Unsolved
1 Rookie
•
3 Posts
•
2 Points
0
151
February 15th, 2026 03:31
CPU0704 CPU1/CPU2 Machine check error detected on Front LCD Panel
Error like "CPU0704 CPU 1 MACHINE CHECK DETECTED ERROR : SYSTEM POWER cycle system check" what could be the issue we provides new termal paste still issue persistent on iDRAC WAS pass
No Events found!


user_395b33
1 Rookie
•
3 Posts
•
2 Points
0
February 15th, 2026 05:23
IDAC and stress test didn't noticed any issue.For reference
Dell-Martin S
Moderator
•
3.8K Posts
•
25.5K Points
0
February 17th, 2026 14:34
Hi,
Update to the latest: BIOS, iDRAC, Lifecycle Controller, and if applicable, backplane and PSU firmware, since firmware as a common cause for CPU machine check events and has known CPU0704-related defects on some generations.
Check SEL / LCLOG around the CPU0704Look for related entries: CPU0000 “Internal error”, DIMM errors on a specific slot (often A1/B1), voltage/VRM warnings, or “problem detected related to previous boot”.
If you see DIMM or internal CPU errors close to CPU0704 timestamps, those usually point to the real root cause.
Hardware reseat / isolationPower drain procedure Dell recommends: power down, remove AC and network, hold power button ~15–20 seconds, then reconnect and boot.
Reseat: all DIMMs, especially those on the CPU indicated (or both CPUs if unclear),both CPUs, check for bent pins or contamination in the socket,PSU connections and any risers near CPU power.Remove non-essential PCIe/third party cards while testing, because add in cards can provoke MCEs indirectly via power or bus errors.
Swap components to see if the error follows:
Swap CPU1 and CPU2 and check if the CPU0704 message changes to the other socket, as Dell support themselves often suggests.
Test with minimum to POST config: 1 CPU, one DIMM in the primary slot for that CPU, no extra PCIe devices, then observe if CPU0704 still logs over several cold boots.
Check if it’s a known logging quirk vs real faultOn 17G systems with early iDRAC10, CPU0704 may point at the wrong CPU index, but the event is still a real MCE; you just must not misinterpret which socket is implicated.
If after firmware updates and reseating, the machine is stable, logs show only a single historic CPU0704, and no new events appear over repeated cold boots, it may be a one off and safe to monitor.
user_395b33
1 Rookie
•
3 Posts
•
2 Points
0
February 19th, 2026 01:40
Thanks for the help! I’ve gone through the BIOS/iDRAC updates and the hardware isolation steps you provided. I'm now seeing a VLT0204 code Voltage out of range appearing.