CapCharger 3000

After throwing my boost converter away because it was so weak and the failure of my second boost converter which was in fact a flyback converter) I decided to build the ultimate capacitor charging device to solve my cap charging problems once and for all. The design goals were simple:

• It should be powerful
• It should be reliable (ie. no blown parts after a while)
• It should be capable of charging caps to 400, 800 or even 1200 volts
• It should stop charging when the selected voltage is reached and recharge if needed
• It should be ran of my lab power supply (which is capable of giving 35 V at 10 A)

Enter the CapCharger 3000.

General circuit description

The power section is just your standard Mazzilli flyback driver (commonly known as “ZVS driver”). It drives a self-wound flyback transformer which transforms the 35 V input* to about 400, 800 and 1200 volts (so the transformer has three “taps”). Depending on the selected output voltage the output of one of these taps is fed into a full-wave rectifier bridge with eight 3 A, 1 kV ultrafast diodes (each leg of the rectifier uses two diodes in series). The output of this rectifier bridge is connected to the capacitors to charge but also to a resistive voltage divider which divides the voltage by a factor 100. This voltage is then fed into a schmitt trigger comparator which compares it to a reference voltage which depends on the selected output voltage. When the voltage over the caps reaches the set value, power to the flyback driver is cut off. As soon as this voltage has fallen about 20 V under the maximum, power is restored, the caps are recharged and the process repeats itself. To select the right transformer tap and the right reference voltage, a system with three relays is used. Additionally, there’s a switch which can be used to disable the flyback driver section completely and a little LED that lights green when the driver is working and orange when power to the driver is cut off.

There is, however, one “but”: you can’t change the voltage setting when the flyback driver is working and there’s high voltage present on the transformer! In fact you can, but it will destroy the relays as an electrical arc will be drawn between the contacts.

* At first I wanted to use 24 V input instead of 35 V, but I forgot to calibrate the 10x mode of my oscilloscope probe so the voltages I measured while winding the transformer to determine the right amount of windings where in fact lower than I believed them to be. Rewinding it would be a lot of work so I compensated for this mistake by increasing the input voltage.

Pictures

During the building process:

Completed device:

Results

This device is able to charge my 7200 μF capacitor bank to about 400 V within a few seconds. I haven’t tested the 800 and 1200 V outputs on large capacitances yet. The voltage control is a lot more reliable than that on my previous converters which had the tendency to charge larger capacitors to lower voltages than smaller ones before cutting off the power. This is probably due to the fact that this converter has a lot more overhead than my previous ones. Component heating is very little except for the LM7815 used to generate a stable 15 V for the control circuitry: it also has to drive the relays and dissipates about 2,5 W doing that.