3
\$\begingroup\$

I am designing a high voltage DC/DC converter that is for an aerospace application and therefore I am focusing on reducing the size and weight, so switching at a high frequency possibly integrating wide band-gap semiconductor devices. However, switching at high frequencies with the typical wire wound transformers has its limitations. Therefore I started to research the use of planar transformers.

However with high voltage applications (in the kV range, with low current in the 100’s mA range), planar transformers are not typically used because of the insulation ratings and the like.

My question is: how viable is it to design both high voltage and high frequency transformers? What will be the main things to be worried about?

If anyone has any references/guides for designing high voltage, high frequency planar transformers, or any reference designs, please do guide me to them. Any general tips for their design, how to insulate them, manage eddy currents and proximity effects at these levels etc will also be appreciated.

The SMPS topology will likely be a form of parallel resonant converter to enable reduction of switching losses.

\$\endgroup\$
  • \$\begingroup\$ How high is high frequency? There are lots of wound transformers made to operate at kilovolts in the hundreds of kilohertz range. Flyback transformers, ignition coils, etc. \$\endgroup\$ – evildemonic Jul 3 '19 at 22:07
  • \$\begingroup\$ Insulation stops being a good insulator at high voltage high frequency... \$\endgroup\$ – MadHatter Jul 3 '19 at 23:29
  • \$\begingroup\$ you can make it void of EMI harmonics with Current pumped resonance and no shielding requirements or make it hard switched Resonant ZVS synchronous Current rectification for high efficiency at high frequency but still noisy without shields and critical layouts. start with ALL your specs for environment , P,V,I EMI E field , temp, altitude. consider best transformer custom designers in HK. I did this for the linear -3kV to +5kV laser printer requirements for a design regulated with 1% kV accuracy using PWM using cheap quad Op Amps and transistors. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jul 4 '19 at 0:47
  • \$\begingroup\$ I'd like to exceed 300kHz. The issue is with flyback transformers and the like is their size, weight, and volume. I am working on an aerospace application therefore these need to be minimized as much as possible, however the planar transformer admittedly is conceptual at this point, although I have seen similar ones done before. Possibly a toroidal core will be better suited? Yes insulation will be my biggest headache, but there are some materials that can handle the HV insulation requirements such as high performance epoxy. \$\endgroup\$ – jvnlendm Jul 10 '19 at 15:36
  • 1
    \$\begingroup\$ Maybe you can run the switcher at high frequency (eg. 1MHz) and multiply the voltage up from perhaps 80-200V. \$\endgroup\$ – Spehro Pefhany Jan 12 at 6:51
0
\$\begingroup\$

We need a lot more information to give you good advice. For example, how high of a voltage makes a huge difference on the size limitation, what frequency are you thinking about, whatever info you can share makes this easier. The numbers can turn this from difficult to impossible very easily.

The main thing to be worried about is straight up breakdown, you need to cover the clearance distance to for the high voltage you are working with, for example you make a 6kV DC/DC with air insulation you will need at least 2cms for it not to breakdown instantly when the voltage is applied(you will of course want more than this distance). You can of course use different materials

You essentially consider insulators as capacitors, and frequency affects them quite a bit so what frequency you work with also has an effect on this element.

The other main thing you have to worry about is your transformer saturating, this will kill your output on the DC end and even if you clear the distance the converter will be useless. So first you choose what material your transformer will be and see if the size needed for that transformer to work suits your needs

you mainly look at the BH curve, some makers do not provide them so it can be hard to design sometimes.

you mainly look at these expressions:

$$ B=\frac{E}{4.44NAf} $$

where B is magnetic field E is rms voltage applied(not the high voltage internally at what voltage you are working with), N is number of turns A is area of your core and finally the frequency.

$$ H=\frac{IdcN}{Le} $$

Idc is the DC current, you can use rms if you are working with sine waves or other stuff, N is number of turns and Le is effective magnetic path length, you will have to look up how to measure this.

then you can go to a ferrite catalog like this one

look at the frequency curves, make sure it even works at your frequency and make sure it is not on the elbows of the curve(you must have an idea of what are your input and output waveform while you check this)

Even if you have a ferrite material that actually can do the job, then you must check that it can also handle the high voltage you are applying by itself. In this case high voltage breakdown will always be your highest priority.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for this reply. I will be limiting the output voltage of the actual transformer to less than 1kV or so, but the insulation will probably still have to withstand over 10kV or so with all the secondaries combined, for safety reasons. I am hoping to drive the transformer over 300kHz. I believe the best transformer material is a soft ferrite MnZN for this range? \$\endgroup\$ – jvnlendm Jul 10 '19 at 15:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.