HDMI cable causing phantom power back to Kawari (large) #5
Comments
|
It sounds to me like your monitor is sourcing current which is not supposed to be the case for HDMI sinks. The spec says only display sources are supposed to provide current. There is a load switch on the board that prevents back-feeding. This case would have powered the 64 on the previous version of the board which is no something you want. But even though I added the load switch, it could be the board is not playing nice with your monitor for some reason. Is your monitor's HDMI port MHL compatible by any chance? It could be that your monitor does not detect the 5V on pin 18 and starts providing its own current to power the attached device. Can you share the model # of the monitor? I'm not sure why swapping out the cable would make a difference though. |
|
Asus VE248H, only lists HDMI 1.3 as the port, no extras. It's a standard computer monitor. Here's a link to the manual. https://dlcdnets.asus.com/pub/ASUS/LCD%20Monitors/VE247VE248VE249/VE247-VE249_English_V1.pdf I've put things away for the night/weekend, but I'll think on if I have a way to pinout the cables later. |
|
I've ordered hdmi breakout boards and will be able to find out the different cabling. |
|
I did a little more testing tonight with a small micro hdmi male to standard hdmi female adapter, along with the 2 full cables. Link to the adapters. All 3 of the following monitors show the phantom power. Asus VE248H, Westinghouse L2410NM, Samsung LS27E310HSG/ZA. Voltage across the led when fully powered is 1.8v Phantom power persists even when the kawari is removed from the motherboard. So 2 out of 3 cables are showing phantom power. What cables are you using? |
|
Also, is this something that may be involved? |
|
Yeah, that sounds identical to this issue. I am using 4 pairs of TMDS
lines to drive my DVI signals directly from the FPGA. I have an AP2331W
load switch connected to the 5V line which prevents sourcing current into
the board from pin 18 but from this article it sounds like this can still
occur due to the pull ups. The pullups are on the sink side (monitor) as
per the HDMI spec. I just don't understand why I'm never seeing the issue
myself. I don't use an adapter. I have a single cable with micro-HDMI on
one end and HDMI on the other. You said one of your cables works okay?
Which one was that?
…On Mon, Aug 29, 2022 at 11:30 PM dabonetn ***@***.***> wrote:
Also, is this something that may be involved?
https://support.xilinx.com/s/question/0D52E00006hphYzSAI/unintentional-backpower-spartan6-through-io?language=en_US
<http://url>
—
Reply to this email directly, view it on GitHub
<#5 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAI3HKAFELQUNDSYNRHX52DV3V545ANCNFSM57333ZSA>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Randy Rossi
- "There are only two things that are hard about computer science; Naming
things, Cache Invalidation, and Off-by-one errors."
|
|
The one cable I have that doesn't feed back is a generic one I've had for awhile. Can't remember where I got it. |
|
Would you be able to check the HDMI connector for solder bridges? I'm wondering if there might be a solder bridge on the back pins of the connector somewhere. It's hard to tell sometimes. But perhaps there is a short on the receptacle end and older cables don't matter because that pin may not be connected. But newer cables might have a connection to the monitor port which may explain this. Just a thought. In the meantime, I will try to reproduce it here. |
|
Just looked under the microscope, and didn't see anything. But it's a very hard thing to see on that connector. |
|
Edit... Removed garbage image.. The breakout board was horrible for pinouts. |
|
Can you describe the setup? You have two breakout boards and are testing
the cable connections between micro end and the full end through the
full-micro adapter? The numbers don't really make sense to me. I think
the numbers on the board are supposed to match the pin numbers.
…On Fri, Sep 2, 2022 at 10:39 PM dabonetn ***@***.***> wrote:
[image: This is an image]
<https://camo.githubusercontent.com/6a17c0eacee1d719504e07c99f22b45dbb86f5da740007fcae43e6ac56cfc668/68747470733a2f2f692e696d6775722e636f6d2f4572644b6f33502e6a7067>
First 2 columns provide phantom power, the 3rd does not.
The numbers are what is on the breakout boards, I'm not sure if they
actually mean the hdmi pins. (Cheapest breakouts I could find on Amazon.)
