Problems updating the flash on the FPGA CopPro

discuss PC<>Acorn file transfer issues & the use of FDC, XFER, Omniflop/disk etc.
User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Problems updating the flash on the FPGA CopPro

Postby tricky » Sat Mar 14, 2015 9:09 am

I'm trying to use iMPACT to update the version of the CoPro software on my board and am having a problem.
"Cable Auto Connect"

Code: Select all

GUI --- Auto connect to cable...
// *** BATCH CMD : setCable -port auto
INFO:iMPACT - Digilent Plugin: Plugin Version: 2.4.4
INFO:iMPACT - Digilent Plugin: no JTAG device was found.
AutoDetecting cable. Please wait.
*** WARNING ***: When port is set to auto detect mode, cable speed is set to default 6 MHz regardless of explicit arguments supplied for setting the baud rates
PROGRESS_START - Starting Operation.
Connecting to cable (Usb Port - USB21).
Checking cable driver.
 Driver file xusb21.sys found.
 File C:/Xilinx/14.7/LabTools/LabTools/bin/nt64/xusb21.sys not found.
 Driver file xusbdfwu.sys found.
 Driver version: src=1027, dest=1027.
 Driver windrvr6.sys version = 10.2.1.0. WinDriver v10.21 Jungo (c) 1997 - 2010 Build Date: Aug 31 2010 x86_64 64bit SYS 14:14:44, version = 1021.
 Cable PID = 0008.
 Max current requested during enumeration is 280 mA.
Type = 0x0605.
 Cable Type = 3, Revision = 0.
 Setting cable speed to 6 MHz.
Cable connection established.
Firmware version = 1100.
File version of C:/Xilinx/14.7/LabTools/LabTools/data/xusbdfwu.hex = 1100.
Firmware hex file version = 1100.
PLD file version = 0012h.
 PLD version = 0012h.
PROGRESS_END - End Operation.
Elapsed time =      0 sec.
Type = 0x0605.
ESN not available for this cable.
I'm not sure about:
Driver file xusb21.sys found.
File C:/Xilinx/14.7/LabTools/LabTools/bin/nt64/xusb21.sys not found.

or
ESN not available for this cable.

"Initialize Chain"

Code: Select all

Attempting to identify devices in the boundary-scan chain configuration...
INFO:iMPACT - Current time: 14/03/2015 08:53:05
// *** BATCH CMD : Identify -inferir
PROGRESS_START - Starting Operation.
Identifying chain contents...'0': : Manufacturer's ID = Unknown
INFO:iMPACT:501 - '1': Added Device UNKNOWN successfully.
----------------------------------------------------------------------
----------------------------------------------------------------------
'1': : Manufacturer's ID = Unknown
INFO:iMPACT:501 - '1': Added Device UNKNOWN successfully.
----------------------------------------------------------------------
----------------------------------------------------------------------
'2': : Manufacturer's ID = Unknown
INFO:iMPACT:501 - '1': Added Device UNKNOWN successfully.
----------------------------------------------------------------------
. . . many more until the SW gives up!

Does anyone have any ideas?
Maybe my clone cable isn't a very good clone?

I think I connected the cables correctly, although the TMS pin did disconnect initially, I pushed it back and re-soldered it.
JTagConnections.jpg

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Sat Mar 14, 2015 12:44 pm

Hi Tricky,

Your "Cable Auto Detect" log looks pretty much identical to mine.

I wouldn't worry about the xusb21.sys not found, it's probably just looking in several locations. I don't see this, but I'm running on Linux.

I do get the "ESN not available for this cable" message. ESN is the Electronic Serial Number. This hasn't caused me any problems with programming.

I think you just have a JTAG signal connection issue.

I've tried removing one signal at at time, and when I remove the white connection (TDO) I get exactly the same error ("Manufacturer's ID Unknown" multiple times).

Some more data points:
- TDI being disconnected doesn't seem to affect the "Identify" stage, but does affect the scan chain integrity test.
- TCK being disconnected gives a "A problem may exist in the hardware configuration" error
- TMS being disconnected gives a "A problem may exist in the hardware configuration" error
- TDO being disconnected gives a "Manufacturer's ID Unknown" error.

I'd look very carefully at the TDO connection. Maybe also check with a multi-meter whether any of the connections are shorted to ground.

Dave

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Sat Mar 14, 2015 5:56 pm

I rechecked the connections and none seem to be loose. none are shorted to GND except GND. I checked the cable connection to its mini breakout board - ok. I tried 32bit iMPACT, but the cable drivers don't work with 32bit, only 64bit.
Any other suggestions welcome.

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Sat Mar 14, 2015 6:11 pm

Tricky,

I have a couple of further suggestions.

Can you check the connectivity of the four JTAG signal pins (TDO, TDI, TCLK, TMS) back to the FPGA? It's a bit fiddly to do, but possible if you are patient.

