| ??? 12/18/03 07:42 Read: times |
#60928 - Some calculation... Responding to: ???'s previous message |
Hallo Mahmood,
I have thought about your problem again... If you can garuantee, that voltage is below 1350V, then you can configure the following array: Take 4700µF/450V electrolytic capacitors. From formula of earlier post Rsym resistor to be connected in parallel to each capacitor should be 4k3. Capacitor and resistor form a 'unit'. Connect three of such units in series, forming a 'row'. Connect 32 of such rows in parallel. This gives you an 50133µF/1350V equivalent, paralleld by 403 Ohm. In total 96 capacitors and 96 resistors are needed. Each capacitor will weight about 600-800g, depending on manufacturer. So, your board containing 96 capacitors will weight about 70kg. When each capacitor has a diameter of 75mm, then your board will need an area of 0.54qm. What about heat dissipation? 1200V dropping across 403 Ohm gives a heating power of U x U / R = 1200V x 1200V / 403Ohm = 3.57kW, which makes 37W per resistor! Also at capacitors heat dissipation will be observed: Each capacitor has about 0.01 Ohm equivalent series resistance. Makes 0.03 Ohm per row and about 1mOhm for total array. 1200A produces a heat dissipation of I x I x R = 1200A x 1200A x 1mOhm = 1.44kW. But this will take place only for some milliseconds! Per capacitor only 37.5A is then flowing, heat dissipation is about 14W. Will this abnormely heat up the capacitors? From dQ = m x c x dT and dQ = P x dt it follows: dT = P x dt / m / c dT is temperature rise, dQ is produced heat energy, m is heated mass, c is specific heat capacity, P is heat dissipation and dt is time duration of heating. If we assume that heated part of capacitor is only the aluminium foil wrapping and involved mass is only about 100g, then we get: dT = 14W x 5msec / 0.1kg / (900 J / kg / K) = 0.0008K Means, heating of capacitors due to short lasting discharging current is negligible! Only heat dissipation of symmetrizing resistors is annoying high. But I found a very easy methode to avoid them totally: ![]() In above schematic, a series circuit of 4 capacitors was assumed. In 'transformer solution' each shown capacitor is in real a parallel circuit of many of them, so that total capacitance finally gives again 50000µF. Probably, it will be helpful to insert a current limiting resistor in series with each diode. This prevents the transformer and diodes from being overrated when your 1200A puls is drawn. Also, inrush current at power-on is highly decreased. Of course, after each discharge it needs some time to recharge the capacitors again. The bigger the transformer, the more current can be drawn, and the less time has to be waited for 'firing' the next pulse current. Another advantage of transformer solution is, that no longer 'long life' grade electrolytic capacitors are needed. This can highly decrease costs... Hope this helps, Kai |




