OpenScope MZ Calibration Issue

[CatFoodMan]

I received my new OpenScope Mz – very exciting!

I installed the Agent (V0.3.7), hooked up the device, started WFLand it upgraded the firmware from 0.185.0 to 1.4.0 without issue.

I launched the calibration, connected the indicated wires with short (3”) jumpers, and it ran the calibration, but couldn’t find any storage in which to save it – “No Locations Found.”  I retried the calibration but it just ran the blue bar all the way to the right and then just sat there.

OK, close the browser, release the port, and try again.  Still can’t find any storage.  Hmmm.

Reboot!  OK, now it sees the Flash Storage Location, but when I click save it says “Error saving calibration.  One or more channels fell back to ideal values.  Rerun calibration.”  Clicking Done sends me to the main WFL screen.  Opening the Device Manager again pushes me back to Calibration.  If I rerun the calibration in the same session, I again get “No Locations Found.”

Disconnect, release, restart, gotta calibrate again, now I see Flash, but still cannot save: One or more channels fell back, etc.).  ~sigh~  If I don’t save it I go back to the configuration screen – but it identifies the board as “uncalibrated."  Evidently if I don’t save the values it’s not considered calibrated.

Installing an SD card doesn’t help – If it sees Flash it’ll see SD0 (and likewise if it doesn’t see flash it doesn’t see SD0,) but can’t save there successfully, either. 

Let’s try something else – WIFI!  Click “Setup WIFI”.  It doesn’t detect any wireless networks, yet there are 5 SSIDs originating from 2 access points and a router just 10 feet away.  My tablet and phone can see 10+ networks in range: OpenScope doesn’t seem to see any at all.  If I manually try to add one and put in one of the 2.4GHz networks, I get “Error setting wifi parameters.”

Help!

-Bruce

[Kristoff]

Hey CatFoodMan,

I split this into a new thread since this calibration issue sounds different than the issue in the other thread.

Can you uninstall the Digilent Agent (if you don't you may get duplicate entries in programs and features due to a bug in some beta builds) and then install v. 1.0.0 from here.  I updated the link that pointed you to 0.3.7 in the getting started.  I don't really expect this to resolve your calibration issue but you should definitly upgrade to 1.0.0.  I'll talk to Dharsan and Keith and see if they have thoughts on what might be causing the calibration issue.

If you go to the device configuration page (click the ... button and then configure on the device card) does it report that the firmware is 1.4.0 (I want to confirm that the firmware update did indeed complete successfully)?

Thanks!

-Kristoff

 

[sudharsan.sukumar]

This sounds like the device is not responding to the commands WFL is sending to list storage locations and save the calibration. What browser and OS are you using? I just uploaded 1.4.0 and calibration seems to work. Like Sam said, I'd update the Agent and then remove and re-add the device to WFL and make sure the firmware version displayed on the device configuration page is 1.4.0.

If you're still seeing issues after that, go to the main page (Device Manager Page) and then open the side menu and click settings. Open the advanced menu and click change console log and select "Local Storage". This saves debug logs to your local storage. Now, try and calibrate the device until you see the error. Once you run into the error, get back to the settings page and under advanced click "Export Log". This should download a .txt file that contains some of the debug information we collected and should give us some info on what is actually happening. Also, the local storage debugging will only run during the session you enabled it. If you refresh WFL and still want to store the debug info, you'll have to go back to the settings page and set it to local storage again.

Thanks,

Dharsan

[CatFoodMan]

Thank you for your prompt replies!  Well, things are a little better.

I uninstalled agent 0.3.7 and installed 1.0.0.  The Configuration page does report firmware 1.4.0

First attempt at Calibration followed the same pattern - "No Locations Found".  Second try never completed.  When I go back to the configuration page I see that it has connected to the wireless, using the same settings I tried and (seemed to have ) failed to set yesterday.  Another new thing - along with the solid blue LED instead of blinking (Yay, wireless) after the calibration attempt the other three LEDs blink once, then #2 blinks 5 times, pause, repeat.  A power cycle of the device  shows the #4 LED blinks for a few seconds, then comes on solid, then the other three start their blinking pattern.

So upgrading the Agent to 1.0.0 must have done something: I have Wireless now and get the patterned LED blinks (I'm not sure how...) but still no calibration happiness.  I've attached a pair of logs - one after a hung second calibration attempt (#1) and one after a Calibration attempt that found no storage (#2.)