The other thing is that if you disconnect TDO and TDI from the LX9, and simply connect them together with a short piece of wire, and then try initializing the scan chain. You should get a different error "A problem may exist in the hardware configuration" rather than "Manufacturer's ID Unknown". If the error doesn't change, then there is something wrong with your programmer. [ I've just tried this with mine to confirm the error changes ]

Dave

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Sat Mar 14, 2015 6:57 pm

With TDI and TDO connected to each other but not the board, I get:

Attempting to identify devices in the boundary-scan chain configuration...
INFO:iMPACT - Current time: 14/03/2015 18:56:02
// *** BATCH CMD : Identify -inferir
PROGRESS_START - Starting Operation.
Identifying chain contents...done.
ERROR:iMPACT - A problem may exist in the hardware configuration. Check that the cable, scan chain, and power connections are intact, that the specified scan chain configuration matches the actual hardware, and that the power supply is adequate and delivering the correct voltage.
PROGRESS_END - End Operation.
Elapsed time = 0 sec.
// *** BATCH CMD : identifyMPM

Off to follow the tracks now.

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Sat Mar 14, 2015 7:10 pm

Whilst following the tracks, one of the pins popped off, so it looks like my soldering skills haven't improved with age! I have re-soldered them and it identifies xc6slx9 SPI/BPI... So back to the instructions ;)

Thanks again.

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Sun Mar 15, 2015 2:54 pm

I have followed the instructions and think I now have the latest posted .mcs on board. The beeb hangs on power-up, but is OK after a CTRL-BREAK. I'll add a decoupling cap and check if changing the DIP-switches changes the message.

User avatar
DutchAcorn
Posts: 1618
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: Problems updating the flash on the FPGA CopPro

Postby DutchAcorn » Sun Mar 15, 2015 4:39 pm

tricky wrote:I have followed the instructions and think I now have the latest posted .mcs on board. The beeb hangs on power-up, but is OK after a CTRL-BREAK. I'll add a decoupling cap and check if changing the DIP-switches changes the message.

Well done! What instructions are you following? I have been struggling to get the image on it. I can't seem to find my way around the tool.
Paul

RobC
Posts: 1774
Joined: Sat Sep 01, 2007 9:41 pm

Re: Problems updating the flash on the FPGA CopPro

Postby RobC » Sun Mar 15, 2015 5:22 pm

tricky wrote:The beeb hangs on power-up, but is OK after a CTRL-BREAK. I'll add a decoupling cap and check if changing the DIP-switches changes the message.

This is the behaviour I see - it always hangs on start-up but is fine after CTRL-Break.

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Sun Mar 15, 2015 6:11 pm

Instructions from page 20 of the main thread. I installed the lab tools and ran iMPACT, some of the words are slightly different to the instructions, but essentially the same.

firthmj
Posts: 227
Joined: Tue May 26, 2009 8:37 am
Location: Ipswich, UK

Re: Problems updating the flash on the FPGA CopPro

Postby firthmj » Sun Mar 15, 2015 6:12 pm

DutchAcorn wrote:
tricky wrote:I have followed the instructions and think I now have the latest posted .mcs on board. The beeb hangs on power-up, but is OK after a CTRL-BREAK. I'll add a decoupling cap and check if changing the DIP-switches changes the message.

Well done! What instructions are you following? I have been struggling to get the image on it. I can't seem to find my way around the tool.


Hi,

Dave wrote up a pretty good description of the procedure in this post:

viewtopic.php?f=44&t=8852&p=105623&hilit=Impact+spi#p105623

How far are you getting - have you managed to get as far as having Impact running and detecting the FPGA?

Regards

Michael
Had fun at the
Image
Meeting 13th May 2017

User avatar
DutchAcorn
Posts: 1618
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: Problems updating the flash on the FPGA CopPro

Postby DutchAcorn » Sun Mar 15, 2015 7:38 pm

firthmj wrote: ... Dave wrote up a pretty good description of the procedure in this post:

viewtopic.php?f=44&t=8852&p=105623&hilit=Impact+spi#p105623

...

Thanks for the link Michael! That worked well. :D
Paul

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Tue Mar 17, 2015 2:55 pm

After the update, do you get a hang on start-up, which is then OK with a CTRL-BREAK?

User avatar
danielj
Posts: 5216
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: Problems updating the flash on the FPGA CopPro

Postby danielj » Tue Mar 17, 2015 3:07 pm

Usually :D

d.

RobC
Posts: 1774
Joined: Sat Sep 01, 2007 9:41 pm

Re: Problems updating the flash on the FPGA CopPro

Postby RobC » Tue Mar 17, 2015 6:55 pm

tricky wrote:After the update, do you get a hang on start-up, which is then OK with a CTRL-BREAK?

