I've had essentially flawless performance since my new build in April 2024, until 24H2 was installed, and now unable to roll back to 23H2 as I've gone past the 10 days or whatever was the limit.  After 24H2 was installed noticed a few audio cutouts during flights.  They have steadily worsened and just recently noticed it's not just audio, video stammers with it.   These cutouts seem like a hardware issue at this point as it's not standard/normal 'stuttering'.  I do use high quality Sony studio monitor headphone and wiggled the case front jack to see if any cutouts with that to no avail.  Of note, it's very random in that never know when it will start during a flight except it's never when pre-flight and initial climb is happening and in general it tends to worsen over the course of a flight.  I've been fooled into thinking an intervention made a difference only to have the issue resume.  Very random, but again, tends overall to be more frequent and worse.  Here are all the things I've done with no success so far:

1. Clean installs of the latest Realtek drivers off my mobo's website, latest chipset drivers, network adapter drivers and GPU drivers.  I've tried a few recent NV drivers not just the latest.
2. LatencyMon shows no issues with DPC latency and says I should not be experiencing audio cutouts.
3. ChatGPU told me to swap out the latest Realtek driver for Windows HD audio driver which I did, to no avail.
4. Updated the BIOS which I had never done since the initial build.
5. I did two PSU tests, one with a tester I've had for several years and it indicated a PG value of 80ms, falling out of the normal range of 100 to 500ms, and its instruction guide says if <80ms the PSU needs replace.  Yet looking at voltages under load while in flight everything looks normal.  I'm told PG value mainly has to do with boot up, but have never had any issues with boot up.  The other test was per Corsair's guide:  jump two pins on the 24pin connector and if the PSU fan spins it's fine.  If it spins, it's still good.  It did spin initially then quickly stopped and stayed stopped.  Someone said this might be normal.
6. Ran OCCT for an extreme stress test, just for 15 minutes, zero errors, good temps, no issues detected.
7. The RAM I have has worked with EXPO settings and memory test checks out that it's okay.  I did upgrade to 64Gb a few months ago and didn't have these issues then. 

My best guess now is some weird issue with 24H2 unless it was a total red herring.  And yet when I posted about this haven't had any corroboration from others that this was an issue.   I did initiate a request to RMA the Corsair RM1000x PSU and am awaiting to hear back.

So at this point run dry of potential solutions and this has pretty much killed being able to use MSFS 2020 as it can get pretty bad.  Much appreciate any insights I haven't considered.   

 

Edited December 23, 20241 yr by Noel

Actually that was a mistype.   It said if <100ms it needs replaced.  And yet PG is supposed to most impact bootability at least that was how I interpreted what I read.

New PSU seems to have solved the issues.  Two full 2-3h flights into dense scenery in complex planes and not a one cutout.  I hope I'm not celebrating too early! 

15 hours ago, Noel said:

ChatGPU told me

Are you sure? 😊

Happy that you solved your problem, Noel 👍

I don't want to spoil anything for you - but I thought the same thing after replacing the sound card - 2 days without dropouts - and then sound dropouts again ! this is so inconsistent and incomprehensible but it is due to the MSFS !

But I hope it doesn't cause any more problems for you! !

BR

Edited December 24, 20241 yr by AUA425

4 hours ago, AUA425 said:

I don't want to spoil anything for you - but I thought the same thing after replacing the sound card - 2 days without dropouts

Just starting test flight #3 so we'll see if it holds up.  I have been fooled many times thinking we were good to go only to have the issue reappear, but usually it was the next flight.  But at least we have corroboration in a failed PSU test.  Corsair contacted me and will replace the unit so they too are buying my testing procedure as shown to them.  🤞

8 hours ago, David Roch said:

Are you sure? 😊

Happy that you solved your problem, Noel 👍

100% sure because in addition to PSU as a potential culprit it covered every other possibility I'd done and then several others.  Now, in mid-3rd flight I've heard the first of maybe 3 brief cutouts...AGAIN.   This makes me wonder if my mobo might apply here to one of CGPT's points:

• Edge Cases: Certain motherboard chipsets, especially high-performance ones with precise power delivery requirements (e.g., for overclocking), may enforce stricter timing tolerances.

