TUSB2036: Connection problem

Danilo

How is 5V being supplied to the two downstream ports?

Is the hub itself being properly recognized?

Thanks

David

Hi David,

Please see the response of our customer below.

Thank you for your message. I designed a board where I have a CM4 module from raspberry. The TUSB2036 is supposed to communicate directly to it.

On my board, there is 220V going into a converter AC DC with 5V output. This voltage also powers the 2 USB ports.

The CM4 doesn't seem to see properly the TUSB2036.

Regards,

Danilo

Danilo

So the raspberry CM4 module is connected to the TUSB2036 upstream port?

With a scope, can you verify there is a 6MHz clk input to the TUSB2036? 

Can you also see if upstream DP is pulled to 3.3V when initially connected to the raspberry CM4 module? 

Thanks

David

Hi David,

According to our customer,

CM4 is connected to the upstream port.

6MHz clock is checks and is perfect. I made some measurements this week-end on the systems. Signal is 3v3 on DP upstream, and DM is 3v3 on downstream port at the beginning when the keyboard is connected.

I made a lot of tries and after some lost times on the keyboard, I saw some strange things.
I saw that actually, USB HUB is OK for mouse, USB key, but not for the keyboard. I checked the signals for each, and USB D+ and D- are ok for all of them. What could cause the hub to accept everything but the keyboard?  Would you have any lead for me to debug that issue?

Regards,

Danilo

Danilo

The keyboard needs to be recognized by the raspberry CM4 module through the hub. If they see 3.3V on the DM from keyboard, then the keyboard indicates it is a low speed USB device. 

Does this issue happen with a particular keyboard or every keyboard? 

If they plug the keyboard directly into the raspberry CM4 module, does it get recognized?

Thanks

David

Hi David,

Here is the response of our customer.

It happens with every keyboard. I don't think it's the CM4 at fault because when I plug the module that use another mainboard with another HUB, any keyboard is detected, without changing anything software wise on the CM4.

On my board with the TUSB2036, keyboard is never recognized.

Regards,

Danilo

Danilo

Looking at the schematic again, I noticed RESET is pulled to 3.3V. For REST, a pulse width between 100 μs and 1 ms is recommended after 3.3-V VCC reaches its 90%. Clock signal has to be active during the last 60 μs of the reset window. Can they add a 1.5k pullup and 0.1uF pulldown cap on the RESET pin?

Thannks

David

Hi David,

Our customer implemented your suggestion and here is the result.

I just made the modification with the pull up resistor to 3V3 and with 100nF pull down capa, nothing changes. Works for mouse/USB key, but not for keyboard again.

Regards,

Danilo

Danilo

Both the USB mouse and keyboard belong to the USB HID class, so if the USB mouse is properly getting recognized, then I would expect the keyboard to be recognized as well.

Has he tried to plug only the keyboard into the hub and see if it is getting recognized? He has two USB ports in the design, did he try both ports?

Can he also probe the 5V being supplied to the keyboard and make sure there is no voltage drop on the 5V?

Thanks

David

Hi David,

I just received this update from our customer.

I have some news. I tried with a wireless keyboard, with a USB dongle, and it worked. I think the thing is, the keyboard is self powered with it's own battery. So we can say that it's an alimentation issue here.

What are the advice on this?
I have the 100 nF capacitor on the VCC pin for the TUSB2036, and on each USB port, I have a tantalum capacitor 100uF on the 5V.

Regards,

Danilo

Danilo

Please see my previous response, with a bus-powered keyboard which gets power from the 5V, there might be a VBUS droop issue when the keyboard plugged into the hub. 

Can he change the tantalum capacitor from 100uF to 220uF? Can he also VBUS with a scope when plugging in the hub?

Thanks

David

Hi David,

I just received this update from our customer.

I got the board back from the customer.
I added one tantalum capacitor of 220µF on each USB to make the test. It still doesn't work.
The system works for the wireless keyboard and wireless mouse with a usb dongle, but not with wired keyboard.

Any other idea?

I checked VBUS and it's stable when the connection is done.

Regards,

Danilo

Danilo

Can he provide a scope waveform with VBUS, DP, and DM when the wired keyboard is plugged into the hub downstream port?

Thanks

David

Hi David,

Please see the scope waveform from our customer,

I took some measures now. I only have a 2 channels oscilloscope with me so I had to do a little bit more to show you everything.
Here are the results :

1. Firstly, with the USB key dongle - wireless mouse, Vbus is stable. D+ goes from 3V to lower value and D- goes from zero to the opposite of D+, mouse works perfectly.

