to create a device with HDMI out i need a $50k scope?

[rambling@midnight]

does the bandwidth of the scope have to match the speed of what i am trying to measure? (sampling speed isn't enough?)
how does anyone create a hdmi-out device without an insanely expensive scope?
Can it be done without observing the signal to make sure you have it right/debug?
most of my project ideas have HDMI-out in them.

[tv84]

Are you developing the HDMI chip? Or are you using an off-the-shelf for HDMI output?

[rambling@midnight]

Are you developing the HDMI chip? Or are you using an off-the-shelf for HDMI output?

I'm trying to output HDMI video with am MCU. If there's a way to do that with a stable premade unit that i can pass data off to, and don't have to debug, that'd be cool.

[exe]

I'd expect the datasheet to explain this. Particularly pcb layout, additional components, impedance matching where it is needed, etc. Chances are, it will "just work".

[tggzzz]

does the bandwidth of the scope have to match the speed of what i am trying to measure? (sampling speed isn't enough?)
how does anyone create a hdmi-out device without an insanely expensive scope?
Can it be done without observing the signal to make sure you have it right/debug?
most of my project ideas have HDMI-out in them.

You need to ensure adequate signal integrity[1], and then it is probably sufficient to flip into the digital domain with logic analysers or printf statements.

Scope bandwidth has to exceed that of your signal. Don't forget that the probe and probing technique are part of the measurement system.

Sampling rate is irrelevant, and for signal integrity purposes can be arbitrarily low.

Make sure you understand how to capture and interpret eye diagrams.

[1] by design, or by test, or by using a pre-built module

[noreply]

how does anyone create a hdmi-out device without an insanely expensive scope?

Not sure exactly what you are asking ... but the 'acquisition and signal processing' aspects of what you are analysing with your 'scope' are separate to the 'display' electronics.

Think of the display as being an artist's canvas (it can be of various size, shape and quality) - the ultimate image you see (scope output) is a function of how well the artist paints the canvas with their 'data' which is in the artist's head, with the scope - it’s the same but the data is 'painted' into an appropriate 'display subsystem' by the acquisition and signal processing interface which already knows what should be displayed (painted) onto the LCD (& HDMI output if appropriate) canvas.

So, from HDMI point of view - in theory, its totally detached from the data that it is displaying
- in the sense that it does not need to be involved in the 'acquisition and signal processing' aspects

[Someone]

So for those with no experience or exposure to HDMI, they make it pretty clear:
https://www.hdmi.org/resource/testing
Its complicated enough (and uses such specialised equipment) that you can buy that sort of testing as a complete service just like EMC.

Like USB, you can ignore all that and push a product out into the market without their "seal of approval" and costs, but you can't use their branding to advertise your capabilities.

[Someone]

how does anyone create a hdmi-out device without an insanely expensive scope?

Not sure exactly what you are asking ...

Not exactly sure why you are posting....
https://www.keysight.com/us/en/assets/7018-06444/flyers/5992-3556.pdf

As transmission speeds of HDMI standard continue to increase, validation engineers need a measurement solution with high bandwidth and <25 picosecond rise-time specifications

For bleeding edge data rates 25ps rise time scope + differential probing is not cheap as the OP will soon discover. Even for mid/low end 1080p signal you need an expensive scope as just one of the tools on the long list.

[Mechatrommer]

Are you developing the HDMI chip? Or are you using an off-the-shelf for HDMI output?

I'm trying to output HDMI video with am MCU. If there's a way to do that with a stable premade unit that i can pass data off to, and don't have to debug, that'd be cool.

there are specialized IC for that you can google... https://www.google.com/search?q=HDMI+IC i cant name specifically they all have different functionality, IO and features. the easiest route is buy a product closest to what you want as a sample to study and start from there. if you want the hard way, you'll need 5-10X the data rate, so for example designing 10Gbps data, you'll need 50-100GHz scope and they are not 50K, maybe you can get them at the price in used market.

[Kean]

Are you developing the HDMI chip? Or are you using an off-the-shelf for HDMI output?

I'm trying to output HDMI video with am MCU. If there's a way to do that with a stable premade unit that i can pass data off to, and don't have to debug, that'd be cool.

