MIPI woes, good design ideas and headaches


EvilDragon

Administrator
Staff member
Joined
Mar 4, 2003
Messages
29,986
Age
46
Location
Ingolstadt
Quite a lot has happened during the last week!
 
Nikolaus (gta04 at the boards) was mostly busy dealing with MIPI.
 
1. LCDs can drive you crazy!
 
MIPI is the standard interface LCDs are using these days - and while "standard" seems to imply it's something easy to use, it sure isn't.
"Standard" is only the communication protocol - getting the LCD to do something is something else.
You need to calculate and set size, timings, etc. everything yourself...
 
After a lot of reading into MIPI and how it generally works and many many patches later, he still didn't get the LCD to show a picture - but at least he can communicate with it now.
 
What's missing now is the startup sequence to tell the LCD to actually do something.
That's something the manufacturer should be able to tell us.
More information on that soon.
 
 
2. A good idea: What we showed at FOSDEM
 
You probably have all seen the small video at the FOSDEM where I showed off what you could've seen and tested there.
If not, here it is:
 

https://www.youtube.com/embed/ruFyq7fbfpU?feature=oembed
 
Designing a PCB takes a lot of time.
Not just for the design, but for debugging as well - and what's worse: You spend a lot of time simply WAITING for the prototype PCBs.

A PCB like we need it usually needs 8 layers.
The normal production time for an 8 layer bare PCB is 6 weeks - and it's pretty expensive as well (especially for just doing prototypes).
 
Aditionally, populating them with all of the complex parts as well (like the OMAP) costs a lot of money. Both for parts AND the population.
 
That's why we're doing a different approach here.
The Pyra PCB right now is a simple two layer-PCB. Can be created in a couple of days for little money.
 
All of our custom parts (keyboard controller, LED controller, nubs, switches, UMTS module, etc.) are populated on the PCB.
However, the complex part (the SoC) is still on the devboard.
 
The devboard and the Pyra PCB are connected via a special connector Nikolaus created, which includes an i²c bus, gpios, etc.
So basically, our custom parts are connected the exact same way as they will be connected on the final PCB to the SoC.
 
That's pretty helpful, as we can now use low-cost Pyra PCBs we can get within a couple of days to test all the hardware we want to include.
And once everything works as it should, "only" the SoC with components like RAM, power circuits, storage, etc. needs to be transplanted to the Pyra PCB - which will then be the first REAL prototype.
"Only" sounds like it's a quick thing to do - but in reality, this is A LOT of work that will need more time than the rest of the design!
 
That way still saves a lot of time and money though.
 
In case you're wondering why this hasn't been done with the Pandora back then:
Well, Michael designed the first prototype when no devboard was available (the BeagleBoard came out after that...) - an impressive amount of work!
 
I'm happy it's more easy this time though!
 
 
3. Headaches
 
One thing that's good and bad about the above mentioned way of designing the PCB:
The SoC comes last - so we are not yet fixed on the OMAP5 (we were, but the last week has shown different options as well...)
 
Thanks to the blogpost of PandoraLive, we got in contact with Qualcomm.
It could very well be that we can get a Snapdragon!
 
However, there's also the new Intel BayTrail SoC that some people here are in favour.
And Allwinner has announced the A80 which could be interesting...
 
There's been A LOT of discussions after my last Blog Post regarding which SoC to choose...
 
This is giving me a lot of headaches right now, as there are new options - and finding out which one is better than the other is not an easy thing.
 
While it will be me having to do the final decision, here is some poll which might be helpful to me!
 
Last edited:
I've got a bottle of extra strength Tylenol for that headache :D

Good updates. I'm sure that you'll make the right decision.
 
The video n°2 (nice carpet btw) is both exciting and frustrating because it's both a Pyra and not a Pyra ^^.
 
It's possible to get a panel of boards 8 layers or more manufactured in under a week, or at least the company I work for has done it in really time critical situations.. but it is of course much more expensive.
 
Wow - that must be really expensive, as I haven't heard of that before :eek:

Probably not every company offers that...
 
the board looks really really sexy! good job gta04! ; )

I've never heard of such an approach. It's very interesting! ED its cool that you share the processs of development *thumbs up*
 
Great video thanks.

That is clever stuff, the way of testing the SoC like that.

I was going to request that the LED Christmas tree boot up sequence be optional, as I would quite fancy that myself... then you followed up stating that the LED would be configurable in user space. So I guess I can have it?

Hopefully people will create and share nice little sequences using those lights :)

Anyway, good video, I love these type of vids from ED
 
I have seen it when looking at PCB plants in the EU. It costs thereafter.

EDIT: http://www.leiton.de/en-kalkulation-preis.html#ergebnis for example, €700 for one 8-layer pandora size PCB, 2 workday leadtime to shipping :p
Okay, that's more than we paid for 4 of our PCBs :)

I've never heard of such an approach. It's very interesting! ED its cool that you share the processs of development *thumbs up*
That's the idea of the Pandora and Pyra :D

I was going to request that the LED Christmas tree boot up sequence be optional, as I would quite fancy that myself... then you followed up stating that the LED would be configurable in user space. So I guess I can have it?