2. Secondly, with the wired keyboard. Vbus is stable.It varies from 4,85V to 5,05V. When a USB device is connected, there is not a further drop in any kind.
The 2 scopes were on the D+ and on Vbus. D+ goes to 3V after about 10ms and then there are a lot of spikes to 0V.

3. Tests with D+ and D-, wireless mouse with usb dongle. Everything is ok, D+ is the opposite of D-

4. Tests with D+ and D- on wired keyboard. D+ goes to 3V and goes multiple time to 0V. D- doesn't move from 0V.

By the way, I made the test on an older board, with the same Vbus that on this board. I didn't have a usb hub but a multiplexer that allow the user to go from 1 usb to the other, software wise. I tried it again this morning and have no issue with the wired keyboard.

Do you have any idea where comes this issue from?

I have no picture for case 1, but it's similar of case 1 so I figured it was not useful.
For every picture, first letter is for the case, and then it's just numbering.

So for picture 4-3, it's for case 4 and it's the 3rd picture. You can see that the D+ drops but D- stays at 0V.

Hi,

By the way, I made the test on an older board, with the same Vbus that on this board. I didn't have a usb hub but a multiplexer that allow the user to go from 1 usb to the other, software wise. I tried it again this morning and have no issue with the wired keyboard.

*** Is this with the same CM4 module? 

Tests with D+ and D- on wired keyboard. D+ goes to 3V and goes multiple time to 0V. D- doesn't move from 0V.

Most keyboard will be a low speed keyboard, so I would expect D- goes to 3V while D+ stays at 0V, this would indicate the device is a low speed device. It is strange that DP will go high. Can you check the DP and DM with the multiplexer board?

Thanks

David

Hi David,

Please see our customer's response below.

Yes it was the same CM4 module, same software inside between the 2 boards I tested, the one with the multiplexer and the other one with the TUSB2036

Sorry I think I did a mistake while talking about the wired keyboard. I think it was the D- going to 3V and D+ not moving at all. Actually, I had to switch between channels during the test that's why there was a D+ in green or yellow the next time, I'll label it better next time.

I will make the test on the older board and on the new one.

If it was the d- like intended that went 3v and d+ stayed at 0V, do you already have a recommendation for that, so I can test a solution tomorrow morning?

Regards,

Danilo

Hi David,

Please see this update from our customer.

I made the tests again. Once again, everything is ok with the USB multiplexer. Wired keyboard or usb dongle mouse work well.

There's all the picture on the board with the multiplexer.

Vbus board V0.2

Wired keyboard board V0.2 -1

wireless mouse board V0.2-1

wireless mouse board V0.2-2

After that, I tried the new board, mounted with the TUSB2036. Wireless mouse works great, keyboard doesn't work.

D- Is ALWAYS on the yellow line on those scopes from today.

D+ not moving

D+ not moving 2

As you can see, Vbus stays at the same level. When I try to connect the wired keyboard on the new design, D+ is not moving with the wired keyboard.

Here's the picture with the D+ and D- with the USB wireless mouse on the new board : everything normal

Any idea how to work this chance? To be honest I'm close to forget this component and start this part again with another one because I had no issue whatsover with the older board.
I hope you can at least give me a direction where to look for.
If you need anything else please let me know.

By the way, one last thing.
If I put the wired keyboard in the system when the usb wireless mouse IS NOT CONNECTED, then it kind of crashed. I tries to connects, D- goes up a few times, then it stops. After that, when I connect whatever USB device, it doesn't pull up anymore on connexion.
You can see that on D+ D-.png and D+ D- 2.png files.D- went 160mV higher but not more, instead of going to 3V3.

D+ D-

D+ D- 2


Maybe after the TUSB doesn't recognize a device it blocks the downstream ports?

Regards,

Danilo

Danilo:

    David is out for couple of days.

     Can you take waveform with Vbus, D+, D- which is triggered by Vbus when you plug keyboard to hub?

Regards

Brian

Danilo:

For some reason, the raspberry CM4 module isn't sending the expected resets to the ports of the TUSB2036, which is causing the low speed device issues.  Without being able to see the UBS traffic between the raspberry CM4 module and the hub it is difficult to tell what is happening.

I went searching online and see if I can find similar issue being reported, and this is what I found, https://raspberrypi.stackexchange.com/questions/43628/pi-2-keyboard-mouse-not-recognized-through-usb-hub. Can they try this command "dwc_otg.speed=1"?

Again, we have to look at the USB log file between the raspberry CM4 module and the TUSB2036 to understand why raspberry CM4 module is not resetting the ports.

