I don't have much knowledge of electronics and unfortunately I must trust what they tell me

Since they are authorized by zeusbtc then I would still trust them, what the main branch says does not really contradict with what the branch in your country told you, it only weakens the probability of the latter, also I must make clear that it's not just because I tend to trust the main branch more - I base my points on that alone, it's more like common sense, what the Chinese branch said has much higher probability of happening.
but is it also possible that if one fails it will send incorrect information to the other 3?
Indeed, you are right, it's very possible, however, it's unlikely, the way the board is designed should not let a single sensor stop the other 3, the proof is we are aware of some hash boards that work with less than 4 sensors, which means the probability of a single sensor stopping all others is kind of low.
In my case the 4 sensors of the 3 hashboards are failing, so I read this failure is related to the PSU?
Yup, your problem is power related, I am 99% sure it is, and let me explain why:
Reason 1:
You now have a total of 12
(3*4) broken sensors, does that make sense? NO
Reason 2:
You have at least 1 bad chip/heat sink on all 3 boards, how possible is this scnerio? i would give it 1% at best, provided they all failed at the same time.
Reason 3:
Something that has an effect on every aspect of the miner, we have two things, the control board and PSU, from the various reports and personal experience I doubt the control board will cause such an issue, so what is left? the power supply.
I would still suggest that you flash the recovery file using an SDcard, and then flash another firmware, see if that helps,
Something that also disconcerts me is that my T17 in the cooler hours of the day works normally.
Which is why I thought it made sense that the temperature sensor.

This actually adds more strength to the power supply being the issue, the sensors are designed to work at temps between -40c to 125c, it doesn't make sense that 10 or 15 degrees change in the ambient temp will make them work or not work, however, a 10 degrees change makes a world of difference to the PSU and the chips, I am suspecting that your PSU is overheating at noontime and fails to deliver the DC voltage needed by the boards, also during peak hours the voltage at your house/farm might actually drop to below 200v and thus causes the miner to not hash, make sure you measure that too.
If you have a spare PSU, I suggest you try it, or at least measure the DC voltage coming into the boards from the PSU
you say you solved the problem by pressing on chip 1
Did your t17 fail on just one sensor or all of them like mine?
4 sensors 1 board, the other two boards were fine and reading all sensors, so it's a bit of a different issue from yours.
This could actually be true. My 2nd T17+ got broken a minute after i tried to overclock it. Worked fine for 2 months before that. So please do not go over 800 Mhz on default cooling.
overclocking has more effect on the chips than on the temp sensor IMO, so it's more likely that some of the chips got toasted, you can test them with a multimeter by the way.
try lowering the frequency.
That's an excellent idea, even when the PSU is having some issues, going with a lower frequency might fix it.