SATA wont awake

Hello,

There is a problem on your schematic:

You are connecting HUB's USB_SS_TX pair to the SATA's USB_SS_TX pair.

You have to connect:

      HUB                                                   SATA

DN_USB_SS_TX----->AC caps----->UP_USB_SS_RX

Regards.

Thank you very much, we will look into it right away! I will reply when we are succesful :)

Hello Elias,

First of all many thanks for your time...

My engineer just had a look at the scheme again and TX to RX is actually what we are doing there, so one of you guys is not looking good.

I know I might sound stupid, but I will ask anyway :) What we actually did was, that we just put all the stuff together on one PCB according to the guides from TI and did nothing more. Now maybe there is some process of loading some firmware to the SATA chip or some process of I don't really know what... Basically what I am asking is, if you connect the chips together correctly, will they simply work? Or do I need to do any additional steps?

I am sorry for my unqualified questions, as you must see I am no engineer :) But I am the one sufficient enough in English :)

Filip

Hello,

Yes, you have to program the external EEPROM connected to the TUSB9261, the default firmware is available here 

http://www.ti.com/lit/zip/sllc416

We have a windows based tool that you can use to program the eeprom, the tool is here

http://www.ti.com/lit/zip/sllc414

And the User's Guide here

http://www.ti.com/lit/pdf/sllu125

Regards.

Hi,

Thank for this, we tried this one, but when we connected the PCB via USB, we only saw the USB2 device on it. Windows did not see the bridge so he could not burn the FW :-/ We are still doing something wrong. Maby the HUB is not alright.

F

Hey Elias,

Would you please be so kind having one more look on our scheme? Or for that matter anyone... We cant seem to awake the SATA HDD drive connected to this device :-/ Maybe the problem is hardware, but I would like to roll out design fault first.

We had a 6-layer PCB made and had it embedded with all the parts. We connected a USB device to it via PCI-E interface and it worked just fine, but when we connected a SATA HDD to it, it just stayed dead. No enumeration :-/

Than you advices us that the bridge we use need to have a FW burned onto it, but we never got to it because when we connected to PC via USB it just stayed silent.

Maybe something was embedded wrong, that might be the case but it would be very helpful if someone helped us rule out our design failure or maybe we missed some step later. Our engineer built many devices in his career, but this is new for him so we are running out of ideas.

A help would be greatly appreciated guys.

Filip

Hello,

Please check again what I mentioned on my first post:

Hub's terminals USB_SSTX_DN0 are connected to the TUSB9261's USB_SSTX terminals, this is not correct.

Hey Elias,

Thank you for the answer, this is also in the scheme:

We believe we have the cross OK, or are we still wrong?

Hello,

Please see the picture I posted above, can you clarify that? In the picture I posted you have TX-->TX

Regards.

Hey Elias,

Thank you very much for your patience! This is going to be the last picture from me :) I see what you are telling, but our engineer insists he is crossing it but that still could not be enough. Please have a look at this one. I believe we will be clear now. Maybe the cross he did was not correct. This might be the thing:

Once again thanks for your time.

Ok, I didn't know that was a dongle.

Can you confirm that the below labels are only "text" and not "net labels" ?

If those are "net labels" the layout software will be take them as the same net and will be directly connected each other.

Disconnect the SSD from the TUSB9261 and try to burn the firmware.

Regards.

Hey Elias,

It was just text. We made some progress, now we know TUSB8040A hub is not working. Now I have the same question as before: Do we need to burn some FW into the USB3 hub? It still might be a problem of embedding it to PCB, but lets debug this first :)

Cheers.

Just trying to get attention for this one :-/ We are stuck because we need an answer.

Hi Filip,

Is this entire design powered by VBUS?  That isn't going to work unless you have a closed application where VBUS > 900mA.  2.5" SATA HDDs can draw quite a bit of current at power on. 

Can you attach the hub portion of the design to a USB host and see if two hub drivers are loaded in Device Manager under Universal Serial Bus controllers, or just send a screen cap?

Thanks!

JMMN

Hi JMMN,

Before we created this PCB we actually rigged together dev kits with both chips and HDD and it spinned up and worked powered by USB2 (which has a bit less power I guess).

On this prototype the HDD wont even try to spin up. We have enough power. Actually the other USB device has 5sec cool down to even start pulling any power. The HDD gets the juice first just to make absolutely sure it will not crash. But the problem is not even the first happens...

As I sayed it still can be mechanical failure, but we are trying to find a fault in our design. JMMN I do not believe what you are suggesting is actually the issue.

Hi Filip,

Please send information found under Device Manager to help us find the issue.

Regards,

JMMN

Good request JMMN :)

There is no enumeration, there is no Device Manager data. That is THE problem we are having.

We actually tried to directly bypass the USB hub and the SATA itself actually enumerated for few seconds and than crashed (which is expected for such approach). So now we are guessing the problem might be in the hub chip.

F

Please check that the USB_VBUS input is around 500 mV at the hub.  I'll assume you already checked the power, reset and clock.  Also, make sure that 4G_D+ and 4G_D- are not presenting a SE1 (single ended one) to the TUSB8040A...this would result in the TUSB8040A entering programming mode. 

A few other things to check...check the resistor value of RF13, make sure your layout did not pull B18 low (it is unconnected in the schematic - that is correct).  Also, please try removing RF 9 and RF10 in the design.  There is an errata on the TUSB8040A/A1 where it does not take unused ports into account when reporting the non-removable ports.

JMMN

I confirm 500mV.

I confirm reset is OK.

I confirm 4G.

Resistors are according to the datasheet. We will try to unplug them tomorrow. I will get back to you as soon as I have results.

One little additional question from our engineer: Is there any SW which is able to look into or modify the register of  TUSB8040 it might clear things out a little and maybe allow us to modify "Port Used Configuration Register (offset 8h)" without HW intervetions (maybe).

Thank you very much for your time btw. :)

The only way to modify registers is through EEPROM or SMBUS on the device.  One other question, are you using a standard USB cable to connect from the hub to the host?

Thank you for the answers...

Yes, we are using standard USB3 cable.

Filip,

Does the clock start oscillating?  Also, please ensure the ground pad is well connected to the PCB.  The device won't operate if the ground pad has a poor connection.

Thanks,

JMMN

Hey JMMN,

We just made another prototype board. We fixed few minor glitches but the main problem stays still the same. But now I might be able to write a better description.

Now here it is. What we want is to connect 2 devices to the hub and connect it to a computer via USB SS.

1. One device is USB HS and is connecting to the hub as USB HS. This device we see when the dongle is connected to PC.

2. Second device is HDD connected through SATA-USB bridge to the hub as fixed USB SS, this one does not errate.

3. The whole set connects to the PC as USB HS, not SS. We really believe (and RTG shows) that all connections are manufactured well. So this should be OK.

So we do not see the bridge, we can not burn the software to it and so the HDD stays dead.

Does this say anything to you?

JMMN said:

The only way to modify registers is through EEPROM or SMBUS on the device.  One other question, are you using a standard USB cable to connect from the hub to the host?

Hi,

We are connecting a regular USB3 cable to the device. But notice, that the plan shows that from the hub, only USB SS line goes to the USB female. So the device uses only the USB3 lines on the way to the computer.

Might this affect the Flashing?

Hi Filip,

Just to review the current program status, the USB 2.0 portion of the tree is running, but not the USB 3.0, correct?  Is the USB 3.0 hub loading in Device Manager?  There should be two drivers loaded for the TUSB8040A/A! device one for the USB 2.0 side and one for the USB 3.0 side.  There is no conversion between USB 2.0 and USB 3.0, the protocols run on completely separate data lines in the cable.

Is it possible that the SATA drive does not tolerate the absence of the USB 2.0 lines?  Per the USB specification, all USB 3.0 devices must have USB 2.0 signals connected for backwards compatibility.

Thanks!

JMMN

Hi JMMN,

One step further, still not too happy :D We finally connected to the bridge, saw it in the Device manager as the Flash burner, the burner software saw the bridge, we choose the correct driver, we burned it, we got success message and than nothing. It remained and still is in Bootloader mode... Any idea? Some screens here:

Just a friendly ping. Still stuck :-(

Hello,

Can you try in another system with no previous versions of the TUSB9261's driver/firmware?

Can you take some USB protocol traces between the hub and the tusb9261?
Regards