Hopefully people will create and share nice little sequences using those lights :)
It works the same as in the Pandora - you can set the events and brightness for each of the LEDs.
 
That is a very interesting approach. I have a... well lets call him a friend... that was in a position to do a similar type of design where the SoC was going to be designed as interchangeable with a family of SoC's (from a Chinese manufacturer). He has pretty much abandoned that idea for that of a dedicated PCB for one specific chipset. We also had talks in creating a new type of standard connection interface and market that as well so other manufacturers could use the idea and see SoC's being close to the type of upgrade you would see on a processor in your desktop. I'm not sure if anything will ever come of his plans anymore as he's more or less broken off contact with the rest of the world. You might know of him but I'm not going to name names. ;)

Point is, it's a great idea considering drop in replacements and reducing development times. Especially with someone like yourself that has a incredibly solid idea and near perfect implementation. This in it's self will future proof you, but I guess you already know that ;)
 
Last edited by a moderator:
Nice video! Good to see some actual development progress :)
 

We also had talks in creating a new type of standard connection interface and market that as well so other manufacturers could use the idea and see SoC's being close to the type of upgrade you would see on a processor in your desktop.
Sounds a lot like EOMA-68. It's currently implemented in, for example, recently released Improv devboard. Will probably be used in the Vivaldi tablet if/when it manigfests :)
 
i have to say i love your vids can listen to you all day talking about things i know very little about:) but really thanks for everything and never giving up and 3 cheers to gta04 for coming onboard can't wait to see what around the next bend
 
That is a very interesting approach. I have a... well lets call him a friend... that was in a position to do a similar type of design where the SoC was going to be designed as interchangeable with a family of SoC's (from a Chinese manufacturer). He has pretty much abandoned that idea for that of a dedicated PCB for one specific chipset.
I assume ED's doing that for development purposes, so he can test all the peripherals without the expense of a fancy board and populating the complex components, as he says in the video.  Once it's all working it should be less work to move the SoC etc. over to the real board and get it booting independently (as he says in the text), but I don't think the idea is for it to be replaceable after market - that would add cost and reliability, make it harder to deal with heat build up and generally be a non-starter for the Pyra I believe.

A couple of questions of my own regarding the video:

- The volume wheel appears to be mounted on it's own, away from the headphone port, like on the Pandora.  If that deliberate?  I'd have thought keeping as much of the audio hardware together as possible would reduce noise and interference, but I'm no audio engineer.

- You mention the LEDs are programmable.  Are the backlight LEDs also multicolour currently?  Some people seem to be keen on that.  Awesome to see them on board already though!

It's cool that you've kept the door open to evaluate any Snapdragon processor or the A80 that Exo spotted, but at the same time I hope you get this settled soon.

And thanks from me for keeping us all informed with these posts and videos.  They really inspire confidence!
 
Nice video! Good to see some actual development progress :)

We also had talks in creating a new type of standard connection interface and market that as well so other manufacturers could use the idea and see SoC's being close to the type of upgrade you would see on a processor in your desktop.
Sounds a lot like EOMA-68. It's currently implemented in, for example, recently released Improv devboard. Will probably be used in the Vivaldi tablet if/when it manigfests :)
Yes, almost exactly like that actually but for a gaming device.


Sent from my SPH-L900 using Tapatalk
 
That is a very interesting approach. I have a... well lets call him a friend... that was in a position to do a similar type of design where the SoC was going to be designed as interchangeable with a family of SoC's (from a Chinese manufacturer). He has pretty much abandoned that idea for that of a dedicated PCB for one specific chipset.
I assume ED's doing that for development purposes, so he can test all the peripherals without the expense of a fancy board and populating the complex components, as he says in the video.  Once it's all working it should be less work to move the SoC etc. over to the real board and get it booting independently (as he says in the text), but I don't think the idea is for it to be replaceable after market - that would add cost and reliability, make it harder to deal with heat build up and generally be a non-starter for the Pyra I believe.
 
Exactly.

A couple of questions of my own regarding the video:

- The volume wheel appears to be mounted on it's own, away from the headphone port, like on the Pandora.  If that deliberate?  I'd have thought keeping as much of the audio hardware together as possible would reduce noise and interference, but I'm no audio engineer.
No, that's not a problem.

We're using symmetrical audio lines (don't pick up noise).

- You mention the LEDs are programmable.  Are the backlight LEDs also multicolour currently?  Some people seem to be keen on that.  Awesome to see them on board already though!
No, the keyboard LEDs are not planned to be multicolor - we would need additional LED driver chips for that (as basically, one multicolor LED is like three LEDs at once), so it would just add costs.

The LEDs in the front row will be multicolor as will a special surprise LED. 
 
So, do those fancy lights in the keyboard area mean we'll have a backlit keyboard this time?
Or did I just miss an earlier confirmation of that and that's why no one's surprised.
 
So, do those fancy lights in the keyboard area mean we'll have a backlit keyboard this time?


Or did I just miss an earlier confirmation of that and that's why no one's surprised.
It was announced long ago..
 
Back
Top