So perhaps the shiny new PSU improved things a bit but ultimately is a mainboard issue relative to 24H2.  Who knows.  Now well into flight I'm hearing a few more 🙁

I remember I had some Asus board with a 12900k build.  I had the weirdest issues with electrical type random noise and even sometimes USB cutouts.  Tried so many different things along the way.  Ended up just selling off the components and doing my current AMD build on a gigabyte MB.  These things are maddening Noel and I feel for you.  Hope it ends up working out. 

1 hour ago, Noel said:

100% sure because in addition to PSU as a potential culprit it covered every other possibility I'd done and then several others.  Now, in mid-3rd flight I've heard the first of maybe 3 brief cutouts...AGAIN.   This makes me wonder if my mobo might apply here to one of CGPT's points:

• Edge Cases: Certain motherboard chipsets, especially high-performance ones with precise power delivery requirements (e.g., for overclocking), may enforce stricter timing tolerances.

So perhaps the shiny new PSU improved things a bit but ultimately is a mainboard issue relative to 24H2.  Who knows.  Now well into flight I'm hearing a few more 🙁

No, no, maybe the quote doesn't appear on your browser.
I was asking if you were sure about ... ChatGPU!!

😂

1 hour ago, David Roch said:

No, no, maybe the quote doesn't appear on your browser.
I was asking if you were sure about ... ChatGPU!!

😂

Haha I thought maybe you were hinting YOU suggested it might have been PSU related!  I find ChatGPT SUPER useful for pretty everything.  Look at my initial question after doing the PSU test on my PSU:

In computer power supplies, what does the value, "PG" in milliseconds refer to when testing the power supply?

ChatGPT said:

ChatGPT

The "PG" (Power Good) value in milliseconds refers to the Power Good Signal or Power OK Signal delay time in a computer power supply. This value measures the time it takes for the power supply to stabilize its output voltages after being turned on before signaling the motherboard that it is ready for operation.

Key Points about PG:

1. Purpose:

   • When you power on a computer, the power supply performs an internal check and stabilization of its output voltages (e.g., +3.3V, +5V, +12V).
   • Once the voltages are stable and within acceptable ranges, the power supply sends a "Power Good" signal to the motherboard via a specific wire (commonly the gray wire in ATX power supplies).

2. PG Delay Time:

   • The PG value indicates the time from when the power supply is turned on until the Power Good signal is sent.
   • Typical values range from 100ms to 500ms.
   • If the PG time is too long or too short, it may indicate an issue with the power supply or its compatibility with the motherboard.

3. Importance in Testing:

   • Ensures the power supply stabilizes properly before the system starts booting.
   • A power supply with an incorrect or inconsistent PG delay may cause issues like failed POST (Power-On Self-Test), random restarts, or system instability.

Measuring PG Time:

PG is typically measured using a power supply tester or oscilloscope, where the exact timing of the Power Good signal can be observed relative to the application of power.

Let me know if you need more details!

4o

 

You said:

Thank you for that, as usual. I'm not sure how a PG time that is below 100ms, in my test case 80ms, could have anything to do w/ compatibility with the specific motherboard. Can you elaborate?

ChatGPT said:

ChatGPT

You're absolutely right to question how a shorter PG time could impact compatibility with a specific motherboard. Let me clarify and elaborate.

Why PG Timing Matters

The PG signal timing exists to ensure that the motherboard only starts initializing components (like the CPU, RAM, and other peripherals) once the power supply has stabilized its output voltages. If this timing is incorrect, the motherboard might try to boot before the power supply is ready, leading to instability or boot failures.

