Weird monitor bugs people sent me in the last 5 years (2022) | Hacker News Reader

Weird monitor bugs people sent me in the last 5 years (2022)

notes.alinpanaitiu.com

187 pointsfanf282 comments

When the monitors connected have the same serial number the OS could just use the connection side to distinguish the monitors, i.e. where is the monitor plugged in. I gets bad when that is not possible, for example when using a bluetooth headset, where suddenly you hear someone elses conversation because you have the same MAC adress. This requires a workaround on the users end: manually change the MAC adress.

This is for a Mac app, I would imagine the most regular occurrence of this are two screens plugged into a MacBook. I know I have observed this in the past.

Since these will all be Thunderbolt/USB connections, even HDMI/DP would be a dongle, the port numbers would be enumerated on wake, and therefore shuffled.

> when using a bluetooth headset, where suddenly you hear someone elses conversation because you have the same MAC adress.

That's not something I would have expected! I thought that both sides had to handshake. Just having a MAC address conflict wouldn't mean that the headset listens, does it? That's mad!

Even when enumerating on wake there should be a way to know the actual physical port the device is plugged into. This must be very easy. If not, it is a design mistake that even a child wouldn't make.

Something like the systemd's predictable network interface names.

I completely agree that there should be a way to determine which device is which, and I know my Linux device is fairly reliable at that, even with dongles.

Apple decided that the correct way for them was to use the display unique identifier, which unfortunately for budget displays is not unique.

That is the fundamental problem here.

You are now arguing from the position of someone who is trying to make the system usable under more scenarios, while Apple doesn't care if your 3rd party equipment doesn't work. The child here is the monitor manufacturer, and you should have bought good equipment in the first place.

It's like complaining about their half arsed HiDpi scaling options. If you had bought a monitor that had the correct pixel count for the size of the screen then it wouldn't be an issue in the first place.

A fun one I once had.

I had a monitor early on in the MST days, a Dell UP2414Q. It had a fun firmware bug, where any time it changed resolution while using MST, there was a chance the monitor would crash or do one of several very strange failure modes.

In particular, it would sometimes do what I would describe as "crashing" (go black and require physical power cycling to come back), or become confused on what I assume was the state of one of the two "streams" in MST - that is, the left or right "half" of the display would be incorrectly displayed, either by being entirely black, or as though it thought that "half" should be scaled to the entire width of the display, but only for that half.

e.g. if we think of the left side as 0-50 and the right as 51-100, sometimes it would only display 0-50's content in 0-50, sometimes it would show 0-50 and the content that should have been 76-100 scaled to the width of 51-100 (e.g. it scaled the right "half" as though it were intended to be the full span of the screen, and then only displayed the portion that would have been visible).

My best guess about how that could happen was that the decoding for the two sections was handled independently and there were synchronization bugs.

Apparently there's an A01 rev of the monitor that does not suffer from these, but Dell says the FW isn't upgradable, so I ended up trading it to someone else (with full disclosure of the reason I was getting rid of it, ofc).

Discussed at the time:

Weird monitor bugs people sent me in the last 5 years - https://news.ycombinator.com/item?id=32629669 - Aug 2022 (227 comments)

Somehow I skipped it but I'm glad it made HN again, as I think it has an important clue I'm facing for at least couple of years with this Apple Thunderbolt Display - I'm using it with a HP Elitedesk 800 G4 with a Titan Ridge controller and a PCI thunderbolt card. Works great, except (1) I cannot suspend the computer and making it to turn off the screen, else it won't turn on again; and (2) and the most irritating one, at random times at boot, the screen will refuse to turn on. I've done all sorts of weird tricks to make it turn on but everything involved restarting the computer and never thought about unplugging/plugging the monitor again.

Thank you so much for the heads up and to the person who posted this again.

> Dual monitors swapped positions

It’s can happen even when both monitors are not the same and it only ever happens in macOS. Never a problem in Windows nor Linux.

Thankfully Apple seem to have fixed the issue in a relatively recent macOS update but it was super annoying for years.

Yeah I get the feeling this article should be titled 'wierd MacOS bugs'

This could very much also happen on windows. Tho rarely and not in the same way exactly. Its mainly the monitor IDs that could get swapped on XP, 7 and 8.x.

I was supporting an application that needed the ID to be able to show Ads on specific monitors. I don't have data for 10 and 11.

>controlling every aspect of your monitor using DDC and other obscure protocols

When clumsy or obscure protocols are involved, I estimate that the hardware and software have not been fully engineered to actually work properly to begin with.

Always a tough and uncertain workaround.

