Everyone,
I have an MSP430F5638IPZ on a custom board that I have been programming for the past couple months. Everything has been working great until one day I tried to "green wire" a high-speed crystal onto two pins. After that, I plugged in the battery and noticed that my power plane had gone from 3.0V to 0.92V. I ordered more MSP's thinking that I blew up the processor. I installed a new MSP, double checking orientation and solder bridges with another co-worker and still had the same problem. Thinking I fried that chip as well, I cleaned my work space, laid down an ESD mat and leashed myself in. I also turned down the soldering irons to approximately 260 C (I am hand soldering under a microscope). I soldered the chip and reviewed my work again, same problem - same voltage! I am drawing 60mA from my regulated power source.
To give more background, when I disconnect the MSP from my board, the power plane is at 3.0V. This means that the MSP is the source of the problem right? It is not static, it is not heat, what is it? Has anybody had any experience with these symptoms?
David Engineer when I disconnect the MSP from my board, the power plane is at 3.0V. This means that the MSP is the source of the problem right?
It's also possible that the crystal causes a massive increase of current. What about the load capacitors? maybe there is a short? Are there load crystals (if not, the crystal might overdrive the oscillator circut, causing excess current too. not likely, but possible)Also, some high-speed crystals (usually above 10MHz) may be overtone crystals. They aren't easy to handle and tend to fall back to baseband (f/3), with increased current.
_____________________________________Before posting bug reports or ask for help, do at least quick scan over this article. It applies to any kind of problem reporting. On any forum. And/or look here.If you cannot discuss your problem in the public, feel free to start a private conversation: click on my name and then 'start conversation'. But please do so only if you really cannot do it in a public thread, as I usually read all threads. And I prefer to answer where others can profit from it (or contribute to it) too.
Jens-Michael,
Thanks for responding, I know it is not my code because the MSP's are new out of the package and have yet to be programmed by me. Also, from your description it is not a cracked trace because the voltage for the whole system goes down to ~0.9V @ 60mA (measured). To clarify what I mean by that, I have a 2 AHr Li-poly battery connected to a switching voltage regulator connected to the Vcc plane, I tested the current between the voltage regulator and Vcc. It also isn't the high speed crystal (FYI 20 MHz) because I removed that after the original chip failed.
I designed the board so I could isolate the various components from the rest of the Vcc supply, I have done my tests with and without the components "active" and get the same results. I am at a complete loss on this issue.
David
One more symptom, I soldered a new MSP to the board and before applying power I determined there was a short between Vcc and ground. There is a problem in the board for sure! Does anybody know what pins would be shorted to cause the symptoms I am seeing?
Before you solder the chip, check the all 100 pin pads on the PCB and list all Vcc and all Vss pins.
I think 11, 25, 64 and 89 should be Vcc. 12, 15, 26, 29, 63, 83, and 90 should be Vss. See Data-Sheet.
Did you remove the added connections to the crystal?
As long as no sufficient supply current is provided, VCC to GND forms a virtual short. There should be only a small voltage (a typical P/N diode voltage) left. Depending on how much test voltage your multimeter applied and how it 'detects' a short, it may be a false alarm.
You should test the VCC pins against each other without any MSP applied.
David Engineer the whole system goes down to ~0.9V @ 60mA (measured)
Thanks everyone for the replies. I completely disconnected the 20MHz crystal after the original working MSP stopped working. I also tested all of the pads after I removed the MSP and found that everything looked correct. I tested the pads by using the continuity test on my multimeter by touching the negative probe to ground and the positive probe to the pads. After testing all the pins, I then switched the negative probe to the Vcc supply pin and touched the positive probe to the pads once again. All of the pads seemed to be operating correctly with no shorts.
The way I have been verifying if the MSP is working (other than Vcc voltage) is to measure the resistance between the power and ground planes. When the system was working correctly, the resistance was approximately 700 ohms, when the MSP is not working correctly, the resistance is 24 ohms. I think I have described everything I have done to troubleshoot my problem in these posts, and based off of what I have found, I am either applying too much heat during soldering or I am receiving defective chips. I have a hard time believing that I would receive so many "defective" chips and I have an even harder time believing that I am applying too much heat. I have looked at the maximum temperature during soldering and I am not exceeding that rating.
I am going to continue to try and figure this out, and if I find a solution, I will be sure to post it. Thanks again to everyone posting, I appreciate the help.
Just a hint , Try soldering couple of PINS at a time and keep checking VCC , you may find the problem.
We once bought a whole tray of Atmel micros. All were bricks. Don't assume you have good parts.
David Engineer All of the pads seemed to be operating correctly with no shorts
24 Ohms is indeed quite low (even though a resistive metering result is onla valid for resistive loads - for complex loads and semiconductors, a sweeping signal and scope view is more 'telling')
Well, this reminds me of an oddity I encountered some 25 years ago with an EPROM board for the C64. One of the PCBs, out of a batch of 200 pieces, was stopping the 'realtime' clock of the C64 when inserted, even with no parts soldered. Strange thing: the rest of the chip that provided the clock count was working 100% okay, and the clock count source was just a rectified 50Hz signal from the transformer that also supplied the whole device. And there was no electrical connection of this signal and the slot where the PCB as inserted.I never figured out what happened, but I think I stil have this PCB somewhere.
I wanted to let everyone know that I figured out the problem with my project. When I was creating the MSP in my schematic program, I mislabeled pin 83 as AVCC3 instead of AVSS3. When I cut the power trace and grounded the pin it worked!
Just a side note - I thought it might have been my DC-DC converter, so I connected a power supply to my power plane and dialed it in to 3.0V. I had ~1.025 Amps running through the MSP, it got hot very fast!
Thanks to everyone that helped!
David Engineer I wanted to let everyone know that I figured out the problem with my project.
However, I wonder how this could ever have worked, as you said the board was working until you wired the crystal. Except, of course, if this mislabeled pin was also missoldered (and maybe isolated from the PCB) until you touched the board for the crystal.
David Engineer I had ~1.025 Amps running through the MSP, it got hot very fast!
I had the same question, I am soldering a lot of components by hand - even the MSP. I am guessing that what you said is exactly what happened. I haven't tried to use the MSP that I passed 1 Amp through, but I will give it a try.
David Engineer I had the same question, I am soldering a lot of components by hand - even the MSP. I am guessing that what you said is exactly what happened.