Hi Guys, When you first threw the idea out there (part 1) , my head went to an Arduino driving three cascaded 74HC595 feeding the pins of one HDMI plug and three 74HC165 reading the pins on the other plug back into the Arduino. From the programming side, three 8-bit integers sent out would be bit-wise compared to the three 8-bit integers returned. Do that several times and the test is done. I think you are doing basically the same, although your chips have a few more connections. Cheers, EJ
Hi Ellie - yeah, that's basically what we'ew going to do. my choice of the MCP23017 gives me more options: I can do a measuement against Pullup Resistors and every pin can go tri-state. And I love the I2C multiplexer :) It will be interesting, if I overthought the whole shebang - waiting for real hardware now!
Gr8 video… the hdmi port tester you chatting about making sounds like a CMIZapper HDMI tester allows you to quickly test HDMI circuits on the fly. Whether you're working on a games console, TV, DVD player, laptop or something completely random, The HDMI tester gives you a visual representation of your HDMI circuit to ensure functionality by lighting up LEDs to represent different signals and lines including The power LED which will light up red when you have 5v connected and present. If you guys make an all in one board that can check cables and ports id buy 👍 anyway keep up the gr8 work n video.
Hi MostGearZer0Idea - don't get your hopes up too high: as I said in the video, for me, this is a project I don't have any use for. If you wanna do this right, you have to do this with actual differential signals (and the according frequencies); I'd guess, to get this rolling this is 4 weeks development. I'm a controller-guy - I have basic understanding how LVDS and stuff works, but would need a deeeeep dive into a shipload of tec - quite frankly: nope. And even then this would be build on textbook knowledge - no real live experience. So - expect more a basic tool; I hate to disappoint, but currently I'd aim for that one.
Hi Peter - yeah, the longer we talked about that, the more sense it makes to do the deep dives beside the project videos. We're probably try this out, when we got the hardware - on Rich's Channel will be the project, and over at mine I'll geek out over hardware :)
Hi Karthor - yeah, didn't see that one comming, but while we were doing the planning, this seems the right thing to do. We probably couple the projects with deep dives on our two channels - that way, you can decide if you want the nitty gritty stuff.
Heya, this electronics is over my head (for now, still like to learn more of it) but I still like to see this and then the 2 of you working to gether and olso the people in the chat giving there oppinion. that a real commonity
I'd like to suggest an addition to your pcb layout that would provide better oscilloscope probing... If you put a ground via next to each high-speed signal, you could use a "spring-type" ground on your probes....you should have those accessories with the probe, a little "spring with a leg", which dramatically improves the response, and reduces apparent noise on many signals, especially SMPS, but certainly LVDS...with the 'normal' scope 'ground leads', you get a strong resonance peak around 100-130 MHz. (I have many years experience as a scope AE for Tek and R&S)
Yeah. Still using a giant GND loop for probing and even no GND lead on the second would be my biggest complaint. Probing the I2C lines for the EDID handshake would make more sense here with a slow scope. Using AC coupling leads to some strange effects here, too.
I've been razzing electronics RUclipsrs about terrible grounding of scope probes for some time now. The probe grounding in this video isn't even adequate for signals in the low megathertz range, much less those in HDMI. Depending on the screen resolution the ground lead is multiple pixels long. FET probes help if you are stuck with long ground connections, but a pair of good FET probes will certainly cost much more than the oscilloscope being used in the video. In any event, the scope is almost certainly grossly inadequate for HDMI signals. When I was doing SMPS design and had top-side ground plane I'd sprinkle small void dots in the solder mask around so I could "conveniently" (not that it is EVER what you'd really call convenient) probe tip grounders. A small mound of solder with an indent made with a very fine awl helped to keep the ground from slopping around. Sometimes I would use contacts from precision machined IC sockets for a really secure ground, though they worked better where I could put ground vias.I couldn't even begin to count the hours I've spent trying to figure out if features I was seeing on waveforms were real or artifacts. It is testament to the resilience, for lack of a better term, of HDMI that such a crude breakout with nothing even remotely resembling impedance matching or delay matching can be stuck into the path and an intact image is still produced.
Hi DigitalToyz3629 - thanks! Yeah, this will be a thing. We probably start with accompaning videos to complete projects, so you guys can decide if you wanna watch the gritty details over on my channel of be OK wirh the project.
I'm a tight Yorkshire man, I like getting my monies worth. I told Det 😉. Look between 2:24 and 3:15 of the video, look top right quarter of the screen you will see it says $5.00 for 10 pcs 1-2 layer. When you get to the order details screen , it defaults to 5. Change it to 10. I might have added jumpers on the board to link the nano I2C to the In and Out I2C, to listen-in on the communication (it's the only thing I know anything about, haha) I don't know if this the case with HDMI, it's just a thought. If moving on to testing live cables, frankly I may be talking out the wrong end, haha. But could some systems using differential pair, use one of the signals negative to GND? Not good good for a microcontroller. If both the signals are positive and are relatively positive and negative to each other, then that is fine.
