Author Topic: HDMI cables: Why would I own several cables with pins 3, 9, 12, 15 unconnected?  (Read 487 times)

0 Members and 1 Guest are viewing this topic.

Offline FinalCatchTopic starter

  • Newbie
  • Posts: 5
  • Country: gb
Greetings. This is my first time posting on the forum. So I apologize in advance for any breach of etiquette.
I wasn't quite sure where to post this but as I'm a beginner this seemed a valid option.

I've been having some issues getting one of my monitors working with one of my PCs (an old Mac actually) over HDMI. So I invested in a cable checker to check some cables to remove one point of failure. I've found some duds and some with all connections working (according to the checker).

However, I've also found several cables (3) with pins 3, 9, 12, and 15 unconnected. Stranger, at least one of them still works.
A Google shows the channels are:

3: TMDS Data 2-
9: TMDS Data 0-
12: TMDS Clock -
15: SCL (I2C Serial Clock for DDC)

I can't find any documentation online about these pins being commonly unused. The cables are all different ages from different manufacturers. So I'm wondering what I'm missing. It's too big a coincidence for these pins to have all failed on their own. Does anyone know what is going on here?

I went so far as to do a breakdown of the cable checker. (I might have to add the other pictures of its PCB in another post)
I'm not 100% sure how it works. Given there is an unmarked IC (SAM8458) which I can't find any info on. Could anyone guess why it uses a Schottky Diode / led combination and only checks the pins one by one? In any case, I doubt the cable checker is at fault.

Is it possible that the pins marked on the cable checker are misordered?
If they are reversed:

 3 --- > 17. DDC/CEC Ground
 9 --- > 11. TMDS Clock Shield
12 --- > 8. TMDS Data0 Shield
15 --- > 5. TMDS Data1 Shield

I could understand the cable operating without the shields connected I suppose. Has anyone come across anything like this before?
 

Offline FinalCatchTopic starter

  • Newbie
  • Posts: 5
  • Country: gb
Other pcb screen shots.
 

Offline FinalCatchTopic starter

  • Newbie
  • Posts: 5
  • Country: gb
last one..
« Last Edit: June 09, 2024, 06:48:29 pm by FinalCatch »
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 3610
  • Country: ua
Some time ago, I bought this HDMI to DVI cable from aliexpress:
https://www.aliexpress.com/item/1005005644920167.html

It shows the picture and can read EDID, but for some unknown reason the display starts to show "Check signal cable" error when computer is going into sleep mode.

I tried different displays and tried different HDMI sources (several PC, tablet, etc), this issue is present with any display and any HDMI source.

When I use different HDMI to DVI cable it works as expected and when PC is going into sleep mode the display also going to sleep with no error message.

So, I'm interesting what is wrong with this broken cable from aliexpress? Is it missing some required wire?

 

Online wraper

  • Supporter
  • ****
  • Posts: 17350
  • Country: lv
If it's all the same pins on multiple cables it could be that tester has crappy connector. As of half of HDMI differential pair missing, sometimes it may work, although unstable.
 

Online wraper

  • Supporter
  • ****
  • Posts: 17350
  • Country: lv
So, I'm interesting what is wrong with this broken cable from aliexpress? Is it missing some required wire?
That it's a substandard garbage. Never buy these from China. I had DP cables where they did not connect the outer shield to connectors although it was present in the cable. And run an extra wire from pin 20 that shouldn't be there according to the standard as it's used for powering adaptors from both sink and source sides. As the result this garbage was back powering GPU from the monitor while PC was off and causing some issue, don't recall what exactly. And image disappearing while plugging devices into mains sockets due to unconnected shield.
 

Offline FinalCatchTopic starter

  • Newbie
  • Posts: 5
  • Country: gb


I spent some time tracing paths on the tester PCB to confirm which light corresponds to which pin on the HDMI cable. Assuming I'm correct in thinking this pin out from Wikipedia https://en.wikipedia.org/wiki/HDMI labelled "HDMI type A receptacle" is what I colloquially call a "female" connector (pictured below) then the pin out becomes:



Tester 10 --> Pin 1   TMDS Data2+
Tester 09 --> Pin 2   TMDS Data2 Shield
Tester 11 --> Pin 3   TMDS Data2−
Tester 01 --> Pin 4   TMDS Data1+
Tester 12 --> Pin 5   TMDS Data1 Shield
Tester 02 --> Pin 6   TMDS Data1−
Tester 13 --> Pin 7   TMDS Data0+
Tester 03 --> Pin 8   TMDS Data0 Shield
Tester 14 --> Pin 9   TMDS Data0−
Tester 04 --> Pin 10   TMDS Clock+
Tester 15 --> Pin 11   TMDS Clock Shield
Tester 05 --> Pin 12   TMDS Clock−
Tester 16 --> Pin 13   Consumer Electronics Control (CEC)
Tester 06 --> Pin 14   Reserved (HDMI 1.0–1.3a) Utility/HEAC+ (HDMI 1.4+, optional, HDMI Ethernet Channel (HEC) and Audio Return Channel (ARC))
Tester 17 --> Pin 15   SCL (I2C serial clock for DDC)
Tester 07 --> Pin 16   SDA (I2C serial data for DDC)
Tester 18 --> Pin 17   Ground (for DDC, CEC, ARC, and HEC)
Tester 08 --> Pin 18   +5 V (up to 50 mA)
Tester 19 --> Pin 19   Hot Plug Detect (all versions)  HEAC− (HDMI 1.4+, optional, HDMI Ethernet Channel and Audio Return Channel)
TesterGND --> Outershield

It looks like the numbers on the tester don't correspond to pin numbers but rather what was convenient for the PCB designer.
A pattern now emerges of just the Shields being disconnected (3,9,12,15 --> 2,5,8,11). (16 CEC  was missing on another cable)
I don't understand/remember enough about twisted pair signal transmission to know what implication not having the shields for each pair. However, I can imagine the cable working over short distances and low speeds. Maybe?

