16

I am working on a magnetic sensors design and noticed that my measured magnetic field bias shifts over time. I have studied the bare PCB and measured that it can be magnetized with a strong magnet.

I tested two blank (unpopulated PCBS). The first PCB I tested had an ENIG plating and could be slightly magnetized. I degaussed the PCB and confirmed that the magnetism was removed, and repeated the experiment to confirm. After a while I realized that the nickel in the ENIG may be playing a part in that.

So I tested a PCB with an immersion silver plating, but it also exhibited the same behavior. Once again, I degaussed it, confirmed the magnetism was removed, and re-magnetized it to confirm that the behavior I was seeing was real.

I don't think it should matter, but the ENIG and immersion silver PCBs were on Isola P95 substrate.

I then tested two PCBs with HASL finish, and they could not be magnetized. These PCBs were on FR-4 substrate.

Any ideas?

Nullified
  • 161
  • 3
  • 5
    Wow, interesting find! I'm looking forward to any answers :) – bitsmack Jan 29 '18 at 17:04
  • You should see if there are any contaminants in the silver plating, since silver itself is not ferromagnetic. – Ignacio Vazquez-Abrams Jan 29 '18 at 17:42
  • Do you know how I would check for contaminants? This is from a reputable and high quality US PCB manufacturer. – Nullified Jan 29 '18 at 17:47
  • 2
    With a sensitive enough magnetometer, almost anything that's not designed to be totally free of magnetisable stuff may contain enough ferromagnetic impurities to do what you're seeing. I had some magnetic brass once! PCB material only has to be 'good enough' to be processed, to be insulating, to be strong. There's no specification, unless you need it and pay money for it, for it to be non-magnetic. – Neil_UK Jan 29 '18 at 18:07
  • 6
    Wonder if maybe they nickel plated, then did the silver immersion. – JRE Jan 29 '18 at 18:09
  • I did confirm with the board house and they stated that there was no nickel plating before the immersion silver. So at least that can be hopefully ruled out. It's going to be difficult to rule out contaminants and other small trace materials though. – Nullified Jan 29 '18 at 19:06
  • Don't know what your level of motivation is or funds but when we want a quick answer for Reach environmental testing we have a board ground up into dust and analyzed. Perhaps you could get a test lab to do something similar and look for traces of other metals. Maybe scrape off a trace and have that analyzed. – Some Hardware Guy Jan 29 '18 at 19:39
  • You could try checking with the board house and see if they use tap water. – John Birckhead Jan 29 '18 at 19:42
  • 1
    There is a process with silver above nickel plating too, [see](https://www.sciencedirect.com/science/article/pii/S0026057613701715). – Uwe Jan 30 '18 at 17:51
  • 2
    You need to change only one variable at a time. Try HASL with Isola P95 to isolate if it is the ENIG. – Joel Wigton Jan 30 '18 at 18:20
  • What levels of magnetization are we talking about? – Attila Kinali Feb 12 '18 at 13:12
  • Nickel plating. A fair number of components have leads or leadframes that are magnetic too, seemingly randomly. You can buy non-magnetic resistors easily but it's a bit of a crap shoot with the other parts. – Spehro Pefhany Feb 14 '18 at 10:58
  • 1
    Update on my findings. I got a hold of a sample of Arlon 85N with ENIG plating from another PCB manufacturer, and was not able to detect any significant amount of magnetization. I have since ordered my PCB design on Arlon 85N in immersion silver and will update here with my findings once I receive it. – Nullified Feb 14 '18 at 16:23
  • To expand on Sphero's comment; the vast majority of TI parts have a finish of NiPdAu which clearly can be magnetised. This is not an unusual finish and has the advantage of being compatible with both leaded and lead free processes. – Peter Smith Feb 15 '18 at 17:31

1 Answers1

4

ENIG plating translates to "Electroless nickel immersion gold" (for neophytes).

This means gold plated copper surfaces using nickel as intermediary layer between copper and gold.

Then, if your PCB is slightly magnetized, could be the nickel plating, because nickel is a ferromagnetic material.

https://terpconnect.umd.edu/~wbreslyn/magnets/is-nickel-magnetic.html

EDIT with more info: With Silver Immersion, the type of silver used is silver sterling: http://www.multicircuits.com/assets/content/files/immersion_silver.pdf

This kind of silver mainly contains silver alloy with copper (7.5%), but there is no industry standard saying that sterling silver has to be this specific alloy, some metal processors will use small quantities of other metals that could be magnetic.

Hope it helps.

José Manuel Ramos
  • 625
  • 4
  • 9
  • This sort of misses the point. He KNOWS that ENIG has nickel and is therefore magnetic. The question is why the immersion silver plated boards are also magnetic. – JRE Feb 14 '18 at 12:58
  • https://www.etsy.com/forums_thread.php?thread_id=6190245 `Pure silver would not be attracted to a magnet, but with sterling the other 7.5% can be any undisclosed (legal) metal. There is a high possibilty that your silver was simply an alloy of silver and a ferrous metal. This would explain the attraction to a magnet. It is still sterling silver as there are no specific rules stating that sterling silver HAS to be a mix of silver and copper.` – José Manuel Ramos Feb 15 '18 at 16:31