pmodDPOT PN: 410-239

[deez2009]

Hi,

I am using an ESP32 and a pmodDPOT to control the output of an inverter.  This has worked fine for me for the last few years, except that about once a year, I have to replace the pmodDPOT.   This has happened twice now, the first time, I didn't think much of it because I assumed it was just bad luck.

 

There are no obvious faults with the device (no magic smoke, discolor, smell, etc).  In fact, it still works, except the output is random--if I set the value to 128, then increase to 129, the output can double or triple.  If I power cycle it, and try again, it might not happen immediately, but usually within a few mins, it happens again.

 

If I swap the pmodDPOT for a fresh unit, with no other changes, it works fine again...until next year.

I just ordered a bunch more of these, but would like to know if there is a way to fix this.

 

Thanks

[JColvin]

Hi @deez2009,

I personally have not heard of or seen this particular issue you are describing with the Pmod DPOT (or heard of any issue with the Pmod DPOT to be honest) on the Digilent Forum or on Analog Devices FAQ (https://www.analog.com/en/education/education-library/product-faqs/ad5160.html). The AD5160 uses volatile memory rather than the non-volatile memory, so there presumably shouldn't be any wearing out of the internal latches or something like that. The "Pmod" portion of the DPOT simply places the IC on a PCB and adds the recommended capacitors from the AD5160 datasheet to keep the power supply stable so there shouldn't be anything to degrade on that end of things.

Presuming you are operating at nominal operating conditions specified by the datasheet, https://www.analog.com/media/en/technical-documentation/data-sheets/AD5160.pdf (supply voltage between 2.7 V and 5.5 V, respecting SPI timing, not attempting to drive current through the terminals {datasheet makes an effort to warn a few times about not subjecting the device to pulse current which could lead to degradation or destruction of the internal switch} reasonable ambient temperature, etc), I can't think of anything plausible that would result in what you are experiencing.

Maybe there's something I'm missing from your setup that could be leading to this issue you are describing, but it's a digital potentiometer with volatile memory so there isn't very much that can be debugged to begin with.

Thanks,
JColvin

[deez2009]

Yeah, I can't think of anything either.  I am driving it with an ESP32, output pin voltage is 3.3V so it's within spec.   This lives in my shed in California, so the temps are fairly mild and is in a case and not exposed to elements.   What else could be missing from my setup, it's literally an ESP 32 connected to the Pmod, the output of which is connected to my inverter.

The first time it happened, I assumed it was just bad luck.  The fact that a second one died made me reach out to you guys for some help.

 

[deez2009]

correction (not that it really matters), I am using a ESP2866 not ESP32.

[JColvin]

Agreed that nothing seems abnormal. You are operating within the voltage spec, presumably operating it for purely a digital output (well technically analog, but the device only supports continuous current of less than 1 milliamp), and are in a clean environment.

Following the 'once is a accident, twice is a coincidence, three times is a pattern' rule of thumb, we're getting real close to the pattern, but from a practical point of view, I don't know what Digilent could do to improve the design. I suppose there is the option of switching to a different digital potentiometer IC, but based on the product page for the chip and revision history for the associated datasheet, Analog Devices has not made any hardware changes to this IC since 2009 which is before Digilent started making the Pmod and ADI continues state the product is appropriate for new designs, so they at least believe there is no issue with the IC.
From a similar practical point of view, if the product works in your system for around a full year, I would have expected any weird issues with timing or something to have manifested sooner, though if your system gets periodically reset throughout the year, such theories about the source code drifting off into an undefined state don't hold up.

All that to say that I am uncertain of what the root cause might be, but at least from Digilent's perspective, the likelyhood of us dedicating a setup to run and monitoring a continuous test on this particular peripheral for many months is next to none.

I wish I had better news, but unfortunately I'm at a loss of what else you can try.

Thanks,
JColvin

[deez2009]

Okay, we can leave this post here if it happens again next year, or if others are having the same issue, but just haven't bother to report it.