Whoa! I’ve had the Lenovo P360 Ultra SFF PC for a week now, and I FINALLY got the discrete Nvidia RTX A2000 GPU working. It showed only a black screen with the Acer XR382CQK monitor. With a Dell 2717 from my wife’s PC as a stand-in, it would run (briefly) then fall over (AppCrash on NvidiaContainer.exe). My suspicion of driver issues were confirmed by the ace Lenovo engineering team. And I was happy to learn that an old school driver repair still works.

What Old School Driver Repair Still Works?

Good question! Having just written a story for TechTarget about fixing black screens, this was chapter and verse for me. If the current GPU driver falls over, received wisdom goes “roll back a version. Keep going till it works…” I’m actually not sure how far that would have gotten me.

But what the Lenovo engineering folks told me falls in line with that approach. They simply said “install version 511.65” and furnished me with a Lenovo download link for same.

Long story short: I installed the older driver. When I rebooted the machine, the previously non-functional XR382CQK monitor worked like a champ in the miniDP port. I didn’t even have to lug my wife’s Dell 2717 into position instead.

A Further Bulletin from Engineering…

Here’s what one of the engineering team emailed to the group assembled to help me over this hump:

 I checked with our lab and there is a known recent issue with Nvidia’s latest driver 513.12 and later. There will be a P360 Ultra BIOS release by end of month to address the issue. However, the workaround in the meantime is to use driver 511.65.  The symptoms are similar to what Ed is seeing – driver crashes.

Given that insight, a quick confirmation that I was running 516.94, and a link to the download for that older driver version, I got straight to work. Problem solved! Nice to know the old school repair still works. Even nicer to get pointed at the last known working version by the Lenovo team.

There are two flavors of Realtek Audio drivers for Windows 10 and 11. The most prevalent are the High Definition Audio (or HDA) drivers. The less prevalent but slightly more capable are the Universal Audio Drivers (UAD). Confusingly, these show up in Device Manager as Realtek(R) Audio. More properly that would be Realtek® Audio, but DevMgr apparently doesn’t do metacharacters like the registered trademark symbol (®). Whatever you call it, the Realtek Audio Console Goes MIA in the MS Store.

Knowing When Realtek Audio Console Goes MIA

One used to be able to access this app through the Microsoft Store. No longer. Confusingly, the app says Realtek Audio Console in its title bar, but the Store listed it as Realtek Audio Control. Thus, for example, if you visit it at MajorGeeks.com (a usually safe and reliable download source), its Microsoft Store download link is broken. Likewise, a direct search at the Store produces no results. Ditto for a search at the Realtek downloads page.

Thus it looks to me that it’s at least possible that Realtek is de-emphasizing the UAD side of its audio drivers. In the absence of statements of direction or intent, it’s only possible to speculate. But it looks to me like UAD drivers and the app console may be orphaned, and no longer supported.

A Driver Search May Tell…

In looking at UAD drivers for Realtek, I see only Nahimic variants for the last half-dozen versions at Station-Drivers.com. None of these work with the plain vanilla FF00 audio codecs on my now-aging Z170 Skylake motherboard. I do have a B500 AMD rig that supports this Nahimic stuff, though.  In a couple of weeks, I’ll probe this mystery further and see if the Audio Console is available (and working) for that set-up.

Right now, I have a working UAD set-up with drivers that are now about a year old (version 9215.1, dated 8/3/2021). I have been unable to find any newer variants that work. Ditto for a newer version of the Realtek Audio Console (or Control). Very interesting!

My production desktop’s dual monitor setup gets a little wonky from time to time. For some odd reason, the right-hand (primary) monitor will start blinking on and off. It’s annoying, but not overwhelming. When it happens, an apparently flaky video driver forces fix revisits. Basically, I keep trying stuff until something works. By no coincidence, that’s a decent operational definition for troubleshooting.

Items Checked When Flaky Video Driver Forces Fix Revisits

It usually goes something like this:

1. Use the Winkey-Ctrl-Shift-B key combo to reset the graphics driver. It does work, sometimes…
2. Check GeForce Experience to see if a newer driver is available; if so, install it.
3. If using the Nvidia gaming driver, switch to Studio driver, or vice-versa.
4. Uninstall, then reinstall the Nvidia driver. I also recommend using the freeware DDU tool to remove all traces of the old before installing the new.
5. Visit the Nvidia Driver Downloads page, and start trying older drivers, going back one version at a time… The recent entries in that list for my GeForce RTX 3070 Ti appear as the lead-in graphic for this story.

