Hi Loren,
I only have an SP4, and most of the people that I've corresponded with also have SP4's, so it's a bit difficult to pull together what may be going on with SB's, but your analysis is very valuable. Would you mind contacting me at ******@pmFAQtory.com? If you have the time, perhaps you might be able to help us to do some testing on your device, put the finishing touches on my document, and help us to drive Microsoft to resolve this for all of us.
I've corresponded a bit with David Catuhe, Principal Program Manager for Windows & Devices at Microsoft. He initially told me that he had found the bug reported somewhere within Microsoft, presumably in an internal database, and that they had reproduced the bug and someone was investigating it. He even mentioned that he thought that a fix was targeted for the Creators Update in 2017, and asked that I ping him every three weeks for a status update. Subsequently, I went to a Microsoft Store and was able to swap my 36W original SP4 power adapter for a 65W one, and that immediately resolved (or masked) the bug on my SP4. I also ordered one of these 36W Kingdo clone power adapters: https://www.amazon.com/gp/product/B01GPFM5GA/. It, too, worked, either resolving or masking the bug. I told David that swapping out power adapters had fixed the problem for me.
But I then considered the possibility that the power adapter might not be the whole story. I've been collecting data behind-the-scenes from various users plagued by this bug. I'm cautious to draw conclusions, so please remember that my document is just a draft, and could contain erroneous conclusions.
If you're saying that you're confident that you've been able to send 15V of steady DC with no voltage spikes that could trigger power state changes to your SB, then I have to wonder why the 36W Kingdo adapter resolved the bug on my SP4. It's very clear that PBT_APMPOWERSTATUSCHANGEmessages cause the focus to be usurped from the taskbar, which causes thumbnail previews to disappear. The question is, What causes these messages?
One possibility is voltage spikes from defective power adapters. Another is a design flaw in the Surface device's power management circuitry or battery (batteries). Another is a bug in the ACPI driver. Possibly another is a firmware bug. Whatever the problem is, it's power management-related. We can only do so much. We need Microsoft to figure this out. Since the bug has existed on the platform ever since the SP3 was released, but it *hasn't* affected most SP3's/SP4's/SB's, from what I gather (and I've asked many users), why us? What's different on our devices or power supplies?
Our crude debugging efforts are limited by the fact that we don't have a deep knowledge of the Surface architecture. We have no idea what all of the interacting drivers, firmware, and other mediating and moderating variables are. So given this, how can we collect enough data to get Microsoft to not drop this and pretend that it's not a real and serious (show-stopping) bug?
Nothing has aggravated me more than Surface Support saying that they couldn't reproduce the bug, until David told me that someone else within Microsoft had.
All I can tell you for certain is that on my SP4, the nanosecond that I plugged in the 65W Microsoft adapter, and a few days later the Kingdo 36W adapter, the bug disappeared instantly.
I invite you to contact me if you'd like to help us close this nightmare out.
Thanks,
Steve