On PC's the default "Standard" W10 & W11 display driver doesn't seem to work with dual monitors very well if at all. Whichever monitor does display, may turn out not to be the "primary" monitor if the OEM display driver is installed. Once you get your multi-monitor layout finalized and saved, it usually doesn't change unexpectedly though, and you should be able to turn on or off monitors in any combination after that.

I think with XP, W7, etc there were differences in multi-monitor graphics adapters and their drivers over a period of time where sometimes the default primary connector would be VGA, other times HDMI. For this reason when multibooting different Windows versions, I would connect a monitor to each of the display jacks at the back of the PC so I made sure to see something on one of the screens and not just a blank.

A lot of times if you boot a PC with only one monitor plugged in, and it's not the recognized primary connector at the time, it boots just fine but you see nothing. And if you plug in the secondary monitor you see nothing there either. Until you reboot. Seems like unless it's proven to work otherwise, you need all monitors plugged in and turned on before you boot. If not the resulting performance or lack thereof can be disappointing.

Different Linux live distros are fairly consistent with each other and usually work for multi-monitors by default without connecting to the internet or installing any display drivers. If you do a lot of configuration to get your desired monitor layout the way you and your mouse like it, and it fails to persist for some reason, that would be about equivalent to booting a live OS which is by nature handling hardware in a default way as newly-discovered devices every time you boot. Kind of like the Mac that wakes up unlike the way it went to sleep.

Like the article hints at, sometimes it's best to carefully wire everything so it's closer to what you want by default, before you start doing any tweaking.

For me nothing is off the table. First of all I like to have a PC for every monitor, while still being able to use dual (or more) monitors with a single PC whenever.

In a multi-desk lineup with 5 monitors from left to right, every monitor will have an average of 2 or more cables or adapting cables connected in the back, which lead up to where the PC's are.

First of all you have to really be able to get more out of 5 than you were 2 or 3, you may know who you are.

That can entail a bit of manual plugging and unplugging of display cables, as well as OS & software settings for different workflow configurations, and the idea is to try and accomplish all that using the connectors at the back of the PC's. And minimizing the need to change cords behind the monitors if that is the least bit more inconvenient. But doing it anyway whenever needed.

This can get kind of tedious, but eventually you end up with a fairly optimized procedure for extracting maximum utility from minimal rewiring. And since it started out a bit better than having fewer monitors to begin with, that's when it feels really better. This is the point it may be possible to add some hardware switching for your display wiring (although a complicating factor) which could replace occasional manual plugging with manual button-pushing after that.

> This could very much also happen on windows.

Concur.

It also happens on my Raspberry Pi 5 with 2 screens connected over microHDMI. They are dissimilar screens, too: one is an HDMI 27", one is an old HDMI 24" via a converter cable, in portrait orientation. So, both fully digital, and both should be easily distinguished by the OS.

Happens in both RasPi OS and MX Linux.

I have had this problem several times on Linux, with different monitors, so it seems that it can actually happen on every OS.

Needs more naming and shaming. If the product is broken, say so. If multiple instances of the same monitor have the same serial number, that's a product defect. If the monitor's firmware can crash so hard that it stops responding to anything, that's a product defect.

Author here, all monitor vendors are guilty of the same serial problem. It’s not present on all models, and most have transitioned to writing the Alphanumeric Serial Number in the last 2 years, which is thankfully unique.

The firmware crashing, I could have created a list of those monitors, but people don’t always send me the exact monitor model so I might not have that info. But it’s still not a single vendor having this problem.

Weird that the physical port number is not part of that unique id used to “hash” the monitor. It is like me not being able to figure out how to leave my kitchen for the lounge because the carpenter installed the same make, model and serial number door on all the frames.

Do you mean the port number on your PC?

To me, it is a feature, and not a lazy implementation. Now I don't have to keep track of which HDMI port I plugged which monitor into.

Yes, the port number should only be used for disambiguation if both monitors have the same ID.

That would seem best since the port numbers only change when you move. If you move your laptop a lot a docking station might help ease the pain so you don’t need to figure out what way to reconnect. If you move your entire setup a lot the inconvenience of switching monitors is probably a rounding error!

It's not that you can't go to the lounge, you just have a 50/50 chance when you pass through a doorway that it will be the lounge :P

Inscribing identical numbers on a pair of doors will, obviously, not have that effect.

Why would we care if monitors have identical serial numbers?

As the article says (“After standby or restart, the monitors get swapped by the OS, so moving the cursor to the left monitor will actually appear on the right monitor, and vice-versa.”) you might have told your computer that one of them shows the left half of your desktop and the other the tight half.

