24" Widescreen CRT (FW900) From Ebay arrived,Comments.

Alright cool, i have an old Intel-based Dell with integrated DP out I can try this with. Thanks for your help.
Did you ever get around to trying fwupd?

It's been updated recently to allow working with AMD GPU. Still, sudo fwupdtool get-devices won't be able to see the MST Hub firmware unless the MST Hub is connected to Intel GPU (with or without display), or to the AMD GPU with a display connected to the MST Hub. So in this case it may be better to connect the MST Hub to Intel GPU because you don't need to connect a display to the MST Hub which might cause a hang.

These commands should be enough to install the latest fwupd from source code (using Ubuntu):
Code:
# update stuff
sudo apt update
sudo apt-get install python3 python3-pip python3-setuptools python3-wheel ninja-build
pip3 install meson
pip3 install --upgrade markdown
pip3 install --upgrade meson

# get source code and dependencies
git clone https://github.com/fwupd/fwupd
cd fwupd
./contrib/ci/generate_dependencies.py -o ubuntu | sudo xargs apt install -y
meson build

# build it
ninja -C build

# install it
sudo ninja -C build install
sudo ldconfig

Then in a new terminal window you can check that the version is 1.6.1
Code:
sudo fwupdtool --version
sudo fwupdtool get-devices
 
joevt

Good news! My Sunix isn't dead. FWUPD detects it and reports the last firmware I installed.

But the problem is that last firmware installed seems to be causing fwupd to think it's an incompatible board version, so the original Sunix firmware isn't installing, and is giving me this message:

Installing on VMM2322… [ - ] Decompressing… [ - ]board ID mismatch, got 0x0011, expected 0x0203

I used install-blob and even tried --force and neither would let me proceed.

Maybe there are some other commands I don't know about? All I need to do is get the thing to light up again, and at that point I think the Windows app will talk to it again and I could force install the old firmware.
 
Last edited:
joevt

Good news! My Sunix isn't dead. FWUPD detects it and reports the last firmware I installed.

But the problem is that last firmware installed seems to be causing fwupd to think it's an incompatible board version, so the original Sunix firmware isn't installing, and is giving me this message:

Installing on VMM2322… [ - ] Decompressing… [ - ]board ID mismatch, got 0x0011, expected 0x0203

I used install-blob and even tried --force and neither would let me proceed.

Maybe there are some other commands I don't know about? All I need to do is get the thing to light up again, and at that point I think the Windows app will talk to it again and I could force install the old firmware.
I think you should open an issue in the fwupd GitHub page. The project is very active. Attach the firmware you want to use, and the output of sudo fwupdtool get-devices --verbose
Also, do you have a dump from the VmmUpdate.exe of your Sunix before you tried updating it? Actually, I think you posted it at #18,347
Maybe the people in the fwupd project can tell you how to patch the code to bypass that check if there isn't already an option to do that.

Well, before doing that, we can look at the source code ourselves. The error message is from https://github.com/fwupd/fwupd/blob.../synaptics-mst/fu-synaptics-mst-device.c#L812
There's a couple flags that can bypass the check: FWUPD_INSTALL_FLAG_IGNORE_VID_PID which I think is a generic flag? and FU_SYNAPTICS_MST_DEVICE_FLAG_IGNORE_BOARD_ID which is specific to the device.

FWUPD_INSTALL_FLAG_IGNORE_VID_PID is set according to the ignore_vid_pid flag at https://github.com/fwupd/fwupd/blob/19abf996c72bfffe31c01c19fb58de0699b3136d/src/fu-tool.c#L3349
The ignore_vid_pid flag is set by the "ignore-vid-pid" command line argument at https://github.com/fwupd/fwupd/blob/19abf996c72bfffe31c01c19fb58de0699b3136d/src/fu-tool.c#L2940
Typing man fwupdtool tells you to use fwupdtool --help to list the options. Doing that lists --ignore-vid-pid as an option. So maybe that's the answer.
 
Ok cool, I'll give that a shot soon.

I did save all my log files from the Windows programs so I should be good there if I make a thread at github.
 
Thanks so much joevt!

So I used --ignore-vid-pid and it started installing. Took way longer to install than the Windows app, but it eventually gave me this error:

Code:
failed to update-cleanup after failed update: failed to wait for detach replug: device 7c2dd454e6f29453896a4824e8234317d237f8c5 did not come back
Firmware update failed: checksum 0 mismatched 57c1d9


BUT I noticed that my desktop flashed when I unplugged it this time, which is definitely a sign of life.

Hooked it back up to main Windows PC, and it was detected and recognized as a vmm2322 monitor, but didn't have any output on VGA at least. Also, no light was on.

However, the Synaptics Windows app could see the adapter again!

So I dumped the ROM and the information text file, and I saw that it recognized the chip, but reported it has having no firmware. Just blank lines next to "firmware version"

That seems to indicate that fwupd wasn't able to fully or correctly install my eeprom backup for some reason.

So after that, I just tried installing the same firmware via the app, and it worked! I didn't even need to force it, Synaptics recognized it as having the correct chipset.

And now I'm back in business! The information file reports I'm back on the orignal firmware (2.15.002). Light turns on. VGA output restored!

So I guess I should report a condensed version of this on github now?
 
Thanks so much joevt!

