with 42 turns on the primary (Got teh same
inductance so I called it close enough). TDKis probable using N45
material which is similar to 3c96.

The application is a small half-bridge SMPS topology running at 200KHz
and 50Vin. The secondary is attached to a 4 ohm Ni-chrome wire that I
need to drive with 1A RMS for a few seconds at a time at low duty
cycle. Seconds on, minutes off.


I understand the transformer resistance is far from ideal for this kind
of load and high pulse current but I need galvanic isolation low cost
and short non-custom design. It is still reasonablly efficient for
what i need.


My question is, I am able to apply almost double the 50V-us I measured
in the initial test. with 100Vin I can apply a 2uS pulse which should
be 50V * 2us or 100V-us, and not see the primary current spike upward.

The lower I make the secondary resistance, the higher I can go in
voltage and still not have the primary current show signs of spiking.

If course I need to limit the time to 10s of mS when testing to not
burn the transformer up sice the load is experienceing much higher
currents than 1A at this point.

If Increase the resistance on the secondary, the primary will saturate
as a lower v-us product. Open circuit on the secondarty would have
it saturate at near the 50Vus.


Why does a load on the secondary change the amount of V-us that can be
applied to the primary without saturating? I thought the peek flux
density was only determined by the Volt seconds applied to the primary,
number of turns, perm, and Ae.

Since the currents are so high is this bucking the flux generated by
the primary voltage?


Vt = BNA

V=volts
t=seconds
B=teslas
N=turns
A=cross-sectional area of core in meters^2

B is limited by saturation, so deltaBmax = Vt / NA

That is only inductive component. There is also ohmic part.
Denoting by V1 (time dependent) voltage on primary we have

V1 = B_tNA + IR

where B_t is time derivative of the magnetic flux. Flux
is produced by both primary and secondary winding. With
load on secondary magnetic fields from primary and
secondary mostly cancel, so primary current is much
larger than in case of unloaded secondary. Conseqently
resistive voltage drop is much larger and to get the
same flux one needs larger primary voltage. Of course,
there is also voltage drop in secondary, so if goal
is to get maximal power to the load then one needs to
balance increasing current with resistive losses in
transformer: both unloaded and shorted transformer
deliver no power to the load and all power goes into
losses. Assuming reasonable transformer model
optimum is closer to 50% increase in primary
voltage than to 100% increase.

--
Waldek Hebisch