early V2 shield, noise on display (SOLVED).

[realpinz]

Howdy,

Thought I would chat to the experts on the issue I am having with a real pin interface.

Have a V2.1 shield (I think) constructed as per the schematic that has noise or hashing when used on a real WPC pinball.

I have tried the latest firmware and rolled it back to V2.30 with no difference.

STM32F704, configured for williams (also tried wpc95 with earlier version).

have tried different:-

   shields, DMD ribbon cables, power-supplies (5V Dc-Dc and smps's).

adjusted the voltage from 4.8->5.0 measured at the shield.

tried it in a WPC-95 and WPC89 system.

with a stern LED display it is clean and clear, the Pin2DMD splash screen is clean, with real color files it locks up the system, with no pallet or color files it is noisey.

just have the license and the config on the SD, config states "Williams" so the key is also working o.k.

 

Any hints?  Used the same hardware on a Simpsons PP and it was perfect.  Different discovery boards is all I can think of at the moment.

 

cheers

[realpinz]

I think I answered my own question.... re-checked the STM32F407's and although they have all the right hardware I think they are still fake.

re-ordering some from  a reputable dealer and see what happens.. If these work will look into what is wrong with the others once I have a benchmark to compare to.

 

Any other ideas, things to try please don't be shy.

[realpinz]

Ok, whilst waiting for the new (hopefully) genuine discovery boards I finally had some time to investigate further.

The boards can be made to work properly as will be described below, very simple process.

Of coarse this will only work if

----you have the same type of counterfeit board

---have corruption or hashing on the display from a real pinball, splash screen or anything generated on the Pin2DMD is fine like menus.

---the issue is corrupted communication between the real pin and the (fake) STM32F40X discovery via any version of shield.

---I had discovery boards with a green audio jack and a black audio jack, both types definitely fake.

 

The issue I was experiencing was due to the data on "pin_DMD_sd" PIN2 of the pinball dmd header which goes to PA7 on the discovery getting corrupted.

The corruption was being created by U5 on the discovery that shares PA7 (and is wired directly to it).  Thee fake boards are using a different U5 and for what ever reason it is trying to send data out when it shouldn't clobbering the DMD serial data and creating the corruption.

 

My proof of concept was to cut the trace at U5 (via that goes to it anyway) from PA7 of the main (large) stm chip.

Make sure obviously you leave the connection between the header where PA7 comes onto the discovery and to the large chip.

 

Didn't have any issues with ST-link or any other functions in this state (with the trace cut).

hope this may help someone stuck with a fake discovery.

 

Cheers

[lucky1]

Thanks for the info. The fake Discovery boards and the problems with components not needed for pin2dmd on it were one of the reasons why I decided to change to the Nucleo design. It is cleaner and more future proof because it is available in different CPU configurations allowing easy upgrade by swapping the Nucleo board.