Can you please show me which chip is the hybrid chip on the board ? I have attached a photo of the acquisition board.
Did you manage to correct the channel 2 and 3 on your 54831 ?
See those four metal covers with the red sticker on each? Remove the cover and you'll see a white chip, those are the hybrid chips (I don't have the part number handy, but can look later if you need me to).
I'm not saying the hybrids are the issue, but it was on mine.
Also, I didn't mean switch to 50ohms during calibration, I meant to take measurements in both high Z and 50ohms. Mine may have worked in high Z (don't remember now), but it was clearly bad in 50ohm mode.
With my unit, I took measurements on a good channel versus a bad channel, traced the circuits, drew the schematic as I found connections, and discovered suspected components were replaced by someone else; but not the same values. With the help of others on here, comparing good channels, etc... I replaced the components, but no luck. The end result was having to replace the hybrid in both ch2 and ch3.
Assuming they are the same hybrids, they have a heat sink underneath (not a BGA) and require a hot air to remove. Too much heat and you'll blister the PCB (this happened to me on ch2), but you need the heat to melt the solder on the underside. The rest of the chip is all soldered to pads on the side of the chip, but they are half circles. Too much solder and the solder runs underneath and to the adjacent pin (also happened to me).
You can't find replacements anymore and I resorted to someone selling used ones on eBay for around $175 each (it was an expensive gamble on my part, but I bought them)
Also note, when you remove those metal covers, the tiny sockets they sit in can easily be caught on clothing, acid brushes (if you need to solder and do any cleaning), etc... and bend out of shape. Reseating the cover can be difficult as any bent sockets can bend inwards or you may not catch all the sockets (they are basically little cups), have to remove the cover again, try reseating, etc...
Update: Also wanted to note, if you didn't figure it out, all four channels are identical. So if you know another channel is good and one is bad, you can compare. I THINK a possibility exists a few components may be orientated differently, but don't quote me on this.
In my case, I took a few measurements on random points and quickly discovered different voltages. This led me to trace further. Between trying to fix the hard drive that crashed within the first few weeks of owning the scope, researching what causes the scope to fail in 50ohm mode, replacing the chips, the components that were replaced with wrong values, a ZIF connector that was bad, and one or two other issues, it was literally close to a year of work (on and off).