TF Riser Revision 0 Arrives

68030 + SDRAM + IDE

Moderators: terriblefire, Terriblefire Moderator

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Fri Feb 14, 2020 8:58 am

terriblefire wrote:
Mon Feb 10, 2020 7:03 pm
arkadiusz.makarenko wrote:
Mon Feb 10, 2020 3:37 pm

Looked at datasheet of f105 and f730 pinout of LQFP64 package there is small difference only on pins 25-30 and different markings on Vss. So unfortunately it would not be direct replacement, but I don't think it would be major change. Can't see any requirement difference for oscillators (but I know shit about it, but Kier Fraser Greaseweezel works on 8Mhz crystal - Pierce oscillator circut), boot1, PB9 etc are in the same spot. They provided compatibility sheet in docs, but not for f10x series but only for f4x1 and F405/F415 series.
This would give second USB port basically for free. Just a thought.

MAX chip could be easier as porting code from MiST might be actually the quickest.

Another suggestion. By having access to PB6 and PB7 you could provide another mean of programming the chip (and way of getting serial feedback), just by adding GND and 5V next to it could help a little. This would allow of programming chip by USB FS port (need of USB A - USB A cable), serial via PB6 and 7 or ST-LINK SW.
Sure. Can you sketch out the circuit you want.
Sorry I cound't put much time into this lately.

Riser Rev 0 Build - still unsucessful.
I connected my cheap £5 logic analyser to usb port, and I can see that USB is being reset and when communication starts it is only for few ns and all stops. Both d+ and d- do their job, but everything stops. From chip perspective I can see that device is being discovered, but any means of communication fail after device request. So stm32 never gets descriptor data from device.

I did go through some information about usb library, and I did find out that changing two sections in library should help with amount of devices which can be used. Issue is that library expects that all sections of usb descriptors are populated, even non mandatory ones. Some devices do not provide it, so code loops over those like device was not supported. Easy fix by changing sections to check only if ubs is not busy, and this should help a little.

I am learning more on USB communications, this is primer to start working on controller support. It is going slowly just now as I have parents in law visiting just now.

Looked at f730 replacement, and it is most liekly just drop in replacement. Only one pin is being used on changed pins, and it is used to drive LED. So this is not an issue at all. I should get hid2ami pcb soon and I will test if I just can install f730 and if it can do basic stuff on USB FS.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Sat Feb 15, 2020 5:11 pm

I really really don't know what is going on.
I have just built HID2AMI, and I can see exactly the same issue.

I have just few items on the HID2AMI board. Regulator, MPU, crystal, 2 caps for oscillator circuit, few resistors, LED and one or two decoupling caps and USB connector.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Sat Feb 15, 2020 9:47 pm

F**K YEA!

OK.Partial success.

I changed chip to F730 (it is drop in replacement btw)

I managed to build firmware, and start device and get it registered!

But it can be done only on external source of power. If I run it out of 5V only I got weird behaviour with the chip. Like issues with debugger etc. It is not 100% stable.

EDIT.

New regulator, caps and complete erase of chips seems to solve all my issues.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

sampedenawa
Posts: 7
Joined: Thu Mar 14, 2019 4:22 pm

Re: TF Riser Revision 0 Arrives

Post by sampedenawa » Sun Feb 16, 2020 8:42 am

arkadiusz.makarenko wrote:
Sat Feb 15, 2020 5:11 pm
I really really don't know what is going on.
I have just built HID2AMI, and I can see exactly the same issue.

I have just few items on the HID2AMI board. Regulator, MPU, crystal, 2 caps for oscillator circuit, few resistors, LED and one or two decoupling caps and USB connector.
Did you check against proper clock settings for USB ?
USB stack needs a well defined clock configuration.

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Sun Feb 16, 2020 9:19 am

sampedenawa wrote:
Sun Feb 16, 2020 8:42 am
arkadiusz.makarenko wrote:
Sat Feb 15, 2020 5:11 pm
I really really don't know what is going on.
I have just built HID2AMI, and I can see exactly the same issue.