So I used --ignore-vid-pid and it started installing. Took way longer to install than the Windows app, but it eventually gave me this error:

Code:
failed to update-cleanup after failed update: failed to wait for detach replug: device 7c2dd454e6f29453896a4824e8234317d237f8c5 did not come back
Firmware update failed: checksum 0 mismatched 57c1d9

BUT I noticed that my desktop flashed when I unplugged it this time, which is definitely a sign of life.

Hooked it back up to main Windows PC, and it was detected and recognized as a vmm2322 monitor, but didn't have any output on VGA at least. Also, no light was on.

However, the Synaptics Windows app could see the adapter again!
When I used fwupd, I remember getting some kind of error but I don't remember what it was. The update seems to have worked anyway (may need to disconnect the adapter and reconnect it to be sure).

I wonder how the checksum works? I don't think fwupd has the ability to read the rom from synaptics MST hub, so it can't get the checksum that way.

So I dumped the ROM and the information text file, and I saw that it recognized the chip, but reported it has having no firmware. Just blank lines next to "firmware version"
What app are you using? I don't think VmmUpdater.exe has the ability to dump the ROM or maybe I'm missing something.

That seems to indicate that fwupd wasn't able to fully or correctly install my eeprom backup for some reason.

So after that, I just tried installing the same firmware via the app, and it worked! I didn't even need to force it, Synaptics recognized it as having the correct chipset.

And now I'm back in business! The information file reports I'm back on the orignal firmware (2.15.002). Light turns on. VGA output restored!

So I guess I should report a condensed version of this on github now?
I suppose reporting the adventure might be useful or at least interesting. Maybe include a copy of the files (including the eeprom backup - maybe there's something weird about the format if it didn't come from the manufacturer?).
 
What app are you using? I don't think VmmUpdater.exe has the ability to dump the ROM or maybe I'm missing something.

VmmDPTool.

I found it in the Windows Store. Not sure if it's available elsewhere. It has a bunch of debugging stuff that I'm not familiar with, some of it only accessible to Synaptics devs.
 
VmmDPTool.

I found it in the Windows Store. Not sure if it's available elsewhere. It has a bunch of debugging stuff that I'm not familiar with, some of it only accessible to Synaptics devs.
I could not get it from the Windows Store. Search from the Microsoft Store app does not find it. I can find it using google which leads to the Microsoft Store website, but clicking "Get" goes to the Microsoft Store app which shows the info for the app but says "is currently not available". I found the app in the download at the bottom of https://kb.cablematters.com/index.php?View=entry&EntryID=147
It says it's version 8.4.0. Is that the same as yours?
I used the "Save to file button" to save my firmware. Works but it takes a few minutes. In the case of devices where the source firmware file contains duplicate ranges, VmmDPTool only reads one range (I haven't tried all my devices yet).
 
Didn't expect to come here so soon for a technical problem with my 454. Sadly here I am.

A couple months ago I tried turning it on for the first time in a while, and was immediately met with this. The monitor has low hours (likely sub 1000), but it was sitting unused for years prior to me using it.


I've never had any issues prior to this, does anyone know what the issue is, and if it's fixable? Any help would be greatly appreciated, miss using her.

Also, another CRT monitor of mine has a problem where there's an audible "click" and the image refocuses or looks like it degausses almost. Any ideas on that too? Might be able to get a video of it sometime but it's in storage right now.

Not sure what's going on with your 454, looks strange. To clarify, those moving lines are visible to you right? (i.e. they're not an artifact of refresh rate interacting with camera framerate).

As for the click and refocus with your other crt, many people here, including myself, have experienced that with aging trinitrons - might have something to do with an impurity temporarily shorting out some part of the electron gun, or the flyback transformer, and then being burnt off. Speculating here.
 
OK, the Ori-co PTAV adapter finally came in (doesn't actually have a hyphen in name, I'm just getting around the censor).

I'm having a few problems, but I think it probably has the Lontium LT8711X-B with 4 DP lanes. I'm curious what you guys think. Especially Derupter

My main problem is that my 5700xt, even though it is connected through a Sunix upd2018 DP>USB-C converter, is detecting it as a VGA adapter, and won't let me create custom resolutions above 360mHz or (or maybe 400), either through CRU or the built in tool.

So as of right now, I can't test maximum pixel clock. BUT I do know it has 4 lanes. At 2560x1600 it shows 4 lanes at 2.7Gbps, and at 1024x768 it shows 2 lanes at 1.67Gbps
********** compare.jpg


So you guys think this is the real deal, right? I don't know if this can be used easily, at full speed, with an AMD card though, not through displayport or USB-C anyway. I might try that HDMI 2.0>USB C converter that joevt mentioned to see if I have better luck with that.

I'm also having an issue where I'm seeing 20 degree yellow diagonal line going across the screen at the bottom of all resolutions. At higher resolutions it's two lines. So I'm not exactly sure if its the VGA adapter, or the Sunix card doing that.

I have another USB-C to HDMI/VGA adapter on the way, one that I don't expect to be using all 4 lanes, that I can at least use to test to see if the Sunix card is the thing having issues.

So this is "to be continued" until I get that HDMI 2.0 adapter. Wish I could borrow a 2000 series RTX card to test with.
 
Last edited:
In other news, I've learned a bit more about Sunix DPU3000 firmware, and I've tried out a low-spec Vention HDMI>HDMI/VGA adapter.

