|
|
Re: CB030, A 68030 SBC for hobbyists [message #7485 is a reply to message #7482] |
Thu, 09 April 2020 20:58   |
mikemac
Messages: 250 Registered: March 2017
|
Senior Member |
|
|
plasmo wrote on Thu, 09 April 2020 20:08Oops, the pictures I provided were not the latest. There are additional modifications beyond the 8 data lines. Because the data port is now 16-bit wide, the address lines need to start from A1 to A3 instead of A0-A2 of the original design. This changes the memory map associated with CF interface. I also need to make change in CPLD to tell 68030 that CF is 16-bit wide port. So you'll need to reprogram the CPLD and reprogram the EPROM.
Give me a day or so to come up with an engineering change detailing all the changes needed to convert 8-bit CF interface to 16-bit.
Bill
Likewise the Linux kernel and boot loader will need to be updated before they'll work with the modifications.
Mike
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7486 is a reply to message #7485] |
Thu, 09 April 2020 21:06   |
mikemac
Messages: 250 Registered: March 2017
|
Senior Member |
|
|
mikemac wrote on Thu, 09 April 2020 20:58
Likewise the Linux kernel and boot loader will need to be updated before they'll work with the modifications.
Hmm. It may still work but in 8 bit mode. The boot loader does issue the command to put the CF in 8 bit mode. So it "should" work just without any benefit.
Mike
|
|
|
|
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7494 is a reply to message #7493] |
Sat, 11 April 2020 08:22   |
quarterturn
Messages: 86 Registered: April 2018
|
Member |
|
|
I got a TTL866CS EEPROM programmer and an Altera "USB Blaster" clone on order. I just went with aliexpress since apparently all the eBay sellers just drop-ship from China anyway, and I was able to get both items from the same store.
Interesting you can't find a super-cheap parallel EEPROM programmer, as compared to all the other programming things. I guess it's a low-interest item now and the sellers know they're being compared to old-fashioned $300 parallel-port programmers.
[Updated on: Sat, 11 April 2020 08:33] Report message to a moderator
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7496 is a reply to message #7493] |
Sat, 11 April 2020 22:13   |
explodinglemur
Messages: 4 Registered: March 2020
|
Junior Member |
|
|
plasmo wrote on Sat, 11 April 2020 06:55Added instructions on how to modify CF interface to 16-bit. Also added the corresponding CPLD equations. Still need to upload the CP/M68K BIOS, modified memory map, and modified schematic.
https://www.retrobrewcomputers.org/doku.php?id=builderpages: plasmo:cb030:cb030_rev1:16-bit_cf_upgrade
Bill
I'm incorporating those changes into the 4-layer design I messaged you about. With pin 24 on the CPLD repurposed for the CF card interrupt, I assume the memory size selection jumper is deprecated?
Also, probably stupid question on the SDRAM, how does addressing/mapping work out with RAS0/RAS2 and RAS1/RAS3 tied together? I can lay out PCBs, and I can write some code, but I've got a big knowledge gap in the middle I'm trying to bridge
|
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7548 is a reply to message #7493] |
Fri, 17 April 2020 21:47   |
mikemac
Messages: 250 Registered: March 2017
|
Senior Member |
|
|
plasmo wrote on Sat, 11 April 2020 06:55Added instructions on how to modify CF interface to 16-bit. Also added the corresponding CPLD equations. Still need to upload the CP/M68K BIOS, modified memory map, and modified schematic.
https://www.retrobrewcomputers.org/doku.php?id=builderpages: plasmo:cb030:cb030_rev1:16-bit_cf_upgrade
Bill
Argh! Spent the day pulling my hair out as none of the results I was getting from the new 16 bit CF interface made sense to me. In 8 bit PIO mode, it should "just work" but it was failing with the data never ready after the read command was sent. I finally switched out my Transcend 4G card for my 64M. Same problem. Swapped it out for the 256M card that has CPM on it. Works like a charm! Swapped that out for another SanDisk 8M card and it works too.
Enable the 16 bit mode and both SanDisk cards work. The 4G and 64M ones still fail. Both of those should have worked at least in 8 bit mode. I'm at a loss as to why they don't. Unless the fact that the IRQ line is now connected to something [GND via a 10K resistor] causes the card to intialize into a different mode.
That's all in the boot loader, BTW.
Using the 8M card, it does boot into Linux. Quickly too! Linux currently craps out once the CF device driver gets initialized because of the new IRQ that the old driver isn't expecting. That should be pretty easy to fix. [Knock on wood!]
I'm not sure I'm going to be able to produce a single set of binaries that'll work on CB030s with and without the 16bit CF mod. At least not without some major work to the CF driver in the boot loader.