Now, when your system wakes up from sleep, how does it know which one is which? The workaround is to use the port they’re attached to, but that isn’t always might even both be attached to the same port.

It does if you give directions to the lounge by saying go through door marked XYZ.

Because unlike rooms and doors, you can switch what port a monitor is connected to.

It might from your perspective. If you work in an office where everything looks the same except for the numbers above the doors that would be the situation.

I'm not being facetious, that's the monitor issue precisely (while also being completely not).

Would this software actually have access to that information? I can imagine it would be protected by the OS because on paper, it shouldn't matter what port a device is connected to.

Second, as a user you shouldn't need to make the distinction; yeet a screen into any port and I'd expect it to know what screen it is and what settings to apply to it. There's also the case of using a multiport adapter, or a chain of them if you're sadistic.

The operating system should be the one using that information in the first place—if you see two identical monitors connected, shouldn't you try and identify them based on the only discriminating factor you have?

Make a torture tool that triggers all these breakages, and send it to people who review monitors.

This reminded me of a problem I had with my monitor last year - an actual bug had crawled inside the screen and died. I assumed it was a cluster of dead pixels and was planning to replace it. Eventually I realized what it was and learned you can use an electric toothbrush to migrate the dead bug to the edge. No luck in my case despite my best efforts. But later when I returned from a month-long trip I noticed the bug was gone and my monitor was once again clear.

Mine is still there. He decided to actually die inside my monitor

https://manuelmoreale.com/a-moment-with-a-hardware-bug

I have two identical monitors (Lenovo ThinkVision P27h-20) that I can reliably reproduce the random position switching under macOS when waking from sleep

The solution I've come to is to use USB-C with one, and HDMI with the other. This way they both have a unique footprint and macOS keeps them in the correct position. The main downside, I lose use of the USB ports on the one monitor

On dumb monitors, the EDID information is on a tiny flash chip. My previous employer accidentally shipped a monitor where we (actually our CM, which we failed to catch) failed to disable writes, so the computer could actually overwrite it. Fortunately, this rarely happens, but we did get some reports of it happening - possibly due to a dodgy connection (you shouldn't treat data flowing over any unpluggable cable as 100% reliable, including the 'this is a read' bit ).

Let me also share my weird bug. Back then, I was connecting my CRT TV via an S-Video cable to my PC's video card, as a secondary monitor. I noticed one day that the GPU fan is spinning slowly, even when the computer is turned off, or even weirder, when unplugged! So, the investigation began, and I concluded that when I connect the TV, the fan gets spinning.

No issues while working whatsoever, just this little quirk.

Two weird ones I see here in the office.

I have an old HP 1440p monitor that will retain the screen image when I unplug the HDMI cable from my laptop. (My solution is to plug it into a power strip, and I just shut it off when I leave.)

There's a new Dell monitor that somehow produces a sort of halo of short vertical lines a few to several pixels long around text.

That can be fixed with a diode

My Dell monitor has a bug where it can hold the last frame or elements of it indefinitely, even if you turn it off, switch inputs, or take other measures. Which implies that somewhere inside is a framebuffer where it stored what was about to be displayed, which has some edge-case exotic privacy implications.

Lunar is so good - rock solid. Along with BetterDisplay, these have been a comprehensive solution to my various MacOS external monitor woes over the years.

Lunar dev here, thank you so much for this! As the dev, I see the app as very fragile because of the dozens of emails per week I get, each describing a different very specific issue.

It’s good to hear the app is actually stable and useful for most people.

Can I ask you a question about the lunar app and DDC?

I have an old monitor with (what appear to be) broken buttons, meaning I can't control the brightness/contrast etc via the OSD. Would lunar be able properly control such a monitor over a VGA or HDMI cable? Would this control also be possible if using a hackintosh? Thanks a lot!

In theory, yes Lunar should be able to do that, both on Mac and Hackintosh. But it depends on the hardware if it allows it, some hubs/GPUs/monitors might block the DDC command.

Anyway Lunar has a free offering, you can try it easily to see if it works.

Thank you!

ClickMonitorDDC should be able to control such a monitor without needing to install a hackintosh setup

https://www.majorgeeks.com/files/details/clickmonitorddc.htm...

Thank you, that worked! The app is a bit fiddly but it works :)

Unrelated to Lunar but here is my fun monitor bug: Sometimes when I connect my Intel macbook to my external ASUS ROG monitor via display port, a strip of pixels about 1cm wide are "moved" from the far right of my external monitor to the far left. My monitor is on the left and my laptop is on the right, so if I move my mouse from left to right, my cursor starts on the left (just right of the error strip), moves off the right edge of the monitor and then back onto the strip at the left edge of the monitor, and then further right onto my macbook display.

