I found the detailed report on the Apollo 12 lightning strike. Summarizing with added comments:
The fuel cells were knocked offline by rapid false triggering of the SCRs in the fuel cell overload sensors. (An SCR is a silicon controlled rectifier that behaves like a solid-state latching relay.) The triggering had to be false because the fuel cells were simply incapable of producing enough current to trip them so quickly.
The trips are best explained by lightning induced voltage rises >500 V/μs on the anodes of the SCRs. dV/dt triggering is a well-known property of SCRs.
When the three fuel cells dropped off the buses, the entire CSM load of 75 amps was thrown onto entry batteries A and B. That's a very heavy load for batteries rated at only 40 amp-hours at a 25A rate. (Battery C was offline in reserve.) The bus voltage dropped momentarily to 18-19 V but recovered to 23-24V within a few milliseconds.
"The low DC voltage on the main buses resulted in the illumination of the undervoltage warning lights, dropout of the signal conditioning equipment, and a lower voltage input to the inverters. The momentary low voltage to the inverters tripped the ac undervoltage sensor and caused the ac bus 1 fail light to illuminate. The transient that affected the silicon controlled rectifiers in the fuel cell disconnect circuitry also affected the silicon controlled rectifiers in the ac overload circuits in the same manner."
The next section discusses the permanent burnout of 9 sensors in the service module, but they did not materially affect the mission.
There were actually five restarts of the computer, most likely caused by that same momentary +28V bus voltage drop when the fuel cells disconnected.
(end of paraphrase)
I see nothing in the diagrams to suggest that the two SCE power supplies weren't identical. Normal SCE power consumption was about 35W, enormous by today's standards but reasonable for the 1960s. Each produced the same set of output voltages: +20, -20, +10, +5. All four output voltages were themselves telemetered since the accuracy of all telemetry readings depended on them. I can't think of any reason to design only one supply to operate over the widest supply voltage range possible. I can't say what that range would be without knowing the specific design of each supply, but it seems reasonable that the 20V supplies would require an input of at least 20V. We do know that the +28V DC buses dropped below 20V when the fuel cells dropped off but recovered to voltages well above that in just a few milliseconds.
Today it would be easy to design a telemetry system to be much less susceptible to this sort of power supply glitch. Switching power supplies have pretty much displaced linear designs for most applications and they are inherently able to withstand a much wider range of supply voltages -- to the point that many "universal" designs can handle anything between 100 and 240V without any operator reconfiguration. They can usually store enough energy to ride through fairly long supply outages.
