I tried some additional tests today. This time without a USB-C to DisplayPort adapter/cable from the CAC-1336. Instead, I connected the CAC-1336 to a CalDigit Element Hub (Thunderbolt 4 hub) using a Thunderbolt 4 cable. The display is connected to one of the downstream Thunderbolt ports of the Element Hub using a Moshi bi-directional USB-C to DisplayPort cable. The GPU Is Radeon RX 580, macOS Monterey 12.4. The EDID returned by the adapter appears to be correctly taken from the display.
Tested modes:
Only a couple modes were stable:
720×480 60Hz 27MHz (there are some small amount of flickering pixels on the right edge)
3840×2160 30Hz 297MHz.
Actually, these stable modes can sometimes be unstable like the following modes:
Usually, the signal appears for only a couple seconds. Then goes black:
640×480 60Hz 25MHz.
720×576 50Hz 27MHz.
1280×720 60Hz 74.25MHz.
1280×720 50Hz 74.25MHz.
1920×1080 50Hz 148.5MHz.
1920×1080 60Hz 148.5MHz.
3840×2160 30Hz 263MHz.
3840×2160 24Hz 297MHz.
3840×2160 25Hz 297MHz.
1920×1080 120Hz 297MHz.
1920×1080 144Hz 333MHz.
2560×1440 120Hz 498MHz.
3840×2160 60Hz 533MHz.
2560×1440 144Hz 592MHz.
Sometimes the signal is distorted when it appears, then goes black:
3840×2160 50Hz 594MHz.
3840×2160 60Hz 594MHz. Sometimes reappears distorted.
Sometimes the signal is always black:
1920×1080 30Hz 74.25MHz.
1920×1080 25Hz 74.25MHz.
The modes support HDR.
I don’t think a Thunderbolt 4 hub is a good solution since it’s probably limited to only 2 lanes of HBR3. Actually, if that were true, then the max pixel clock would be 540MHz for 8bpc RGB or 4:4:4. So I checked the modes that macOS accepts using my AllRez app and all the 594MHz modes are 10bpc 4:2:0 (max pixel clock for HBR3 x2 for 10bpc 4:2:0 is 864MHz). The 533MHz modes are 8bpc 4:4:4.
But the 2560×1440 592MHz mode is 8bpc 4:4:4 so maybe the CalDigit really is getting 4 lanes of HBR3? I didn’t think USB-C hubs supported 4 lanes of DisplayPort when they support USB 3.x or later, but maybe the Thunderbolt 4 hub is smart enough to see that the CAC-1336 doesn’t have a USB 3.x mode.
Question: is the CAC-1336 able to negotiate a two lane DisplayPort output mode? Or is it always 4 lanes?
One good thing about the test is that the CalDigit Element Hub is appearing as a USB 2.0 hub which means the CAC-1336 can pass a USB 2.0 signal like the CAC-1332 can. I don’t know why this feature isn’t mentioned in the specs. It’s important for USB-C displays with USB features such as audio, camera, USB ports, settings. One strange thing is that I can connect USB devices to the Thunderbolt ports of the Element Hub but not the USB type A ports of the Element Hub. The type A ports are connected to a secondary hub inside the Element Hub. I suppose that’s just an issue with the Element Hub’s USB-C (non-Thunderbolt) compatibility mode.
I replaced the Moshi cable with a CAC-2068+CAC-1567. The 3840×2160 60Hz 533MHz is stable but all the other modes go to black after a couple seconds. Actually 3840×2160 60Hz 533MHz is not stable either. It’s strange. I don’t know why a mode is sometimes stable. There’s not any errors or flickering that you might see with a bad connection. It’s kind of unusable like this.
I tried 5120×2880 39Hz 594MHz so it seems the CAC-1336 can do 5K but I don’t have a HDMI 2.1 port or a 5K display such as the Apple Studio Display to test 5K60. This mode can be stable for a while, go black, but then reappear again.
I tried 6016×3384 28Hz 585MHz but the image was distorted. Either the RX 580 can’t output 6K to its HDMI 2.0 port, or the CAC-1336 doesn’t like it. I should try other HDMI 2.0 ports that I have (Nvidia Maxwell and Pascal and Kepler)
I noticed the CAC-1336 continues to work even if the USB power is disconnected. If it requires USB power to function, then shouldn’t it turn itself off? Or maybe USB power is required for higher bandwidth conversion? It appears to use 5V 0.43A for the 5120×2880 39Hz 594MHz 8bpc 4:4:4 mode.
