I found a trace from each of the USBH-N (D23) and USBH-P (E22) lines to 15K ohm pulldown resistors just above the CPU, under the shield frame.
At this time, anyone can carry on the search to see if the traces are actually connected outside of the CPU shield. Refer to the photo below with the test points.
There are two pulldown resistors. One is completely under the top corner of the shield. There is a testpad for the USBH-N line so I'm guessing there is probably one for the USBH-P line - maybe on the backside of the PCB.
I still need to find two more lines though - the USBPEN (D22) enable line and the USBHPWR (C23) power supply enable. This is not as simple as the HOST pins as these two GPIO pins can be defined for other purposes. Specifically, the GPIO pin D22 can also be used for AC97_SYSCLK which I think is sound related.
If you plan on soldering a couple of wires to the resistors to make it easier to trace the circuits, make sure to secure the wires to the board using a drop of 5 minute epoxy or some other method. This needs to be done to prevent you from ripping the parts off of the PCB if you pull on the wires. The traces on these boards are only a mil thick and a few mil wide and tear REALLY easily. (One of those "been there - done that" experiences).
Stay tuned :approve:
ps: the references D22, C23, etc refer to the rows and columns of pins on the CPU chip. Refer to the PDF file in Afarre's 1st post for more info.
I found a trace from each of the USBH-N (D23) and USBH-P (E22) lines to 15K ohm pulldown resistors just above the CPU, under the shield frame.
At this time, anyone can carry on the search to see if the traces are actually connected outside of the CPU shield. Refer to the photo below with the test points.
There are two pulldown resistors. One is completely under the top corner of the shield. There is a testpad for the USBH-N line so I'm guessing there is probably one for the USBH-P line - maybe on the backside of the PCB.
I still need to find two more lines though - the USBPEN (D22) enable line and the USBHPWR (C23) power supply enable. This is not as simple as the HOST pins as these two GPIO pins can be defined for other purposes. Specifically, the GPIO pin D22 can also be used for AC97_SYSCLK which I think is sound related.
If you plan on soldering a couple of wires to the resistors to make it easier to trace the circuits, make sure to secure the wires to the board using a drop of 5 minute epoxy or some other method. This needs to be done to prevent you from ripping the parts off of the PCB if you pull on the wires. The traces on these boards are only a mil thick and a few mil wide and tear REALLY easily. (One of those "been there - done that" experiences).
Stay tuned :approve:
ps: the references D22, C23, etc refer to the rows and columns of pins on the CPU chip. Refer to the PDF file in Afarre's 1st post for more info.
You got it!!!!
The worst part is done, now we have the main signals!!!
:cheers: :cheers: :cheers:
If we were to use the USB host all the time with a powered HUB,
do you think we need USBHPEN and USBHPWR?
I know for sure a USB hub does not require to get power from host when used self-powered. I know if because I have a PC with one USB host power burned and it still works when I place a powered hub.
My doubt is if it will be required to force USBHPWR to some state all the time just to let CPU know there is no fault condition.
...At this time, anyone can carry on the search to see if the traces are actually connected outside of the CPU shield. Refer to the photo below with the test points...
You are right, any one can do it.
Now I am far away from the place were I have my shop, I will not be there until next month.
But I would like to have with me the required complementary IC's the day I open again my Ax.
But, who knows, maybe nothing else is required
If we were to use the USB host all the time with a powered HUB,
do you think we need USBHPEN and USBHPWR?
I know for sure a USB hub does not require to get power from host when used self-powered. I know if because I have a PC with one USB host power burned and it still works when I place a powered hub.
My doubt is if it will be required to force USBHPWR to some state all the time just to let CPU know there is no fault condition.
Close - check the USB host schematics from the Intel eval sheets (pg29) and you will see that there are two control lines going to the MAX1693 USB power control chip. One is the over current sense output to the CPU (USBHPWR) and the other turns on (or off in the case of an overcurrent condition) the 5V to the external device (USBHPEN).
We probably don't need those signals. However, we DO need a signal conditioning/transient-protection device similar to the STF201 USB interface chip between our external device and the CPU.
One interesting observation from those schematics on page 29 - note the 15K pulldown resistors...:approve:
Close - check the USB host schematics from the Intel eval sheets (pg29) and you will see that there are two control lines going to the MAX1693 USB power control chip. One is the over current sense output to the CPU (USBHPWR) and the other turns on (or off in the case of an overcurrent condition) the 5V to the external device (USBHPEN).