So back when I thought I killed my Sunix DPU3000 adapter, I ordered a clone from Icy-Box, the IB-SPL1031, to replace it.

That finally arrived, and I saw it had newer firmware: 2.27 versus the Sunix's 2.15. The firmware actually had "sunix" in the name, so I'm guessing it was also in later releases of the Sunix DPU3000.

I flashed this new firmware to my Sunix to see if it performed any better, and as far as I can tell, it didn't. Still get jitter at some resolutions and side-swapping at resolutions slightly above 2048x1536.

But 2.27 was actually worse: it reduced the max pixel clock! I couldn't run 2880x2160 or 3200x1800 any more. So I went ahead and flashed the older firmware back to my Sunix.

If anybody wants to experiment further with these firmware, I'll attach them below. My guess is 2.27 has fixes for the DP and HDMI outputs, not the VGA.

And please, if anybody else has a Sunix or the Delock or Icybox rebrand, upload your firmware if you have a different one from these two.

And for the Vention AFVHB adapter, i put it through some tests. It can hit 1600x1200 at 85hz, but gitches out on 2048x1536 @ 60hz. So it has a max pixel clock somwhere between those two. It would be fine for 17" monitors and lower frequency 19" monitors.
 

Attachments

  • Sunix DPU3000 firmware.zip
    73 KB · Views: 0
Last edited:
OK, the Ori-co PTAV adapter finally came in (doesn't actually have a hyphen in name, I'm just getting around the censor).

I'm having a few problems, but I think it probably has the Lontium LT8711X-B with 4 DP lanes. I'm curious what you guys think. Especially Derupter

My main problem is that my 5700xt, even though it is connected through a Sunix upd2018 DP>USB-C converter, is detecting it as a VGA adapter, and won't let me create custom resolutions above 360mHz or (or maybe 400), either through CRU or the built in tool.

So as of right now, I can't test maximum pixel clock. BUT I do know it has 4 lanes. At 2560x1600 it shows 4 lanes at 2.7Gbps, and at 1024x768 it shows 2 lanes at 1.67GbpsView attachment 377840
The Lontium LT8711X-B is supposed to be able to do HBR2. Maybe it's choosing HBR because that is more likely to have a cleaner signal.

You need a dump of the DPCD registers to learn the maximum link rate. The AGDCDiagnose command on an Intel Mac will get you the info. I'm not sure if there's a method to get the values from Linux. AGDCDiagnose also interprets most of the registers in human readable format. You can get a list of the registers from the DisplayPort 1.2 spec (google it) but it will be missing the DisplayPort 1.4 registers (can't find that spec on the web). The DPCD also includes a 3 byte OUI that identifies the manufacturer (there's a list online to translate that) and a 5 byte product ID (I don't think there's a list online for that part but it may contain values related to the chip as is the case for Synaptics chips).

If the max link rate is HBR, then that would explain why you can't exceed 360 MHz. 360MHz is exactly the maximum for HBR x4 8bpc RGB. I would have thought that some kind of overhead would need to be included so that the max would be more like 350 MHz? I suppose the adapter doesn't advertise 4:2:2 or 6bpc to allow for higher pixel clock? I would test that with an EDID override.

I don't know if this can be used easily, at full speed, with an AMD card though, not through displayport or USB-C anyway. I might try that HDMI 2.0>USB C converter that joevt mentioned to see if I have better luck with that.
I think the UPD2018 would have better behaviour than a HDMI 2.0 to USB-C adapter since the UPD2018 just does DisplayPort pass thru. Both the UPD2018 and the HDMI 2.0 to USB-C adapters are limited to DisplayPort 1.2.

I'm also having an issue where I'm seeing 20 degree yellow diagonal line going across the screen at the bottom of all resolutions. At higher resolutions it's two lines. So I'm not exactly sure if its the VGA adapter, or the Sunix card doing that.
I expect digital adapters to not have such artifacts. But this is VGA. Probably there is some interference with the DAC or analog side. If it's the Sunix card (maybe it's not providing clean USB power) then you'll want to test other USB-C DisplayPort Alt Mode sources.

So this is "to be continued" until I get that HDMI 2.0 adapter. Wish I could borrow a 2000 series RTX card to test with.
You could try a Thunderbolt add-in card for DisplayPort to USB-C DisplayPort Alt Mode conversion. They don't require a motherboard with Thunderbolt support to do USB and DisplayPort Alt Mode.
AMD Navi (5000 series) is another possibility.
 
would test that with an EDID override.

That's what I was getting at in the first post. AMD cards will just ignore EDID overrides with a lot of external adapters. They do this with the Sunix DPU3000 hub as well. And it won't let me create higher resolutions via the built in tool. My best guess is that it's a forced limitation in the driver when it detects a DP>VGA adapter.

With the Sunix DPU3000 though, I'm able to get around it by using an external EDID emulator in the chain between the monitor and DPU3000, maybe because its a MST hub. With this straight USB-C>VGA adapter, the drivers seem to be more strict and just ignore the higher values it reads on the external EDID.

No access to a Mac, I'll see if I can find something to expose the DPCD either in Linux or Windows.
 
That's what I was getting at in the first post. AMD cards will just ignore EDID overrides with a lot of external adapters. They do this with the Sunix DPU3000 hub as well. And it won't let me create higher resolutions via the built in tool. My best guess is that it's a forced limitation in the driver when it detects a DP>VGA adapter.

