alternatevista wrote:What value resistor are people using to discharge the capacitor C1?
I can't answer that question, but, 1k ohm would be OK.
A more important question is: How may I bleed charge from C1 without first dismounting the PCB? Typically, C1 is an end-mount cap with its terminals out of reach on the underside of the board. To answer my question, one must know what the HV-circuit looks like, in advance, so one may choose some top-side leads to touch. Roomba/Scooba SMPSs have evolved over time, so part-labels marked on the PCB may be different, but if I refer to the original Fast Charger's construction the cathodes, (K), of FW-bridge diodes D1 & D2 connect together and to C1(+); and D3 & D4 connect together and to C1(-). However, there is a problem with top-side touching the D1K, D2K, C(+) node, so its handy to also know that caps CY1 & CY2 each have one of their leads connected to that node, and bare lead-wires are exposed top-side.
More recent PCB layouts appear to make available all diode leads (for measurement).
So, one may measure residual voltage top-side, and apply the bleed-resistance to the same points.
But, all of that is somewhat moot if these internal events are considered:
a) Upon removal of mains voltage, switching will momentarily carry on until C1's voltage is drawn down to below some assigned low-line voltage value (not sure what that level is, but I think I recall reading '80VAC', somewhere, hence V_C1 would have been taken down from 170Vdc, to about 113Vdc by de-powering the unit) level. Still dangerous, but slightly less threat.
b) Then, C1 will will be further discharged via its own leakage resistance working in parallel with any circuit shunt resistance(s). Shunt paths are not obvious by looking at the PS's circuit, but there might be one through the TOPSwitch's controller. Since that is not clear, I decided on an empirical approach. I charged C1 (still in the SMPS-circuit, a PWB with no line-cord) to 29Vdc, then disconnected the charging cable from the dc-power source. Ten seconds later I touched the charging cable ends to a DVM's input and noted about a two-volt level! Almost 30V down to 2V in ten seconds! Based on that, I expect C1 will have very little charge in it by the time a modifier gets the screws removed from the unit's base, so its cover may be taken off.
I am not saying 'don't worry about it'. Do play it safe by measuring voltage across C1, before separating the PWB from its base casting.
What I mean to say is: Don't be concerned about choosing a bleed-resistance until you know there is some charge that needs to be bled off, and you know where to apply the resistive shunt.