A look at HDMI's website https://www.hdmi.org/resource/cables suggests to me that what I may be looking at is an old cable that corresponds to the old "Standard HMDI Cable" standard. Something like HDMI 1.0 which required only `3.96 Gbit/s of video bandwidth (1920 × 1080 or 1920 × 1200 at 60 Hz) and 8-channel LPCM/192 kHz/24-bit audio`



Alternatively, after shifting through reems of marketing BS™ websites I found this YouTuber who has tested a few cables recently



, and one of the suppositions they had was that "bundled cables", those that come with devices etc, can only have the pins/performance that match the features of the device they are bundled with, as there is no requirement for them to work with anything else.
(Incidentally, if anyone can explain to me what the x and y axes are in the "eye diagrams" produced by their tester in the video (example attached below) I would be grateful!)
So possibly whatever device these cables came from they didn't need the shield pins connected?

Some time ago, I bought this HDMI to DVI cable from aliexpress:
https://www.aliexpress.com/item/1005005644920167.html

It shows the picture and can read EDID, but for some unknown reason the display starts to show "Check signal cable" error when computer is going into sleep mode.

I tried different displays and tried different HDMI sources (several PC, tablet, etc), this issue is present with any display and any HDMI source.

When I use different HDMI to DVI cable it works as expected and when PC is going into sleep mode the display also going to sleep with no error message.

So, I'm interesting what is wrong with this broken cable from aliexpress? Is it missing some required wire?



@radiolisterner I think it's entirely possible your cable is missing some connected pins. Maybe pin 13 - CEC if this is the one responsible for telling the monitor to sleep. But bear in mind I know nothing.

If it's all the same pins on multiple cables it could be that tester has crappy connector. As of half of HDMI differential pair missing, sometimes it may work, although unstable.

@wraper I think the tester's connectors are ok. I've confirmed that multiple "good" cables, multiple times, light up all the tester's LEDs. I've also confirmed that a "bad" cable doesn't have pins 2,5,8, and 11 connected. I did this using my BM789 DMM and inserting stripped 30AWG wire into each of the cable's pins in turn. My DMM returns the same results as the tester. (at least after I mapped which tester light corresponds to which HDMI pin as shown above).


« Last Edit: June 10, 2024, 07:43:22 am by FinalCatch »
 

Online wraper

  • Supporter
  • ****
  • Posts: 17350
  • Country: lv
@wraper I think the tester's connectors are ok. I've confirmed that multiple "good" cables, multiple times, light up all the tester's LEDs. I've also confirmed that a "bad" cable doesn't have pins 2,5,8, and 11 connected. I did this using my BM789 DMM and inserting stripped 30AWG wire into each of the cable's pins in turn. My DMM returns the same results as the tester. (at least after I mapped which tester light corresponds to which HDMI pin as shown above).
As you figured out there is pin number mismatch, it's now obvious those are missing shields. It's still possible they could be there but connected to the shell. I have a thought those must be shields but then I read you wrote those were data pins.
 

Offline FinalCatchTopic starter

  • Newbie
  • Posts: 5
  • Country: gb
@wraper I think the tester's connectors are ok. I've confirmed that multiple "good" cables, multiple times, light up all the tester's LEDs. I've also confirmed that a "bad" cable doesn't have pins 2,5,8, and 11 connected. I did this using my BM789 DMM and inserting stripped 30AWG wire into each of the cable's pins in turn. My DMM returns the same results as the tester. (at least after I mapped which tester light corresponds to which HDMI pin as shown above).
As you figured out there is pin number mismatch, it's now obvious those are missing shields. It's still possible they could be there but connected to the shell. I have a thought those must be shields but then I read you wrote those were data pins.

I just checked with a DMM and 2,5,8, and 11 are not connected to the shell (outer shield). -Though the shells of each cable end are connected.
 

Online wraper

  • Supporter
  • ****
  • Posts: 17350
  • Country: lv
@wraper I think the tester's connectors are ok. I've confirmed that multiple "good" cables, multiple times, light up all the tester's LEDs. I've also confirmed that a "bad" cable doesn't have pins 2,5,8, and 11 connected. I did this using my BM789 DMM and inserting stripped 30AWG wire into each of the cable's pins in turn. My DMM returns the same results as the tester. (at least after I mapped which tester light corresponds to which HDMI pin as shown above).
As you figured out there is pin number mismatch, it's now obvious those are missing shields. It's still possible they could be there but connected to the shell. I have a thought those must be shields but then I read you wrote those were data pins.

I just checked with a DMM and 2,5,8, and 11 are not connected to the shell (outer shield). -Though the shells of each cable end are connected.
I meant individual twisted pair shields, not pins.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf