Power supply issues


MikeyMike

Recommended Posts

Hi everyone,

 

I recently purchased a Papilio One 500K. I have successfully managed to install the "hello world" bit file. However not without issues. One issue seems to be very similar to andyfives post here http://forum.gadgetfactory.net/index.php?/topic/2047-question-about-power-supply/.

 

Both the 2.5V (U5) and 1.2V (U4) regulators are getting extremely hot. After testing the outputs neither are maintaining their rated output voltages. Both the 5V (U2) and 3.3V (U6) regulators appear to be fine. If I connect an external 12V power supply the entire board is drawing much more than the suggested 800 mA.

 

With the 1.2V regulator removed from the board the 2.5V regulators output voltage is fine.Which makes sense since the 1.2V regulator is fed from the 2.5V regulator. Also the 1.2V regulator appears to be fine when tested off the board.

 

Does anyone have any ideas ? Could the FPGA be faulty?

 

Thanks,

 

Mike.

 

 

Link to comment
Share on other sites

I'm pretty sure the Papilio isn't designed to be used with a 12V input. The 5V regulator on the Papilio is rated for 6.5-15V but it isn't capable of dissipating excess heat without a heatsink when using higher voltages. The reccomended voltage for the papilio one is 6.5-10V as seen here. Have you tried powering it with a lower voltage or powering it just from USB? With USB power or using a lower voltage you might have better luck.

Link to comment
Share on other sites

Hey MikeyMike,

 

Yes, 12V is too high for powering the Papilio. I recommend using the power from the USB port, there are no projects for the Papilio at this point that need more then the 500mA that the USB port provides. If you are going to use the external power jack then keep the voltage as low as possible. 6.5-7V is best.

 

Jack.

Link to comment
Share on other sites

Hi guys,

 

Thanks for the replies.

 

Just tried powering it again from 7V this time.

 

Measuring the outputs of the four voltage regulators I get :-

 

U2 (5V) reads 4.879V

U6 (3V3) reads 3.284V

U5 (2V5) reads 2.212V

U4 (1V2) reads 0.914V

 

The whole things still drawing more than 1A because the bench supplies current limiter is kicking in. Also the 2V5 regulator is getting far to hot to touch. To be honest I too scared to plug it into my USB port in case I damage it I know it should be current limited but I don't want to risk it since why should powering it from the USB port be any different to using a external supply.

 

Mike.

Link to comment
Share on other sites

Mike,

 

Once again I highly recommend using the USB to power the board, you will be much happier. Very few people use the external power jack, it was only included for the rare case where more then 500mA needs to be drawn. 

 

What are you running on the FPGA when you are drawing more then 1A? The current you draw is highly dependent on what bit file is loaded onto the FPGA. Designs like the Sump Logic Analyzer use less the 50mA and don't heat up the linear regulators while designs like the ZPUino use much more current which does heat up the regulators quite noticeably.

 

The Papilio Pro and DUO both use more efficient, and expensive, switching regulators because of the heat that the linear regulators in the Papilio One produce. 

 

Many, many people have run ZPUino designs on the Papilio One without any problems from the heat generated by the linear regulators, but if you are really worried about it maybe we should upgrade you to a Papilio Pro?

 

Jack.

Link to comment
Share on other sites

Hi Jack,

 

Thanks for the reply.

 

Originally I had the "hello world" bit file downloaded to the FPGA. Which would be lost when the power is removed. So there is nothing running when it exhibits these issues which makes me think it must be a hardware issue. I understand a lot of people power these from a USB port but technically your only bypassing the 5V regulator by doing so and I'm sure the problem lies with the 2.5V and 1.2V regulators so the issue would remain.

 

Nice to see you opted for a switch mode supply rather than using linear regs on the Papilio Pro.

 

I am really worried about it because it isn't working. The two regulators are definitely exhibiting fault conditions and not just drawing a lot of current. This is my first venture into FPGAs and choose the Papilio because it represented really good value for money and I love open source. The problem is I bought the Papilio along with the logic start mega wing from Seedstudio back in June and hadn't got around to playing with it until now. Since more than 60 days have elapsed I would be unable to send it back under warranty. Which means I am stuck with a development board that set me back $60+ that I cant use. I know its my fault for leaving it so long before using it but annoying all the same.

 

Also I am in the UK so getting stuff shipped to and from takes a while. There are sellers here but it still works out much cheaper to order from overseas.

 

As it stands I am toying with the idea of cutting my loses and either ordering a Papilio Pro since I already have the logic start mega wing which I don't want to waste. Or just ordering a Digilent or Terasic development board from Farnell or RS components here in the UK. You don't anywhere near as much bang for you buck with these commercial boards it would be a easily solution.

 

Mike.

Link to comment
Share on other sites

Before attributing any faults to the board it's worth noting that someone at my university had a similar issue with powering the Papilio with the external power jack. He had been using 12V in and was experiencing instabilities on the board (power led flickering, board not running properly). His solution ended up being soldering a wire onto the back of the board at the output to the 5V regulator to power it directly. Unfortunately, somewhere along the line the board was damaged(either from the 12v in for extended periods of time or the heat from soldering). After that the board never worked quite right - we can't always load anything form ZAP onto the board and the regulators are much hotter than the replacement papilios our lab ordered. Depending on how long you were powering it with 12V it is quite possible that it could have damaged something on your board in a similar fashion to the one in our lab. Even then, it sitll works fairly well over usb when it actually lets us program it. You might want to consider the option of powering it with 5v directly from your supply on one of the 5v out wing connectors if you can't get anything else to work - that's one way we have managed to externally power our papilios. Just remember to make sure to remove the power select jumper or at least make sure it isn't set to USB when testing.

