Vivado does not recognize my boards anymore.

[fx62]

 

Hello, I am using a Nexys A7-100T and an Arty a7-35T board with a Windows 10 computer. Both boards worked well. However since yesterday, Vivado does not recognize the boards anymore. Auto Connect only connects to localhost.

 

Installing Vivado again (with cable drivers) didn't help. I've also tried the hints in https://digilent.com/blog/what-do-you-do-when-your-board-is-no-longer-detected-by-vivado/ and it seems that the Windows device manager recognizes the boards.

I tried the boards with another computer and they were recognized.

Has anyone an idea what to do?

Thanks in advance.

fx

[zygot]

Well, I have to admit that what you are describing is different from previous, but similar posts.

Did you do any Windows Update between the time that the boards were working and when they stopped? Do you let Windows automatically update drivers?

It's curious that loss of JTAG connectivity to both boards happened a the same time.

One problem is that you can't have two programs connected to the same USB devices at the same time. The Windows ADEPT Utility is pretty robust. Make sure that VIvado isn't running as a process. Plug in your board. Then fire up the ADEPT Utility and see if a board gets displayed. It has a Device Manager that might help if no board gets found.

FTDI has a utility called usbview that is more informative than Windows Hardware Manager. You might want to try that one.

For example my Genesys2 shows up in usbview as a "USB Composite Device" with iProduct = " Digilent Adept USB Device" and iManufacturer = "Digilent". Of course, in my case the Adept Utility on Win10 finds the Genesys2 as soon as the program is running. Win10 Hardware Manager is pretty worthless in telling me which USB device is the Genesys2 JTAG device. The only way to figure it out using Windows Hardware Manager is to unplug the Genesys2 cable, see what's been enumerated and then plug it in and hope to notice what's new. This is easier on a Linux machine. Microsoft really doesn't want to help users take charge of their own hardware. Microsoft's terms of use basically means that installing Windows on your PC makes the PC Microsoft's PC...

Have you tried a different USB connector on your PC? Have you tried using a powered HUB?

Just because your PC enumerates a USB JTAG device doesn't mean that there's connectivity to the JTAG hardware on the other end. Curiously usbview says that the Genesys2 is "Self-Powered Bus Powered". I don't know what that means because Digilent doesn't want to show the schematic wiring. It's possible that a Bus Powered USB device could get detected by your PC but the device isn't powered so the JTAG chain is broken. Your board power LED is on so the board is powered. If one of your boards lost connectivity I might wonder about ESD or power-line surge that killed your FPGA devices. Seems unlikely...

On some FPGA boards with HSMC or FMC connectors the JTAG chain runs though the connector and not enabling by-pass can prevent configuration... this isn't something that your boards do. Edited September 22, 2022 by zygot

[zygot]

Try connecting to a new target in Vivado Hardware Manager instead of using Auto-connect.

I don't like to recommend products but I do suggest trying the Adept Utility for Windows as an alternate method for configuration. Vivado Hardware Manager can be difficult when using more than one board as a target. Vivado certainly could do a better job of finding and showing available target devices.

[JColvin]

Hi @fx62,

The other thing you might check on the first machine (though if you had reinstalled Vivado, I would think this might have resolved it) would be to see if there is an instance of hw_server still running in the Windows Task Manager and "End Task" that process, since restarting the GUI does not necessarily end that instance for you.

Thanks,
JColvin

[fx62]

Hi,

thank you for your answers

Adept doesn't recognize the Hardware.

Tried "Open new Hardware Target" and connect to local host. No Target is listed there. Stopping hw_server didn't help either. I've reinstalled  Vivado several times. It also didn't help. I've also deinstalled the drivers USB Serial Converter  A and B. I did a CPU reset on the board. No effect.

The board is connected to the computer (The power LED is on) and when I connect the board to the computer the drivers USB Serial Converter  A and B are being installed by the computer and it is recognized as diligent USB device.

