Compute Module Test Board

[Briarios]

UPDATE v2: http://rpi.pmfg.me/ <--- for download files.

UPDATE: Version 2 http://www.raspberrypi.org/forums/viewt ... 40#p609140


Hey Guys,

I was wondering if i can get some feedback on a design i am working on (link is the schematic) i just want to confirm that i have covered the correct areas, Note i haven't connected the GPIO's, HDMI, TVDAC, or anything else. I have only covered the Bare essentials to get the CM up and running and responding via USB and Ethernet. once I get feedback that this will work, then go further with prototyping it properly

http://pmfg.me/rpi/CM_Test_Board_1.x/CM_Test_Board.pdf

This is one of many steps to creating a great idea

Thanks for your help and feedback

note: this is my first crack at doing schematics so may need to polish up on it further.

[Briarios]

Hi again guys,

I have done some small changes to the Schematic (see link Below) and have changed the board design from using right angle sodimm socket to a vertical socket, again the links are below. Please feel free to view and Comments are always helpful

Schematic
http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... .1_sch.pdf

Board view
http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... .1_brd.pdf

[Briarios]

If People are having troubles with the Links I have put in the IP address in case of dns Issues

http://103.4.235.228/rpi/CM_Test_Board.pdf
http://103.4.235.228/rpi/CM_Test_Board_1.1_brd.pdf
http://103.4.235.228/rpi/CM_Test_Board_1.1_sch.pdf

[Briarios]

Ok i just made a quick change to the test board, with this version I am going to be building Bill of Materials list and seeing on what the costs of prototyping this board will be. If anyone is interested Please feel free to let me know as I myself do not have a dev kit yet, so this board was created to be a simple version of that.

again attached are the designs for viewing.

http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... .2_sch.pdf
http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... .2_brd.pdf

[mimi123]

Have you the KiCAD of your design? Much more important than shematics.

[Briarios]

Have you the KiCAD of your design? Much more important than shematics.

No Sorry i've only used Eagle cad in this instance as most of the parts that i have found are for eagle cad, i am not too sure if eagle and kicad can use the same library.

[Briarios]

Ok a Small update i had to do as I had found my differential Pairs for both the USB and the Ethernet were off by about 20% which in the world of electronics is well... bad, So i re did the traces and placed the usb and ethernet with a 2.5% tolerance which is good. Another side note I am still deciding on a fab house to go with as well.

without further ado here is the updated board layout for your perusal

http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... .4_brd.pdf

[ceteras]

Am I the only one who gets an empty pdf at the link above?

[rpdom]

Am I the only one who gets an empty pdf at the link above?

I get a board layout which takes some time to draw but seems to be complete.

[fruitoftheloom]

Am I the only one who gets an empty pdf at the link above?

Views without issues

[ceteras]

It took 25 seconds to draw on a core i7 2600 (3.4GHz). Amazing! At first, I didn't wait that long.
The board itself is interesting.
I would put the crystal closer to the LAN9512, usually it's a requirement, to put it as close as possible. The upper usb connector could be safely moved upwards.
Are you going to publish the eagle files for this project?

[Briarios]

I would put the crystal closer to the LAN9512, usually it's a requirement, to put it as close as possible. The upper usb connector could be safely moved upwards.
Are you going to publish the eagle files for this project?

Thanks for pointing that out i think i'll do that in the next revision same with the USB Ports too .

As with the eagle files, Yes I will publish the files once I know that the prototypes work, otherwise there is no point in publishing something that is not going to work

[tsaG]

Did you keep the resistance of the USB connection in mind? It looks quite thin, so Im just assuming that you missed that out.

[Briarios]

Did you keep the resistance of the USB connection in mind?

[John Cleese voice] ahh yes, it would seem that i have indeed missed it... but i checked the traces and they are rated for about 1.6 amps across the trace, i also checked the resistance and the first usb socket has a voltage drop to roughly 4.9v at 500ma and the second socket has a voltage drop of 4.89v at 500ma, so I think this is well within spec for something like a usb flash drive. [/John Cleese voice]

[6677]

Did you keep the resistance of the USB connection in mind?