With the Sunix DPU3000 though, I'm able to get around it by using an external EDID emulator in the chain between the monitor and DPU3000, maybe because its a MST hub. With this straight USB-C>VGA adapter, the drivers seem to be more strict and just ignore the higher values it reads on the external EDID.

No access to a Mac, I'll see if I can find something to expose the DPCD either in Linux or Windows.
I remember AMD Radeon software custom resolutions don't work well - they don't let you change the bpc so you have to use CRU. But the EDID overrides created by CRU do not get applied to displays connected to MST hubs that are connected to AMD GPUs. But I believe this problem was fixed in a recent AMD driver version. I don't think I've tried an EDID override for a VGA adapter in Windows. I did try it in macOS - I had to create a kext pixel clock patch (for Intel GPU) to get more than 160 MHz for an adapter that supports 330MHz (two lanes of HBR2 which can support up to 360MHz but the adapter produces a black screen after 330 MHz).

EDID emulators are great - too bad no-one makes an inexpensive one for DisplayPort but since you're using VGA, you have the option to put can emulator on the VGA side of the connection.

For DPCD, in Linux, the Intel driver has some DPCD info at /sys/kernel/debug/dri/0/DP-1/i915_dpcd but if the device is an MST hub then you have to make sure you're reading from the DisplayPort to VGA adapter rather than the MST hub. It may be that the MST hub has different link rate / link width than the VGA adapter. If you're not doing Intel or the file doesn't have all the registers that you want to see, then there's maybe an i2c device for reading/writing DPCD? I've never looked into that.
 
OK, the Ori-co PTAV adapter finally came in (doesn't actually have a hyphen in name, I'm just getting around the censor).

I'm having a few problems, but I think it probably has the Lontium LT8711X-B with 4 DP lanes. I'm curious what you guys think. Especially Derupter

My main problem is that my 5700xt, even though it is connected through a Sunix upd2018 DP>USB-C converter, is detecting it as a VGA adapter, and won't let me create custom resolutions above 360mHz or (or maybe 400), either through CRU or the built in tool.

So as of right now, I can't test maximum pixel clock. BUT I do know it has 4 lanes. At 2560x1600 it shows 4 lanes at 2.7Gbps, and at 1024x768 it shows 2 lanes at 1.67GbpsView attachment 377840

So you guys think this is the real deal, right? I don't know if this can be used easily, at full speed, with an AMD card though, not through displayport or USB-C anyway. I might try that HDMI 2.0>USB C converter that joevt mentioned to see if I have better luck with that.

I'm also having an issue where I'm seeing 20 degree yellow diagonal line going across the screen at the bottom of all resolutions. At higher resolutions it's two lines. So I'm not exactly sure if its the VGA adapter, or the Sunix card doing that.

I have another USB-C to HDMI/VGA adapter on the way, one that I don't expect to be using all 4 lanes, that I can at least use to test to see if the Sunix card is the thing having issues.

So this is "to be continued" until I get that HDMI 2.0 adapter. Wish I could borrow a 2000 series RTX card to test with.
Are you sure the chipset is the LT8711X-B?
Your exact model is ORIC0 PTAV?
Oh now i see that, but only on the Chinese website
Maybe you can ask ORIC0 to be sure

It is normal to see 4 HBR lanes instead of 2 HBR2 under 360 MHz for stability purpose.
The inability to go in HBR2 mode can be caused by several things:
-the chipset is not the LT8711X-B, the only chip i know with 4 HBR lanes (not HBR2) is the ITE IT6562
-bad cable-poor PCB design of the adapter, bad passthrough cable (from GPU to UPD2018)
-incompatibility or driver problems with your graphic card
-chipset firmware artificially locked to handle HBR only?
But usually when there are cable-connection problems you can configure resolutions that need HBR2 and bad things happen only when you set it.
In your case is like the HBR2 is not supported instead of stability problems with HBR2 (like the Delock 62967)
I don't think it has anything to do with the UPD2018, that card has a great hardware and the Texas Instruments crosspoint switch is a very good chip (at least on paper)

About the Vention AFVHB
Based on your tests, do you agree with this report?
If all this is true it's not a great presentation for these Lontium chipset
Now the question is:
What the fuck adapter does the Digital Foundry guy have that can do 2304x1440 80Hz without problems, what's the exact model code?
Maybe we discover that it is a Synaptics based and not a Lontium one.

About adapters in general
There is a new chipset that can potentially go over 360 MHz, the Algoltek AG9300 (USB-C to VGA)
Adapters available with this chipset:
-Delock 64002
-Digitus DA-70853
-Lindy 43295
Now keep in mind that the cheap Benfei and Rankie adapters from Amazon are based on the AG6200 with that problem of YCBCR over 180 MHz

Delock has a new adapter which appears to be a copy of the impossible to find Vention CMFHB, the 87776
I asked them what chipset is inside, if is the Lontium LT8712X finally we have an adapter to test

I updated the adapters post with the new entries here
 
-incompatibility or driver problems with your graphic card

I'm almost sure this is why. I have resolutions in my EDID emulator that won't even show up in Windows. And AMD drivers are blocking custom resolutions over a certain point, with the message "resolution not supported by your display". I can make 2560x1600@60hz, but I can't make 2400x1800@60hz.

