ZC706 power problems

[rmelo@inti.gob.ar]

I have a ZC706 board. It was working until the last friday. Today, when I went to use it, no voltage on indicator is ON. Also the fan do not spin.

I did some meassures:

• VCCPINT_EN_B is LOW, but no other enable work.
• Input voltage and reset is ok in UCD90120A.
• PMBUS_ALERT is also LOW.

Is maybe this problem? #AR60702

My assembly number is 0431845-02. I think that it is a rev2.0.

Any idea? Thanks in advance

[gnarahar]

rmelo@inti.gob.ar Before your board can be classified as per that AR, please try reprogramming the Power Controllers

Please refer to below AR's that will help you for reprogramming and getting the TI USB module

https://www.xilinx.com/support/answers/56811.html

https://www.xilinx.com/support/answers/37561.html

https://www.xilinx.com/support/answers/53599.html

https://www.xilinx.com/support/answers/54022.html

To find out which board rev you have, check this AR

https://www.xilinx.com/support/answers/62325.html

[rmelo@inti.gob.ar]

In my case, the revision number is not where AR#62325 said, but I found it: HW-Z7-ZC706 REV 2.0

I read these Answer Records before. The problem is that I have not the programmer and I live in Argentina. I need to import the programmer and wait several weeks to have it. I am running a project with this kit. Are there something else to do? UCD9240 normally lost its programation? I have also a ML605 which I used a lot and I never had this problem.

Regards

[gnarahar]

I understand but unfortunately if it is a issue with Power Controller, the only way to reset is to reprogram them. 

Having said that, even other Power rails are not working?

[rmelo@inti.gob.ar]

The voltage on indicator leds are not working. 12v reach the regulators but the enable signal is not generated by UCD90120A. The Leds ON in the attached image are: DS31, 32, 33 and 34. Also P (12v) and Ethernet in the side. The UCD90120A is energized with 3.3v (I checked pins V33DIO1/2 and V33A/D). I also checked that RESET_B has a '1'. The signal PMBUSALERT_B is in '0', so, I suppouse, something is wrong there.

It is normal for UCD90120A to lose its programming? If yes, maybe I can write the register using I2C and the XML with values.

Regards

[rmelo@inti.gob.ar]

Question:Is it normal for UCD90120A to lose its programming?

I need a response to explain to my client what happend with the board.

We are trying to read the chip through I2C to see if is alive.

Regards

[gnarahar]

---

Question:Is it normal for UCD90120A to lose its programming? 

---

Typically it shouldn't. However these being Eval boards and customers using them for various design tests / functionalities and varying the voltages based on their application along with these being open systems, there may be a chance for it to happen. 

Without knowing the history, wouldn't really be able to predict or answer the query. 

[rmelo@inti.gob.ar]

We are using the board with a FMC164 from 4DSP. We where doing test on Friday and when we tried on Tuesday to use it again, it failed. PMBUSALERT# is Low. In the datasheet of the UCD90120A said (section7.5.3 Internal Fault Management and Memory Error Correction):

The UCD90120A verifies the firmware checksum at each power up. If it does not match, then the device waits for
I 2 C commands but does not execute the firmware. A device configuration checksum verification is also
performed at power up. If it does not match, the factory default configuration is loaded. The PMBALERT# pin is
asserted and a flag is set in the status register. The error-log checksum validates the contents of the error log to
make sure that section of flash is not corrupted.

We are trying to get an I2C response without success (but the device is detected in the address 0x65).

I suppose that the checksum fails :-(

[rmelo@inti.gob.ar]

We were using the board with the FMC164 at least three weeks without this problem.

[rmelo@inti.gob.ar]

We bought the USB-to-GPIO and we need to wait it arrive at least a month. When I used it, I will post the results.

Regards.

[gnarahar]

Missed your earlier updates. Not receiving a response doesn't sound good. But good that the device is identified in the chain. 

Good to hear you ordered the tool and yes, please post the update once you have a chance. 

[brunovali]

Hello, I will be following this thread instead of rmelo.

We have bought the USB-to-GPIO adapter from Texas Instruments and run the tests scripts provided in https://www.xilinx.com/support/answers/56811.html, all of them finished with failure.

I've attached one of the logs and down here there are two pictures of the screen when running the test.

The AR#56811 includes this,

Zynq-7000 AP SoC ZC706 Evaluation Kit: Please see ZC706_Power_Controllers_Reprogramming_Steps.pdf attached (to be used in conjunction with ZC706_Rev1_TI_XML.zip; if using a ZC706 Rev.B, please use zc706_revb_ti_xml_power_script_r9_with_alternates.zip)

It seems that refers to a ZC706 rev. B instead a rev. 2.0 board, is that right? If not, where could I find the configuration/test scripts for the ZC706 rev 2.0 Board? I'm using the Texas Instruments Digital Power Manufacturing App Version 7.0.19.

Thanks in advance.

[brunovali]

Hello, we couldn't solve the power problem issue yet, is there any other thing we can do? We have bought the texas instruments USB Interface Adapter, tried reprogramming the Power Controllers and made some power checks by hand and we still have the same problems.

Steps made in reprogramming the ZC706 Power controllers (just for one of the scripts)
1_ plug the USB interface adapater
2_ turn on the ZC706 board
3_ Load scripts in the digital power manufacturing tool software
   3.1_ Load 1v8_vadj --> Scripts --> ZC706_Rev2_0_TI_Addr101
4_ Run Script
   4.1_ Start
        _ When the device serial number pop-up appears, press OK
        _ When cycle board power now pop-up appears, I turn the board off, then on (here I've tried
           several times with different wait values) and press OK. Here always appears a failure warning.

In the attachments I put again the log file of the  ZC706_Rev2_0_TI_Addr101 1v8_vadj test.

Am I doing something wrong?

Some help would be really appreciated.

Thank you.

[brunovali · Accepted solution]

Finally I've found the problem with the board.

It was because a failure of a component in the ZC706 board we are using, one of the bulk capacitors of the 1.0V power domain was "punched". The power rail #1 was powered off by the UCD90120 because of an overcurrent issue caused by the "punched" capacitor in the start up.

 

C299 was the failed component, after replacing it the board started to work OK.

Are this boards being made with bad quality components? They are quite expensive as to have this kind of problems.