Mike
|
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7798 is a reply to message #7797] |
Wed, 17 June 2020 06:11   |
quarterturn
Messages: 86 Registered: April 2018
|
Member |
|
|
I got the USB Blaster clone to be recognized by the computer and Quartus (it needed a driver). I'm opening the unzipped "cb030_r1_cpld_16bitcf_enablecfirq_100hz" directory as a project, selecting "top.pof" in Quartus, then clicking the "programmer" button.
However, when I click "Start" to (presumably) program it, it fails. I've got the cable with the red stripe next to the "p2" text on the PCB. I've tried it the other way as well, same deal.
My CB030 has a pin missing from the JTAG header, not sure if that's intentional or part of the problem. My CB030 is working, so it not that the CPLD is failed. Also, the board is being supplied power.
[Updated on: Wed, 17 June 2020 06:21] Report message to a moderator
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7837 is a reply to message #7835] |
Wed, 24 June 2020 22:37   |
mikemac
Messages: 250 Registered: March 2017
|
Senior Member |
|
|
Sorry for the slow response, I went on vacation last week.
I have a version of the T030 boot loader that works with the 16 CF interface but Linux does not work. During initialization, Linux does something that causes the CF to go into the error state where it stays. I need to hook up the logic analyzer and twiddle the I/O lines on the 7 segment display to try to narrow down when and where this initialization is occurring.
Unfortunately, I've been busy with other projects like finding a new job. I got furloughed at the end of April. But the Linux 16 bit CF driver is still on my to do list. It's just going to take longer than any of us would have preferred.
Mike
|
|
|
|
|
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7855 is a reply to message #6931] |
Wed, 08 July 2020 11:52   |
Yves-D.
Messages: 21 Registered: January 2020
|
Junior Member |
|
|

@Plasmo
Just to let you know: another CB030 is alive, currently running the 16MB RAM test-routines.
I'm enjoying EhBasic for the moment , the CF-adapter is on order for putting the Linux port on it.
I've slightly overclocked the CPU with the 24.57MHz oscillator I had on hand.
I didn't check but remember reading that the RAM-refresh cycle is already tight with 60nS Rams at 24MHz. But it's stable so far in the RAM check.
I had a hard time getting some output from the serial port as I didn't notice the RTS/CTS pins from the USB-TTL-adapter being soldered to point T4 / T11.
It's visible your pictures, if you know where to look.
You could add a note for people using other USB-TTL-adapters that T11 must be connected to CTS and T4 to RTS.
On the logic analyser I saw the 68681 chip-select being busy and the chip responded with DTACK, i was pretty confident that the cb030 was up and running.
Just something with the output being wrong.
The Aha-moment came after reanalysing the annotated pictures and hardware handshake being mentioned in the BIOS notes. I only had GND,RX,TX connected - so that had to be it.
Thanks for your nice work being offered to the community
PS: the cb030 is sitting near my 68SEC000-Altoirola in Altoids tin-can-format, don't ask - someone had to do it , for footprint comparison
It's a 5V 12MHz 512kb Flash / 128k RAM board, using CMOS 68C681 and low power GAL - yielding around 50mA total current consumption. The GAL being responsible for 40mA alone.
It's quite amazing what they did at the end of the 68000 life-cycle.
I'll also try to construct a 3.3V variant of the same board: the 68SEC000 is 3.3V capable, I've got 3.3V RAM, Flash and GAL.
Then there's a specific Philips DUART (SC28L92) being 3.3/5V capable, 99% pin compatible (not 100%) and 99% software compatible - some changes are necessary
Kind regards,
Yves
|
|
|
|
Re: CB030, A 68030 SBC for hobbyists [message #7858 is a reply to message #7856] |
Thu, 09 July 2020 05:07   |
plasmo
Messages: 916 Registered: March 2017 Location: New Mexico, USA
|
Senior Member |
|
|
Yves,
Excellent!
The monitor (rev 0.6) on the CB030 homepage requires RTS/CTS handshake, but I've found out that RTS/CTS handshake is not really needed, so I'll post an earlier version (rev 0.5) that does not require hardware handshake. I'll also make better notation of the serial adapter interface as you've mentioned.
There is a small plug-in board with a 7-segment display that indicates the software operating status. It is an useful visual indication. It is on my 'to do' list but I'll get it published soon.
Your 68K in Altoids reminds me of Kuno that's based on P90CE201, a 68000 look-alike. It also has 512K EPROM, 128K RAM and a CPLD. The board is 93mm x 53mm, it is designed for the Arduino Mega enclosure. I need to publish the Kuno design as well. Sigh, you'll think I'll have more time in retirement...

The hard part about CB030 is assemble and program EPM7128SQC100. I hope you will consider offering CB030 board with EPM7128SQC100 assembled and programmed at whatever price you like. It will make the boards more attractive to hobbyists, especially in Europe since shipping from US is expensive and slow and there are VAT expense.
Bill
|
|
|
|
|