Hello! Faced a small problem related to the FPGA family Cyclone IV E EP4CE55 in the BGA-484 case. 1 of 2 FPGAs on the same boards and with the same project is heated like an iron. I power the FPGA with separate power sources. There where consumption 30mA
Inquiry: Hello! Faced a small problem related to the FPGA family Cyclone IV E EP4CE55 in the BGA-484 case. 1 of 2 FPGAs on the same boards and with the same project is heated like an iron. I power the FPGA with separate power sources. There where consumption 30mA on 3.3 V is not heated, and 250mA on 1.2 V. There where 800mA on 3.3 V is heated, on 1.2 V did not look. The tester looked the resistance between power and ground. In a normal case on 700 Ohm on 3.3 V and on 1.2 V. There where heats up 5-6 Ohm. At the same time, the overheated FPGA manages to function. Tell me, please, what could be the reason?
I suspect you have a build fault on the board that is heating up. The fact that it is working is (perhaps) not surprising depending on the nature of the fault.
Is it the FPGA that is the source of the heat, not a nearby power supply? Does it heat up even with the FPGA unconfigured?
It won't function for very long if it's overheating in the manner you suggest. I recommend having it removed and replaced. If you can check the board before having the new part fitted I suggest you do.
Alex has a point. This looks like power short scenario.
- You can probe and measure the power rail to check for voltage drop as sign of power short issue
Some of the factors are either
- On board power rail is shorted
- You can try check for series inductor or resistor component that you can removed to cut off the power for ease of debug
- Another typical failure will be on the blow capacitor that sitted between power and ground that caused the short.
- Basically you need to slowly isolate potential failure parts on your board to find out the root cause
- Another possibility is faulty on board power regulator that you need to replaced
- FPGA internal power rail is shorted which you need to replaced the FPGA
Alex and d lim, thank you for answers) FPGA heats up regardless of project and even in moment his configuration. Removed a number of capacitors on feeds 1.2 and 3.3 not help. Watched thermal imager warming up, hoping to determine the area of the short circuit. The heating starts from the FPGA core and diverges in different directions, while the remaining capacitors on the power supply remain cold. As for the resistance measurements, I will correct the values, since I initially made a mistake. In the normal case, in the off state, the resistance between 1.2 and 3.3 is 500 Ohms. Between 1.2 and GND about 125 Ohms, and 3.3 and GND about 150 Ohms. In the case of overheating, in the off state, between 1.2 and 3.3-11 Ohms, between 1.2 and GND about 5 Ohms and between 3.3 and GND about 6 Ohms. Initially, the second Board also worked fine, but it was debugging other chips and there were many other tests. On malfunction drew attention only then, when observed, that nutrition 3.3 with regulator through coil inductance in 330 MGN has fallen until 2.9 Volts.
Many times the Board was shortened and burned different chips, but each time it was possible to restore its performance. It seems to me that in FPGA something burned down as a result there was a short circuit on internal feeding buses. It is difficult to replace it, so it remains to work with it in this mode. If we can still determine the cause, I will report it. Thanks)
Yup, looks like a power short condition as 3.3V power rail now become 2.9V only
- If you have rule out everything on your board and leave with FPGA only then there is still 2 possibility
- Either FPGA internal already damaged/shorted or the bad soldering is causing the short on FPGA balls. You can use X-ray machine to check for solder ball short condition.
Ultimately you still need to fix it as the situation is going to get worsen and eventually kill your board.
- Even if you board still alive the voltage drop is going to affect your design functionality as well as it's dropping beyond normal operating range.
One more dummy question, I hope this is true power short issue and NOT due to
combination of FPGA cooling design issue + on board power regulator exceed current limit issue
- This will also caused voltage drop as well
dlim, we will have the opportunity to do an x-ray next year, I hope that the FPGA will live up to this happy moment) the regulator Itself and the FPGA are spaced quite far and in my opinion the heat dissipation polygons cope with their work.
I am setting this case to closure first as it's idle for long time.
Feel free to take your time to debug and you are always welcome to file new case later to resume the discussion.