Dear all, thank you for the kind comments and the suggested links. I have been able to unsolder the push-pull transformer T948, quite a bit of work, but I managed to keep the PCB intact. I used the desoldering pump to clean the pads and then I applied hot air with a blower set to 400 °C. It went out nicely, I did not even swear : ) I tested the transformer and both primaries have a 2-µH magnetizing inductance which is a clear indication of failure. No wonder the inverter did not start. The transformer sticker indicates the following part number: 120-1594-00. However, the part list on the 2235 service manual says 120-1348-00 or 120-1348-01. It seems to refer to some serial number where some parts were replaced by other. I bought the scope in 1986, I will provide the serial tomorrow. So what would you recommend, get the 120-1348-00 (available at Qservice) or wait until the 120-1594-00 pops up somewhere? Any suggestion will be appreciated, thank you.
Transformer damage in a switching power supply is rare, furthermore in a self-protected version as with the 2235. The transformer does not look burnt and as I said, it peacefully passed while I was working with the scope: no noise, no smoke. Given the primary-side current limit, I doubt a short-circuit on one of the secondaries could cause this. I am more thinking of a dielectric damage given the high-voltage winding that drives the multiplier. Besides this, I don't see.
Regarding the pre-regulator, as both TIPs have been removed, I connected an electronic load between TP940/TP950 and I could power up the device with a regulated voltage around 41-42 V. It starts up ok up to Iout = 600 mA and I can increase the current up to 1 A. At 1.1 A the protection trips and it enters hiccup which seems normal given the 0.2-ohm/200 mV sense/max sense voltage configuration. However, it goes into hiccup when I start up already loaded with 700 mA and above. Is this normal ? Usually, it happens to power supplies in which the aux winding does not take over quickly enough and the controller Vcc prematurely touches the UVLO. What do you think? If you believe it should start full power at a 1-A load then I will look into this tomorrow. However, not sure the scope circuitry at power up draws the full 1-A at once... Thank you!