can i repurpose LOPT/TVflyback cores?

[km-r]

hello guys!

just like what the title said can i repurpose LOPT/TV flyback transformer coils for SMPS? im thinking these are nice and cheap substitute for C cores.
anyone with experience? theoretical advantages/disadvantages?

thanks!

 

[Jagd.Panther]

hello guys!

just like what the title said can i repurpose LOPT/TV flyback transformer coils for SMPS?

Sure, Ive seen few completed projects that had used such cores. Both for SMPS trafo and output inductor (these were welding inverter)

im thinking these are nice and cheap substitute for C cores.
anyone with experience? theoretical advantages/disadvantages?

thanks!

Comparing to U cores stacking LAPT cores wouln't be that efficient (one leg is round in shape, that menas you'll have to use more slightly longer wire due to voids).

 

[km-r]

thanks for the reply.

so someone has done this before... i cant think why i didnt find any. i think, most of them use self-oscillating or royer designs retaining the secondary for use with HV or spark gap fun projects. im thinking about making 12VDC-320VDC few hundred watts inverter for home lighting mostly 220VAC LED bulbs which (tested)worked fine with 320VDC. so i guess the design would need good current limiting and voltage feedback.

so im getting the idea that i can use this like regular ferrite transformer? ill just calculate the core area of the smallest side[the square leg or the round leg) and proceed with the design?

thanks in advance.

 

[Jagd.Panther]

thanks for the reply.

so someone has done this before... i cant think why i didnt find any. i think, most of them use self-oscillating or royer designs retaining the secondary for use with HV or spark gap fun projects.

http://w5jgv.com/hv-ps1/
http://www.edaboard.com/thread57813-5.html
http://www.edaboard.com/thread57813-2.html

im thinking about making 12VDC-320VDC few hundred watts inverter for home lighting mostly 220VAC LED bulbs which (tested)worked fine with 320VDC. so i guess the design would need good current limiting and voltage feedback.

so im getting the idea that i can use this like regular ferrite transformer?

Yep, just remove the spacers/shims and clean up the core using some kind of solvent.

ill just calculate the core area of the smallest side[the square leg or the round leg) and proceed with the design?

thanks in advance.

Nope (unless you know what kind of material is used, I bet you don't).

W/o knowing perm, losses at desired frequency & B it's hard to impossible design anything. However it'spretty easy to measure that, w/o using any fancy equipment (except oscilloscope, hower it's not really needed).

To measure all that you have to build a square wave generator with source/sink capability of like 5-10A+, 10-20V peak-to-peak output and ideally variable frequency and variable voltage. The easiest option to do that is to use IR2153 + two low voltage low Rdson mosfets, like IRFZ44/IRFZ48/IRFZ3205/etc. Then you assemble the core, set up the generator to desired frequency, wind few test turns (like 10-20-30) on the core and connect this winding to the generator.. Then if you don't have an oscilloscope you can either wait for 15-30 minutes and check if the core heats up (if it does, the number of turns is too low for test voltage/freq, you have to add more turns and/or increase frequency and/or decrease voltage) or check magnetising current by measured current drawn on the generator.

If you have an oscilloscope you can check magnetizing current by probing voltage on 0.1-0.2-1omh resistors connected in series with the output of the test winding (if the core is close to saturation, the current will abruptly increase at the end of each cycle, when the core isn't saturing you'll get a nice trainagle wave).

At the end youll get a test winding running at desired frequency and at some test voltage. You use this data to calculate number of turns per volt for the core at given frequcny. Then you proceed with selecting the wire diameters according to available window area/frequency/current/etc.

 

[km-r]

http://w5jgv.com/hv-ps1/
http://www.edaboard.com/thread57813-5.html
http://www.edaboard.com/thread57813-2.html

Interesting... Monstrous projects. Thanks!
Never thought about the deflection coil/core...

Yep, just remove the spacers/shims and clean up the core using some kind of solvent.

The core came off clean from the HV bobbin. There was an airgap/spacers on each leg around 1mm.