What else can I do?

[fx62]

Hi Zygot,

 

thank you for your answer. To avoid misunderstandings: I am not using both boards at the same time. It's just that I teach two different classes at two different schools and each school uses a different board.

Anyway, Vivado works now with my Nexys board. I think the trick was to run the ADEPT utility before starting Vivado as you recommended. Once Adept recognizes the Board, Vivado recognizes the board even without Adept being started before. I can't use my Arty board now for some other reason, but I will test Vivado with the Arty board as soon as I've got a replacement. If I don't post anything during the next days, it also works with the Arty board.

Anyway It's good that the fault happened during testing my boards and not during teaching in class.

Best regards

 

fx62

 

 

 

 

 

 

 

 

[fx62]

Hello,

 

my problem is back. Both boards are neither recognized by Vivado nor by Adept. Screenshots1.docxThe LEDs on the boards are on and I use a powered hub. The Arty board is new and my nexys board worked yesterday. 

I do not use the boards simultaneously

When I connect one of the boards there is a reaction in the device manager and in USB View  it appears as "USB Composite Device" with iProduct = " Digilent USB Device". Interestingly, the device name vanishes after some seconds and "Port#0006.Hub_#0001" appears. The  line remains highlighted in green. (See the attached screenshots)

 

Any Idea what this could mean?

 

 

 

 

 

 

[elodg]

If the device enumerates momentarily with the proper driver loaded, it sounds like there is some other software forcing re-enumeration or configuration change. Maybe a USB debugger or packet monitor?

[zygot]

2 hours ago, elodg said:

it sounds like there is some other software forcing re-enumeration or configuration change

If you are suggesting that something unusual is going on with the rash of users experiencing board connection problems, then I am in violent agreement with you.

5 hours ago, fx62 said:

I do not use the boards simultaneously

There's no reason why you can't have two or more boards connected to your computer at the same time. I frequency do projects involving 2 or more FPGA devices that need to be configured during a session. The Adept Utility for configuration of multiple boards is much more robust than Vivado Hardware Manager.  There will be a problem if, say VIvado Hardware Manager is connected to the same target as the Adept Utility for Windows is.

Running the Adept Utility should have no effect on Vivado Hardware Manager's ability to connect to an FPGA device...unless you are running both at the same time.

As for detachment issues, my Win10 PC has an audible sound for USB attachment/detachment events. Do affected users hear this? I'm not convinced that this is the root issue. I doubt that there are USB debuggers involved. That doesn't mean that there isn't a process running in the background that isn't causing problems. If you PC is connected to a company or University network I'd confer with the IT department.

If I were a product engineer thinking about this I'd want to start getting user information:

• What version of OS are you using?
• Does you OS automatically update driver?
• Are you running a third party malware detector process?
• What's your FTDI Driver version Number? (If you can't find the enumerated device this won't be easy, but a different USB device using an FTDI bridge will use the same driver)
• What version of Vivado are you using?
• Can you find a "Digilent Adept USB Device" using usbview.exe? ( from FTDI )
• What's you board rev?

I'll likely have more to add to the list as this keeps interrupting the things that I'm trying to do. Anyway, this doesn't sound like the usual connection issues. I suppose that it could be faulty parts ( these things happen, especially when inventory issues are bad.

Digilent should be proactive about this.

 

 

Edited September 23, 2022 by zygot

[fx62]

The boards work now. I am deactivated com 3 and 4 for the nexys board or com 8 and 9 for the arty board. (see screenshots). Could it be that drivers are being loaded twice or that there are drivers of a previously installed version are being loaded?

@zygot

I am using WIN 10 pro and Vivado 2022.1. The I cant see any rev number on the board. Both boards were bought one year ago.

The boards are now visible in USB View. Before, they were just visible for some seconds (See my post above)

 

 

 

Screenshots2.docx

[zygot]

@fx62,

I've never seen an interaction between COM devices and Xilinx JTAG devices in many years of doing FPGA development. That doesn't mean that there isn't something different about your Win Pro and my Win10 Pro, or that there isn't something different about your boards.