All the time - both of mine always display this behaviour on power-up but are fine after a CTRL-Break.

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Tue Mar 17, 2015 7:33 pm

The hanging on power up is I think down to the time it takes for the FPGA to load it's configuration, compared to the time it takes the Beeb to reset.

Even with the fastest configuration clock speed, this seems to be not fast enough on many people's systems, mine included.

I'm interested if the latest update seems to have made this worse for anyone. If so, what version of the firmware were you running previously?

Dave

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Tue Mar 17, 2015 8:41 pm

The version that was on mine when delivered never hung on power up on my model B issue 4, but the latest always does. Always and never are about 6.

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Tue Mar 17, 2015 8:49 pm

Tricky,

Which Co Pro model's are have you tried? Are they all hanging on power up?

Dave

User avatar
tricky
Posts: 1875
Joined: Tue Jun 21, 2011 8:25 am
Contact:

Re: Problems updating the flash on the FPGA CopPro

Postby tricky » Tue Mar 17, 2015 8:55 pm

It is the default setting, I have just had a look and there seems to be some sort of protective plastic over the switches.

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Tue Mar 17, 2015 10:13 pm

I would like to get to the bottom of this power-up configuration issue.

I've just checked the .xise files in github, and the configuration rate is set to 26 (MHz) which is the highest speed available in Spartan 6.

The bitstream file is ~3.5MBits, so configuration should take 3.5/26 = 135ms.

The accuracy of the configuration clock is actually only +- 50% so in the worse case this might double it as well, to 270ms.

However, because of the multiboot loader, there will be two configuration cycles, which makes ~540ms worst case.

So I can easily see why this might actually be longer than the time taken for the Beeb to boot, thus causing a hang.

But Tricky is reporting a definite regression since the original version, so there is something we are not understanding here...

A few questions for Jason, when you get a moment:

- Did all the production Co Pros ship with a multi-boot loader?

- Did you ever experiment with using the bitstream compression flag?

- I have a vague recollection of reports that changing the 6502 core made this issue worse (which would point to it not being a config issue but something else entirely). Anyone else recall this as well?

I probably won't have time to dig into this for about 10 days...

Dave

User avatar
flynnjs
Posts: 747
Joined: Tue Jul 06, 2010 9:33 pm

Re: Problems updating the flash on the FPGA CopPro

Postby flynnjs » Tue Mar 17, 2015 10:15 pm

I also have started getting this. I seemed to start when we swapped to the other 6502 core.

I did wonder if using a compressed bitstream might speed the transfer up enough to counter the problem but not had time to try.

User avatar
flynnjs
Posts: 747
Joined: Tue Jul 06, 2010 9:33 pm

Re: Problems updating the flash on the FPGA CopPro

Postby flynnjs » Tue Mar 17, 2015 10:18 pm

lol, I posted while you did :)

I think this might be something to do with the number of cycles the 6502 core takes to come out of reset rather than the configuration time.

The early board shipped with the original 6502 core and only later builds have swapped to Alan's core.

User avatar
hoglet
Posts: 6471
Joined: Sat Oct 13, 2012 6:21 pm
Location: Bristol

Re: Problems updating the flash on the FPGA CopPro

Postby hoglet » Tue Mar 17, 2015 10:29 pm

(Edit) more crossed posts :lol:
flynnjs wrote:I also have started getting this. I seemed to start when we swapped to the other 6502 core.

This may not be a config issue then, as I would not expect the bitstream size to change (unless compression was used, which I don't think it is).

It's possible for some reason the Alan D 6502 core is hanging on a power up reset, and needs a second reset to get it going.

I've just checked behaviour on some of the other Co Pro designs. The Z80 and 6809 are fine. The 6502 and X86 both hang on initial power up. This seems consistent.

I'm now thinking this is something other that the configuration taking too long.

Dave

RobC
Posts: 1774
Joined: Sat Sep 01, 2007 9:41 pm

Re: Problems updating the flash on the FPGA CopPro

Postby RobC » Wed Mar 18, 2015 5:38 pm

Just as an additional datapoint, I see the hanging behaviour on power-up for every co-pro (i.e. 6502, Z80, 6809 and 286).

User avatar
DutchAcorn
Posts: 1618
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands

Re: Problems updating the flash on the FPGA CopPro

Postby DutchAcorn » Wed Mar 18, 2015 7:46 pm

Tried this on mine (sn028), 20150215_1904, decoupling cap fitted, BBC B.
I also see hangups on start-up only on the 6502 and the 80286 cores. Z80 and 6809 are fine.

Also, if you press break instead of ctrl-break you get a coloured "x" and a * prompt.
Paul


Return to “software & utilities for the pc, mac or unix”

Who is online

Users browsing this forum: No registered users and 1 guest