I need to charge a capacitor up to 640 V (from a 3V) source for an embedded systems project I'm working on. I've considered other options, such as voltage multipliers and Hulda-Clark based zapper circuits, and I think the following will be a better solution, given my space constraints.

I plan to use this Linear Technologies chip: http://www.linear.com/product/LT3484 to control the charging, but I'm having some difficulty finding a small flyback transformer with a 1:20 turns ratio suitable for this application.

I have found a current sense transformer that's the right size and turns ratio: TDK Current Sense Transformer B82801B01, but its a current sense transformer and not a flyback transformer.

  • 2
    \$\begingroup\$ The so called "flyback transformer" is not really a transformer per se. It's more of an inductor with 2 windings. Flyback inductor-transformer stores energy in the magnetic field, while a regular transformer doesn't. The core of the flyback inductor-transformer has usually got a significant reluctance, and often it has got a gap. The transformer core has got a low reluctance and no gap, typically. [I'll try to get back to this and convert this comment into an answer, although somebody else will probably beat me to it.] \$\endgroup\$ Commented Sep 14, 2016 at 17:54
  • 1
    \$\begingroup\$ Even if it works electromagnetically, the voltage on the secondary may prove to be destructive long-term. Maybe you can vacuum impregnate it or st. \$\endgroup\$ Commented Sep 14, 2016 at 18:34

3 Answers 3


The LT3484 is quite clear in the data sheet: -

enter image description here

The primary inductance is recommended to be about 10 uH. Take note!

The current sense transformer you have highlighted has a primary inductance of at least 80 uH and ranges up to 8 mH depending on which one you choose. The specific one in your question has a primary inductance of 180 uH.

Now you might get it to work very badly with the output capacitor charging slowly and maybe not reaching full expected voltage but if you read LTI's data sheet they specify these parts: -

  • KIJIMA MUSEN PART# SBL-5.6-1, LPRI = 10µH, N = 10.2
  • TDK LDT565630T-003 LPRI = 10.5µH, N = 10.2
  • TDK PART# LDT565630T-001, LPRI = 6µH, N = 10.4
  • TDK PART# LDT565630T-002, LPRI = 14.5µH, N = 10.2

.... In the example circuits provided.

  • 1
    \$\begingroup\$ There is also the required core volume and cross section area. Using estimates for the duty cycle, \$f=150kHz\$, and on-voltage (taking into account winding resistance and peak current), I figure the ferrite volume must be \$\tfrac{1}{3}mm^3\cdot \mu_r\$, where \$\mu_r\$ is the relative permeability of the core itself. Inductance alone isn't enough. You need the necessary vacuum gap inside the core material to store the energy without also saturating the ferrite "filler." \$A_c \ge \tfrac{1.6 cm^2}{N}\$, with N being the primary turns also seems important here. \$\endgroup\$
    – jonk
    Commented Sep 14, 2016 at 19:33
  • \$\begingroup\$ The problem with the recommended part numbers is the turns ratio. They are all on the order of 1:10 and I need a 1:20 to hit the target voltage. Barring that, I'll have to follow the recommended circuit with a voltage multiplier. The application power-consumption sensitive and i need to minimize losses. So a voltage multiplier at high voltage is not ideal. \$\endgroup\$ Commented Sep 15, 2016 at 19:02
  • 1
    \$\begingroup\$ Getting a core set and winding your own sounds really challenging but, take my word for it, if you use yellow polyester tape to insulate secondary from primary, it is relatively straight forward. \$\endgroup\$
    – Andy aka
    Commented Sep 16, 2016 at 9:44

Digikey list a range of transformers intended for flyback converter use. These include 5 intended for photoflash operation with the same part number series as in the LT3484 examples BUT 4 of these are shown as obsolete and the other is non-stock.

However, there are a range of Wurth transformers which may be suited (the one I looked at in detail only had a 2.5 uH primary) and other brands as well.
This Digikey search may be useful

As the converter uses flyback the exact turns ratio is not overly important - for instance the examples typically show 1:10 ratio with 3V in and 300V out.
Primary:secondary isolation IS important - some show 1000 VDC and others 500 NDC. The latter MAY be OK at 600V. Or not.

Given the smallish inductances (10 uH primary) and eg 1:10 turns ratio, it should not be too hard to wind your own if needed. Something like an iron powder micrometals toroid is liable to suffice.

Relationship: Customer. I have used and been please with Micrometals products in the past. They were good to deal with and helpful with samples supplied along with a paid for irder.

I do not know the relationship to Micrometals Arnold who seem to be the same but different.
Software, application notes, ...
Inductor design calculator - login required - free account.

Micrometals downloadable inductor design software


The current sense transformer datasheet specifies the inductance of the secondary and not the primary. I suppose that you can use a current sense transformer with Ls = Lp x N^2 if you find it. For the LT3XXX chips, you need a primary of about 10 uH. They recommand a transformer with N = 10. So the Ls = 10 uH x 10^2 = 1 mH.

The TDK current sense transformer with Ls = 980 uH and N = 70 will have a primary inductor of Lp = 980/70^2 = 0.2uH = 200 nH. Very far from the specified Lp > 5 uH. If you find a current sense transformer with Ls = 1 mH to 5 mH and N = 10 to 20 then it might works.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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