Short PG Time (e.g., 80ms):

1. Motherboard Initialization Timing:

   • Some motherboards have specific timing expectations for the PG signal, typically assuming a delay of 100ms to 500ms.
   • If the PG signal is too short (e.g., 80ms), the motherboard might interpret it as premature and not ready for stable operation. This can cause issues like:
     • Failed POST (Power-On Self-Test).
     • Unreliable power delivery to critical components during initialization.

2. Voltage Stabilization:

   • A PG time below 100ms might indicate the power supply is signaling "Power Good" before its voltages are fully stabilized. While this could work on some motherboards, others might be more sensitive to fluctuations and expect additional delay.

3. Motherboard Design Tolerances:

   • Motherboards designed to meet ATX specifications (which suggest a typical PG delay of 100ms–500ms) might behave unpredictably with out-of-spec power supplies.
   • In contrast, some motherboards are more forgiving and can handle shorter PG times without issue.

When Compatibility Becomes an Issue

• Edge Cases: Certain motherboard chipsets, especially high-performance ones with precise power delivery requirements (e.g., for overclocking), may enforce stricter timing tolerances.
• Legacy Systems: Older or more basic motherboards may tolerate deviations more flexibly than modern ones designed for tighter standards.

Verifying the Issue

If you're concerned about a short PG time, you can:

1. Test with Another Motherboard:
   • Verify if the same power supply exhibits issues across different systems.
2. Monitor Startup Behavior:
   • Look for symptoms such as instability during boot or power-on failures.
3. Use a Power Supply Tester:
   • Confirm that other voltages are within spec (e.g., no droop during startup) and that the short PG time isn't indicative of an underlying issue.

In your test case (80ms PG time), if the system boots and runs stably, it’s likely not a concern. However, if you notice boot irregularities or instability, this could point to the PG timing being out of spec for the specific motherboard.

4o

 

You said:

Let me add additional details about what we're experiencing. About a month or so ago I noticed some what I thought were audio cutouts while flying in MSFS 2020. My PC is a custom built 8 months ago 7800X3D machine with RTX 4090 GPU and 64GB of DDR5 6000. It had been performing marvelously since built. Of note I believe this began with Windows Update's automatic install of 24H2 in Windows 11 Home. This could also be a red herring. Over the past month I have done all the usual interventions to troubleshoot or resolve these audio cutout including updating motherboard BIOS, updating all chipset, video and audio drivers, and more. Generally a performance-demanding flight scenario will start out fine, for example up to cruise altitude and beyond. Then an occasional audio cutout will ensue. Lately it can be as frequent as every few seconds and what I noticed today was it's not just audio--video streaming cuts out as well. And that is not like a typical stutter per se which I rarely see now in MSFS. Anyway, I thought perhaps this could indicate a failing PSU. And yet in your above comments it would appear more about boot issues, not computationally demanding scenarios during runtime. My PSU's tester said 80ms PG. What more can you add?

ChatGPT said:

ChatGPT

Memory updated

Thank you for the additional details; they shed light on possible causes. Based on what you're describing, this issue seems more complex than just PSU-related, but let’s dissect it step by step:

---

1. Why the PSU's PG Signal May Not Be the Primary Culprit

• PG and Booting vs. Runtime Behavior:
  • The PG signal primarily impacts boot behavior. Once the system is running, it’s less likely to cause runtime issues like audio/video cutouts unless the PSU is struggling to deliver consistent power under load.
• Runtime Power Delivery:
  • If your PSU is under stress (e.g., due to aging capacitors, insufficient wattage, or power spikes from demanding GPU/CPU usage), it could cause intermittent instability. However, given your PSU is relatively new (8 months) and you’ve seen no major power-related crashes (like sudden reboots), this is less likely.

Steps to Check:

• Use software like HWMonitor to watch for voltage fluctuations on the +12V, +5V, and +3.3V rails during demanding MSFS scenarios.
• If significant deviations (>5%) occur, the PSU might be underperforming.