So that's why I think this is simply AMD is software blocking high pixel clocks for VGA adapters.

I ordered the gofanco HDMI 2.0>USB C adapter, a clone of the Club 3D, and I know I'll be able to trick my Radeon into thinking I'm not using a VGA adapter with that. I have a Dr. HDMI EDID emulator to put in between just in case.



About the Vention AFVHB
Based on your tests, do you agree with this report?
If all this is true it's not a great presentation for these Lontium chipset

Yeah, it gets blurry and pixelated above a certain threshold.

Now the question is:
What the fuck adapter does the Digital Foundry guy have that can do 2304x1440 80Hz without problems, what's the exact model code?
Maybe we discover that it is a Synaptics based and not a Lontium one.

Didn't we figure it was the Vention CGMHA? I think he got lucky and bought one before they were discontinued

About adapters in general
There is a new chipset that can potentially go over 360 MHz, the Algoltek AG9300 (USB-C to VGA)
Adapters available with this chipset:
-Delock 64002
-Digitus DA-70853
-Lindy 43295

I have an adapter with the AG9321 on the way: https://www.aliexpress.com/item/1005002898714265.html?spm=a2g0s.9042311.0.0.22864c4dhb1Lf2

It has optional 4 DP lanes, but I guess some of the data bandwidth might be permanently going to the USB outputs? I bought one in case it's like the Synaptics Vmm2322, where you get a higher pixel clock if you only use one output.
 
Last edited:
Didn't we figure it was the Vention CGMHA? I think he got lucky and bought one before they were discontinued
No he showed an Amazon page in the video where you can see the Vention CGKHA which uses the ITE IT6562 + VIA VL100.
With these chipset it can't go over 360 MHz so we thought he pointed to the wrong model, but we don't know what the exact model is.
If it's really the CGKHA then the Vention specs are wrong and we don't know what is the real chipset, also the CGKHA is no more available.
I have an adapter with the AG9321 on the way: https://www.aliexpress.com/item/1005002898714265.html?spm=a2g0s.9042311.0.0.22864c4dhb1Lf2

It has optional 4 DP lanes, but I guess some of the data bandwidth might be permanently going to the USB outputs? I bought one in case it's like the Synaptics Vmm2322, where you get a higher pixel clock if you only use one output.
That's the problem, with VMM2322 you have one chip that handle all the outputs and all the four DP lanes are connected, with those docking stations there are outputs not handled by the main chipset so they use some of the DP lanes for the extra things, in that case the USB 3.0 port.
I don't know if there are cheap switches to handle the DP lanes on those things, in case you don't use certain outputs.
When you test it you'll see if all the four lanes are connected or not
 
Last edited:
I have an adapter with the AG9321 on the way: https://www.aliexpress.com/item/1005002898714265.html?spm=a2g0s.9042311.0.0.22864c4dhb1Lf2

It has optional 4 DP lanes, but I guess some of the data bandwidth might be permanently going to the USB outputs? I bought one in case it's like the Synaptics Vmm2322, where you get a higher pixel clock if you only use one output.
Right. The spec says it has USB 3.0. USB 3.0 requires two lanes of USB-C so only two lanes remain for DisplayPort (USB 2.0 uses USB-C lines that are not used by USB 3.x or DisplayPort). It's limited to DisplayPort 1.2 HBR2 (no DSC) input so the max is 360 MHz (assuming 8bpc RGB).

There are some displays that can switch between USB 3.x with two lanes of DisplayPort and USB 2.0 with four lanes of DisplayPort. I've not heard of any other device (except host USB-C ports) that have that ability.
 
No he showed an Amazon page in the video where you can see the Vention CGKHA which uses the ITE IT6562 + VIA VL100.
With these chipset it can't go over 360 MHz so we thought he pointed to the wrong model, but we don't know what the exact model is.
My best guess is that John bought the CGMHA many months earlier, probably after reading this thread, then by the time he made the video, it was discontinued. But he didn't know the difference, so he just posted the picture for the other adapter that looks similar.

It would be easy to do, they all look so similar
 
Last edited:
Ok guys, so the goFanco (probably identical to Club 3D) HDMI 2.0>USB-C converter gives no signal and isn't detected by my GPU or PS4. I'm assuming this is because it waits for a handshake on the USB-C side, and it doesn't provide the necessary power the oric0 adapter needs to power on and do said handshake.

So the question now is.... should I go deeper?

I'm assuming this thing is designed to output DP-alt mode when it's working. So do I just run the HDMI 2.0>USB-C converter to a USB-C>DP converter? Maybe some of those don't need too much, if any, additional power? Then I would feed that DP signal BACK into my Sunix DP>USB card where I then hook up the oric0 again. Hopefully at this point, with the extra resolutions HDMI 2.0 affords in AMD drivers.
 
Last edited:
Ok guys, so the goFanco (probably identical to Club 3D) HDMI 2.0>USB-C converter gives no signal and isn't detected by my GPU or PS4. I'm assuming this is because it waits for a handshake on the USB-C side, and it doesn't provide the necessary power the oric0 adapter needs to power on and do said handshake.

So the question now is.... should I go deeper?

