k-man
Newbie level 3

If anyone here is familiar with switch-mode power supplies (or Tek's HV P/S), I'd be grateful for a little guidance, as I'm at a loss at this point.
The main problem appears to be a general over-voltage of ~30%. There is a feedback trace from the 1.9kv CRT line to a compensator (opamp mc1458 ) which in turn adjusts the base bias level of a switching trans (a Toshiba 2SA1264N). This regulating process appears to be functional -- the bias is adjust to up to +3.44V when the HV output is at overvoltage (clipping the switch-tran), and ultimately falls to a -1V bias on the undervoltage portion of the power cycle. There are a couple things that have me perplexed:
1) the power cycles are not steady - they're coming in bursts spaced about 24.4mS apart.
2) In between these bursts cycles, the bias-level of the switch-tran is held constant at +1.6V.
3) During a burst, the oscillation period is 22uS. That would produce a ~45KHz signal, rather than the spec'd ~50KHz.
Along those lines, two questions:
1) I've replaced the switch trans (2SA1264N) with a KTB688, which is "supposed" to be functionally equivalent. Even if it was different at some level, would that cause the above symptoms given that it is being driven by a bias-compensator on a feedback loop?
2)I've completely disconnected the CRT plug/lines. So, aside from the board circuitry, the P/S is running without a load. When this problem first started it ultimately produced arcing off of the CRT's aquadag, so I've been hesitant to reconnect it for fear of overcooking the heating coil. Does this generation of Tek's P/S need full load on it to operate properly?
Thanks in advance for any insight or suggestions.
The main problem appears to be a general over-voltage of ~30%. There is a feedback trace from the 1.9kv CRT line to a compensator (opamp mc1458 ) which in turn adjusts the base bias level of a switching trans (a Toshiba 2SA1264N). This regulating process appears to be functional -- the bias is adjust to up to +3.44V when the HV output is at overvoltage (clipping the switch-tran), and ultimately falls to a -1V bias on the undervoltage portion of the power cycle. There are a couple things that have me perplexed:
1) the power cycles are not steady - they're coming in bursts spaced about 24.4mS apart.
2) In between these bursts cycles, the bias-level of the switch-tran is held constant at +1.6V.
3) During a burst, the oscillation period is 22uS. That would produce a ~45KHz signal, rather than the spec'd ~50KHz.
Along those lines, two questions:
1) I've replaced the switch trans (2SA1264N) with a KTB688, which is "supposed" to be functionally equivalent. Even if it was different at some level, would that cause the above symptoms given that it is being driven by a bias-compensator on a feedback loop?
2)I've completely disconnected the CRT plug/lines. So, aside from the board circuitry, the P/S is running without a load. When this problem first started it ultimately produced arcing off of the CRT's aquadag, so I've been hesitant to reconnect it for fear of overcooking the heating coil. Does this generation of Tek's P/S need full load on it to operate properly?
Thanks in advance for any insight or suggestions.