I'm running Windows 10 Version 1703 Build 15063.413.  I'm using Chrome 64-bit Version 58.0.3029.110, although as I was checking it updated itself to 59.0.3071.109.  The update didn't change anything, though.

Tried on other method: I disconnected and connected the device to a external USB power source and connected WFL to it via WIFI.  Ran a calibration, this time it saw the Flash but couldn't save it.  I've attached that log as #3.

Thanks!

-Bruce

OpenScopeLogs - 1.txt

OpenScopeLogs - 2.txt

OpenScopeLogs - 3.txt

[Kristoff]

Hey CatFoodMan,

 

Once you are connected to Wifi the LED's blink the last octet of the OpenScope MZ's IP address.  We found this super helpful when trying to connect from WaveForms Live without a USB connection to check the device IP.  We have a little documentation on this in the reference manual here and adding more.

Dharsan will parse through those logs for you.

-Kristoff

[sudharsan.sukumar]

Hey I got to look at the logs. Definitely looks like there's something going on when the OpenScope is calibrating. I talked to Keith who wrote the calibration and he said this:

"Look at the LEDs while calibrating. It takes about 30 sec to calibrate, for about 28 seconds the RED LED should be solid ON. If the RED LED goes off after about 10 seconds or so, calibration is failing. This would happen if DC is not tied to OSC properly; OR IF some other circuit type PCB error on the feedback circuit; which might slip through our manufacturing test."

Edit: More info:

"The blue LED should be solid ON while WiFi is up and running, or blinking if not connected to a network, but should be OFF during calibration. When he does the save, is the Blue LED blinking or solid ON? if not, the MZ probably either crashed (LEDs stuck in position they were when crash happened), or asserted (All LEDs ON). This could happen if I get a math error (divide by zero) which could happen if one of the feedback circuits are not working and ALL values are the same when reading the ADCs, I'll probably get a divide by zero fault and crash the MZ."

WIll you try that out and let us know? If that's not it, I have a couple of things we can try to figure out what is going on.

-Dharsan

[CatFoodMan]

OK, so here we go.  When I plug in the USB cable, after about ten seconds the blue led will start flashing: A few seconds of that and it turns solid and stays that way.  The other three will blink once, then #2 will blink 5 times, and they'll repeat.  Last octet is 161, so that now makes sense.

So I start Chrome, open WFL, indicate the locally connected device, it says it's an uncalibrated device.  I check to make sure the proper pins are shorted (pin 23 red shorted to pin 29 orange and 24 white to 27 blue, assuming D8 is pin 1 and P5 is pin 2 as per the board labels,) and launch calibration.  The blue light goes out and the red one starts blinking.  It will blink for about 20 seconds or so, until the blue bar is 60% or so of the way across. Then it will go out, L05/L06 will dance a bit, and at about 95% the blue light starts flashing again, then it goes solid and about 15 seconds or so later the screen comes back and wants me to save the calibration, but (usually) doesn't have any locations.  I've used front panel, back panel, and USB2.0 hub USB ports.

At save time, whether it finds the flash or not, the blue light will be on solid by that time and the other three will be indicating address.

if I plug into a USB power source and calibrate via WIFI I have a much better chance of it seeing the Flash, but it still won't save because "one or more channels fell back..."

 

[sudharsan.sukumar]

OK so what's happening is the OpenScope is losing power while trying to calibrate and then automatically restarts.

This is why there are connectivity problems and also why sometimes WFL can read the storage locations and sometimes it can't. While the OpenScope is starting up, it will not be able to respond to the commands. When the OpenScope (OS) starts up, LD5/LD6 flash because the firmware is printing out the OS info. This also explains why WFL is complaining about the calibration falling back since when the OS restarts, the calibration is reset if it wasn't saved. Further proof this is the issue is that when the OS restarts, the Digilent Agent is no longer able to talk to the device and you'd have to release it or restart WFL to talk to the device over the Agent again. I would try a different USB cable and maybe a different power supply (I use the one I have to charge my phone). If you still can't get it to calibrate, it sounds like you might have received a faulty OS and we'll send you a new one 

It sounds like you're connecting the wires correctly, but as a sanity check here's a picture of my set up.

Thanks for your patience debugging this! Very helpful write-ups made it easy to figure out what was happening.

