I am having some troubles with using a STM32F103 bluepill and I have narrowed the problem -I think- to the USB 3.0 interface. I did not found any related post so I thougth in asking you. Let me tell you my tests so far
1) I am using a VMware Workstation 12 VM machine as my main development platform. This machine is on a main PC with only USB2.0 ports. All is working correctly there, when I plug the bluepill it appears as COM11 serial port all the time, no problem at all.
2) I copied the VM to two other VMWare Workstations 12 in two different new PCs with only USB 3.0, I am not able to see the machine as a COMx, only as DFU, and even in that case, not every time. As a workaround I’ve installed the drivers at host level, I can see the Maple 3 at host level but without success in terms of COMx at VM level.
3)When I use the bluepill in the USB 3.0 PCs, when using arduino IDE, making a compilation without seeing the COMx, some times appears as a COMx, not always, but in any case it shows errors when resetting from COMx to DFU. It does not show a uploading error in the ‘orange letters part’ , but basically never upload.
Based on 1-3, the only difference is the USB 3.0 interface.
So my question is: Does someone apart of me have had problems with USB3.0? It can be a matter of the way the reset of the USB is simulated is working only in usb 2.0? Can be VMWare+USB 3.0+whatever?
I have to confirm yet that using my USB2Serial directly to the bluepill in the USB 3.0 machines is working, but as this is in order to have an on-the-road machine I preffer to use it with as less additional hw as possible, if possible.
What do you think?
Thanks in advance
/F
I do not have a 3.0 port PC, but I think the easiest way to isolate the issue is to use a 2.0/2.1 USB hub between the STM32duino and the USB 3.0 ports.
Verify that the 2.x USB port extender works with USB drives or flash memory or some other 2.x device(s). Then try the STM32duino. If there is a timing issue on the 3.0 port, the hub should mask the problem, IMO.
Ray
Will try and post results when possible
/F
Both work fine, maple mini clone (bootloader2.0) has been recognized, uploaded, reset, run.
Important is you set in ide tools/port the number you get for the respective usb port (ie. in Device Manager – Ports (COM&LPT) you see the Maple(COM23)).
It might be the VMware adds a layer..
For me… lots of troubles with USB3 when the device is other than a flash drive or USB drive enclosure.
USB/serial…. only FTDI chip based (not counterfeit) USB/Serial work dependably for me.
I have about 4 USB/Serial cables, some 3.3V, some 5V, and a few more RS232.
Prolific chip set cables – I have trouble with many – depending on age of chipset. Mostly bad OS driver design.
I use a lot of USB hubs in my work. I have about 15, mostly 7 port, most with power switch per port and external power.
USB2 hubs of many varieties generally work. Not all.
USB32 hubs of several varieties mostly DON’T work, meaning the devices other than memory/disk don’t enumerate.
I’ve spent a lot of time on this and have yet to find a reliable product/vendor for USB3 hubs.
The PC mainboard “root” hubs for USB3 tend to work better (ASUS mainboard).
It is all very much hit and miss.
Advice: Buy USB/Serial only which are FTDI and known non-counterfeit, e.g., from FTDIs web site. Mainly because their drivers are well designed and within windows distributions. Other people rant about them costing more. I say caveat emptor and I do NOT want to waste time on trial and error.
hub in use is a powered combo, 3 off usb3 and 4 off usb2, although same when direct to laptop(2 & 3)
i’ve gone completely ‘standard’ ie arduino.cc 1.6.5r5 64 bit binary. AND that has changed from 32 bit ??
anyone know of a linux live 32bit dvd/cdrom version with arduino 1.6.5 and arduino_stm32 installed?
i need to find an much older arduino_stm32, how do you get such a zip from github?
nano’s with ch340 are all quite happily working.
stephen
i need to find an much older arduino_stm32, how do you get such a zip from github?
nano’s with ch340 are all quite happily working.
stephen
I had USB issues on my Gigabtye Skylake chipset motherboard. Which seemed to be fixed by increasing the reset length.
But the Maple mini with its additional reset hardware seems OK.
I didnt isolate this to which port the boards were plugged into. I think it occurred on a normal USB 2 port, but I am not in a position to double check at the moment
thank you, i’ll send a PM.
seems i declared 64 bit’ness about 29 feb.
pretty sure i would have done some checks of functionality, i don’t recall any problems with uploads.
i do know i’ve had will-o-wisp issues for quite a while, inconsistency of output on serial monitor.
it maybe that i got slightly jaundiced, consequently my switch to nano’s, although i did get consistency with low memory available messages 
dfu-utils version is showing 0.8 (??)
@Roger
is there list of the shipped versions for the ‘utility support’ software, st-info, stflash etc ?
stephen
I’m not sure what you mean.
the version i’m seeing of dfu-utils when used displays its version as 0.8
the dfu-sources from Rays link, when compiled and run displays 0.9. that leads to the question of which one should i be using?
so i was wondering if the other tools have version numbers.
i.e. if i suddenly see dfu-utils announce itself as 0.7 and i know it was last 0.8, somethings wrong or substantially changed.
if i did a software acceptance review i’d expect to see all the changes documented with the s/w change notes and the versions of all the software in the build process and the o.s. versions, e.g. it now being compiled with gcc-5.1 and last time with 4.9 or 32bit v 64bit (much, much, much more exacting from q.a. side).
And yes i was on the receiving end of the s/w QA dept a lot, software Q.A dept gets really very, very, very picky. different lifetime.
did someone change the messages coming from dfu-utils; port busy c.f. port not available?
stephen
takes that long to annoy my eyes.
it’s playing hairy fairy again, flummoxed, working … … wait for it, quote arnie.
srp
Considering that other users does not have this problem with directly attached USBs, I am suspecting a unfortunate combination of the VMWare intermediate layer behaviour with the bootloader implementation of the reset procedure. Too unclear or erratic to find the root cause of the error.
In any case as I decided long ago that using VMs to isolate activities is the best way to tame the installation of multiple drivers and dlls in my systems, I think that the best option will be either use an USB2Serial device or get a STLink at aliexpress and forget all those worries
/F
<…>
In any case as I decided long ago that using VMs to isolate activities is the best way to tame the installation of multiple drivers and dlls in my systems, I think that the best option will be either use an USB2Serial device or get a STLink at aliexpress and forget all those worries
/F
That is a reasonable decision, but VM’s do complicate some areas as you have determines with device enumeration. A VM is not a cure-for-all. But, if you are comfortable with it… that is good.
Ray
<…>
…at some point some thing goes south and all starts to crawl even wigh 32GB, or a version update explodes or whatever and you spend a week to reinstall all from the beginning. That, if you are fanatic in terms of backup (I manage 3 copies). If not.. well, add the time to pull your hair etc.
<…>