No idea what resolution you are aiming for, and if you are just wanting to display static images, sprites for retro games, an HMI, or FMV.
I'd suggest you start looking at what others have done.  Generally you'll want to use an FPGA, or maybe a dedicated SoC with GPU/video out.

Some results from 30 seconds of Google...
https://electronics.stackexchange.com/questions/189649/why-are-fpgas-used-so-often-for-hdmi-video-projects
https://www.eevblog.com/forum/beginners/outputting-to-hdmi-with-a-low-power-mcu/
https://www.reddit.com/r/microcontrollers/comments/7ubbom/whats_the_smallestcheapest_hdmi_capable/
https://tinyletter.com/jamesbowman/letters/dazzler-promo-video-first-look
https://www.electronics-lab.com/arduino-hdmi-shield-bridging-gap-small-mcus-full-hd-monitors/

Once you have chosen a solution and have something that seems to work and drives a display, you will definitely need to actually verify signal integrity and pass EMC testing before you can sell it.  You can pay someone to run those tests for you if you don't want to invest a lot of money in the equipment yourself.

[Kean]

Oh, and once you've chosen an mcu/fpga/SoC/hdmi transmitter you will want to carefully read any datasheets, app notes, and reference design docs - and unless you know better you should follow the PCB layout recommendations to the letter.

[edigi]

most of my project ideas have HDMI-out in them.

Most of the FPGA cards (I mean modern ones) have HDMI output, Raspberry Pi has HDMI output as well. Why would you start the hard way? Simplest/quickest is Raspberry Pi, then FPGA and as last would come to my mind starting from scratch with some HDMI chip.

[nctnico]

I'd expect the datasheet to explain this. Particularly pcb layout, additional components, impedance matching where it is needed, etc. Chances are, it will "just work".

That is the best way. With most chips which have an HDMI output there will be a manual on how to route the traces up to doing the layout around the connector. If you stick to the recipy then you are copying a tried & tested design. NVidia has great documentation together with their Jetson TX2 modules for example.

However it is easier said than done because you still need to be aware of trace impedances (both single ended and differential) and do length and phase matching. Get the PCB design right and it is all up to the chip to drive the HDMI signals properly.

[jemangedeslolos]

I have a friend who works only on video ( HDMI, HDSDI, 3GSDI, 6GSDI etc etc ).
He works with an old tektronik TDS3000 but he has tons of experience so it helps.

There are design guides on this kind of component and/or reference design so we don't start from scratch.
I personally designed an SDI to fiber optic video converter without problem thanks to his experience and the documentation available.

The little tip he has is that he goes to exhibitions events to test his design on the dedicated Tektronix devices exposed on the stands.

If you really want to have a device in your lab, there are devices dedicated to video signal analysis or professional broadcast monitor with embedded signal analysis which can already help and which do not cost 50k$

[Elasia]

+1 for reference designs, anyone worth their salt will have everything down to pcb layout rules and structures so it just works with minimal hassle for you to use when it comes to these and other high speed buses

jemangedeslolos... that's just sneaky.. i love it lol

[tggzzz]

The little tip he has is that he goes to exhibitions events to test his design on the dedicated Tektronix devices exposed on the stands.

Last time I tried that the Tek sales person said it was OK - until his boss came over and prevented it.

The LeCroy sales people were more knowlegeable and accomodating...

[Mechatrommer]

jemangedeslolos... that's just sneaky.. i love it lol

its called "social engineering". you can get a legitimate permission by saying that "oh i want to test your device capability, my equipment back home showed some excessive ringing, lets see if your device can show it correctly". you can then pretend you dont know how to use the new device, they can then even show you all the bells and whistles (this is the whole point of the booth afterall). the conversation/session should end up you complementing their device by saying how good and how prospective it is, you will keep this in to buy list (even if in fact you really broke and/or no plan to buy at all). if you have other viewers around you can help the promoters to prove why they should also buy it (play symbion) you also have some friends and companies you can pass the words and experience to them.. everybody will walk happy.

[exe]

I just come to both, say upfront that I'm not an EE, and just want to play with equipment. Most of the time they still help me . Is it possible that both people have a rule to help everyone? Like, the boss wants employees to be busy or something.