> ____________________ > | | > | (Philips logo) | > | 0.013/5/2000 | > | 2000Vpp 16 KHz | > | KP/MKP 357 HQ T4 | > |____________________| > | | > | | > >It looks like to be a 13nF x 2000V - impossible to find. If you can«t find a Magnavox repair center... Let«s build one! What about two 4n7F 2kV(min) and one 3n9F 2kV(min) in paralell ? Or 10nf + (3,3nf or 2,7nF) ? Alternatively, use two 27nF 1kV(min) in series which will give 12,5nF that would probably be within margins. Or three 39nF in series, each min 670V What a about a 15n 2kV, and a 100n 400V in series? For two serially connected capacitors C=(C1*C2)/C1+C2) = 13,04nF OBSERVE: I must also add that if there is any DC component across *series connected capacitors* leakage currents may cause voltage across any capacitor to rise towards the full voltage over the whole series! So we should add a resistive divider (high ohmic to not interfere with cirquit operation) with a tap for every capacitor joint. I have a box of capacitors i use for repair different things this way. High voltage caps in i.e resoant SMPS are often odd values and must not be altered. See what caps you can dig up! Ummm. Also find a cap with reasonably low losses so it do not get hot, this beauti seem to be pecified for 16kHz 2kV. But it also says it is an MKP so those sort should be enough... >...and note the upper case "K" for kiloHertz. that«s bad. /Morgan