rusefi.com

Hello,reen
When I plug in my frankenso board to nothing but my 12v power supply, I get an "overcurrent" (LED8) light on the stm32f4, and it doesn't start up the LCD screen or otherwise operate correctly. When I plug in only my USB, it works fine. Any ideas?

Thanks!

With no Discovery installed and no LCD, etc installed, and when powered by 12V, do you get 5V? My guess is that you are operating at more than 5V which is dumping to the rail clamping diodes and causing the STMPS2141STR to over current. That's just a guess, so might as well start with baby steps, and check the 5V with it being as isolated as possible. Also remove W23 if you can. This will isolate the 5V and prevent wrong voltages from getting to 5V components. Once you know that 5V is good, you can then install W23 and let it power the 5V rail.

I don't see any notes about LD8 here, but this page might offer some insight http://rusefi.com/wiki/index.php?title=Hardware:stm32f4discovery_repair

Has your stm32f4 ever worked while on Frankenso with +12v?

I have to confess it's a shortcoming of my testing procedure - I test power supply without stm32, confirm +5volts with a multi-meter and only then I add W23 to feed the rest of the board and which point I do not re-test it with stm32 again under +12v, just USB power from here on. Should change it to make sure to test with +12v source and brain board simultaneously.

Looks like I have the same issue with the whole latest batch. Will provide more details once I have them.

LD8 is connected to STMPS2141STR which is http://www.mouser.com/ds/2/389/CD00167470-251613.pdf

A workaround would be to remove FB701 - this case the green D351 LED on the red board which should not be powered by the red board is removed from the board, this case looks like there is no over-current and the rest of the board works just fine. The USB on the red board does not work this way obviously.

Jared would have technical details regarding why we have the issue with the latest batch and I have a better workaround and a fix

workaround without changing anything on the board: have the ECU powered via the miniUSB on discovery while restarting it on the car - this case it would start just fine. Once the ECU has started USB could be removed and it would work on just +12v just fine.

A fix which has helped a couple, but unfortunatelly not all the boards: remove D703 diode. Nothing would change functionally, but the board would start OK on +12 volts only.

I got to talk with russian, who has hardware that's exhibits the problem, but I have not seen this hardware in person, so this is the best I can see with a blind fold on.

The STMPS2141STR can "Fault" on over current, but it can also fault on under voltage. I think the fault is happening from under-voltage. I see this chips enable bit will have around a 5ms to 10ms rise time before the output is operational. I know there is a rise time of the power supply, which may or may not be close to this trip time. I suspect that variations in china sourced parts have caused the 5V switching regulator to turn on slower, which may cause these new boards to trip this under voltage fault as it's starting up.

Looking at figure 7.1.3 from here http://www.mouser.com/ds/2/389/CD00167470-251613.pdf I expect that when the voltage rises above the threshold, it would turn off the fault bit. I'm not sure if this fault is latched, or if it would reset as soon as the voltage got above the threshold. I know that russian's board was above the threshold, and was still showing the faulted light. However his STM board was also functional and appeared to be operational.

Hmmm, also the discovery has recently changed, but I thought the new one was basically the same as the old one. Could there be a change on the discovery that's causing this? I wonder if someone has done a comparison of the two board versions.

I'm tempted to say remove U6 --> STMPS2141STR from the discovery. I see it influences the output of the knock circuit, which has been problematic for us. I suspect this chip is driving that output high when we want the knock chip to drive low. Also it's really only needed for USB OTG, which we are not doing and are not planning to do. We are probably best off removing this chip to prevent it from messing with the function of other IO. Also this chip is not part of the equivalent discovery that we have on the PCB.

Perhaps we want to consider a board like this http://www.ebay.com/itm/STM32-64C-STM32F103RBT6-STM32-core-board-/162051785876?hash=item25bb0a0494:g:8JcAAOSw1DtXIMHe which doesn't have all that extra stuff on it. I know the audio jack is also an issue that we work around.

I suspect that when russian removed an LED, it allowed the board to start up slightly faster and prevented this from tripping. I wonder if this chip were removed, would see it go away on all the PCB's that russian has.

The root problem was the STM board we were using. It worked fine when powered through USB, but not on the Frankenso. We replaced it with another STM board and it started working without a problem.

marroyo wrote:The root problem was the STM board we were using. It worked fine when powered through USB, but not on the Frankenso. We replaced it with another STM board and it started working without a problem.

I know there are revisions C and D and at some point I had a theory that that's about the board revision but that did not seem to be the case.
Still, STM board is probably not the root cause - I think it's just both boards not playing together well. Glad you have found a solution!