I'm assuming this thing is designed to output DP-alt mode when it's working. So do I just run the HDMI 2.0>USB-C converter to a USB-C>DP converter? Maybe some of those don't need too much, if any, additional power? Then I would feed that DP signal BACK into my Sunix DP>USB card where I then hook up the oric0 again. Hopefully at this point, with the extra resolutions HDMI 2.0 affords in AMD drivers.
Just to be sure, we are talking about the gofanco HDMIUSBC and not the USBCHDMI? The HDMIUSBC requires USB 2.0 input for power. Is it getting enough power? The USB 2.0 connection should also allow USB connection to USB-C DisplayPort Alt Mode display. The display cannot be of the USB 3.x + two lane DisplayPort Alt Mode variety. I don't know if it works with two lane DisplayPort Alt Mode adapters or with MST adapters.

USB-C>DP cable should work. The thing that requires power is the USB-C output since USB-C is supposed to provide more power than HDMI. Also, all HDMI to DisplayPort adapters use USB for additional power for the converter even if they are not doing USB-C output so it's reasonable that this HDMi to USB-C adapter is no different in that regard.
 
Yeah, it's HDMI 2.0>USB-C with micro USB for power. But I don't know how much power it's providing on that USB-C output. Maybe not 5v, maybe just 3.3v like it's a displayport cable?

Alright, since USB-C to DP cables are so cheap, I might as well give it a shot.

But I'm thinking it's one of two situations with this gofanco adapter: not enough power for the VGA adapter, or it's just no the handshake it's looking for. If the handshake is the problem, then going through the Sunix DP>USB-C probably won't work, because I think that just passes through DP handshakes and all that.
 
Ok guys, so the goFanco (probably identical to Club 3D) HDMI 2.0>USB-C converter gives no signal and isn't detected by my GPU or PS4. I'm assuming this is because it waits for a handshake on the USB-C side, and it doesn't provide the necessary power the oric0 adapter needs to power on and do said handshake.

So the question now is.... should I go deeper?

I'm assuming this thing is designed to output DP-alt mode when it's working. So do I just run the HDMI 2.0>USB-C converter to a USB-C>DP converter? Maybe some of those don't need too much, if any, additional power? Then I would feed that DP signal BACK into my Sunix DP>USB card where I then hook up the oric0 again. Hopefully at this point, with the extra resolutions HDMI 2.0 affords in AMD drivers.
That is the same old problem that arises when you try a daisy chain with multiple adapters.
Look here and here
The way through the Sunix Card can work, it depends on how the gofango adapter handles the displayport signal.
 
Ok, the saga continues, one step forward, one step back.. Got some confusing results.

The new chain is:

5700xt HDMI>Dr. HDMI 4k with custom EDID>gofanco HDMI 2.0 to USB-C>USB-C to DP cable>Sunix DP to USB-C card>Oric0 PTAV USB-C to VGA adapter

And it works kinda. AMD drivers no longer hiding resolutions thanks to the Dr. HDMI.

But, I'm actually getting worse performance overall.

For example, when I just run DP from GPU>Sunix DP to USBC>Oric0, I can display 1760x1320 @ 100hz with a stable picture.

But with the new HDMI setup, I get some unstable output with scanlines creeping back and forth, giving comb-like artifacting. And even though I'm now able to select 1920x1440 @ 90hz from the drop-down list, the output isn't stable enough for my monitor to sync.

So now I'm wondering if the gofanco adapter sucks (it's supposed to be able to handle 3840x2160 @ 60hz, LCD timings), or if the problem lies elsewhere. I don't think it's the USB>DP cable, it's only 2 feet long, and supposedly supports 4k60 speeds.

In other news: spacediver might be interested to hear that AMD drivers are now outputting 10 bit color to the Sunix DPU3000 hub, with most resolutions.

This wasn't the case when I first got the 5700xt, so I don't know when it changed. This applies even to really high resolutions over 400mHz, like 2304x1728@75hz. However, there seems to be a limit at some point, because 2880x2160@60hz was still reporting 8bpc.
 
Last edited:
I'm on the newest driver, but I have no idea when they made this update. There's a good chance they made the change before they dropped Win 7.

You know you can activate a Win 10 install with your Windows 7 key, just FYI
 
Oh, I know, It's just that I neither want a spyware nest nor an OS that does what he wants without asking my opinion on any of my computers. The day I'll be forced to leave windows 7, I'll probably finally switch to Linux. ;)
 
Ok, the saga continues, one step forward, one step back.. Got some confusing results.

The new chain is:

5700xt HDMI>Dr. HDMI 4k with custom EDID>gofanco HDMI 2.0 to USB-C>USB-C to DP cable>Sunix DP to USB-C card>Oric0 PTAV USB-C to VGA adapter

And it works kinda. AMD drivers no longer hiding resolutions thanks to the Dr. HDMI.

But, I'm actually getting worse performance overall.

For example, when I just run DP from GPU>Sunix DP to USBC>Oric0, I can display 1760x1320 @ 100hz with a stable picture.

But with the new HDMI setup, I get some unstable output with scanlines creeping back and forth, giving comb-like artifacting. And even though I'm now able to select 1920x1440 @ 90hz from the drop-down list, the output isn't stable enough for my monitor to sync.

So now I'm wondering if the gofanco adapter sucks (it's supposed to be able to handle 3840x2160 @ 60hz, LCD timings), or if the problem lies elsewhere. I don't think it's the USB>DP cable, it's only 2 feet long, and supposedly supports 4k60 speeds.
Maybe it's a bug with the HDMI output of the 5700 XT, same results without the Dr HDMI?
In other news: spacediver might be interested to hear that AMD drivers are now outputting 10 bit color to the Sunix DPU3000 hub, with most resolutions.