Thanks

David 

Hi David,

According to our customer,

Thank you for the idea. I tried the command "dwc_otg.speed=1" added on the cmdline.txt in the config on the CM4 but nothing changed.

I gave you 4 pictures. The "keyboard connected D- yellow" 1 and 2 is when I connect the keyboard directly to the TUSB20236. The second 2, Keyboard connected after a usb HUB.png 1-2, is when I connect the wired keyboard through an external USB HUB.

keyboard connected D- yellow 1

Keyboard connected after a usb HUB 1

Keyboard connected after a usb HUB 2

When the wired keyboard is directly connected to the TUSB2036, or if there is the external HUB between them, in both cases, the keyboard doesn't work.

On the CM4 IO board, there is also a USB hub, and it's one from Microchip. It's the USB2514B-I/M2.
Do you think there's a difference that could do the trick between this one and the TUSB2036?

Do you have any idea on how to go further in this case? If you need any other scope please let me know.

Finally, I'm sending you two pictures of the lsusb commande on the CM4 if this can help you.

Please let me know if you need me to do anything else.
I'm looking forward to getting the issue of this system.

Regards,

Danilo

Danilo:

     We will review your results and respond soon.

Regards

brian

Hi Brian,

We would like to request an update to this inquiry.

Thank you!

Regards,

Danilo

Danilo:

   Sorry for the late, can you take USB log file between the raspberry CM4 module and the TUSB2036?

Regards

Brian

Hi Brian,

Please see our customer's response below.

Could you give more explanation please? What do you mean with the log? I gave the lsusb command on the cm4 but I'm not sure what more do you need.

Regards,

Danilo

Danilo:

     I mean trace log took by USB analyzer.

      by the way, did customer talk to raspberry about host firmware?

Regards

Brian

Hi Brian,

I just received this email from our customer.

Sorry for the delayed answer, I was on holiday so the project was on standby for the end of the year.
I can check log but just to be sure, are you talking about a software like wireshark or are you talking hardware wise with something like the beagle USB 12?

Regards,

Danilo

Danilo:

   I'm talking about the firmware or USB driver from Raspberry host. All USB host  need driver to make it work.

Regards

Brian

Hi Brian,

According to our customer,

I got the module back from the customer so I'm finally able to try this out. I made 3 test and you can check the log. First sequence is when I connect a wireless mouse with a dongle key. The second one is a wired mouse, but still at full speed. Lastly is when I connect a wired keyboard, which is Low speed. As you can see, low speed is not recognized. If you need any information please let me know. I'm eager to see what we can do to improve this issue.

13h37 - 13h40 : wireless mouse Full speed
13h46 : wired mouse Full speed
13h51 : wired keyboard Low Speed

Jan 16 13:37:03 rpi kernel: [  108.014091] usb 1-1.2.4: USB disconnect, device number 5
Jan 16 13:37:07 rpi kernel: [  111.884107] usb 1-1.2.4: new full-speed USB device number 6 using xhci-hcd
Jan 16 13:37:07 rpi kernel: [  112.031139] usb 1-1.2.4: New USB device found, idVendor=046d, idProduct=c52b, bcdDevice=24.00
Jan 16 13:37:07 rpi kernel: [  112.031162] usb 1-1.2.4: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Jan 16 13:37:07 rpi kernel: [  112.031175] usb 1-1.2.4: Product: USB Receiver
Jan 16 13:37:07 rpi kernel: [  112.031186] usb 1-1.2.4: Manufacturer: Logitech
Jan 16 13:37:08 rpi kernel: [  112.562405] logitech-djreceiver 0003:046D:C52B.0008: hiddev96,hidraw0: USB HID v1.11 Device [Logitech USB Receiver] on usb-fe9c0000.xhci-1.2.4/input2
Jan 16 13:37:08 rpi mtp-probe: checking bus 1, device 6: "/sys/devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.4"
Jan 16 13:37:08 rpi mtp-probe: bus: 1, device: 6 was not an MTP device
Jan 16 13:37:08 rpi kernel: [  112.729224] logitech-hidpp-device 0003:046D:4008.0009: HID++ 2.0 device connected.
Jan 16 13:37:08 rpi mtp-probe: checking bus 1, device 6: "/sys/devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.4"
Jan 16 13:37:08 rpi mtp-probe: bus: 1, device: 6 was not an MTP device
Jan 16 13:37:08 rpi kernel: [  112.887338] input: Logitech M185 as /devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.4/1-1.2.4:1.2/0003:046D:C52B.0008/0003:046D:4008.0009/input/input21
Jan 16 13:37:08 rpi kernel: [  112.887714] logitech-hidpp-device 0003:046D:4008.0009: input,hidraw1: USB HID v1.11 Mouse [Logitech M185] on usb-fe9c0000.xhci-1.2.4/input2:1
Jan 16 13:37:08 rpi kernel: [  112.897543] input: Logitech K270 as /devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.4/1-1.2.4:1.2/0003:046D:C52B.0008/0003:046D:4003.000A/input/input22
Jan 16 13:37:08 rpi kernel: [  112.898315] logitech-hidpp-device 0003:046D:4003.000A: input,hidraw2: USB HID v1.11 Keyboard [Logitech K270] on usb-fe9c0000.xhci-1.2.4/input2:2
Jan 16 13:40:08 rpi kernel: [  293.021117] usb 1-1.2.4: USB disconnect, device number 6