We probably don't need those signals. However, we DO need a signal conditioning/transient-protection device similar to the STF201 USB interface chip between our external device and the CPU.
One interesting observation from those schematics on page 29 - note the 15K pulldown resistors...:approve:
Totally agree.
What I am wondering if I CPU needs a HIGH at USBHPWR (LOW means fault) to turn on data flow. I do not think so, CPU should use that input just to set USBHPEN output.
Yeh!, pull down resistors are the same at page 29.
So, forgive me if I am misunderstanding, but the usb host pins from the PXA270 are actually connected to *somewhere* on the board, we just don't know where yet?
Has anybody tried the points down the side of the motherboard?
So, forgive me if I am misunderstanding, but the usb host pins from the PXA270 are actually connected to *somewhere* on the board, we just don't know where yet?
Has anybody tried the points down the side of the motherboard?
I found the USBH-P testpoint on the bottom of the PCB under the battery (under the label).
Now both the USBH-N and USBH-P testpoints are known. Unfortunately, at this time, they don't appear to be connected to anything else that I can find - but I'm still looking.
It also looks like pins 18 and 19 on the sync connector are NOT connected to anything - nudge nudge say no more :approve:
I found the USBH-P testpoint on the bottom of the PCB under the battery (under the label).
Now both the USBH-N and USBH-P testpoints are known. Unfortunately, at this time, they don't appear to be connected to anything else that I can find - but I'm still looking.
It also looks like pins 18 and 19 on the sync connector are NOT connected to anything - nudge nudge say no more :approve:
Could the IC at the middle of the picture be like STF201?
It has 6 pins and it is very near to USB-P test point and sync connector.
Could the IC at the middle of the picture be like STF201?
It has 6 pins and it is very near to USB-P test point and sync connector.
Unfortunately, I didn't find any 6 pin chips to be connected in any way to the testpoints - but don't stop testing on your own time - I may have missed something.
Chris - We have the driver for WM2003SE - I'm sure it was in the A02 firmware. It was removed from the newer releases. I have no idea what to use for WM5 though. Something would need to be scarfed from another WM5 PDA with a USB host interface.
I'm going to pull apart some cheap USB hubs this weekend to see if I can get the transient suppression interface chip needed for this.
alright.... now that we have the hardware bit down shoukldnt it be tiem to start writting usb host drivers?
chris
Better if we get more information, like if driver circuit is present, power lines, etc.
But I think we already have enough information to test a setup like this.
Of course, sofware will be required :rolling:
Better if we get more information, like if driver circuit is present, power lines, etc.
But I think we already have enough information to test a setup like this.
Of course, sofware will be required :rolling:
You'll KILL the device if you use 5V on the interface chip:exc: :exc:
EDIT: No, you won't kill the device - looking at the specs, pin 1 is used for transient suppression of the 5V line, so yes, you CAN connect it to 5V (But no higher).
Look carefully at the schematics - you'll find that the line labelled USBH0_VBUS is NOT connected. This is VERY important. You need to remember that you will be interfacing 3V and 5V signals.
Let me do some research on the interface chips before you start publishing schematics that someone might use and kill their PDA. I know you want to get this figured-out quickly - but you don't want to kill anything either.
You'll KILL the device if you use 5V on the interface chip:exc: :exc:
Look carefully at the schematics - you'll find that the line labelled USBH)_VBUS is NOT connected. This is VERY important. You need to remember that you will be interfacing 3V and 5V signals.
Let me do some research on the interface chips before you start publishing schematics that someone might use and kill their PDA. I know you want to get this figured-out quickly - but you don't want to kill anything either.
In fact I wasn't sure about VBUS.
Just in case I fixed the drawing.
Have a look at the STF201 docs: http://www.semtech.com/pc/downloadDocument.do?id=139 particularily figure 4. Pin 1 is used to clamp transients on the 5V line. I believe the actual USB data lines are 3.3V differential signals.
Newark Electronics sells the Semtech STF201-22 for around $2 ea ($25 min order). This is a very small surface mount chip and you need specialized soldering equipment and a good magnifying glass to solder to it.
I just had a GREAT idea :approve: - Instead of building the interface into the Axim, purchase a powered USB hub, wire it into an X50/X51 sync connector and let it be a dedicated interface - with several USB ports! Oh - I am so proud of myself
Of course, we'll still need to place a couple of transient diodes inside the case - just not a whole interface circuit.
Follow Us
RSS Feed
Follow on Twitter
Facebook
YouTube
Found traces!!