-Dharsan

[CatFoodMan]

I've been working in various computer-related support and consulting disciplines for nearly 30 years now, so I try to make things as easy as I can for both sides

Well, I've tried 4 different USB cables in various ports, including using an external powered USB 2.0 hub.  I've also used several different stand-alone USB power supplies, from 1A up to 2.1A, across WIFI.  I even used the power supply on my workbench (a GW GPC-3030D) with a USB breakout adapter - without success - same issues as before.  ~sigh~

My initial calibration hookup was very similar to yours, but here's the setup I'm using now: it's easier to move around when I'm changing cables and such (hopefully it's right or I'll be awfully embarrassed...)  Also, I printed a spiffy case for it: the STLs you provide printed up beautifully using the guidelines given.  if only I could put the unit into it and use it...

[sudharsan.sukumar]

You should be getting an email with RMA details. Let me know if the email you used to sign up for the forums doesn't work.

Thanks again for your patience,

Dharsan

[CatFoodMan]

The new one was waiting for me when I returned home this evening – Thank you for your prompt service!

 Just like the original one, it came with Firmware 0.185.0 needed a firmware update to 1.40.0, which was carried out without issue.  However, unlike the original one, this one was able to calibrate and save successfully. YAY!.  Also, this one was able to enumerate WIFI SSIDs instead of having to put one in manually.  I've been able to do the measurement test using the internal waveform generator and capture some external signals successfully, as well.  I think we're doing much better here now.  I've even put it into its case...

I’ve got the old one packed into its original box and ready to ship back to you.

[Kristoff]

Hey CatFoodMan,

I'm very happy to hear that the new unit is working!  I'm sorry that you got a bad unit and Dharsan, Keith and I are very interested to get the bad unit in house and do some more debugging. We've all been saying that if someone had to get a bad unit we were lucking it was CatFoodMan.  We really appreciate your patience and thorough explanation of your troubleshooting and experience.   

Thanks!

-Kristoff

[CatFoodMan]

Glad to be helpful - I've been on your side of the support equation.  I'd be interested in knowing what you find with the unit.

By the way, I've found that if I click on the "Go to this post" link in any of your your notification emails (links to  to https://mandrillapp.com/track/click/...etc...) I get sent to a forum page with a large bang graphic stating stating "Sorry, there is a problem - We could not locate the item you are trying to view. - Error code: 2S136/C"  Not a big deal: i just then hit the message envelope in the corner and can go right to it from there, but something doesn't seem to be working as intended there.

I'm looking forward to seeing just what this little thing can do.  I see there's an WFL Android app?  Hmmm... 

[Kristoff]

For anyone that is interested here is Keith's analysis of the RMA'd board:

----------

 

Hi,

 

I reviewed the returned OpenScope and there is indeed a hardware failure. On the DC output 2 opamp, the voltage at pin 5 on the MCP6H82 was measured at zero volts instead of the VREF1V5 1.5v. On the reference side of R100 there is indeed 1.5v, on the opamp side it is 0v. So either the input pin of the opamp internally shorted out due to a failure (most likely case), or the cap at C68 shorted to ground. I did not unload the cap to determine which as the circuit is clearly failing; but the opamp is providing a constant -1.0v no matter what the PWM signal is, so it is my very strong suspension that the opamp failed.

 

 

 

During calibration to determine the coefficients of the DC circuit, the PWM circuit is run to generate an output of +3 to -3 volts and the difference measured. Since the output is putting out a constant -1v (actually -0.96v), the difference will be zero, and the coefficient for the PWM term will also be zero.

 

 

Later, when the 2^nd OSC channel is calibrated we use the calibrated DC output to set the input stimulus to the OSC channel (channel 2). To determine the appropriate PWM value to generate the requested voltage we need to divide by the DC output coefficient.

 

 

Dividing by zero causes a coprocessor instruction fault and causes the MCU to reset itself; and thus why it looked like the processor browned out, it really didn’t but the effect is the same, the processor reset itself.

 

This is an infant death condition on the opamp. This clearly passed during the manufacturing test as we calibrated and the calibration would have fail if this had occurred at the factory. Clearly the opamp failed shortly after it was released from manufacturing.

 

The opamp itself has both good ESD and short circuit protection.

 

 

It is unlikely the customer did anything to cause the failure.

 

My conclusion is, nothing is at fault other than we got a bad opamp that failed shortly after manufacturing. I am please we replaced the board, and am confident this will not be an ongoing problem in production.

 

Thanks

KeithV