Jan 16 13:46:43 rpi kernel: [  687.172272] usb 1-1.2.2: USB disconnect, device number 7
Jan 16 13:46:48 rpi kernel: [  692.930471] usb 1-1.2.2: new full-speed USB device number 8 using xhci-hcd
Jan 16 13:46:48 rpi kernel: [  693.070388] usb 1-1.2.2: New USB device found, idVendor=04d9, idProduct=a1f7, bcdDevice= 3.00
Jan 16 13:46:48 rpi kernel: [  693.070444] usb 1-1.2.2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Jan 16 13:46:48 rpi kernel: [  693.070457] usb 1-1.2.2: Product: USB Gaming Mouse
Jan 16 13:46:48 rpi kernel: [  693.070468] usb 1-1.2.2: Manufacturer: E-Signal
Jan 16 13:46:49 rpi kernel: [  693.837158] input: E-Signal USB Gaming Mouse as /devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.2/1-1.2.2:1.0/0003:04D9:A1F7.000E/input/input26
Jan 16 13:46:49 rpi kernel: [  693.837833] hid-generic 0003:04D9:A1F7.000E: input,hidraw0: USB HID v1.10 Mouse [E-Signal USB Gaming Mouse] on usb-fe9c0000.xhci-1.2.2/input0
Jan 16 13:46:49 rpi kernel: [  693.846956] input: E-Signal USB Gaming Mouse Keyboard as /devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.2/1-1.2.2:1.1/0003:04D9:A1F7.000F/input/input27
Jan 16 13:46:49 rpi kernel: [  693.910676] input: E-Signal USB Gaming Mouse as /devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.2/1-1.2.2:1.1/0003:04D9:A1F7.000F/input/input28
Jan 16 13:46:49 rpi kernel: [  693.910973] hid-generic 0003:04D9:A1F7.000F: input,hiddev96,hidraw1: USB HID v1.10 Keyboard [E-Signal USB Gaming Mouse] on usb-fe9c0000.xhci-1.2.2/input1
Jan 16 13:46:49 rpi kernel: [  693.914311] hid-generic 0003:04D9:A1F7.0010: hiddev97,hidraw2: USB HID v1.10 Device [E-Signal USB Gaming Mouse] on usb-fe9c0000.xhci-1.2.2/input2
Jan 16 13:46:49 rpi mtp-probe: checking bus 1, device 8: "/sys/devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.2"
Jan 16 13:46:49 rpi mtp-probe: bus: 1, device: 8 was not an MTP device
Jan 16 13:46:50 rpi mtp-probe: checking bus 1, device 8: "/sys/devices/platform/scb/fe9c0000.xhci/usb1/1-1/1-1.2/1-1.2.2"
Jan 16 13:46:50 rpi mtp-probe: bus: 1, device: 8 was not an MTP device




Jan 16 13:51:57 rpi kernel: [ 1001.733328] usb 1-1.2.3: new low-speed USB device number 9 using xhci-hcd
Jan 16 13:51:58 rpi kernel: [ 1002.283348] usb 1-1.2.3: new low-speed USB device number 10 using xhci-hcd
Jan 16 13:51:58 rpi kernel: [ 1002.743818] usb 1-1.2-port3: attempt power cycle
Jan 16 13:51:59 rpi kernel: [ 1003.403288] usb 1-1.2.3: new low-speed USB device number 11 using xhci-hcd
Jan 16 13:51:59 rpi kernel: [ 1003.403640] usb 1-1.2.3: Device not responding to setup address.
Jan 16 13:51:59 rpi kernel: [ 1003.623391] usb 1-1.2.3: Device not responding to setup address.
Jan 16 13:51:59 rpi kernel: [ 1003.943321] usb 1-1.2.3: new low-speed USB device number 12 using xhci-hcd
Jan 16 13:51:59 rpi kernel: [ 1003.943469] usb 1-1.2.3: Device not responding to setup address.
Jan 16 13:52:00 rpi kernel: [ 1004.163383] usb 1-1.2.3: Device not responding to setup address.