Regarding graphics cards: an HDMI tester would be excellent but I would like to see it combined with a riser so that it could be analyzed without having to plug it into a main board each time. And if you really want to get fancy, add a power supply input which can be quickly powered by a bench supply at a single point. Basically making it independent from any main board.
I was just thinking about doing something similar. I need 20 pin or 10 pair cable tracing work, so will have have a transmitter and a receiver module. Hopefully your device will give me a insight into how to make this work. :)
Add a constant current source, then you can measure the voltage using the analog input to measure the resistance of the conductor being tested. Also make the board generic so that it can test different types of cables such as yo olde parallel cables.
A program on a pc that sends clean sine/square wave signals to each of the outputs so you can do a quick check on each output with an oscilloscope maybe of benefit… possibly
Hi Jonathan - this could be done with a little function gen (or a MCU with a DAC) - we'll see along the way if it makes sense; we may need analog measurements anyway when we get into functional analysis of a graphics card. Yeah - some way to go :)
A go / no go LED system is theoretically possible though you would require buffers for the lanes and comparators as a minimum (to avoid loading the lines and to cater for the differential signals and offset voltages). LEDS would be mostly always on as the signals are operating at very high bit rates. You would need some form of clocked gate / latch to produce flashing LEDs. An interesting project, though off-the-shelf HDMI analysers are available (albeit not cheap). I am, of course, guessing what you require and may be totally wrong (usually the case).
Hi Anthony - I'd say I'll get away with the NANO; never really liked the RP2040 stuff. If something smaller, I'd go for a Seeedstudio Xiao, but since everybody has some NANOs lying around, it's probably the build-friendliest solution.
You can NOT check the hispeed lines with an ordinary oscilloscope!!! You need a special differential probe with a very fast scope 2-3 GHz bandwidth. The high speed lines are balance 50 ohms you should keep the PCB traces at 50 ohms. Not the easy stuff that you show us.
Yup. A scope adequate for HDMI is a *very* expensive item. ANY probe adequate for the job will cost more than the scopes owned by most repair shops and hobbyists. You could detect "stuck" signals with a scope of moderate cost, but that's about it.
@@andymouse I agree. About the only thing they tell you is that signals are present. You most certainly would not expect huge differences in amplitude if you had an oscilloscope and probes actually suitable for looking at the signals.
Hi Guys,
When you first threw the idea out there (part 1) , my head went to an Arduino driving three cascaded 74HC595 feeding the pins of one HDMI plug and three 74HC165 reading the pins on the other plug back into the Arduino. From the programming side, three 8-bit integers sent out would be bit-wise compared to the three 8-bit integers returned. Do that several times and the test is done.
I think you are doing basically the same, although your chips have a few more connections.
Cheers,
EJ
Hi Ellie - yeah, that's basically what we'ew going to do. my choice of the MCP23017 gives me more options: I can do a measuement against Pullup Resistors and every pin can go tri-state. And I love the I2C multiplexer :) It will be interesting, if I overthought the whole shebang - waiting for real hardware now!
Gr8 video… the hdmi port tester you chatting about making sounds like a CMIZapper HDMI tester allows you to quickly test HDMI circuits on the fly. Whether you're working on a games console, TV, DVD player, laptop or something completely random,
The HDMI tester gives you a visual representation of your HDMI circuit to ensure functionality by lighting up LEDs to represent different signals and lines including
The power LED which will light up red when you have 5v connected and present. If you guys make an all in one board that can check cables and ports id buy 👍 anyway keep up the gr8 work n video.
Hi MostGearZer0Idea - don't get your hopes up too high: as I said in the video, for me, this is a project I don't have any use for. If you wanna do this right, you have to do this with actual differential signals (and the according frequencies); I'd guess, to get this rolling this is 4 weeks development. I'm a controller-guy - I have basic understanding how LVDS and stuff works, but would need a deeeeep dive into a shipload of tec - quite frankly: nope. And even then this would be build on textbook knowledge - no real live experience.
So - expect more a basic tool; I hate to disappoint, but currently I'd aim for that one.
A 'strange devices explained' session would be a great idea 🤓👍
Hi Peter - yeah, the longer we talked about that, the more sense it makes to do the deep dives beside the project videos. We're probably try this out, when we got the hardware - on Rich's Channel will be the project, and over at mine I'll geek out over hardware :)
strange devices explained sounds like a good idea
Hi Karthor - yeah, didn't see that one comming, but while we were doing the planning, this seems the right thing to do.
We probably couple the projects with deep dives on our two channels - that way, you can decide if you want the nitty gritty stuff.
Heya, this electronics is over my head (for now, still like to learn more of it) but I still like to see this and then the 2 of you working to gether and olso the people in the chat giving there oppinion. that a real commonity
I'd like to suggest an addition to your pcb layout that would provide better oscilloscope probing...
If you put a ground via next to each high-speed signal, you could use a "spring-type" ground on your probes....you should have those accessories with the probe, a little "spring with a leg", which dramatically improves the response, and reduces apparent noise on many signals, especially SMPS, but certainly LVDS...with the 'normal' scope 'ground leads', you get a strong resonance peak around 100-130 MHz. (I have many years experience as a scope AE for Tek and R&S)
Yeah. Still using a giant GND loop for probing and even no GND lead on the second would be my biggest complaint. Probing the I2C lines for the EDID handshake would make more sense here with a slow scope.