[John Cleese voice] ahh yes, it would seem that i have indeed missed it... but i checked the traces and they are rated for about 1.6 amps across the trace, i also checked the resistance and the first usb socket has a voltage drop to roughly 4.9v at 500ma and the second socket has a voltage drop of 4.89v at 500ma, so I think this is well within spec for something like a usb flash drive. [/John Cleese voice]

USB2 spec calls for 5v +/- 0.25v officially so technically you are still within the specifications. Many devices actually run on lower voltages and simply feature onboard regulation anyway and will work over a larger range than spec calls for, heard of the ralink wifi adapters function at as low as a 3.5v vcc

[Briarios]

USB2 spec calls for 5v +/- 0.25v officially so technically you are still within the specifications.

Oh yes the hardest part about this that i had to learn on the fly was getting the USB and Ethernet to meet specifications once that was done, everything else was easy.

[Briarios]

Hello Guys,

Sorry for the delay, as I had just started a new job and all. so here is possibly the last update I will make as it now all seems to be correct and the USB and Ethernet Specs have (what I think) been met. So yeah Enjoy and again thanks for the info and support

http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... nt_1.5.pdf
http://pmfg.me/rpi/CM_Test_Board_1.x/CM ... ck_1.5.pdf

[claudioalmeiida]

Hello, have you considered using the LAN9514 chip? is it possible to use with the compute module?

[Briarios]

Hello, have you considered using the LAN9514 chip? is it possible to use with the compute module?

Yes I have considered using the LAN9514 chip, but at this stage i didn't want the hassle of finding space for 2 extra USB ports on the board at this stage, and yes it is possible to use with the compute module because the CM will see it as a USB Hub.

[gsh]

You can get rid of the crystal for the 9512 and instead output a 25MHz clock from GPCLK1 on 2835... This is low enough jitter for the ethernet and can be initialised automatically by the code.

[Briarios]

You can get rid of the crystal

oooh thats good to know , also is there anything i need to check and make sure with the layout/ schematic or have I covered pretty much most of it?

[tsaG]

Did you keep the resistance of the USB connection in mind?

[John Cleese voice] ahh yes, it would seem that i have indeed missed it... but i checked the traces and they are rated for about 1.6 amps across the trace, i also checked the resistance and the first usb socket has a voltage drop to roughly 4.9v at 500ma and the second socket has a voltage drop of 4.89v at 500ma, so I think this is well within spec for something like a usb flash drive. [/John Cleese voice]

USB2 spec calls for 5v +/- 0.25v officially so technically you are still within the specifications. Many devices actually run on lower voltages and simply feature onboard regulation anyway and will work over a larger range than spec calls for, heard of the ralink wifi adapters function at as low as a 3.5v vcc

Sorry, I meant the 90R differential resistance as meantioned in the IO Board shematic

Route bold-ringed signals as matched
length 90R differential pair

For example: With my Board, I needed about 1.4mm trace width.

BTW: The 64kbit filesize limit in this forum is terrible....

[Briarios]

With my Board, I needed about 1.4mm trace width.

A 1.4mm trace is a huge trace. from what I see with your Calculations on that is it looks like you're making a 2 layer board? if so, I can see why you would need quite big traces.

I am using a 4 layer board and as you can see from the calculations that I have done, I am able to keep the traces smaller while maintaining the 90 ohm diff.
http://pmfg.me/rpi/CM_Test_Board_1.x/pcbtrace.jpg

The same also goes for Ethernet, Ethernet requires a 100 ohm diff to work apparently, i spent many hours looking that info up as well.

[Briarios]

Hello again guys,

I thought I would give you an update on the test board.

I have sent the necessary files to the Fab house in China to get the prototypes done. it should take about 2-3 weeks from now till i receive the end product in my hands. while this is now in the hands of the PCB fab house. I thought I would share with you the cost associated with creating, part sourcing and assembling 5 Prototype Boards. (links Below) some details in the links have been removed to protect both myself and the PCB fab house.

http://pmfg.me/rpi/CM_Test_Board_1.x/Quote.png
http://pmfg.me/rpi/CM_Test_Board_1.x/Parts%20List.png

You will note from both the quote and the parts list that it starts getting quite expensive really fast, then you add roughly 100 for assembly then you get the estimate for just 5 boards, So prototyping something complex like a rpi hat or add on board does add up the costs quite quickly so it's definitely not for the feint hearted, but at the end of it and if the prototype works the effort spent for putting time, money and working your brain overtime is greatly rewarding.

[svrsig]

Wow! Smallest text 32mm high!!