I have just few items on the HID2AMI board. Regulator, MPU, crystal, 2 caps for oscillator circuit, few resistors, LED and one or two decoupling caps and USB connector.
Did you check against proper clock settings for USB ?
USB stack needs a well defined clock configuration.
48Mhz on UBS, HSE from external 8mhz crystal/oscillator.

Anyway it does work on f730 chip. I can start working on firmware!

Ps. I reused f105 chip. I have suspicion that those are not that robust as I initially thought. I think I destroyed 2x f105 and a f730.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

sampedenawa
Posts: 7
Joined: Thu Mar 14, 2019 4:22 pm

Re: TF Riser Revision 0 Arrives

Post by sampedenawa » Sun Feb 16, 2020 11:09 am

arkadiusz.makarenko wrote:
Sun Feb 16, 2020 9:19 am
sampedenawa wrote:
Sun Feb 16, 2020 8:42 am


Did you check against proper clock settings for USB ?
USB stack needs a well defined clock configuration.
48Mhz on UBS, HSE from external 8mhz crystal/oscillator.

Anyway it does work on f730 chip. I can start working on firmware!

Ps. I reused f105 chip. I have suspicion that those are not that robust as I initially thought. I think I destroyed 2x f105 and a f730.
Hummm.. these are my settings for HID2AMI:



#define SYSCLK_FREQ_72MHz 72000000


/* Configure PLLs ------------------------------------------------------*/
/* PLL2 configuration: PLL2CLK = (HSE / 5) * 8 = 40 MHz */
/* PREDIV1 configuration: PREDIV1CLK = PLL2 / 5 = 8 MHz */

RCC->CFGR2 &= (uint32_t)~(RCC_CFGR2_PREDIV2 | RCC_CFGR2_PLL2MUL | RCC_CFGR2_PREDIV1 | RCC_CFGR2_PREDIV1SRC);

RCC->CFGR2 |= (uint32_t)(RCC_CFGR2_PREDIV2_DIV2 | RCC_CFGR2_PLL2MUL10 | RCC_CFGR2_PREDIV1SRC_PLL2 | RCC_CFGR2_PREDIV1_DIV5);

// USB PRESCALER FOR 8MHZ EXT CRYSTAL
// PLL2CLK = ( 8 (HSE) / 2) * 10 = 40 MHZ
// PREDIV1CLK = PLL2 / 5 = 8 MHz


/* PLL configuration: PLLCLK = PREDIV1 * 9 = 72 MHz */
/* USB PRESCALER 1.5 for 8MHZ Crystal */

RCC->CFGR &= (uint32_t)~(RCC_CFGR_PLLXTPRE | RCC_CFGR_PLLSRC | RCC_CFGR_PLLMULL | RCC_CFGR_OTGFSPRE);

RCC->CFGR |= (uint32_t)(RCC_CFGR_PLLXTPRE_PREDIV1 | RCC_CFGR_PLLSRC_PREDIV1 | RCC_CFGR_PLLMULL9);
RCC->CFGR &= (uint32_t)~RCC_CFGR_OTGFSPRE;

terriblefire
Moderator Team
Moderator Team
Posts: 1852
Joined: Mon Aug 28, 2017 10:56 pm
Location: Glasgow, UK
Contact:

Re: TF Riser Revision 0 Arrives

Post by terriblefire » Sun Feb 16, 2020 11:16 am

I had managed to get it all working with CubeMX and the f105... the code is in github with the riser schematics.
———
I get cranky when asked to repeat myself.

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Sun Feb 16, 2020 11:54 am

sampedenawa wrote:
Sun Feb 16, 2020 11:09 am
arkadiusz.makarenko wrote:
Sun Feb 16, 2020 9:19 am


48Mhz on UBS, HSE from external 8mhz crystal/oscillator.

Anyway it does work on f730 chip. I can start working on firmware!