---

2. Windows 11 24H2 Update as a Factor

• Drivers and Latency:
  • Windows updates, particularly major ones, can introduce driver-level changes that negatively affect latency-sensitive applications like MSFS 2020. Even though you've updated your drivers, the interaction between the new OS and drivers could still be problematic.
• Audio Interruptions as a Symptom:
  • Audio cutouts and microstutters often stem from DPC (Deferred Procedure Call) latency. Video cutouts accompanying them point to broader latency or resource contention issues.

Steps to Investigate:

• Use LatencyMon to analyze DPC latency and identify problematic drivers or processes.
• Check for real-time performance issues (e.g., processes spiking CPU usage) in Task Manager or Resource Monitor.

---

3. Performance Under Heavy Load

• Thermal Throttling:
  • Your system’s performance might degrade if the CPU, GPU, or even the VRM (voltage regulator modules) on the motherboard are overheating.
• Memory or Storage Bottlenecks:
  • A demanding MSFS session with high-resolution scenery or add-ons might cause stuttering if paging occurs (despite 64GB RAM) or if your SSDs are under heavy I/O load.

Steps to Check:

• Monitor temperatures (CPU, GPU, VRM) using tools like HWInfo or MSI Afterburner.
• Check for pagefile usage and SSD health. A misconfigured or overburdened SSD could introduce delays.

---

4. Other Potential Causes

• Audio Drivers and Realtek Issues:
  • You mentioned uninstalling Realtek drivers, which were automatically reinstalled. Consider trying a generic Windows audio driver instead.
• Background Processes:
  • Recent updates may have enabled or changed background services (e.g., indexing or telemetry) that interfere with MSFS.

Steps to Try:

• Test MSFS in a clean boot environment (disabling all non-essential startup processes).
• Disable any unnecessary services introduced by the update.

---

5. How to Rule Out PSU Issues

• Stress Testing:
  • Use a tool like OCCT to perform a combined CPU/GPU stress test while monitoring system stability.
• Test with a Different PSU:
  • If you have access to another high-quality PSU, swap it in to rule out power delivery issues.

---

Summary:

Your symptoms seem to point more toward driver-level issues, latency, or OS changes rather than a failing PSU. However, the PSU shouldn’t be dismissed entirely, as runtime instability can sometimes relate to power delivery under heavy load.

If you can run LatencyMon, monitor voltage rails, and perform a stress test, it will help narrow the possibilities further. Let me know how you'd like to proceed!

4o

 

You said:

Thanks. I did use LatencyMon and it did not indicate issues that would create audio cutouts. What value might indicate too high DPC latency?

ChatGPT said:

ChatGPT

In LatencyMon, DPC latency issues typically become problematic for real-time audio when the highest reported values exceed 1000 microseconds (µs) consistently. Here's a breakdown of what to look for:

Thresholds for DPC Latency:

• Below 500 µs:
  • Ideal for real-time audio/video applications. This indicates a healthy system.
• 500–1000 µs:
  • Acceptable but could occasionally cause minor audio glitches in very demanding applications.
• Above 1000 µs:
  • Likely to result in audio dropouts, glitches, or video stutters. This suggests a driver, hardware, or OS-level issue.

Key LatencyMon Metrics:

1. Highest Execution Time:
   • Look at the highest DPC execution time for individual drivers in the "Drivers" tab.
   • Drivers exceeding 1ms (1000 µs) should be investigated. Common culprits are network (e.g., Wi-Fi), GPU, or audio drivers.
2. Kernel Timer Latency:
   • Pay attention to the reported kernel latency, which shows the system's overall interrupt handling efficiency.
3. Processes Tab:
   • Ensure no single process is consuming an unusually high percentage of CPU or interfering with normal scheduling.

---

What to Do if Latency Seems Fine:

• If LatencyMon shows no problematic drivers or high DPC latency, the issue may lie elsewhere, such as:
  • Hardware conflicts (e.g., PSU or RAM stability under load).
  • Application-specific bugs (e.g., MSFS or its audio engine).
  • Windows 11 24H2 introducing subtle compatibility issues.