Today’s Fix Occurred Mid-way in Sequence

I got to Step 4 today before the blinking stopped. That’s a bit further than I usually have to go, but that’s Windows for you. I’m just glad I can concentrate on what’s showing on both displays, rather than how one or the other is (mis)behaving.

Some Windows errors or gotchas can be set aside and ignored for a while. Others — especially when they interfere with normal system operation — demand immediate attention. While today’s gotcha was one of the latter, it was familiar. Thus, I knew what to do, and how to do it, with minimum need for diagnosis and root cause analysis.

I just marched through the foregoing list and found my solution in under 10 minutes. I can only wish that all problems were so easily fixed. And that’s the way things are unfolding today, here in Windows World. Stay tuned: there’ll be more!

Note Added August 13

Yesterday, I installed a new NVIDIA Studio Driver (version 516.94). The problem has not re-appeared since I applied my uninstall/reinstall “fix” earlier this week. That jumps me up from Step 4 (original fix) to Step 2 (current fix). And so it goes…

It’s always the little things that jump up to bite you (or me, anyway). In today’s case, it was my blithe assumption that Intel Integrated Connectivity (aka CNVi) wouldn’t prevent the AX201NGW M.2 Wi-Fi card from working on my AMD B550/Ryzen 3 5800X build. Yeah, right! But when I replaced it with a no-name (REKONG) Media Tech MT7921K module (depicted in the lead-in graphic), Device Manager picked up that non-Intel hardware immediately. After a bit of driver fiddling, this US$29 (tax included) nonCNVi M.2 Wi-Fi Device delivers LAN access as it should. It does have interesting limitations, though . . .

Fiddling Means NonCNVi M.2 Wi-Fi Device Delivers LAN Access

At first, after plugging in the device, I saw only non-working BlueTooth and Network Adapter devices in Device Manager. This informed me that Windows couldn’t find the required drivers on its own. But a quick search on “Windows 11 drivers for MT7921K” quickly turned up what I needed. They’re available from Lenovo, as it turns out, with a separate .exe for each of Bluetooth and Wi-Fi.

As the owner/operator of half-a-dozen (or more) Lenovo laptops, I’m quite familiar with their self-installing drivers. After downloading and installing them, here’s what I see in Device Manager:

With the right drivers installed, the BT components and the Wi-Fi interface all show up. Good!

Just a Few More Things

Wi-Fi behavior on desktops can be interesting. The interface has a tendency to turn itself off upon reboot, I’ve learned. I’m also trying to figure out why I can access the LAN (via the nearby Asus AX6000 router), but I can’t yet get Internet access through this interface. I have a wired GbE connection that works fine, but had hoped to switch over to wireless. So now, I’m researching those two issues in hopes of finding solutions soon.

A little more time put intro troubleshooting the M.2 Wi-Fi card tells me lots of interesting stuff:

• The lack of an external antenna means the device doesn’t see that many wi-fi interfaces as it scans the airwaves. Thus, for example, it doesn’t see the Spectrum-supplied router in my bedroom closet (all of my laptops in the same office see it quite well).
• The fastest throughput I can get on the device is between 250 and 300 Mbps (observed through a connection to Fast.com).
• The 2.4 and 5 MHz connections to the “office router” are flaky in interesting ways: sometimes, I can access one or the other to get on the LAN, but don’t get Internet access. At other times one channel or the other will be inaccessible. Again, I attribute this to lack of an external antenna. My son has a PCIe 802.11ax adapter card with triple external antennae in his bedroom, and he gets up to 900 Mbps from the bedroom closet router, and up to 500 Mbps from my office router.

No External Antenna Is NOT a Plus

I’m increasingly inclined to observe that an M.2 Wi-Fi card makes sense only where close proximity to a WAP is available. It’s probably not a good idea for machines that do lots of heavy upload/download stuff, either. That’s kind of what I wanted to learn more about, so I’m not disappointed by this experience. I feel like I understand the capabilities and limitations of these devices much better now. I will keep my GbE wired connection going forward, too: the M.2-based Wi-Fi is not fast enough for my needs. If I’m ever *forced* to go wireless, I now understand that a PCIe device is my fastest option.