My next step is to remove some components from discovery boards to heal all the boards I've got.

http://rusefi.com/forum/viewtopic.php?f=2&t=1013&p=18923#p18923 mentions another approach

russian wrote:http://rusefi.com/forum/viewtopic.php?f=2&t=1013&p=18923#p18923 mentions another approach

On the one hand, removing SB11 and R68 from the back of the discovery solves the "Frankenso does not start on 12volts" problem.

Unfortunately the down-side is that this case Frankens ONLY starts on 12 volts - no more running the board using mini USB. No more bench-testing with just 5v power source via top miniUSB.

At least flashing firmware via mini USB still works.

kb1gtt wrote:I'm tempted to say remove U6 --> STMPS2141STR from the discovery. I see it influences the output of the knock circuit, which has been problematic for us. I suspect this chip is driving that output high when we want the knock chip to drive low. Also it's really only needed for USB OTG, which we are not doing and are not planning to do. We are probably best off removing this chip to prevent it from messing with the function of other IO.

Removed R50 - no improvement, still does not start on 12 volts.
Removed R46 - no improvement, still does not start on 12 volts.
Removed U6 - still no improvement, still does not start on 12 volts.

Combined with the SB11/R68 kind-of-success which I did on a different discovery/frankenso pair this makes me wonder if STMPS2141STR has nothing to do with this issue altogether?

The STMPS2141STR might not be part of the 12v issue, but it is likely an issue with the knock. So before we say remove or do not remove this U6 chip, I say we first check if it's removal helps with the knock situation.

Did the SB11 thing fix those problematic boards? In the above did it have the SB11 modifications or not? It's possible we are fighting 2 different problems that are manifested with the same results.

i knew it would be confusing
I've modified two boards differently. One board removing stuff around u6 and no difference. Another board sb11 and r68 which fixed 12v and killed 5v

killed 5V, you mean from the discovery board? I'm assuming the 12V generated 5V is still good, but perhaps the USB 5V power is not good.

kb1gtt wrote:killed 5V, you mean from the discovery board? I'm assuming the 12V generated 5V is still good, but perhaps the USB 5V power is not good.

Yes, that's exactly what I was trying to tell. I've just edited the post above to hopefully make it more clear:

russian wrote:Unfortunately the down-side is that this case Frankens ONLY starts on 12 volts - no more running the board using mini USB. No more bench-testing with just 5v power source via top miniUSB.

At least flashing firmware via mini USB still works.

It's not the first time that user clive1 from st forum is saving me - this guy knows everything stm32!

https://my.st.com/public/STe2ecommunities/mcu/Lists/STM32Discovery/Flat.aspx?RootFolder=%2fpublic%2fSTe2ecommunities%2fmcu%2fLists%2fSTM32Discovery%2fSTM32F4DISCOVERY%20vs%20STM32F407G-DISC1%2c%20exact%20differences%20hardware%20or%20software&FolderCTID=0x01200200770978C69A1141439FE559EB459D75800084C20D8867EAD444A5987D47BE638E0F&TopicsView=https%3A%2F%2Fmy%2Est%2Ecom%2Fpublic%2FSTe2ecommunities%2Fmcu%2FLists%2FSTM32Discovery%2FAllItems%2Easpx

There are issues with the firmware that shipped with them, use the current ST-LINK Utilities (v.3.9.0) to update (V2.J27.M15).

And it worked! With latest STM32 ST-LINK Utility v3.9.0.exe I've updated discovery firmware (not rusEfi firmware) and the issue is resolved!

With a little caveat: newer boards take an extra second to start. Still an open question how to put old firmware into the new board.

4 second different ...

How much longer is that from the original time? As well can we make the delay go away if we plan to not use the SWD part of the board. AKA I believe our brain board does not have that kind of delay.

Clive1 on the st forum mentioned above mentions that there is a chance to down-grade firmware version, anyone willing to try that? Also he has a suggestion for a different st_link firmware altogether.

Anyone available to investigate? https://svn.code.sf.net/p/rusefi/code/firmware_archive/STM32%20ST-LINK%20Utility/

Not sure if Drew mentioned it to you or not but we confirmed startup of the Festiva with the newer firmware on Friday night...We need to get the right tune into the board as the car was running really, really, really fat...

I am failing to find the newer thread abouot the same issue

answer is to update ST-Link firmware, do not ask how that's related

Simon welcome to the club!

Huh, weird.

Well I will give it an update and confirm if the board wakes up with just frankenso power.

Sounds like this is an FAQ thing.