If you'd like, I can guide you further on isolating specific drivers or alternative tests!

4o

 

You said:

Yep mine were very low, 100-300 us

ChatGPT said:

ChatGPT

With 100–300 µs DPC latency, your system is well within the optimal range for real-time audio and video processing. This strongly suggests that DPC latency is not the cause of the audio and video cutouts you're experiencing in MSFS 2020.

Given These Results, What's Next?

1. Focus on MSFS 2020 and Related Drivers:

   • Audio Drivers:
     • Revisit your audio driver. If you’re still using a Realtek driver, try switching to the Microsoft High Definition Audio Driver:
       1. Open Device Manager.
       2. Expand Sound, video, and game controllers.
       3. Right-click your audio device > Update Driver > Browse my computer for drivers > Let me pick from a list.
       4. Select High Definition Audio Device (Microsoft).
   • GPU Driver:
     • Reinstall the Nvidia driver cleanly:
       1. Use DDU (Display Driver Uninstaller) in Safe Mode to remove the existing driver.
       2. Install the latest driver from Nvidia's website.

2. Investigate Pagefile Behavior:

   • MSFS 2020 is memory-intensive, even with 64GB of RAM. If your pagefile is too small or fragmented, it can cause streaming or stuttering issues.
   • Recommended Configuration:
     • Set a static pagefile size on an SSD (e.g., 16GB–32GB).
     • Disable the pagefile on slower drives.
   • Steps:
     1. Right-click This PC > Properties > Advanced system settings > Advanced Tab > Performance Settings.
     2. Go to the Advanced tab > Virtual Memory.
     3. Select your fastest SSD, choose Custom size, and enter the same value (e.g., 16384 MB for both Initial and Maximum size).

3. Test PSU Stability Under Load:

   • Since you've already checked DPC latency and drivers, PSU underperformance under load could still explain the progressive audio/video issues during demanding flights.
   • Tools:
     • Run OCCT with combined CPU/GPU stress testing to simulate heavy load.
     • Monitor voltage stability (+12V, +5V, +3.3V) in real time using HWInfo or your motherboard’s monitoring utility.

4. MSFS-Specific Workarounds:

   • Reduce resource intensity during the cutout-prone stages:
     • Lower terrain LOD (Level of Detail) and texture resolution.
     • Test with fewer add-ons or live services (e.g., real-time traffic/weather).

---

If Problems Persist:

The combination of Windows 11 24H2 and your hardware might introduce a rare compatibility quirk. If the issue coincided with the update, consider rolling back to a prior version (if possible) as a last resort to rule this out.

Let me know how it goes or if you’d like help with specific steps!

4o

 

You said:

Thanks again. OK well OCCT combined extreme testing passed. Voltage stability I didn't see any kind of fluctuation indicator but all voltages except +3.3V were slightly above nominal. 3.3 was 3.26V or something like that. I did install the Windows HD Audio Driver and also discovered I had a smallish fixed page file of 4096MB, but with 64GB of RAM figured it wouldn't be an issue. Plus HWInfo said page file in use was 0.

ChatGPT said:

ChatGPT

Memory updated

Your test results and adjustments give us some good insights:

Key Takeaways:

1. OCCT Testing and Voltages:

   • Passing the OCCT combined stress test with stable voltages indicates your PSU is likely not the culprit. The slight undervoltage on +3.3V (3.26V) is well within ATX specifications, which allow for ±5% variation (3.135V–3.465V).

2. Windows HD Audio Driver:

   • Switching to this driver should eliminate any Realtek-specific issues. Monitor for any change in the audio and video cutouts.

3. Pagefile:

   • The small pagefile (4GB) might not have been an issue given your 64GB of RAM, but MSFS 2020 and other apps may still use the pagefile regardless of physical memory. Setting a larger pagefile (e.g., 16–32GB) could improve stability during resource-heavy scenarios.