This wasn't the case when I first got the 5700xt, so I don't know when it changed. This applies even to really high resolutions over 400mHz, like 2304x1728@75hz. However, there seems to be a limit at some point, because 2880x2160@60hz was still reporting 8bpc.
Can it be useful if the DAC is 8 bpc?
Most of the displayport receivers of these adapters support 10 bit on input, the fact that it is in most cases unselectable is a mystery.
Example LT8711X-B and LT8712X should support 10 bpc on input and output(VGA)
Can you select 10 bit only with the DPU3000?
Have you tried with your other adapters?

About the Delock 87776
The chipset inside it is the ITE IT6224 which is like the IT6564 but with native USB-C input.
 
8-bit vs 10-bit doesn't matter on AMD cards because AMD cards do very high quality dithering on all 8-bit displays. This is one of two advantages of AMD cards over Nvidia cards. Second being able to correct gamut. For CRT this requires passing custom EDID to the card which with something like Dr. HDMI should be possible.

The EDID emulation is necessary because FW900 has whitepoint in EDID configured to different value than exactly what it should be for sRGB.
When I had Radeon 7950 I made myself EDID emulator out of I2C EEPROM and DVI-A to VGA converter (just 'hacked' it and put SO8 IC inside it) and programmed it via Arduino and custom SW for it to put actually measured color points for R, G and B and default D65 whitepoint. This resulted in having all colors which fall withing monitor and sRGB gamuts to be displayed correctly. sRGB colors outside monitors color space were clipped as to not change their hue which means overall range of displayed colors was slightly smaller but most content looked much better, especially things like skin tones. I still have my EDID emulator but without AMD card it is useless so I do not use it.

Dithering thing solves one issue 8-bit connection has which is banding issues caused by gamma manipulation. This can be enabled on Nvidia cards but it requires tweaking registry and not most easy to do procedure.

Of course true 10-bit DAC will be better than 8-bit DAC with dithering but only when 10-bit DAC also has dithering!!!
10-bit alone gives more overall error than 8-bit + temporal dithering (I did check this on 10-bit monitor)
I did noise analysis of both AMD and Nvidia algorithms and noise is there but it is imperceptible (temporal algorithm in Nv, the same as AMD uses) in normal conditions.
Dithering noise it is not really perceivable on CRT so it is good enough to not worry about lacking 10-bit output.
AMD card owners should not worry about how much "bits" their converters have.
Nvidia owners should worry and for 8-bit they should enable dithering (unless they do not use gamma correction at all and use native gamma in which case 8-bit is enough). For 10-bit adapters enabling dithering is still advised because 10-bit is just four more brightness levels than 8-bit. It allows displaying all 256 unique levels though and differences are so miniscule enabling dithering should not make big difference. Still I would do it and I do it for my 10-bit displays.

5700xt HDMI>Dr. HDMI 4k with custom EDID>gofanco HDMI 2.0 to USB-C>USB-C to DP cable>Sunix DP to USB-C card>Oric0 PTAV USB-C to VGA adapter
That is quite long snake you got there 😅
If that works there should be no input lag as long as none of adapters add any significant amount of lag. there will probably be some input lag but I doubt it will be even enough of it to count single line worth of latency.

Issues you see might come down to power. Are these adapters (active ones) all powered externally or are they taking power from ports?
 
So I recently picked up a Startech DP2VGAHD20 DP->VGA/HDMI adapter. While the VGA and HDMI outputs work fine with a couple monitors I have tried connected to my Nvidia GTX 1070, I'm having some issues with sleep/standby.

Whenever Windows puts the monitor connected to the Startech into sleep/standby, the adapter appears to go into a disconnection loop. I keep hearing Windows play the sound effect you hear when a new device in being plugged in or detected. This connect and disconnect keeps repeating forever every couple seconds until I touch the mouse or keyboard to wake the monitor. It's extremely annoying and I either need to pull the adapter, turn off my speakers, or disable display sleep in Windows power settings.

Can't tell if this is normal for this adapter or if it's defective. I have tried it with 3 different video cards (GTX 960, 1070, and Intel iGPU) and 2 different PCs running Windows 10 and Windows 8. Also 2 different monitors. Happens with either VGA or HDMI.

Has anyone else experienced such issue? I have another DP -> VGA/HDMI/DVI adapter that doesn't have this problem.
 
Does this adapter have external power source?
On the StarTech website all photos are from the same side so I cannot tell...
Some times these adapters have slot for power plug but work without it depending on how much power they can take from source (DVI/DP/HDMI)
 
So I recently picked up a Startech DP2VGAHD20 DP->VGA/HDMI adapter. While the VGA and HDMI outputs work fine with a couple monitors I have tried connected to my Nvidia GTX 1070, I'm having some issues with sleep/standby.

Whenever Windows puts the monitor connected to the Startech into sleep/standby, the adapter appears to go into a disconnection loop. I keep hearing Windows play the sound effect you hear when a new device in being plugged in or detected. This connect and disconnect keeps repeating forever every couple seconds until I touch the mouse or keyboard to wake the monitor. It's extremely annoying and I either need to pull the adapter, turn off my speakers, or disable display sleep in Windows power settings.