The Nexys A7 and Arty boards have a dual channel USB Bridge, like many of Digilent's boards. One channel should be enumerated as a COM device and the other as something else.

FTDI VCP drivers have had issues with device enumeration in the past. I wonder if this is a possibility. 

There was a dark period, a while ago, when FTDI pushed out drivers that were looking for counterfeit chips. Amazingly, they thought that the solution was to permanently brick hardware that was deemed to be counterfeit. I'd be horrified if they revisited that idea, even if only involved detachment. A driver enumeration issue seems more likely.

If you can see your board in usbview, use Win10 Hardware Manager to see what the driver version is. You haven't mentioned if you allow Win10 to automatically update third party hardware drivers. Personally, I do NOT for a variety of reasons.

** Quartus for Intel-Altera devices has a terrible time differentiating between common FTDI COM devices and USB ByteBlaster devices. You aren't running Quartus are you?

Edited September 23, 2022 by zygot

[fx62]

The version of the FTDI drivers is 2.14.1.2 of 14.9.2021 they are in the device manager under USB serial converter A and B. Interestingly, the USB ports which I had to deactivate (com3, 4, 8 and 9) use the same FTDI driver while COM 6 and COM 7  use Microsoft drivers. (I've attached the screenshot again).

The Windows-settings are made by our IT-department. I don't know them and I can't change them. I've looked  at the update history and they were no recent updates. However, I've updated some drivers manually, but this was some weeks ago and Vivado recognized the boards afterwards 

Screenshots2.docx

[zygot]

FTDI has 2 Windows drivers for its USB UART bridge devices. VCP is for bridge channels set up as UART COM ports and D2XX is bridge channels set up for alternate functionality such as JTAG, MPSEE, etc. VCP is the default driver for Windows detecting FTDI USB bridge devices. It's possible that some product vendors using these devices might use their own driver. I don't know of any that do.

If some process running in the background is scanning hardware, I suppose that this might explain how USB devices are getting recognized as the wrong type of USB device. I've not seen this... except as I mentioned before that Quartus has issues distinguishing between USB ByteBlaster devices and regular VCP or D2XX devices. I only mention this because I do, sometimes, use Quartus and Vivado at the same time on my Win10 Pro box.

Vivado Hardware Manager, when running, seems to be scanning periodically, for possible targets. It would be aware of a connection that disappeared because something caused the device to enumerated as something that is isn't. I do a lot of work with FPGAs and USB bridge devices from FTDI but I can't say that I recall experiencing your particular problem.

I might be leaning toward something running in the background that was installed by your IT guys. If your OS is controlled by IT then I would suggest bringing them into the conversation... at least presenting you issue to them and providing the gist of this thread.

FTDI bridge devices use a configuration EEPROM that initialize the device upon power-on. Each channel has a setting in the EEPROM that sets the driver to either VCP or D2XX. I don't know what arrangement that Digilent has with FTDI that makes Digilent think that it's configuration hardware design is proprietary but it's possible that this is involved. In theory, the OS and the software running on an OS should be able to query any FTDI bridge device and determine the proper driver ( unless it's a 3rd party driver ) for all enumerated devices. Obviously, reality is not always the same as the theory. Edited September 24, 2022 by zygot

[zygot]

11 hours ago, fx62 said:

The version of the FTDI drivers is 2.14.1.2 of 14.9.2021

According to FTDI's website, a minute ago, the latest VCP and D2XX drivers for Win7-Win11 are version 2.12.36.4 from date 07/15/21.

Perhaps you got the number wrong. Perhaps this is a supply side attack. I really can't say for sure. You may be aware that the world is at war.

What I can say is that I've never seen this behavior before and don't currently have an explanation that makes sense.

Edited September 24, 2022 by zygot