I have the Mind Pro and the XPlay. Mind Pro running Windows 11 Pro (OS build 26200.8875) and is at BIOS version 1.01. XPlay running firmware 1.00. Product model A1021.
The problem I continuously face is when putting my machine to “sleep mode” or “hibernate”, upon booting from those states, the Mind Pro gets into what I call “limp mode” and the max core speed is 400 MHz verified by HWiNFO64 “Core Clocks” set to 400 MHz and “Core Effective Clocks” never able to go above 400 MHz. I have to shutdown my computer, disconnect the Mind Pro from XPlay, disconnect the power cable and press and hold the power button on the XPlay for 10 seconds then reconnect everything and boot again to reset.
According to the Windows Mind software application for XPlay, there is a firmware update to 1.01 but every time I try to update, the update fails.
Can this “limp mode” issue be fixed and how do I update XPlay so it doesn’t fail every time?
**Update #1: I was continuously working and all sudden everything slowed down to a crawl. I’m locked at 400 MHz without having put system to sleep or hibernation. Windows Power Plan is at “High Performance”.
To those experiencing the same problem. I was able to narrow it down with iPTAT monitoring (with logging enabled) and Claude performing output analysis. The issue appears to be with how the Khadas firmware is controlling the PROCHOT# pin. The iPTAT logs show that the Khadas embedded controller asserts a PROCHOT# pin even at ~50C thermal temperature and then never releases it. This pin is also locked and controlled by the firmware, and we are not allowed to manually release it via software like ThrottleStop.
There is also an issue with the battery to where I’d be actively working and the system just shutdowns down. According to logs, the reported battery will deteriorate the “full charge capacity” (FCC) and “design capacity” (DC).
(Date - Full Charge Capacity - Design Capacity)
Sep 12 - 48,290 mWh - 47,932 mWh
Sep 13 - 37,672 mWh - 38,815 mWh
Sep 14 (onward) - 1,140 mWh - 7,980 mWh
from 48,290 mWh (FCC) and 47,932 mWh (DC) to 1,140 mWh (FCC) and 7,980 mWh (DC) over the course of a few days. According to a report from Claude, the battery deteriorates and will report a capacity of “2.4%” until it bottoms out. The configuration for the battery’s “cycle count” reads simply as “1”, the serial number reads “123456789” and the manufacturer reads “KHADSA” (yes, that exact spelling) which means that the configuration becomes corrupt.
According to Claude’s analysis, this is a battery hardware issue and not something that can be fixed via firmware.
I’m seeing the exact same failure signature on my Khadas Mind Pro.
I ran a read-only battery inspection and the reported data matches the corrupted-battery state described above:
Manufacturer reported: “KHADSA” (corrupted string, same as reported)
Serial number: “123456789” (placeholder value — configuration is corrupted)
Chemistry: LiP
Cycle count: 7 (far too low for this level of degradation)
Full Charge Capacity (FCC): 1,140 mWh
Design Capacity (DC): 7,980 mWh
Current remaining capacity: 969 mWh (reported as ~85%)
Given the original design capacity of roughly 48,290 mWh, the battery is now reporting a state of health of only ~2.4%, despite having only 7 charge cycles. No physically healthy battery degrades to ~2.4% of its rated capacity after just 7 cycles, which strongly points to a corrupted battery/gas-gauge configuration (likely firmware/EC-level) rather than normal wear — fully consistent with what has been described in this thread.
— Assisted by model: DeepSeek-V4-Flash (official release)；English isn’t my first language, so please excuse any unclear wording. Thanks!
A new finding: In one instance, it suddenly locked to 0.4 GHz again, but a 2-hour cold reset followed by power cycling and re-enabling did not restore it. After disabling the 80% battery capacity protection in the MIND software, it suddenly returned to normal and has been operating safely for one day.
Precisely the same random CPU speed degradation issue occurred twice for My Mind Pro this morning.
On resume from hibernation this morning the system behaved in “slow-motion” and the Event log showed numerous errors and warnings. After finally reaching the desktop any activity/actions at all performed/occurred VERY slowly.
Given the numerous failure issues I have had with wake from “Sleep”, I assumed that something had finally corrupted the OS.
I made a system image (for fallback protection) and then restored a system image from a week ago.
Even the Recovery Environment behaved extremely slowly during the image restore.
