DC-DC step up from 3V to over 100V

Question

I'm currently building a DC-DC step up circuit to step up 3V to 100V. I don't want to use transformer because i want to keep the circuit as small as possible. My thought is ... building a 30V-output circuit using MIC2250, then build a boost circuit to boost the voltage 3-4 times. The 30V output has been achieved, but the boost doesn't work. The diagram is as below.

I was thinking of repeating this circuit 2 to 3 times in order to step up the overall voltage 2-3 times. but it doesn't work.

My another design is to use IC solely. some ICs like LT3757 and LT3783 seem to fit my design, but I just don't find any suggested circuit to boost from 3V to 100V. I tried MCP1652 before by following the circuit in the datasheet. Although the datasheet clams that the circuit can step up the input to 100V, I just can't do it. The maximum I got was only 40V.(I emailed the technician from MCP...he asked me to find the solution on the internet myself...poor aftersale service...) the only suggested circuit from 5V to 350V in LT3757 requires transformer, but i just try to avoid transformer.

Can any body share you idea with me about the circuit stepping up from 3V to 100V? Iout = 0.1A will be good enough. slightly smaller will be acceptable too.

Answer 1

1N4004 diodes are inappropriate for this application because they have a very long \$ t_{rr}\$ reverse recovery time. Use an ultra fast UF4004 here.

Assume output voltage is 100V and current is 100mA. Then ton/toff is about 2.5, and peak inductor current will be 700mA (the inductor must not saturate - so that's a big inductor). On time will be about 30usec and off time 12usec, so about 24kHz.

It's feasible if the parts are properly rated- but the most likely culprit if it does not work when fed with a high capacity 30V bench supply is the inductor- use one that will not saturate with 1A or more.

Answer 2

Avoiding a transformer to keep the circuit small ... doesn't work so well.

See Spehro's answer : to avoid saturating the inductor in a flyback convertor, you need to keep the current low, or use a big inductor. The problem is that you are storing energy in the inductor core at low voltage, and releasing that energy (in a fast pulse) at high voltage. And it will only hold so much energy before it saturates.

With a transformer, while the primary winding is storing energy, the secondary is already pulling energy out (at higher voltage but lower current) so very little energy is actually stored in the core. This means you can transfer much more energy for the same size of core. Or use a smaller core for the same power delivery.

Something like this might be close to what you want, about 1.6*1.6*1.6cm to deliver about 6W (0.06A at 100V). You could drive its 3.3V secondary from an H-bridge at 100kHz and extract power from its primary, with a bridge rectifier made from those fast diodes. Or wind a custom transformer on a core like this or this smaller one to get exactly what you need. I haven't done the math to design a 10W supply around either of these cores. It may be better to look at application notes for some of the ICs you mentioned for guidance on winding suitable transformers.

Answer 3

Im guessing that your 3V power source was unable to deliver enough power/current and thats why you had 40V output instead of 100V on converter built on MCP1652.

You should measure input voltage when converter is running.

Answer 4

It is conceivable to use tapped inductor (auto-transformer). The mosfet switches the tap' the right side goes to the diode. The idea is to avoid exhausting duty cycle control range limitations. As bonus MOSFET sees lower voltage.