   Suggested Test:

   • Increase the pagefile size to 16GB–32GB (static size) on your fastest SSD and recheck during a demanding MSFS session.

---

What to Monitor Now:

• Post-Audio Driver Switch: Check if audio cutouts persist, especially during demanding MSFS flights.
• Performance After Pagefile Change: Look for improvements in both audio and video stability during intensive scenarios.

If the issue continues, we may need to explore MSFS-specific logs or experiment with reducing in-game resource demands. Let me know what happens!

4o

 

You said:

Thank you kindly once again. I'm in the middle of one test flight and so far so good, but this happens in the first half of flights often.

ChatGPT said:

ChatGPT

Memory updated

You're very welcome! The second half of a flight, especially as MSFS processes more scenery and simulates in detail, is when resource demands can spike. Keep monitoring, and feel free to share results or observations after your test flight. If the issue recurs, we’ll continue narrowing it down. Best of luck with this flight!

4o

 

You said:

Once again, pre 24H2 never had audio or video cutouts and experienced superb performance. I'll keep my fingers crossed!

ChatGPT said:

ChatGPT

Memory updated

That’s a crucial detail—pre-24H2 performance being flawless strongly points to the update introducing a variable we haven’t yet uncovered. Fingers crossed indeed! Please keep me posted on the outcome, and I’ll be here to help analyze or tweak further if needed. Safe virtual skies! ✈️

4o

 

You said:

Well, no luck so far once again. Corsair describes how to test their PSU and it goes like this: jump pins 15 and 16 on the main motherboard connector, then turn on the PSU. If the fan spins, it's okay. Yes, the fan spun, but only for a few seconds then stopped spinning so not sure how to interpret that.

ChatGPT said:

ChatGPT

Memory updated

The behavior you observed during the PSU test—where the fan starts spinning but stops shortly after—is actually common in modern PSUs. Many PSUs have a zero RPM mode, meaning the fan only spins when the PSU detects a significant load or temperature threshold. This is part of their efficiency and noise-reduction design.

How to Interpret the Test:

1. Fan Spinning Initially:
   • This confirms that the PSU powers on and is capable of supplying basic voltage to the motherboard (at least in idle conditions).
2. Fan Stopping:
   • Likely normal if there’s no load. The PSU assumes no significant power is being drawn, so it stops the fan.

---

To Be Sure:

If you suspect the PSU could still be causing issues under load, you might consider these steps:

• Load Test with a PSU Tester:
  • Use a dedicated PSU tester that applies a simulated load and measures all voltage rails for stability.
• Swap with a Known Good PSU:
  • If possible, test with another PSU of equal or higher wattage to see if the problem persists.
• Monitor Voltage Behavior Under Load:
  • Run a stress test like OCCT Power Test while using HWInfo or similar tools to ensure stable voltages.

Let me know if you'd like help with any of these steps!

 

😊
Yes, I must say that I use it a lot too.
For all sorts of subjects.
Every time, I secretly hope that he won't know the answer, just to challenge him.
And every time, I'm amazed.
I have 45 years of experience in aviation, this word not allowed manages to correct me on hyper detailed technical questions, including on system models or parameters which are no longer used at all.
It's great and scary at the same time.
In a few years, millions of people will be unemployed because they are less qualified than these machines.
Including lawyers, consulting engineers, journalists, historians, etc... and there will only be ONE truth, that of AI.
Well, for now, we don't really care, it's very useful especially when we don't have much time or desire for long and messy research.😜

That said, you made me really laugh with your chatGPU , which shows how obsessed we are with our PCs!
🍻

55 minutes ago, David Roch said:

That said, you made me really laugh with your chatGPU , which shows how obsessed we are with our PCs!
🍻

Shoot!!!!!!!!!  I'm screwed unless I, what, install Windows 10?  I do have a clone done a few months ago I may have to resort to that which was done in 23H2.  But then I have to permanently prevent 24H2 from going in again.  🙁