Nope (unless you know what kind of material is used, I bet you don't).

W/o knowing perm, losses at desired frequency & B it's hard to impossible design anything. However it'spretty easy to measure that, w/o using any fancy equipment (except oscilloscope, hower it's not really needed).

To measure all that you have to build a square wave generator with source/sink capability of like 5-10A+, 10-20V peak-to-peak output and ideally variable frequency and variable voltage. The easiest option to do that is to use IR2153 + two low voltage low Rdson mosfets, like IRFZ44/IRFZ48/IRFZ3205/etc. Then you assemble the core, set up the generator to desired frequency, wind few test turns (like 10-20-30) on the core and connect this winding to the generator.. Then if you don't have an oscilloscope you can either wait for 15-30 minutes and check if the core heats up (if it does, the number of turns is too low for test voltage/freq, you have to add more turns and/or increase frequency and/or decrease voltage) or check magnetising current by measured current drawn on the generator.

If you have an oscilloscope you can check magnetizing current by probing voltage on 0.1-0.2-1omh resistors connected in series with the output of the test winding (if the core is close to saturation, the current will abruptly increase at the end of each cycle, when the core isn't saturing you'll get a nice trainagle wave).

At the end youll get a test winding running at desired frequency and at some test voltage. You use this data to calculate number of turns per volt for the core at given frequcny. Then you proceed with selecting the wire diameters according to available window area/frequency/current/etc.

This is gold. Thank you! Can i clarify a few things though? I have a small 5-20V bench supply with current limiting adjustable up to 10A.
So i hook up an IR2153 with variable freq + two IRF3205 in push-pull configuration?
I have a DMM with freq/duty and temperature. How hot is hot?

Set freq > run for 15-25min while observing temperature; adjust winding; repeat until suitable temperature?

We have a scope at school's lab but its not that easy to get permission especially with a "student's experiment" id like to see the current trace as you suggested.


Thank you.

 

[Jagd.Panther]

Never thought about the deflection coil/core...

Me too.

The core came off clean from the HV bobbin. There was an airgap/spacers on each leg around 1mm.

Mating surfaces should be free of any kind glue/resin. Usually spacers are glued to the core (acetone should be good to remove almost any kind of glue or resin, at least thats what I used in the past with good results).

This is gold. Thank you! Can i clarify a few things though? I have a small 5-20V bench supply with current limiting adjustable up to 10A.
So i hook up an IR2153 with variable freq + two IRF3205 in push-pull configuration?

It's up to you. I personally prefer half-bridge for such experiments (easier to work with one winding), sure thing you have to use a cap to block DC for half-bridge.

I have a DMM with freq/duty and temperature. How hot is hot?

One option is to split copper/core losses 50/50, so for example if you design a transformer for cca. +40degC temp rise under full load you should be after < 20degC core temp rise on idle (keep in mind that bare core has lower Rth than a core with windings). However if you don't plan to use the transformer 100% of time at near full load it makes sense to reduce B to reduce losses (and magnetization current) under light load. Also both perm and corelosses depend on temperature. Study datasheets for popular materials like n27,N87,N97,3C90,3C94,3F3, etc to get an idea what to expect at lower/higher temperatures. WRT magnetizing current you usually aim Imag < 1/10 of primary current at full load.

Set freq > run for 15-25min while observing temperature; adjust winding; repeat until suitable temperature?

Exactly, it's that simple, then adjust number of turns to make some room for higher input voltage/operating at lower temp/etc (if needed).

 

[Jagd.Panther]

WRT to switching frequency I guess something in 30 kHz ... 60kHz range should be OK for the LOPT cores.

 

[km-r]

maybe ill have push pull, low voltage halfbridge isnt really for me hehehe.

i appreciate the help of the nazi war machine thank you. hopefully ill get some decent use for these busted flybacks.

 

[Jagd.Panther]

It's a complex question, it depends on target conditions, acceptable loss, etc.

Without going too deep into that: < +10...+15deegress.C temperature rise on idle.