18 Replies Latest reply on Sep 26, 2021 9:04 AM by scottiebabe

    What to do when a review doesn't go as planned

    scottiebabe

      I am currently having an issue reviewing the MAX77714 EVK where the 32k XO isn't running. Its not really a hardware issue as I can get the XO to run just with the capacitance of 2 10x scope probes. The issue has to do with configuring the PMIC to enable its on die load capacitors for the 32k oscillator.

       

      I can't even try to enable the load capacitors because the evaluation software doesn't have the correct register map for the PMIC. The evaluation tool locates the 32KLOAD bitfield in register 0x31:

       

      When actually the bitfield is in register 0x30 (from the datasheet: https://datasheets.maximintegrated.com/en/ds/MAX77714.pdf ):

       

      Actually as I write this, even if the 32KLoad register was addressed correctly in the evaluation software it is readonly so I still wouldn't be able to change it. So why isn't it booting with a value of 0b00 instead of 0b10? Wrong OTP config?

       

      So, my questions is what do reviewers typically do in this situation?I don't want to write a scathing review over something so silly.

       

      I suppose I will start by seeing if maxim will confirm the issue with the EVK and see if they can provide a software fix or want to send new hardware.

       

      One of the uncertainties I had with my last review was determining if I am reviewing the silicon product or the evaluation platform. In this case, am I reviewing the MAX77714 or the EVK and software for the MAX77714? Or, a bit a both? Or, is it up to the reviewer?

       

      Thanks

        • Re: What to do when a review doesn't go as planned
          Jan Cumps

          When I have a road test with issues, I usually log a ticket on the supplier's support system. I then write about the progress of that ticket in my review blogs - good and/or bad experience.

          if it goes really bad, I send  heads-up to the road test organisers - so that they can inform their contact at the supplier.

          edit: I add a link to the road test or my blog in the ticket, when I create one specific for that.

          9 of 9 people found this helpful
          • Re: What to do when a review doesn't go as planned
            dougw

            It is certainly frustrating when something like this stalls a review or causes a lot of time to be spent on something trivial.

            Manufacturer involvement would seem to be the only real solution here, as Jan points out.

            Interesting work-around though - running with scope probes attached... (Does this mean you could kludge on physical caps?)

            3 of 3 people found this helpful
              • Re: What to do when a review doesn't go as planned
                scottiebabe

                Yes, normally this is a good way to drown out a 32kHz oscillator, lol. The board designer did include not fitted load caps, so I can make the on board 32kHz XO function, just not what I was expecting from a Maxim EVK...

                 

                At this point I'm still not sure who made the mistake.

                 

                There are also a number of silly unit typos in the datasheet (which does happens, just annoying to see):

                 

                1 of 1 people found this helpful
              • Re: What to do when a review doesn't go as planned
                jc2048

                 

                The datasheet says the setting can be changed through a test register for prototyping purposes. So it's possible that the software tool you're using is meant to be doing that behind the scenes, though it does rather look like they might have taken the tool for another part and not changed all the things that they should have done.

                6 of 6 people found this helpful
                  • Re: What to do when a review doesn't go as planned
                    scottiebabe

                    Nice discovery in the datasheet! Hmm, I wonder where this test register is. Also interesting is they don't list the 32KLOAD_OTP value in the OTP table, so the load caps may be disabled by default in this OTP version (CID4=0x01).

                     

                    This is certainly not what I had expected to see, the 32k oscillator fail to start. Fortunately, there is an on die ~32kHz RC fail-safe oscillator. 

                    1 of 1 people found this helpful
                  • Re: What to do when a review doesn't go as planned
                    Navadeep Ganesh U

                    I too see the same thing happening wherein the register values are incorrectly mapped against the capacitance value choices. But I am curious, why that value needs to be changed/configured? Seems like internally it's taking default capacitance value and clocks fine. Simply for testing purposes? Alright.

                     

                    Apart, there are minor bugs where the whole 'register explorer' tab GUI minimizes when the configuration tray is pulled down(i saw it happening twice. now, not finding where exactly was it,......murphy is real)

                    1 of 1 people found this helpful
                    • Re: What to do when a review doesn't go as planned
                      14rhb

                      Like many other roadtesters and yourself, I too have run into issues with development boards before. Progress is very slow and each step seems to be a struggle - but keep trying. Use the manufacturer's forums and log a question early (as some take a while to gather useful replies) - perhaps drop a line to state you are evaluating this product for another company. Posting into E14 as well may get useful responses from your fellow roadtesters - after all you are testing the same product and not in a competition against each other - or others that have tested similar technologies may also reply helpfully.

                       

                      If you keep trying but get no where then I believe that is still useful in your final roadtest review. I've seen a few roadtests where people have obviously given up far too early which were disappointing to read and other roadtests that fail but it is obvious much effort was expended in trying to get the system working. If I were a manufacturer I'd find it useful feedback from end users on what issues they had. Maybe they decide to review their online beginners tutorials? Maybe a driver needs reworking?

                       

                      If your report end up like that and later you become aware of the steps needed to progress, even after the roadtest period, I believe it is nice to do a follow up blog linked from the original roadtest.

                       

                      ...and good luck :-)

                      5 of 5 people found this helpful
                        • Re: What to do when a review doesn't go as planned
                          Jan Cumps

                          14rhb  wrote:

                           

                          If I were a manufacturer I'd find it useful feedback from end users on what issues they had. Maybe they decide to review their online beginners tutorials?

                          If I were the manufacturer and I had a campaign running here,  I'd read every post on element14 during that road test, and act.

                          1 of 1 people found this helpful
                          • Re: What to do when a review doesn't go as planned
                            fmilburn

                            14rhb  wrote:

                             

                            If you keep trying but get no where then I believe that is still useful in your final roadtest review.... If I were a manufacturer I'd find it useful feedback from end users on what issues they had. Maybe they decide to review their online beginners tutorials? Maybe a driver needs reworking?

                             

                            Agreed...  Unfortunately, on one of my RoadTests, the administrator of the technical forum responded promptly to my issue and stated he was forwarding it to the appropriate people but I never got a technical response even after a reminder and several months.  I then contacted Randall for assistance and got a sales person's response about sorrow for the issue and they would look into it. I decided to publish after waiting a while and offered to rewrite the RoadTest if they addressed the issues.  But I never got a technical response other than the forum administrator finally saying they had other pressing issues.  The product is no longer being sold.  The manufacturer has other well-regarded products.  About all you can do is document the issues and rate them fairly whether bad or good.

                            3 of 3 people found this helpful
                              • Re: What to do when a review doesn't go as planned
                                14rhb

                                I've found the same where a 'less mature' product is very soon superceded; once I recall before I'd even finished roadtesting it! The urgency of getting a roadtest completed in the allocated time period is no different to a manufacturer who may have bought this board to plug a new quick-turnaround development opportunity. Good manufacturer assistance for both cases is highly valued, and any manufacturer that can provide that level of support will outshine the others.

                            • Re: What to do when a review doesn't go as planned
                              scottiebabe

                              I am still a little bit salty that I received an EVK for an RTC (+PMIC) where the 32k XO doesn't run, but time passes.

                               

                              Thank you for the words of wisdom and support.

                               

                              At least I got to make a gif of my frequency counter, never done that before:

                              32K_CLKOUT on GPIO6:

                               

                              Running with 10x scope probes on either side of the 32k crystal serving as load capacitors:

                              I had to use a tripod, I needed both hands for the 2 probes...

                               

                              I will take the time to write up a proper blog post and will link to the blog post when I contact Maxim/ADI support.

                               

                              Thanks

                              2 of 2 people found this helpful
                              • Re: What to do when a review doesn't go as planned
                                colporteur

                                My response may just be an echo of what you have already heard.

                                 

                                If I have exhausted all public available resources for the product I to send the RoadTest review coordinator an email. In the email I describe the issue, what I have done to resolve it, what impact the issue has on the RoadTest Review and how I should proceed in light of current situation. I identify that I'm willing to work the issue if I can get contact details for the vendor of the product. I also suggest depending on response an extension to complete the work might be needed if I am working with vendor resources to resolve the issue. If I'm trashing myself it rather difficult to nail down a completion date.

                                 

                                Here are my assumptions. The RoadTest review community is in contact with the individuals working for the company who are looking for the results. They know or should know who to talk to about the specific product. Maybe you were sent a product in error (that has happened to me). They were really glad to get back the more expensive unit than the one I was testing.  As much as you hope opening a publicly available support ticket would get results, you need fast turn around. Your deadline is the RoadTest Review deadline. That is the same deadline the company is expecting. You need the experts and they should want to help.

                                 

                                I have had two RoadTest that needed vendor support. The contact for both was provided by the RoadTest support team.

                                2 of 2 people found this helpful
                                • Re: What to do when a review doesn't go as planned
                                  Navadeep Ganesh U

                                  scottiebabe  wrote:

                                  One of the uncertainties I had with my last review was determining if I am reviewing the silicon product or the evaluation platform. In this case, am I reviewing the MAX77714 or the EVK and software for the MAX77714? Or, a bit a both? Or, is it up to the reviewer?

                                  About this, I think it's expected to be a full-stack review covering the whole device platform(the chip, evaluation hardware, software). Getting all the features into test and review might not be very practical for all the RoadTests, given specified time, but exploring one or two unique ones (probably the one's which other manufacturers don't provide) and probably covering a project realated to that feature putting device to edge cases. Afterall, nothing is bounded. Reviewer is free to plan the flow of RoadTest.

                                    • Re: What to do when a review doesn't go as planned
                                      scottiebabe

                                      I'm glad I have taken the time to collect my thoughts on this issue. I will also define what I expected to be a reasonable response and level of support from Maxim. If they are able to:

                                      • Provide a software fix for setting the 32KLoad register
                                      • Acknowledge and update the datasheet to reflect what the 32Kload register is with OTP=1 silicon

                                       

                                      Then I believe they would have done everything they should do given the issue at hand. If it does turn out to be the datasheet isn't correctly documented, I can see how the EVK board designer made the "mistake".

                                       

                                      So far I haven't seen any issues with the MAX77714 silicon per se. Its been with the EVK, documentation, and software. I suppose the review is a bit of everything.