Ps. I reused f105 chip. I have suspicion that those are not that robust as I initially thought. I think I destroyed 2x f105 and a f730.
Hummm.. these are my settings for HID2AMI:



#define SYSCLK_FREQ_72MHz 72000000


/* Configure PLLs ------------------------------------------------------*/
/* PLL2 configuration: PLL2CLK = (HSE / 5) * 8 = 40 MHz */
/* PREDIV1 configuration: PREDIV1CLK = PLL2 / 5 = 8 MHz */

RCC->CFGR2 &= (uint32_t)~(RCC_CFGR2_PREDIV2 | RCC_CFGR2_PLL2MUL | RCC_CFGR2_PREDIV1 | RCC_CFGR2_PREDIV1SRC);

RCC->CFGR2 |= (uint32_t)(RCC_CFGR2_PREDIV2_DIV2 | RCC_CFGR2_PLL2MUL10 | RCC_CFGR2_PREDIV1SRC_PLL2 | RCC_CFGR2_PREDIV1_DIV5);

// USB PRESCALER FOR 8MHZ EXT CRYSTAL
// PLL2CLK = ( 8 (HSE) / 2) * 10 = 40 MHZ
// PREDIV1CLK = PLL2 / 5 = 8 MHz


/* PLL configuration: PLLCLK = PREDIV1 * 9 = 72 MHz */
/* USB PRESCALER 1.5 for 8MHZ Crystal */

RCC->CFGR &= (uint32_t)~(RCC_CFGR_PLLXTPRE | RCC_CFGR_PLLSRC | RCC_CFGR_PLLMULL | RCC_CFGR_OTGFSPRE);

RCC->CFGR |= (uint32_t)(RCC_CFGR_PLLXTPRE_PREDIV1 | RCC_CFGR_PLLSRC_PREDIV1 | RCC_CFGR_PLLMULL9);
RCC->CFGR &= (uint32_t)~RCC_CFGR_OTGFSPRE;
It was 100% hardware issue. I had damaged 2x f105 somehow.
I managed to set it up on f407 and f730 on dev boards and different devices without any issues at all.
I was just banging my head on this for so long, as it took me like 3 weeks to make this riser board to work.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

terriblefire
Moderator Team
Moderator Team
Posts: 1852
Joined: Mon Aug 28, 2017 10:56 pm
Location: Glasgow, UK
Contact:

Re: TF Riser Revision 0 Arrives

Post by terriblefire » Sun Feb 16, 2020 12:03 pm

arkadiusz.makarenko wrote:
Sun Feb 16, 2020 11:54 am
It was 100% hardware issue. I had damaged 2x f105 somehow.
I managed to set it up on f407 and f730 on dev boards and different devices without any issues at all.
I was just banging my head on this for so long, as it took me like 3 weeks to make this riser board to work.
Where did you get the Chips? I got mine from RS.
———
I get cranky when asked to repeat myself.

User avatar
arkadiusz.makarenko
Posts: 264
Joined: Wed Jun 19, 2019 7:36 am
Location: Edinburgh

Re: TF Riser Revision 0 Arrives

Post by arkadiusz.makarenko » Sun Feb 16, 2020 12:30 pm

terriblefire wrote:
Sun Feb 16, 2020 12:03 pm
arkadiusz.makarenko wrote:
Sun Feb 16, 2020 11:54 am
It was 100% hardware issue. I had damaged 2x f105 somehow.
I managed to set it up on f407 and f730 on dev boards and different devices without any issues at all.
I was just banging my head on this for so long, as it took me like 3 weeks to make this riser board to work.
Where did you get the Chips? I got mine from RS.
First one from Mouser, but I damaged it very quickly. As I couldn't connect to flash at all.

Second once which was partially working was from eBay (Thinking about it, it could be fake! )

Working f730 is from Farnell.
Do not trust people. They are capable of greatness.
~ Stanislaw Lem

Post Reply

Return to “TF330”