@tnt @ewenmcneill In the eighties, a Dutch consumer advocacy TV program bought 20 Philips TVs, waited until warranty expired, cut the same wire, and sent them to official Philips service centers for repair.

They then compared the bill to fix it, which varied wildly.

A Philips spokesman argued that finding an issue like that depends on the technician’s experience and some luck and that they charge by the hour.

I still think he had a point.

@joshua The "About Infiniium" screen now identifies itself as a 54832B.

People do crazy things with this scope: replace the LCD panel from 640x480 to 1024x768, add newer motherboards which requires sheet metal surgery, install fast SSDs, update the OS to Windows 10 etc.

But that doesn't really change the way the scope works: it's already on the latest version of "scope.exe". I'm happy enough with the way it is right now. Let's call it a day...

The resistor mod works!

A rise time of 331 ps down from 481 ps, good for 1.05 GHz.

This is within spec, but on the low side compared to the 1.2 GHz that others have seen. It could just be my sample or it could be due to the rise time of the 10 MHz AUX out port being just that.

@joshua has offered me his Leo Bodnar 60ps pulse generator. I'll test that sometime next week.

So converting an Agilent 54831B into a 54832B and thus increasing the bandwidth from 600 MHz to 1 GHz requires some serious PCB surgery: you need to remove 1 resistor...

Left: before, right: after.

The scope is now fully working. The only thing that's left to do is check the bandwidth (which should be > 600 MHz) and then upgrade the scope to >1 GHz.

But first measure the before case. I'm feeding in the 10 MHz AUX output of the scope into channel 1. Without a pulse generator, I have to hope that the edges of this output are sharp enough.

Using the 0.35/t_rise formula, a rise time of 481ps gives 729 MHz, well above the spec of 600 MHz.

Let's remove a resistor.