Using AC coupling leads to some strange effects here, too.
I've been razzing electronics RUclipsrs about terrible grounding of scope probes for some time now. The probe grounding in this video isn't even adequate for signals in the low megathertz range, much less those in HDMI. Depending on the screen resolution the ground lead is multiple pixels long.
FET probes help if you are stuck with long ground connections, but a pair of good FET probes will certainly cost much more than the oscilloscope being used in the video. In any event, the scope is almost certainly grossly inadequate for HDMI signals.
When I was doing SMPS design and had top-side ground plane I'd sprinkle small void dots in the solder mask around so I could "conveniently" (not that it is EVER what you'd really call convenient) probe tip grounders. A small mound of solder with an indent made with a very fine awl helped to keep the ground from slopping around. Sometimes I would use contacts from precision machined IC sockets for a really secure ground, though they worked better where I could put ground vias.I couldn't even begin to count the hours I've spent trying to figure out if features I was seeing on waveforms were real or artifacts.
It is testament to the resilience, for lack of a better term, of HDMI that such a crude breakout with nothing even remotely resembling impedance matching or delay matching can be stuck into the path and an intact image is still produced.
👍 to the strange devices explainer
Hi DigitalToyz3629 - thanks! Yeah, this will be a thing.
We probably start with accompaning videos to complete projects, so you guys can decide if you wanna watch the gritty details over on my channel of be OK wirh the project.
I'm a tight Yorkshire man, I like getting my monies worth.
I told Det 😉. Look between 2:24 and 3:15 of the video, look top right quarter of the screen you will see it says $5.00 for 10 pcs 1-2 layer.
When you get to the order details screen , it defaults to 5. Change it to 10.
I might have added jumpers on the board to link the nano I2C to the In and Out I2C, to listen-in on the communication (it's the only thing I know anything about, haha)
I don't know if this the case with HDMI, it's just a thought. If moving on to testing live cables, frankly I may be talking out the wrong end, haha. But could some systems using differential pair, use one of the signals negative to GND? Not good good for a microcontroller. If both the signals are positive and are relatively positive and negative to each other, then that is fine.
Regarding graphics cards: an HDMI tester would be excellent but I would like to see it combined with a riser so that it could be analyzed without having to plug it into a main board each time. And if you really want to get fancy, add a power supply input which can be quickly powered by a bench supply at a single point. Basically making it independent from any main board.
I was just thinking about doing something similar. I need 20 pin or 10 pair cable tracing work, so will have have a transmitter and a receiver module. Hopefully your device will give me a insight into how to make this work. :)
Add a constant current source, then you can measure the voltage using the analog input to measure the resistance of the conductor being tested. Also make the board generic so that it can test different types of cables such as yo olde parallel cables.
Excelente content my friends! It is 5 months already, are you guys planning to publish the Part 3? :)
A program on a pc that sends clean sine/square wave signals to each of the outputs so you can do a quick check on each output with an oscilloscope maybe of benefit… possibly
Hi Jonathan - this could be done with a little function gen (or a MCU with a DAC) - we'll see along the way if it makes sense; we may need analog measurements anyway when we get into functional analysis of a graphics card. Yeah - some way to go :)
@@DetBuildsStuff great project. Looking forward to getting them when released
Thanks great project, mis LED in the system, over each line. Is that posibel? Thanks
Anything is possible, can you explain in more detail what you meant
A go / no go LED system is theoretically possible though you would require buffers for the lanes and comparators as a minimum (to avoid loading the lines and to cater for the differential signals and offset voltages). LEDS would be mostly always on as the signals are operating at very high bit rates. You would need some form of clocked gate / latch to produce flashing LEDs. An interesting project, though off-the-shelf HDMI analysers are available (albeit not cheap). I am, of course, guessing what you require and may be totally wrong (usually the case).
Arduino or Raspberry pi?
Both have similar functions.
Hi Anthony - I'd say I'll get away with the NANO; never really liked the RP2040 stuff. If something smaller, I'd go for a Seeedstudio Xiao, but since everybody has some NANOs lying around, it's probably the build-friendliest solution.
Are you designing pcb's for manufacturing at PCBway with the software from JLC?😅
Oh no we would never do that 😉
Cmi zapper has already hdmi port tester.
Binge watching LER… again lol
Interesting..
You can NOT check the hispeed lines with an ordinary oscilloscope!!! You need a special differential probe with a very fast scope 2-3 GHz bandwidth. The high speed lines are balance 50 ohms you should keep the PCB traces at 50 ohms. Not the easy stuff that you show us.
Yup. A scope adequate for HDMI is a *very* expensive item. ANY probe adequate for the job will cost more than the scopes owned by most repair shops and hobbyists. You could detect "stuck" signals with a scope of moderate cost, but that's about it.
I think the wiggly lines we see are of no real use..IMHO.
@@andymouse
I agree. About the only thing they tell you is that signals are present. You most certainly would not expect huge differences in amplitude if you had an oscilloscope and probes actually suitable for looking at the signals.