Regards,

Danilo

danilo:

    Is this issue only happend for CM4 model? Did you try connecting  keyboard to Cm 4 directly?

regards

Brian

Hi Brian,

According to our customer,

It happens with every keyboard. I don't think it's the CM4 at fault because when I plug the module that use another mainboard with another HUB, any keyboard is detected, without changing anything software wise on the CM4.

Regards,

Danilo

From above messsage, there is a Microchip hub between CM4 and TUSB2036?

Regardds

brian

Hi Brian,

Here is the response of our customer.

I can't connect directly to the CM4 on my board because I can't switch the routing from the CM4 to a usb port and bypassing the TUSB2036. There is a IO board for the CM4 and on it, there is a USB hub from microchip, the USB2514B? With this board, there is no issue with the low speed and full speed USB.

On my board, there is nothing between the TUSB2036 and the CM4. I send you the schematics for the connection on the TUSB2036. The DP and DN go directly to the CM4 port, with termination resistors.

Any other idea?

Regards,

Danilo

got it. so there is no hub between CM4 and TUSB2036 in your  real system.

Hi Bian,

Do you have any solutions to this issue?

Regards,

Dailo

Can you increase the cap to 1uf on reset pin?

Regards

brian

Hi Brian,

I received this update from our customer.

Tried to change the capa, nothing changes.

If I connect the wired keyboard it's not detected, and then I can't even connect my wireless mouse as it's not detected anymore.

Can it be an incompatibility issue between the tusb2036 and gnu Linux?

Regards,

Danilo

Danilo:

          we do see some issue for TUSB2036 for some new USB device, probably not Linux related.

Regards

Brian

Hi Brian,

According to our customer,

OK for the answer but do you have any advice next or should I seek another ic to do the job then? The keyboard I have the issue with is not that new, maybe 4-5 years old.

Regards,

Danilo

Can you send keyboard to us or more analysis? or youo may need to use TUSB4041 instead.

Regards

Brian

Hi Brian,

According to our customer,

I could send you the keyboard but it's a typical wired keyboard. I'll try tomorrow with 2 or 3 other one but the result will be the same I fear. I can send you the datasheet of the keyboard if you want.

For the the TUSB4041, the issue is that it's at least twice as expensive as the TUSB2036 with no added features for our projet, and it's not available on the market right now.

Here is the update of the tests.

I tried 3 wired devices :
Logitech M90 (mouse)
Logitech K120 (keyboard)
Logitech Deluxe 250 (Keyboard)

Everytime, the device sees that I connect a low speed USB, but I always have the same issue in the log : "Device not responding to setup address."

Does it mean that there is no way to use a lows peed USB with the TUSB2036 ? Any workaround with this issue?

Regards,

Danilo

we have an issue before that software where the USB host is not reading the status from the hub on disconnect / reconnect. Can you check?

Regards

Brian

Hi Brian,

Please see the response of our customer below.

Could you tell me a little bit more on what you need me to do? I'm not sure there.

If it's just needed to remove the usb device and put it in again, I tried and it does nothing but failing to recognize again.

Regards

Danilo

Danilo:

   it looks wireless keyboard is FS (dP high), wired is LS (DM high). so it seems LS issue.

Regards

brian

Hi Brian,

Here's the comment of our customer.

I know that the issue is about the low speed and the full speed works fine.
But correct me if I'm wrong, the TUSB2036 is supposed to accept both kind of usb and then use the maximum speed to the upstream port.

Regards,

Danilo

 Is USB2514B HS hub? low speed always working after HS hub. are these wired keyboard working if directlty connect to CM4?

regards

brian

Hi Brian,

Please see this update of our customer.

USB2514B is a HS hub, so is TUSB2036 so it should work also with the TI hub?
On the board, I removed the TUSB2036 and I wired the DP0-DM0 to a downstream port directly with straps soldered on the pads. Wired and wireless mouse work, same for keyboard, I can connect everything.

So this issue is on the TUSB2036 side as far as I know, maybe with the CM4 module. But the CM4 module on itself is able to accept every kind of USB without an issue.

No need to send the keyboard, it's the same for every wired keyboard on the market.

Please note that sometimes, TUSB2036 fails to make the wireless mouse recognized.

Regards,

Danilo

TUSB2036 is FS hub.

Since it's working after removing  TUSB2036, can you take waveform with wired keyboard  with the baord removing TUSB2036?

Regards

brian