Link to comment
Share on other sites

>> Or just ordering a Digilent or Terasic development board

one disadvantage of these boards is that the programming can be very slow, unless using an external JTAG interface. Also, watch out for hidden shipping fees (such as: what you pay is "handling fees" and shipping doesn't show because the carrier will charge you for it). Maybe some vendors are more honest than others.

 

I'd simply replace the regulators. They are cheap. If I remember correctly, Farnell even offers free shipping for UK orders. Or ebay.

 

Papilio Pro is a good choice, IMO.

If you manage to revive the other board later, you can turn it into a fairly decent logic analyzer with a couple of mouse clicks.

Link to comment
Share on other sites

There seems to be some confusion on what the maximum input voltage on that jack is. 7V, 10V or 15V. Looking at the LD1117 datasheet although the device is rated at 12W max the junction to ambient thermal resistance would only allow for around 1W before the device started to over heat. Which at 12V equates to around 142 mA. Not sure how much that "hello world" bit file would cause the FPGA to draw.

 

Getting back to my issue I first tried powering the board with 5V directly onto the 5V supply line. Still the same problem. Next I removed all of the regulators and tested them off the board. All of them appeared to work fine under no load. Then I began fitting them back onto the board one at a time. No issues arose until the 1.2V regulator was fitted. So I removed it again. I then supplied 1.2V from my bench power supply with the current limit set to 500 mA. Soon as the power was applied the current limiter tripped in.

 

The next step would be to replace the FPGA but those Spartan chips are around £30 on Farnell. I don't have a re-flow oven so it would have to re-soldered by hand.

Link to comment
Share on other sites

Tough luck... So the regulators were OK, after all.

 

Sounds like a broken FPGA. For example, I've seen it happen that one IO bank was blown and caused overcurrent, but the FPGA still worked on the other banks.

 

>> The next step would be to replace the FPGA but those Spartan chips are around £30 on Farnell. I don't have a re-flow oven so it would have to re-soldered by hand.

For me, the "C"onsumer grade version shows as GBP 15.08:

http://uk.farnell.com/xilinx/xc6slx9-2tqg144c/fpga-spartan-6-lx-9k-144tqfp/dp/1876228

Still, not sure if I'd want to do the solder job.

 

Anyway, these setbacks aren't unusual. If the budget is an issue, keep in mind that you don't actually need hardware. You can do a lot of self-study with the "ISIM" simulator that comes with the ISE tools, free download.

Hardware is good for motivation and as "reality check" but not mandatory. Usually it just tells me "this doesn't work... go simulate some more..."

Link to comment
Share on other sites

offroad that link is for the XC6SLX9 Spartan 6 used on the Papilio Pro not the XC3S500E Spartan 3 used on the Papilio One.

 

Just checked Farnell again and neither of the parts they have available are the 100 pin QFP package. Managed to find some on Digikey for around £15 each but minimum order quantity of 90!!!

 

Getting the chip off the board wouldn't be an issue its just getting it back on  :)

 

Budget isn't really an issue either just a bit annoyed at having shelled out £40 for a board that is faulty. Lesson learnt test all new hardware the minute you receive it while still under warranty. Even if your not planning on using it any time soon. Starting to think its time I ordered a Papilio Pro and see how I get on with that. 

 

Mike.

Link to comment
Share on other sites

>> that link is for the XC6SLX9 Spartan 6

uh, right... here's the reason to upgrade: cheaper spares :)

 

The soldering is maybe not even that hard when the board is prepared properly and you use solder wick (never worked on that board myself, though).

But, takes a lot of time.

 

As said, P. Pro is a good choice, IMO. Nowadays there is even the upgrade path to the IO-compatible Pipistrello if you need more horsepower later.

Link to comment
Share on other sites

MikeyMike,

 

I'm sorry that I didn't catch this conversation earlier, I checked out for the Labor Day weekend and wasn't looking at forum posts.

 

But don't worry, even if you ordered the Papilio One from Seeed 60 days ago we will make it right for you. Please send us an email at support@gadgetfactory.net and a link to this forum thread. We will replace the Papilio One or work out a way that you can upgrade to the Papilio Pro by paying the difference. Just let us know what you prefer in the email. 

 

Sorry for the hassle,

Jack.

Link to comment
Share on other sites

  • 6 months later...

Mike, this is weird since I’ve used it with a USB port for so long and I don’t think it ever exceed the 450mA limit of my port since I’d have gotten an error then. And with a 12V power supply, ideally it should require even less current from the supply itself. Sometimes, when the voltage regulators are not properly sinked, they tend to get hot which produces an error in voltage regulation as well as drawing more current. 

pcb manufacturing and assembly

Link to comment
Share on other sites

This is an old thread that looks to have been resolved already. The problem was almost certainly a damaged FPGA loading down the regulator.

 

The Papilio Pro uses a switching regulator so higher input voltage will cause it to draw less current but the Papilio One uses linear regulators so the current will not change, the excess voltage is simply burned up as heat. That is why you have to derate the input voltage or add additional heatsinking.

 

For the record, desoldering and soldering QFP IC's like the FPGAs on these is a piece of cake once you get the technique down. I typically mask off the surrounding area with aluminum foil and then use a heat gun to carefully heat the IC until the solder melts and it can be lifted off with tweezers. To solder the new one, put some flux on the board and set the IC in place, align it carefully and touch the iron to one of the corner pins to flow it. Check the alignment again and then press down lightly with the iron on groups of a few pins at a time and drag the iron away from the IC. If you get any bridges or nasty looking joints, add a bit more flux and use desoldering braid to wick off excess solder. Very little solder is needed, usually the tinning on the board is just about enough.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.