I shut down the machine after the image restore rather than going directly to reboot on the concept that “perhaps” a cold boot after a power unplug “may” resolve any remaining stale hardware (i.e eprom) leftovers.
After power unplug > 30 second wait > power re-plug > cold boot the restored image was performing at normal CPU speed then after about an hour the system suddenly slipped back into “slow-motion”.
Turning off the Khadas Mind “Smart Charging” setting resulted in the system immediately returning to normal CPU speed
As a test, I restored the “slow-motion” image that I had made this morning…after rebooting the system was performing at norma CPU speed…and, the Smart Charging setting was showing as turned off in the Mind app (meaning that the eprom had retained the “off” setting and that the app in the restored image read the eprom and not a value that was previously saved in file).
Takeaway: Khadas drivers and/or firmware are the issue here as well as in the Task wake issues I note in my thread “Mind Pro - Crashes & Driver Failures (updated)”. There are definitely some unresolved driver and firmware issues that have not been resolved as yet for the Mind Pro…the workaround for the CPU “slow-motion” issue (for now) appears to be what nineyuan discovered !!!
I appreciate everyone chiming in here! Though I will admit, I’m cutting my losses and going elsewhere. My stability issues with my Mind 2 and Mind Pro and this lack of support is enough perspective for me to take the Khadas investment as a net loss and move on. Best of luck to you all! Hope you all get your issues resolved.
I fully get the frustration! Stability and reliability are critical to trusting work and data on any system.
I have managed to achieve a stable Mind Pro configuration after many hours of configuration changes and testing…although, I still am wondering when a new “gotcha” will crop up. The power management firmware and drivers do have issues that have not been resolved. There are some workarounds that “appear” to stabilize the system or, at least, my system configuration.
These generally involve circumventing or neutering various items to prevent or minimize the power management that occurs during or as a result of the “S0” Modern Standby “Sleep” states. The longer an “S0” system is in “Sleep” the more power saving transitions there are occurring as the system goes deeper into “Sleep”. Those transitions are apparently involved in many, if not all, of the power issues…at least that is what I am seeing. The shorter the system is in “Sleep” the fewer the number of power transitions that occur…thus, get out “Sleep” and into “Hibernate” within some short period of time and the power transitions stop and the power issues disappear. This requires accessing the “Hibernate after” capability of the system which the OS does not display in “Settings > System > Power & battery > Screen, sleep & hibernate timeouts”. The lack of the “Hibernate after” setting appearing is “probably” related to the system being identified as an “S0” capable system or due to an OEM system capability that is presented or not presented in their configuration strings for the system. The configuration strings are defined by the OEM for their hardware not by users or by the OS.
But, the “Hibernate after” UI can still be displayed thru Control Panel via a registry setting and, most importantly, is still fully supported and fully operational by the OS and, even more importantly, still functional on the Mind Pro. Thus, a path to circumventing the power state transitions that start to occur due to Modern Standby deep sleep is still available. This is how my two Mind Pro systems are configured and have achieved a stable and reliable “Sleep > 10 minutes > Hibernate” transition and along with a few other power related tweaks are avoiding the crashes and wake from sleep failures. I will post a full recap of the entire configuration in my thread “Mind Pro - Crashes & Driver Failures (updated)”.
The CPU speed change appears directly related to the “Smart Charging” behavior and turning that off “appears” to have resolved that issue…although, the app notes it will automatically turn it on again after 5 days…which means I’ll turn it off again until Khadas fixes this issue!. There are other issues that show up in the Event log and also appear to be power management related i.e. “WHEA-Logger” which is associated with the PCI root hub. Turning off “Link State Power Management” in “Control Panel > Power Options > Change plan settings > Change advanced power settings > PCI Express” appears to have resolved (or at least radically “calmed” down) those issues. Again, there appears to be some firmware/driver issue with proper powering/behavior of the ports that results in various system and/or attached hardware errors. The attached hardware issues that a number of folks are having “may” be directly related to the “WHEA” issues…?
Generally, it seems that the aggressiveness of the power management is emerging as an issue for the Mind Pro. And, that aggressiveness is manifesting itself in a number of odd behaviors, failures and crashes. You may well have been caught by them…I certainly was! Getting out of “Sleep” and into “Hibernate” before the deep sleep transitions start accumulating “appears” (for me) to have significantly resolved most, if not all, of the current power related issues…if and until Khadas resolves the firmware and driver issues.