Can't tell if this is normal for this adapter or if it's defective. I have tried it with 3 different video cards (GTX 960, 1070, and Intel iGPU) and 2 different PCs running Windows 10 and Windows 8. Also 2 different monitors. Happens with either VGA or HDMI.

Has anyone else experienced such issue? I have another DP -> VGA/HDMI/DVI adapter that doesn't have this problem.
Yes this problem has been mentioned here with a dual monitor configuration.
Maybe this should be fixable disabling the secondary monitor in device manager, but with single monitor you can't do that.
I don't have that adapter so i can't test, my Delock 62967 with Analogix chipset doesn't have that problem.
You can disable device connection-disconnection sounds in Windows options, or shut off the monitor manually.
 
Last edited:
Yes this problem has been mentioned here with a dual monitor configuration.
Maybe this should be fixable disabling the secondary monitor in device manager, but with single monitor you can't do that.
I don't have that adapter so i can't test, my Delock 62967 with Analogix chipset doesn't have that problem.
You can disable device connection-disconnection sounds in Windows options, or shut off the monitor manually.

Google search and online reviews didn't turn up anything. Should have searched this forum before posting.

Thanks for the link. At least I now know my unit is not defective. Unfortunately it's happening even with a single monitor attached to the system. I opened a tech support ticket with Startech. Maybe they can do some kind of firmware update on existing units or come up with a new revision to fix it. A real shame because this adapter supported higher resolutions and refresh rates than any other DP adapter I've tried. Turning off the Windows sounds is one workaround but disconnects will still happen in the background and cause other issues like mentioned by others. Unfortunately this adapter doesn't have an external power input but neither did others I tried without this problem.
 
Google search and online reviews didn't turn up anything. Should have searched this forum before posting.

Thanks for the link. At least I now know my unit is not defective. Unfortunately it's happening even with a single monitor attached to the system. I opened a tech support ticket with Startech. Maybe they can do some kind of firmware update on existing units or come up with a new revision to fix it. A real shame because this adapter supported higher resolutions and refresh rates than any other DP adapter I've tried. Turning off the Windows sounds is one workaround but disconnects will still happen in the background and cause other issues like mentioned by others. Unfortunately this adapter doesn't have an external power input but neither did others I tried without this problem.
Good idea, if there are news from Startech please report here.
I can understand the problem with dual monitor because if you continue to use the computer with the primary display you have that thing in background that brake your balls.
But with a single display, when it goes to standby, it is assumed that you are not using the computer exept for downloading and other automated stuff or am i wrong?
What happens to the system, does it become unstable?
Does that thing consume resources?
About the adapters, those with the Synaptics chipset have better performance than yours, here you can find a list.
 
Last edited:
The Startech DP2VGAHD20 tops out at about 370MHz which in the case of the resolutions I use is 2184x1365 85Hz (367.37MHz). But as you can see from the attached pic I use a lot more than that on my GTX 980Ti with the highest (limited by the videocard's 400MHz RAMDAC not monitor's horizontal bandwidth) beeing 2736x1710 60Hz which is quite sharp in 60 fps capped games (i.e. idTech 4 engine based games). To me, the only acceptable way to use the Dell P1130, Sony GDM-FW900 and F520 that I have is with the 980Ti only.
 

Attachments

  • Annotation.png
    Annotation.png
    53.7 KB · Views: 0
Last edited:
There are converters which even higher pixel clock than 400MHz so for these cases where it is the pixel clock that is the limiting factor they would be even better.

Though I personally do not see any reason why anyone should bother to run more than 1200 lines on these small CRT's when it is already hitting ability of the tube to resolve each pixel and when each line must be compensated by reducing refresh rate to stay within horizontal refresh rate limits. 97Hz at 1200p is the lowest I am willing to go and I usually use 1080p@107Hz when I use FW900 anyways.

Of course 60fps locked games can use higher resolution without lowering refresh rate because it is already low. The issue here however is how much should few 60fps locked games matter? Does it really make such a big difference if you play Doom3 or Rage or whatever at higher or lower resolution? You can always use DSR to put more load on the GPU if it being bored bothers you
 
Still using my FW900 as my main monitor at home for everything as I have for many years. (More work in that mix of course with the current situation. Though on the flip side, the phone took over some of the hours that would have been on the CRT.)

I run 1880 x 1200 at 85 Hz for work. That same resolution at 72 Hz for Cyberpunk. (Highest I could lock or almost lock in with settings maxed. Though with DLSS Quality engaged. Game seems to benefit overall from a higher resolution over running at a lower resolution with DLSS off.)

1600 x 1024 at 100 Hz for other stuff.

Highest I could get at 100 Hz is 1800 x 1152 with GTF timings. However, I don't really like maxing things out. (And pushing it to the limit there did introduce a little blur.)

I have done some much higher resolutions in the past and it's fun how high it can go and still be legible. However, as far as true clarity goes I think the physical grill is only around 1920 across.

Still using the StarTech DP2VGAHD20. Seems ok. Picked up a spare or two as well.
 
Last edited:
Still using my FW900 as my main monitor at home for everything as I have for many years.
You are one of the last CRT-only Mohicans my friend :dead:
Are you using single monitor setup or have two FW900?
If single monitor then I recommend getting 21" Trinitron (or two) to accompany FW900 and that would be pretty sweet setup 🤗
 
Back
Top