I got confused. You said CAC-1576 which is a USB-C to USB-C cable. I want to use a CAC-1567 which is a USB-C to DisplayPort Adapter. Or I want to use a USB-C to DisplayPort cable. Club-3d tech support says the USB-C port of the CAC-1336 doesn’t work with the CAC-1567 or any active adapter which probably includes USB-C to DisplayPort cables even though the cable or adapter just needs to pass the DisplayPort signal unchanged.
Click the Export button and post the result. A copy of the hex bytes as text is sufficient. The info for the display modes is probably hidden in the CTA-861 or DisplayID extension blocks.
Is there only one HPN36C1 in the menu? That may indicate that it’s only one display device (SST instead of MST)
Did you install the billboard driver? (if it was included?)
Open a Windows cmd window.
Change the directory using the cd command to the directory containing the .bat files.
Then try 01_ReadVersion.bat command to see if it will show the current firmware version of the adapter. It replaces the Version.txt file in the IspTool directory with a new Version.txt file that contains the current version. Save a copy of this file somewhere else or rename it to “Version_original.txt” so you have the before info.
Then try the 00_Update_FW.bat command to do the update. It updates the Version.txt file with the new version info that was updated to the adapter. Compare the version numbers.
Does the headset appear as a display when used with the CAC-1335?
When connected to DisplayPort without the CAC-1335, does the headset present itself as a 4320 x 2160 display or as two 2160 x 2160 displays? Or the operating system might not show it as a display. I’ve never tried VR before so I’m not sure.
Can you extract the EDID(s) from the headset? Using AMD or Nvidia software? Or using CRU?
If it’s two 2160 x 2160 displays then it won’t work since that would require DisplayPort MST which cannot originate from HDMI.
I’m not sure the adapter was tested with any resolutions other than 3840×2160 or 7680×4320, so you should do your own tests if possible.
Do you have a normal display that you can test the CAC-1335 with? Can you create a 4320×2160 custom resolution for that display using AMD or Nvidia or CRU and test it without the adapter? I have a 4K display that can accept resolutions up to 4K120 or 5K60 or 6K50 or 8K30 as long as the pixel clock does not exceed ≈1080 MHz. Once you know you have a display that can support 4320×2160, you can test the adapter with that mode. You may need to recreate the custom resolution when you switch the connection from DisplayPort to HDMI port.
You said “The same cable will connect to an older laptop with a mini-displayport connection to the discrete graphic card, so the cable it good.” The CAC-1335 has DisplayPort, not Mini DisplayPort. Please explain exactly what is connected to the laptop (adapters, cables, display).
I would try this connection order: USB power, DisplayPort connection to display, HDMI connection to computer.
They are sold on Amazon. I received a CAC-1336. It didn’t work. Baak says he successfully tested it in some situations but not all the situations I would test. https://insights.club-3d.com/thread/cac-1336-operation/
I understand you want to use HDMI 2.1 to get 4K120 4:4:4 HDR. If you do get 4K120 you have to know a method to determine if the TV is receiving the expected signal, and not 4:2:0 8bpc as can happen on some computers, such as those that have GPUs that support only HDMI 2.0 or those that have crappy drivers such as macOS that don’t support HDMI 2.1 dongles properly.
I believe the LG has an onscreen menu that shows the type of signal its receiving. There’s probably YouTube videos that show how to do that.
I said the Vizio EDID looks normal. It is not missing 4K60 modes. They are listed in the CTA-861 Extension Block in both the Video Data Block and the YCbCr 4:2:0 Capability Map Data Block which means these modes support both RGB/4:4:4 8bpc and 4:2:0 (8, 10, 12bpc) like many HDMI 2.0 TVs.
[code]
VIC 97: 3840x2160 60.000000 Hz 16:9 135.000 kHz 594.000000 MHz
VIC 102: 4096x2160 60.000000 Hz 256:135 135.000 kHz 594.000000 MHz
VIC 96: 3840x2160 50.000000 Hz 16:9 112.500 kHz 594.000000 MHz
VIC 101: 4096x2160 50.000000 Hz 256:135 112.500 kHz 594.000000 MHz
[/code]
I think the main problem is with the adapter. My suggestion was to connect the adapter in a certain order:
1) turn on TV and computer
2) connect HDMI cable from adapter to TV.
3) connect adapter USB-C to computer.
