Arty A7 -35T gets hot, even when unprogrammed

[Deskarano]

Hello,

I bought an Arty A7-35T a few months ago. I'm a student in computer science, who wanted to gain some experience in hardware design. I never had any issues with the Arty until a few days ago, where I was experimenting with some configurations involving MicroBlaze and the onboard DDR3. As I was running some code on the MicroBlaze, I noticed that the FPGA chip had become very hot. I figured that this was normal, considering that there is a whole soft processor programmed in the fabric. However, when I unplugged the Arty and re-programmed it with a different, much simpler configuration, it quickly became just as hot. Now, even when the FPGA is unprogrammed, the Artix 7 chip quickly becomes too hot to touch.

Is there any way to solve or debug this problem? As I said, I am studying computer science (not electrical engineering) so I'm a bit lost as to how I could solve this issue.

Thanks!

Gregor

Edit: forgot to mention, the Arty still works! It runs any configuration that I program on it, it just also overheats very quickly and I don't want to cause long-term damage to the device.

[Deskarano]

Hi again,

Sorry for bumping this old topic, but I finally managed to get some good readings on some of the power rail capacitors and resistors. I used this diagram to identify the components to measure:

5V Rail

C170: .12V

C171: 1.79V

C176: 3.26V

C177: 3.26V

VCCINT Rail

C172: 1.78V

C173: 3.25V

DDRVCC Rail

R253: 1.42V

C174: 1.24V

C175: 3.26V

1.8V Rail

R304: couldn't find

C180: 0V (missing? see attached)

C181: 3.26V

3.3V Rail

R254: 4.03V

C178: 3.26V

C179: .8V (missing?)

 

As I mentioned above, some of the capacitors in this circuit appear to be missing. Here's a picture of what they look like:

As I look around the board, there are other components (both capacitors and resistors) that look like this. Is this what they are supposed to look like?

Thanks again for all your help,

Gregor

[JColvin]

Hi @Deskarano,

The only other thing that Jon and I can reasonably think of to try (to see if we can get different results to hopefully point to a root cause) would be to load the program (such as the button and switch project from earlier in the thread) onto the flash memory (a guide for the Arty is available here) and then power the board with an external power supply and see if the FPGA is still hot to the touch (since viewing it in Vivado will end up using the cable and USB connection again). If you happen to have a JTAG HS2 or something similar we can try that as well (since we can also view the temperature profile that way), but it's fine if you do not have that.

Thank you,
JColvin

[Deskarano]

Hey @JColvin

I tried a few different USB ports (on the same computer, no hub) and also some 5V outlet to USB adapters with various cables:

USB port 1: 4.86V

USB port 2: 4.90V

Wall adapter 1: 4.91V

Wall adapter 2: 4.95V

[JColvin]

Hi @Deskarano,

Thank you for the information; the voltage all look pretty good to me as well aside from the 5V. How are you powering your board? If you are powering it via USB (direct to PC or USB hub?), could you test that voltage? You can access it by measuring between the underside of J13 (the regulator vs USB power jumper) and J11 (an unloaded single pin next to the Ethernet port); this value should be pretty close to 5V (mine is 4.98V).

Also, did you get the opportunity to try a different USB cable or computer (to see if that affects power delivery)?

Thanks,
JColvin

[Deskarano]

Hey @JColvin 

Here's the requested measurements:

C143: 4.84V

C64: .94V

C162: 1.34V

C159: 1.78V

C152: 3.27V

All look pretty close to me except the 5V rail

[JColvin]

Hi @Deskarano,

I apologize for the delay; I thought I had sent this off, but apparently not. In the meantime I relooked at this and realized that we inquired about capacitors that are for a newer revision of the Arty, rather than the Rev C board. The newer revisions have some additional protection and capacitors for the DDR memory, so a lot of the numbering system was shifted around, while the board between revisions functionally remained the same. The correct schematic for the Arty Rev C is available here: https://reference.digilentinc.com/_media/reference/programmable-logic/arty/arty_sch.pdf.

With that, here are the correct capacitors that line up with the voltage rails we were intending to have you measure:

C143 (should be 5V):

C64 (should be 0.95V):

C162 (should be DDR 1.35V):

C159 (should be 1.8V):

C152 (should be 3.3V):

I have attached a picture for your convenience showing the location of C143, C162, C159, and C152. C64 is on the back of the Arty underneath the letter 'N' in the "Xilinx" silkscreen label.

I apologize for the confusion and let me know what you find.

Thanks,
JColvin

[Deskarano]

Hi @JColvin, @jpeyron

I was wondering if you guys had any updates the overheating issue?

I put a heatsink on my Arty and that lowered temperatures 10-15 degrees right off the bat, and its even better with airflow. I would still like to fix the underlying issue if possible though.

Thanks,

Gregor

[Deskarano]

Thanks @JColvin! Let me know if I can provide any more information to help out. I'm measuring the resistors and capacitors involved in FPGA power and power regulation right now to see if any of these are at fault.

[JColvin]

Hi @Deskarano,

I happened to also have a Rev C to test and I'm getting the same results (within 20 mV) as you for testing the requested capacitors and I know my Arty is working and does not over heat, so we will take a further look into this as to why all of the voltage measurements are not as expected.