This has happened 5 times in as many months, and is fixed by unplugging and replugging the usb-c end of the display port cable a few times.

Could you have accidentally set a screen arrangement in your preferences -- you do not mention what OS -- that overlapped the two screens?

sounds similar to a bug in my predator x27 monitor where the middle strip of pixels moves to the far right side until I power cycle it.

The good news is it's just a firmware issue and they've already fixed it.. the bad news is to get the update they want me to ship a heavy ultra wide monitor to a service center.

so id check if your monitor has firmware updates and more importantly if you can apply them yourself.. and ensure any future monitor you buy actually allows users to apply updates to the firmware.

My weird monitor issue (since we’re talking about it, not that anyone asked) is sometimes when I wake my M1 MacBook Pro, my monitor (AW3821DW) has a cyan overlay across the whole screen.

I unplug/plugin the usb-c cable and it fixes it.

I have a usb-c to DP dongle, then connect via DP to the monitor.

I find apple's treatment of monitors kind of weird.

PC graphics cards just have a plentiful supply of video outputs, say a mix of HDMI and DP.

But apple has always used dongles, many of which are basically small computers that can basically convert one video signal to another on-the-fly.

(I don't know if USB-C to DP is an electrical conversion or has logic)

But the complexity involved is probably a steaming heap of corner cases for the driver. And apple has probably tested and prioritized bugs for their hardware, but 3rd party monitors are probably hit or miss.

USB-C to DP is an electrical conversion. The monitor has to negotiate an alternate mode with the USB controller (then the high-speed data pairs in the USB-C cable are used for Displayport signalling and not USB).

(There are such things as USB monitors which actually take pixel data over the USB protocol, but I don't think they ever caught on.)

> There are such things as USB monitors which actually take pixel data over the USB protocol

DisplayLink. It caught on for a while at least, especially in the case of some dock products that would output to a more conventional display interface but were USB DisplayLink adapters themselves. You’ll see tons of forum and Reddit posts about suboptimal CPU load and performance related to it.

https://en.wikipedia.org/wiki/DisplayLink

It more recently was resurrected to overcome M1 limitations.

https://danielcompton.net/apple-m1-displaylink-multiple-disp...

> But apple has always used dongles

Most the Mac laptops I’ve owned have had display port or hdmi, the dongle lifestyle is a new and unpleasant one.

The laptops with dedicated ports might have been the last common-sense laptops they made.

I laughed (and grumbled) at the article "Benjamin Button Reviews The New MacBook Pro":

https://blog.pinboard.in/2016/10/benjamin_button_reviews_the...

(The Benjamin Button movie was about a man who lived life in reverse, starting as old and wrinkled and gradually getting younger)

The current laptops have a HDMI port. DP over USB-C doesn't require active conversion. I think you're slightly jumping at shadows here.

(Apple's _really_ problematic video dongle was the thankfully largely dead lightning to HDMI one, for phones; this was essentially an external USB video card.)

Active or not, it’s still a dongle (but a dongle or a cable for an external monitor is less of a rage concern and they have added back HDMI as mentioned). However, seriously give me just one fucking USB-A port on a “pro” device. There’s enough people in that segment that use flash sticks and need to interface to all kinds of various hardware (MIDI controllers, pro audio equipment (not every use is a rack)), all kinds of random special purpose hardware, etc. It’s a pro device it should cater to a wider range.

wow...

https://en.wikipedia.org/wiki/MacBook_Pro_(Apple_silicon)#/m...

no usb-a ports sucks, but I guess it's a step in the right direction.

sort of amazing how out-of-touch they (still) are with "pro" users. who cares if you need usb-a only once a year.

People might only need the bed in their pickup truck once a year, but they buy pickup trucks, get stuff done and are fiercely loyal.

Thank you for this kind of information. I have been pulling my hair out for years working with usbC hubs and monitor problems. There are bugs all the way up and down the stack: windows, the hub, the monitor itself. It seems there is fairly sparse testing of external monitors with windows for sure.

I have used external monitors for decades with Windows and I have not run into problems. One thing which may help is I always plug the monitors into the video card. I typically use an NVIDIA professional video card. I have also not had problems with laptops plugged into external monitors.

Yeah, it's interesting to see some of these bugs, and the duplicate serials makes a lot of sense.

However my Linux laptop can handle multiple dongle attached screens without getting confused for layouts that my Mac gets confused about. I wonder what it's doing differently?

Regarding the monitors with identical EDIDs. Isn't it possible to assign them at the port level? I get this shouldn't need to happen, but it's certainly still a better option than a script which lets you quickly fix it every time it happens.