Regarding 4K120Hz displays, it is possible for HDMI 2.0 to output 4K120 using 4:2:0 8bpc. Don’t bother with HDMI 2.1 cable in that case since the bandwidth of HDMI 2.0 is limited to 18 Gbps anyway. But HDMI 2.1 cable should be compatible with HDMI 2.0 signals (6 Gbps per line) and are higher quality to support HDMI 2.1 signalling rates (6,8,10 Gbps per line)
Upscaling is invisible to adapters and the host computer. The computer outputs a signal with a certain amount of pixels (720×480, 1280×720, 1920×1080, 3840×2160, 4096,2160, etc.), and the display puts the pixels on the screen however it likes.
The SYN3000 EDID is probably some kind of default EDID from the adapter when it can’t read the EDID of the display properly?
SYN = Synaptics Inc = the manufacturer in the EDID
3000 = the product number in the EDID.
The manufacture date in the EDID is 2013.
The EDID only goes up to 1920×1080 60Hz.
Why didn’t they bother to make an EDID that matches the description and capabilities of the Adapter?
The EDID of the Vizio looks normal. It has a couple minor issues according to edid-decode but I don’t think that should cause these problems.
When the screen goes black, you have another display connected so you can continue testing different settings? Or use a screen sharing app so you can control the computer from another?
For 4K60, try 8bpc RGB or 4:4:4. Also try 4:2:0 in case the connection is being limited to HDMI 1.4 speed. Check the Vizio manual for settings that might affect bandwidth (color depth, refresh rate, etc.).
When connecting the adapter, with the TV on, try connecting the HDMI first, then the USB-C second.
The screenshots are missing some details. Can you export the EDIDs and post a link to them? Or paste the hex bytes into a post.
What list of modes does the Nvidia control panel show? Does it show chroma subsampling mode 4:2:0? It will probably be necessary for 4K60.
Did you try exporting the EDID from the Vizio, then import it into the Syn3000 using CRU?
Can you show the list of modes?
Can you extract the EDID? Maybe with CRU (Custom Resolution Utility)?
I talked to club-3d support Tuesday. They said to use a USB-C to USB-C cable even though I told them I don’t have a USB-C display. I want to use the CAC-1576. Haven’t heard back from them since.
I would try the CAC-1336. It might support 4K. It might even support 5K or 6K or 8K30 depending on the GPU and drivers (maybe from a PC but probably not Xbox). I don’t think 5K and 6K have been tested by anyone yet. None of the earlier HDMI to DisplayPort adapters support width > 4096. The CAC-1336 might allow USB data which would enable the USB functions of the display but I don’t think this has been tested (The CAC-1332 had that function even though the product description doesn’t mention it). The CAC-1336 I received produces no signal – I probably need a replacement – so I was not able to test any resolutions, 4K or otherwise, and I saw no indication of USB functionality (no power from the USB-C port and no USB device connection in the OS).
[quote author="Baak"]I tested this with an RTX3070 + AORUS-FV43U (4K120Hz)and it worked. Using the CAC-1336 + CAC-1576.
And an RTX2080 (HDMI2.0) + MSI optix mag321curv (4K60Hz)and this worked too. [/quote]
Did you try testing some custom timings with your displays? First, using a DisplayPort connection, try to get 5K60 or 6K50 or 8K30. Basically anything under 1080 MHz might be possible (8bpc RGB). Those modes work on my Acer XV273K. You can use Nvidia control panel or CRU to create the custom timings. Then see if those modes work using the CAC-1336. Note that EDID overrides created by CRU are per GPU port and display vendor and product ID, so changing the GPU port will require recreating the EDID override. You can save an EDID override and copy it to the other ports. Give the override a unique display name for each port.
I did more testing.
I tried the CAC-1336 in Windows with a USB-C to DisplayPort cable. The CAC-1336 is connected to the HDMI port of a Radeon RX 580. The display is detected as CHD-LT2610 which means it is overriding the EDID of the display (XV273K) or it’s not reading the EDID from the display. Is it supposed to override the EDID? What chip does the CAC-1336 use? If it’s the LT6711GX then wouldn’t LT6711 be a better name for the EDID? Why is it using an EDID with manufacturer ID 4d79 = SKY = SKYDATA S.P.A., product ID = 0x9001, year 2004? It seems kind of lazy to use a random EDID from a 2004 display in a brand new adapter…
The options for the LT2610 EDID only include up to 1080p.
The display is not detecting a signal. The screen is black and the LED remains yellow instead of turning blue. I disconnected the adapter and reconnected it and it didn’t appear again.
I put the PC to sleep using the Windows 10 start menu, and when I woke it, the CHD-LT2610 EDID from the CAC-1336 is detected (but the display does not detect a signal). I can get the same results using the CAC-1567 instead of the USB-C to DisplayPort cable. I get similar results from a PC that has a GTX 1070 using a Dell P2715Q.