Thank you for your patience,
JColvin

[Deskarano]

@jpeyron I have the rev C model

@xc6lx45 Okay, thank you. I will try to verify the power nets

[jpeyron]

Hi @Deskarano,

What REV of the Arty-A7-35T do you have?  Its on the back by the bar code.

thank you,

Jon

[xc6lx45]

as a general comment, it's fairly common to have unplaced components on a PCB, especially a sparse one where there is no pressure to reduce size.

Your readings are all over the place - I doubt that with 12 V (0.12 V?) on the 5V net and 3.3 being the new 1.8 (C181) there would be any magic smoke left inside the board...

But with a proper multimeter (make sure you have a good connection to GND on the black lead), I see a good chance that you'll be able to spot the problem, basically if any net is significantly off its designated value.

[jpeyron]

Hi @Deskarano,

Since the extra heat issue has not been narrowed down to an IC or power rail I would suggest to currently use either or both a heat sink and a small fan to handle any thermal concerns you might have.

thank you,

Jon

[Deskarano]

Hi @xc6lx45 and @jpeyron

I managed to get an XADC configuration working! Here's what the measurements are:

5V power supply voltage: ~4.69V

5V power supply current: ~.588A

External supply voltage: not used (~0V)

FPGA core current: ~.03A

Here is the raw data from the XADC itself, along with the configuration that I used:

 

[xc6lx45]

it's possible that your multimeter is in AC mode (need DC). Also double-check that the "plus" probe is not accidentally plugged in for ampere measurement (separate socket on most models) otherwise it's a short circuit. Usually you connect the minus lead to GND permanently - the interesting caps have GND on one end.

Be aware that probing "accidents" (short circuits) can damage a board.

 

[Deskarano]

Thanks so much for your help!

I have tried measuring voltage across these capacitors but have gotten 0v measurements for all of them. I'm also not 100% sure that I'm measuring right; I just placed the DMM's probes on both sides of the capacitor. Is this correct?

I can get a basic XADC monitoring project working by tomorrow, probably. I'll follow the instructions on the site you linked.

I have not tried using a different cable or PC, but I can try both of these tomorrow as well. I am not using a USB hub.

[jpeyron]

Hi @Deskarano,

I agree with @xc6lx45 that this seems to be a hardware issue. I ran my Arty-A7 with the above hdl and all the switches/LED's on for  around 10 minutes with an average of 31 °C.  I would suggest to check the power rails. 

If you have a DMM please check:

C170 ( should be 5V)

C178( should be 3.3V)

C180( should be 1.8V)

 

These capacitors are around IC11 which is on the top of the board just above and to the left of SW3(switch).

If you are also wanting you could run an a couple of xadc projects that are described in the reference manual here under 3.2 FPGA Core Supply Current Monitoring and in 3.3 5V Supply Power Monitoring. 

Have you tried using a different USB cable or a different PC?

Are you using a USB hub?

thank you,

Jon 

 

 

[xc6lx45]

Hi,

just as a reality check, an unprogrammed Artix FPGA does not run hot. It's marketed as "lowest power and cost at 28 nm", this would turn back time a decade.

My first impression is there is something wrong with your board (not necessarily the FPGA, those tend to be very robust). If you have a multimeter, you could check the voltage regulators by locating a capacitor on the net from the schematic. This is where I'd start looking for a hardware fault.

Now, statistically speaking, most suspected board failures have a completely different root cause, so I wouldn't be surprised if this turns out to be the case.

Remove all other connections. Then, the USB cable is a main suspect.

[Deskarano]

Thank you, I've confirmed that the Arty is in JTAG mode. I booted it in both modes, and it heated up the same way in both. I've attached ~10 minutes of temperature data; as you can see, it seems to level off at around 71, 72 degrees C. Still well within the temperature range, but hotter than I would like. And no, only the Artix 7 chip itself is hot.

[jpeyron]

Hi @Deskarano,

Here is the reference manual for the Arty-A7. Under the Section "4 FPGA Configuration" in the "Figure 4.1 Arty A7 Configuration" it shows the two modes for the Arty-A7 on JP1. Make sure the Arty-A7's mode  is set to JTAG. Does this change the FPGA's temperature in the  temperature graph? 

The  XC7A35TICSG324-1L is the FPGA on the Arty-A7. The C means  it is commercial grade for temperature . The  commercial operational temperature range is 0 to 85 °C. The temperature you are showing(68 °C) is still within the thermal tolerance range of the specific Artix-7 on the Arty-A7.

Are any of the other components on the Arty-A7 hot to the touch?

thank you,

Jon

[Deskarano]

Thanks @jpeyron

Attached is a screenshot of my temperature graph, when my Arty is programmed using the provided HDL code. I didn't set the boot mode to JTAG since I've never touched this setting before and I'm not sure where it's located.

[jpeyron]

Hi @Deskarano,

Open the XADC graph in Vivado Hardware Manager to see the FPGA die temperature. Make sure boot mode is set to JTAG. It is normal for the board to get warm. Please attach a screen shot of you temperature graph. I have included a screen shot of when I configured the Arty A7-35T with the attached hdl code.

arty_sw_btn_led (6).zip