MY weird monitor bug: Sometimes a vertical strip of ~3-5mm that ought to be in the middle is actually displayed on the far right.

The solution? Wiggle the Displayport cable.

I've replaced the cable entirely so it's not actually the cable itself.

Interesting, I have the same thing on my rog swift. I fix it by turning it off and on again though.

I have the same bug. I usually power cycle the monitor to get rid of it.

I used to have an Asus monitor at work that would freeze completely if I had a macbook plugged in when it woke from sleep. I had to always remember to first press some buttons to wake it up then plug in my laptop. If I forgot, I would have to unplug it from the wall, unplug my laptop (so it didn't just immediately freeze again), plug it in and let it boot up, then plug in my laptop.

That and a very poorly made Asus laptop left me with the impression that while they might build good motherboards, that quality didn't necessarily extend other products.

I had a very similar issue with a rather expensive LG monitor. The issue was reduced after a firmware update to the monitor, but input auto detection still does not work at all

Yeah, my current LG monitor had some bugs with DP 1.4 that were resolved through a firmware update. It was a replacement for a different LG monitor that falsely advertised 60W USB PD "laptop charging", when it could only actually deliver 15W. (Which is not enough to charge most laptops.)

(I think the current one also has some false advertising, because they claim it supports a 75hz refresh rate, but I've never been able to get more than 60hz out of it. Before the DP 1.4 firmware fix, it was stuck at 30hz sometimes.)

The weirdest monitor bug I ever encountered was a bug, a very small bug/fly, that had crawled between the layers of the LCD panel. It was like a moving dead pixel. Really weird and confusing at first until I understood what it was. I was afraid to touch the panel where it crawled as it might be squashed and stay there forever.

I've seen that fairly often in summers wherever there are thrips around

> Dual monitors swapped positions

God I hate this.

The ridiculous 30 second long seizure that any computer suffers when the set of attached displays changes (before often ending up with them not working properly) is something that the industry should feel deeply ashamed of.

Not a problem on ChromeOS. Honestly the one Linux everyone should try just to get a feeling for what it feels like when Linux works.

Pretty much. As much as I hate proprietary spyware like Android and ChromeOS it's crazy how well they can nail stuff like wayland-like-compositor, VRR, anti-tearing, HDR support, low latency, etc on top of the Linux kernel, many years before the FOSS community for PCs, when there's a massive corporation behind it steering it in one direction instead of the project being a house of cards of 5000 different independent devs each pulling in their own direction.

Like I see how smooth and HDR Android is and has been for many many years while having HW acceleration always working out of the box, and I ask myself why can't it be exactly like that out of the box on Linux on my PC? Why is PC Linux so much more janky than on Android or ChromeOS when it comes to video and graphics?

Yeah, I know, Linux is getting better, Wayland support is getting better, VRR is being merged to Gnome and KDE compositors, but all this was the norm and working seamlessly out of the box on Android for like 10 years now just because Google put it the effort and we're still not there on PC, where Linux still has a fair bit of jank once you get over the honeymoon phase.

There's a big difference between open source projects driven by professional organizations and the rest of them. In the general open source world you could be locked in an argument with a pre-teen who has never had a job and who is for some reason the gatekeeper of a git repo. In the corporate world there's a top-down mandate that says make this laptop boot in one second and you will get paid a bonus. Big difference for outcomes.

>identical EDID

Isn't this an EEPROM that can be written to over the wire?

Yes, but most monitors disable writes to them. Some monitors have service menus that can enable that but it’s increasingly rare to see that.

I thought about simply writing a random serial number to one of the monitors as the solution, sadly it’s not possible: https://notes.alinpanaitiu.com/Decoding-monitor-EDID-on-macO...

Who's going to make the first open source monitor?

Who's going to buy a monitor with outdated technology at 10x the price of competitors? Just look what happened with Openmoko.

Why would it have outdated technology, I'm suggesting a major display company open sources their firmware, saves money on debugging, adding new features, etc

The same people that will make a “dumb” TV that people will actually want to buy because it has table stakes display features: no one.

Never worked out why people who hate Smart TVs don't just not connect them to the LAN

Quite unrelated, but I am also in the processing of writing a Mac app and I was wondering which Macbook do you use to develop Lunar? I am looking to get a refurbished model so not so sure. Thanks!

> That means that on reconnection, the monitors will look identical from the OS side, so the stored settings will have to be assigned randomly for each monitor.

Huh? So macOS stores two different settings for the same ID and then assigns them randomly? That seems very odd.

I'm confused why anyone would want anything to do with MacOS, but to each her own

Reply on news.ycombinator.com