—
Reply to this email directly, view it on GitHub
<#5 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAI3HKGOJ5VLMYBXSBIZKBTV4K26DANCNFSM57333ZSA>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Randy Rossi
- "There are only two things that are hard about computer science; Naming
things, Cache Invalidation, and Off-by-one errors."
|
|
Ok, getting out a needle and figuring out the actual hdmi pinout for the break out boards, the cable that doesn't feed phantom power has the following pins NOT connected. 2, 5, 8, 11, 14, These are the grounds for TMDS 1,2,3 clock, and then HEAC -. |
|
And just for sanity check, using a one to one cable, I see that 2,5,8,11,15 & 19 are connected together on the Kawari board, correct? |
|
Also, if the VGA connector is also connected to the monitor along with the HDMI, the cable that works HDMI only still provides phantom power. |
|
It makes sense that the cable that 'works' doesn't have those pins
connected. Because without those pins there is no return path. So if you
provide one (like VGA cable to the monitor for example), then the return
path is present. But it doesn't make sense for those pins to not be
connected (cheap cable?).
So I think the issue is that if the monitor is pulling up the TMDS lines,
it will power the core as mentioned in the Xilinx article. With my
monitor, I can reproduce the phantom power issue but only before the
monitor gives up on finding a signal. Once it gives up, the voltage is
gone and I can boot again. If you wait a while, does the dimmed LED
eventually turn off?
…On Sat, Sep 3, 2022 at 3:18 PM dabonetn ***@***.***> wrote:
Also, if the VGA connector is also connected to the monitor along with the
HDMI, the cable that works HDMI only still provides phantom power.
—
Reply to this email directly, view it on GitHub
<#5 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAI3HKELCZDT3D74STGWESTV4OQBTANCNFSM57333ZSA>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Randy Rossi
- "There are only two things that are hard about computer science; Naming
things, Cache Invalidation, and Off-by-one errors."
|
|
Asus monitor goes into standby mode, but never kills the pull ups. |
|
I went and grabbed a Acer V227bmix and a Asus VE228 from my stock to try. But for using the monitors, none I have tried allows for scaling the HDMI input, but all allow for the scaling of the VGA input, so vga is looking like the winner for a proper display. |
|
It looks like whether this happens is going to be at the mercy of the monitor. If the monitor times out and drops the power, you can boot by just waiting a few seconds. Another work around is to use an HDMI switch. My switch does not power the core even when it is the display selected by the switch. So it looks like the work arounds are: 1> Wait for monitor to give up its scan for a signal Unfortunately, the proper way to deal with this is to add an HDMI buffer IC which will add more space/cost and potentially trigger license issues. I"ve read that even if you use an IC that is licensed for HDMI, your device still requires an HDMI license. (Kawari uses only DVI protocol and no HDMI). But I wanted to stay clear of even a hint of any HDMI license issues. Do you think these workaround are acceptable? If not, I might just drop the HDMI connector entirely from my design as it is proving to be quite a pain. |
|
IMO, I would drop the DVI support, since you don't have a easy way to force 4:3, and maybe just concentrate on RGB/VGA and maybe optional component output instead of the DVI. Or different firmware to pick between RGB and component. There are so many devices like the RetroTink, etc, getting the highest quality out of the 64 we can while keeping the correct aspect ratio seems to be the way to go in my book. The 15K RGB out looks great on my Sony PVM. |
|
I opened a poll on the discord channel. So far people like the idea of easy HDMI connection even with the boot issue and sometimes lack of perfect 4:3. BTW: Did you have any issues closing up the 250466 machine with the kawari installed? Someone is asking on the discord and I don't have one to try. |
|
I have a few loose boards that I've throw in a breadbin case to test. |
As you can tell by the reports, I finally had a chance to sit down and test some.
I've got 2 micro hdmi to hdmi cables (6' generic) that I bought at different times for pi4 and another project.
One works fine with the kawari, the other does the following.
After the first boot, when I turn off the 64, LED1 on the kawari is faintly lit.
Turn on the 64 after this doesn't boot, until I either unplug and then plug the hdmi cable back in, or power cycle the monitor.
This causes the led to completely go out, and then the 64 boots normally.
The led doesn't light up until after the first boot, so monitor seems to wait until after it has a valid signal to provide voltage back down the hdmi cable.
The text was updated successfully, but these errors were encountered: