MAX5389 resetting by noise

[aless2056]

Guys, I made a simple project with a PIC16F877 and a MAX5389, which is a digital potentiometer. I'm just connecting the UP/Down and Inc terminals directly to the PIC.

Both the PIC and the MAX5389 have a 100nF capacitor next to the VDD and GND pin.

The PIC also drives a transistor, which drives a relay to activate a contactor.

It doesn't happen every time, but sometimes it happens that when activating the contactor, the MAX5389 resets and returns to the average resistance value it has when it is initialized (on power-up).

How can I resolve this?
Swap the 100nF capacitor for another one?
Also include capacitors in the two connections between the PIC and the MAX5389 (On the connections with Up/Down and Inc).

 

[barry]

Obviously, the problem is the relay. Is it sucking the supply down? Is your routing terrible? without a schematic, Pcb layout or any other useful information, it’s impossible to say.

 

[stenzer]

Hi,

how does your voltage distribution looks like? Does the relay and the MAX5389 share a common SINGLE trace? As @barry mentioned, a picture of your layout as well of your schematic would be of help, to give any feedback.

BR

 

[KlausST]

Hi,

I just did a search for "reset" in the datasheet.

* there neither is a reset pin
* nor a reset function.
Please explain what really happens.

***
The only function I see is "set to midpoint at power up".
If you expect this to happen, then you need to verify this. How exactly do you know it is set to "midpoint"? You have to exclude that it is any other value than "midpoint", else you (and we) are focussed in the wrong direction.

If really "midpoint" then (as long as you treat the device according datasheet specification) the only reason is a failure in power supply of the IC.
(In opposite to this IC a microcontroller may have a several sources for RESET, like power supply, software reset, external reset, watchdog reset, ...)

***
But I'm with the others: I think the root problem is a schematic or PCB layout issue.

Klaus

 

[barry]

Hi,

I just did a search for "reset" in the datasheet.

* there neither is a reset pin
* nor a reset function.
Please explain what really happens.

***
The only function I see is "set to midpoint at power up".
If you expect this to happen, then you need to verify this. How exactly do you know it is set to "midpoint"? You have to exclude that it is any other value than "midpoint", else you (and we) are focussed in the wrong direction.

If really "midpoint" then (as long as you treat the device according datasheet specification) the only reason is a failure in power supply of the IC.
(In opposite to this IC a microcontroller may have a several sources for RESET, like power supply, software reset, external reset, watchdog reset, ...)

***
But I'm with the others: I think the root problem is a schematic or PCB layout issue.

Klaus

The problem could also be a burst of noise on the inputs to the MAX5389 that drive It to one limit or the other.

 

[KlausST]

Hi,

I fully agree, thus I wrote "as long as you treat the device according datasheet specification".
Still I´m with you regarding schematic an PCB layout.

Klaus

 

[aless2056]

I noticed that on some occasions it goes back to the average bridge value, which is mentioned in the datasheet.


There is a 100nf capacitor in the power supply, on top of the IC, and today you have to change it to other values without success.

Finally, I placed two capacitors in parallel, one always with a very small value and the other with a much higher value. At the end placing a 100nF miltilayer capacitor in parallel with a 1uF tatalo capacitor. The problem almost doesn't happen.

Can I do better so that the problem never occurs?

 

[Aussie Susan]

Definitely sounds like noise on the supply lines.
You need to use a scope to check how bad the noise (spikes?) is and when it occurs.
Once you have found the cause of the noise (possibly the relay as mentioned above) you can start to fix that, rearrange the PCB tracks, add more bypass capacitors (for very high frequency spikes you may want 10nF caps) and/or look to provide a larger capacity PSU.
With the information you have provided we can only guess.
Susan

 

[barry]

It sounds like the supply lines are a big part of the problem, but there's no way of knowing that noise isn't also coming in on the signal lines, or if you've also got a grounding problem.

Does your PCB have a ground plane?

 

[KlausST]

Hi,

Can I do better so that the problem never occurs?

Yes, like requested several times: show us your schematic and PCB layout.

Klaus

 

[d123]

Hi aless2056,

"Can I do better so that the problem never occurs?"

One approach is to look for e.g. 'holdup capacitor calculations' or 'calculate DC bulk holdup capacitor', you will find it takes a few searches to find a formula you are comfortable with/want to attempt to understand.

When I find the pdf I need to, I'll post it here as it was easy enough to understand the idea and the formulas it provides.

--- Updated May 20, 2022 ---

Found it... pages 18 to 24 might help with regard to calculating bulk capacitance and high frequency decoupling.