Prototype 1 Objectives

300px 300px

Since this is the time we worked with a COM express module, realistically, we just want to see if we can see if the module works on our baseboard. We also want to test the KVM circuits.

In the current design, the high speed signals are spread out, minimal number of vias, and with no overlapping, except for the USB.

The KVM itself gets a VGA and USB signal.

The list of things to test: * Working COM Express. * Power system works (3.3V, 5V, 12V Amperages) * USB: short and long traces * KVM (switching) board * GBE: short and long traces * Working SATA: install OS * Express Card

Prototype 1 Results

The logs are written below, but this is the summary

* KVM USB and VGA works → no need for the extra filtering circuitries. * Incorrect COM Express footprint. Can't test anything

Bug list

Start with the architecture first. That is start at main: * COM Express module out ports. Make sure the connections are routed out properly (no switching from data- to data+ on USB, for example)

Parts: * Part list: are they in stock? * Pad size, location (maybe the order is flipped) * pad ↔ pinout matching

POST Baseboard PROTOTYPE1 BUG LIST

General: * -Add 4 stand-off holes on the corners of the board

KVM side: * F -Propeller Crystal ECS-50-20-1, which has a 'metal casing' must 'not' have any 'vias' within the border * F -Take out every floating pin on the KVM schematics * F -USB Mux(MAX4999) (Pin2,20) needs 3.3Vcc power instead of 5Vcc. * F -VGA Mux enable pin need GND to enable * F -USB choke pad is too small * F -MAX4999 pin 22(D4+),23(D4-) is flipped. * F +USB works without choke. * F -USB control signals needs 3.3Vcc, not 5Vcc * F -The Propeller pin map is totally wrong between DIP and QFP package.

COM-Express Side: * F -COM-Express connector row order is reversed (DCBA from the top) * F %SATA Hard drive mount holes needs to be added & draw (overlay) border lines around the SATA HD * F -Express Card connector left mount hole size is incorrect (too small 1.6mm → want 2mm). * F -SATA Connector mount hole (the one for the clip) needs a ring for soldering * F -Vert SATA Connector mount hole also needs a ring for soldering * F +Add copper pads to the corners of COM Express

POST KVM_PROTOTYPE2 BUG LIST

+Good thing to know/have, -error must be fix, #just note,%some improvement

* -The VGA & USB connectors are not align with the edge. * -The mount hole may hit connector * %The QFN pad should be double length * %Add power switch to control 5Vcc control line * -VGA LED2 & LED3 is fliped * +Computer side VGAmux M needs go to 5VCC * +Monitor Side VGAmux M needs to go to GND * -C1 & C2 VGA_EN# & VGA_SEL are fliped * -Impedance for the control signal is to high, need to use thicker trace(20 mil) * -Propeller P29 connect to EEPROM SDA pin need a 10K resistor with 3.3VCC which is missing * -EEPROM type is wrong. It should be I2C type, not SPI.

Logs

'January 12, 2009 Test COM Express on Evaluation Board' After we screw in the COM-Express heat spreader to the module, we start by testing the COM-Express (to see if it works at all) on an 'evaluation board' from Kontron first. Make sure that the spreader is cooled with a fan. With this evaluation board, we have to power it using 'both ATX and AUX auxiliary (2×2) connector'. Check the board Design Guide for the jumper information to take out the need to use the auxiliary connector.

The first test is VGA, and Keyboard and Mouse (KVM). We have a bit of a problem using the PS/2 KVM. The USB KVM works just fine. We use the Ubuntu Live CD to load an Operating System. It works great.

Ethernet works as well. There is a link LED to indicate that a connection is established (good design).

SATA Hard drive works.

Firewire and camera.

'FIX image of setup'

'January 13, 2009 Test COM Express on Prototype 1 board'

Now that we know that the COM Express module is working, we can try it on our base-board. Without the COM Express board, the Prototype 1 board power system (powered by a power supply) is working properly (standby LED turned on). However, when we plug in the COM Express module, the power is shut down.

We probe the system using a multimeter and found that the 12V power and GND are shorted together. As it turned out, the 'COM Express FOOTPRINT is incorrect (it is vertical mirror of the correct one)'

This is it, we are back to the drawing board.

Back to Electrical System

beobot


Navigation
QR Code
QR Code beobot_2.0_electrical_system_prototype_1 (generated for current page)