I have all the cable types.
I’m not sure what a passive USB3 cable is. Aren’t they all passive except for optical cables? Can your passive USB3 cable do 10 Gbps of USB? Or is it a 5 Gbps USB cable? HBR2 is 5.4 Gbps per line. It might work with a USB 5 Gbps cable.
Any USB-C cable that can do 10 Gbps USB can do 32.4 Gbps DisplayPort 1.4.
A USB-C cable has 4 SuperSpeed lines.
– 10 Gbps USB uses two of those lines, one for receive and one for transmit.
– 20 Gbps USB uses four lines, two for receive and two for transmit.
– DisplayPort HBR3 (8.1 Gbps per line) can use 1, 2, or 4 SuperSpeed lines to transmit up to 32.4 Gbps.
– USB4 40 Gbps (20 Gbps per line) uses two lines for transmit and two lines for receive). A cable rated at 40 Gbps is required (a USB4 or Thunderbolt 3 or Thunderbolt 4 cable).
DisplayPort uses 8b/10b encoding, so 32.4 Gbps on the wire can transmit 25.92 Gbps of data. DisplayPort HBR2 is 17.28 Gbps of data.
HDMI 2.0 (18 Gbps) is three lines at 6 Gbps each using 8b/10b encoding = 14.4 Gbps of data.
HDMI 2.1 is four lines at 6, 8, 10, or 12 Gbps using 16b/18b encoding (for 21.3, 28.4, 35.6, 42.7 Gbps of data). What output rates does the CAC-1336 support?
The amount of bandwidth required for a display mode depends on the pixel clock (MHz) and the bits per pixel (bpp). Just multiply them together.
4K120 is between 1075MHz (CVT-RB2) and 1188MHz (HDMI).
8K30 is between 1019MHz (CVT-RB2) and 1188MHz (HDMI).
between 8 and 16 bpp: DSC
12bpp: 8bpc 4:2:0
15bpp: 10bpc 4:2:0
16bpp: 8bpc 4:2:2
18bpp: 6bpc RGB/4:4:4 or 12bpc 4:2:0
20bpp: 10bpc 4:2:2
24bpp: 8bpc RGB/4:4:4 or 12bpc 4:2:2 or 16bpc 4:2:0
30bpp: 10bpc RGB/4:4:4
32bpp: 16bpc 4:2:2
36bpp: 12bpc RGB/4:4:4
48bpp: 16bpc RGB/4:4:4
With 12bpp, even HBR2 or HDMI 2.0 can do 4K120 and 8K30. Of course, people with a HDMI 2.1 output would prefer not to use 4:2:0.
What DSC target bpp does the CAC-1336 support? slices per line? pixels per slice? pixel rate per slice? slice height? DSC bpc? I’ve seen one DisplayPort 1.4 MST Hub which wouldn’t do DSC with 10bpc – it was limited to 8bpc which means it couldn’t do HDR.
Can the adapter compress or decompress DSC? A DisplayPort 1.4 MST hub can decompress.
What pixel types from the above list does the CAC-1336 support for HDMI input and DisplayPort output? What pixel type conversions can the adapter perform if any?
Can you connect a mouse to the USB-C end of the CAC-1336 and get power? mouse movement?
USB-C DisplayPort Alt Mode supports two connection methods:
– 2 lanes of DisplayPort + USB 3.x
– 4 lanes of DisplayPort + USB 2.0
The CAC-1336 has a micro USB connection so it can only support the latter (if it supports USB data at all, which the CAC-1332 did).
The AORUS-FV43U has a USB-C input and USB type A ports. Did you test the USB type A ports? You would have to switch the display to prioritize resolution instead of USB data since the CAC-1336 can’t provide USB 3.x data. The switch sets the USB connection to USB 2.0 only to allow all 4 SuperSpeed lines of the USB-C cable to be used for DisplayPort. Actually, I can’t find mention of a switch in the manual. The manual says USB-C can be used for the USB upstream (part of the KVM feature), but it doesn’t say if it’s limited to USB 2.0 or not…
The MSI optix mag321curv doesn’t support USB data on its USB-C connection. Seems neither display is a good test of USB data functionality of CAC-1336.
There are gaming monitors that have the 4 lanes DisplayPort + USB 2.0 connection type option.
Also, the LG UltraFine 4K, LG UltraFine 5K, Apple Studio Display (5K), and Apple Pro Display XDR (6K) support the 4 lanes DisplayPort + USB 2.0 connection type (for USB 3.x, they need a Thunderbolt connection instead of USB-C DisplayPort Alt Mode).