BBB Broken? SPI signal is at 1.8v

I’ve been building a custom cape for a robotics project and one of the chips I’m using is controlled via SPI. I’ve used an oscilloscope to validate that the SPI is working as expected. However, two days ago I noticed that the chip stopped responding and after scoping the SPI signal I can see that the BBB is sending the SPI data pulses at 1.8v. The SPI signal is still happening… and the clock signal is still at 3.3v. It’s just the SPI data line that is only peaking at 1.8v.

So, I’m wondering if I’ve done something bad to my BBB or if I’ve somehow triggered a feature that I don’t know about yet. I’m attaching a photo of the oscilloscope screen that shows the issue.

DS1Z_QuickPrint3.png

No way for me to tell what you may have done, but 1.8V is not good. Any chance you can provide more information like the pin number and connector you are using?
What do you have connected to this pin?
How is that device powered?

Gerald

Thanks for responding Gerald. The scope capture above was done with no cape. It was taken from SPI0. The chip has one-way communication so I’m only using SPID1 (P9_18 from memory). The clock is coming from P9_22. Chip select is P9_17. When I get home tonight I’ll try the same from SPI1 and see if it’s affected as well.

Forgot to add info about how the cape is powered. The chip I’m using on the cape is an AD5206 digital potentiometer (10k). I’m using the 3.3v rail to power the SPI side and an external 5v (4.9v measured) supply powers the pots. However, I’ve been getting voltages from the pots that aren’t what I’d expect, so a few days ago I disconnected the 5v supply and ran the BBB 3.3v rail to a single pot just to see if the resulting voltages would make more sense. Maybe that was what did it?

I've been building a custom cape for a robotics project and one of the
chips I'm using is controlled via SPI. I've used an oscilloscope to
validate that the SPI is working as expected. However, two days ago I
noticed that the chip stopped responding and after scoping the SPI signal I
can see that the BBB is sending the SPI data pulses at 1.8v. The SPI
signal is still happening... and the clock signal is still at 3.3v. It's
just the SPI data line that is only peaking at 1.8v.

Had a chip (Epson S1D13781) with an SPI interface. Had a similar
problem when sharing the SPI interface with another chip. Bottom line
was that, unlike the standard SPI chips where the chip gets completely
off line when not selected, this chip *always* (the Epson) drives MISO
line low.

I was getting about 1.8 volts or so maximum voltage out of the
paralleled chip because it was trying to pull up a driver that was
stuck at zero.

Not sure that you have exactly the same problem, but is the chip
you're driving somehow trying to drive this line (and shouldn't)?

Harvey

Power sequencing is the key. How are you doing this?
If you are applying the 5v to the pots first you may be causing the
issue and blowing the processor pins.
All pins on the processor need to be isolated from everything until
power rails are stable.
Drive a input pin when the processor is not fully powered on and you
risk destroying the drivers in the chip.

Schematics of your design would help identify the problem.

There are no other SPI chips on that bus. Since the AD5206 is permanently a slave on the SPI bus I would think it would not try to drive the line.

I suspect you’re exactly right Wulf Man. I was not really paying attention to the order I applied the power and I was assuming that the POT voltage would be isolated from the digital control side of the chip. But I also assumed that if I blew the drivers that I would no longer get any signal from that pin. Is it normal to get a reduced voltage? I suppose I could almost fix the issue with an opamp (except I’d be worried that the defective pin would completely go dark someday.)

If I really did blow those pins then I’ll get a good signal out of SPI1 when I test it tonight. How would I prevent this in the future? Maybe some clamping diodes to keep the voltage on the digital pins in range?

Walker

SPI1 shows the same 1.8v output. I will attempt to move to another BBB (a BBG actually) and if things work there then I’ll assume I toasted the SPI on this board.