RetroBrew Computers Forum
Discussion forum for the RetroBrew Computers community.

Home » RBC Forums » General Discussion » KISS68030 + MF-PIC
KISS68030 + MF-PIC [message #4734] Thu, 31 May 2018 05:54 Go to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi Folks

Working towards getting my KISS68030 running and have a couple of questions based on the fact that I can get nothing out of the serial port yet...Further, experimenting with various DRAM sometimes I get a solid green RUN light and I can see what looks like good activity on the bus monitor. Other times it wont RUN on power up or flips over to STOP within a few seconds and looks like it locks up.

So the questions
1. Which ROM file should I look to use? kissbios.rom or the 68K firmware version 10.2 file ? Will the firmware come up even if there is no DRAM?

2. I found the RAM Compatibility list https://www.retrobrewcomputers.org/doku.php?id=builderpages: tschaaps:kiss68030-ram-compatibility and reading this with the board page I take it that I can only use single sided 16MB, double sided 32MB, single 64MB or double sided 128MB 72pin, 60ns DRAM. The 4MB and 8MB DRAM that I have just wont work ? Any more detail here?

3. Is there jumper documentation for the KISS68030? Have setup as per the photos but I would like to understand that better if possible.

4. I see reference to flash codes on the CPU board in various threads. Is there a more formal list or a video showing correct operation etc...?

for reference my config is as follows
- 12 slot backplane (has been working nicely with the SBCV2 and a lot of other cards)
- ECB Bus Monitor
- MF-PIC
- KISS68030 with the modifications as per https://www.retrobrewcomputers.org/lib/exe/fetch.php?media=b oards:ecb:mf-pic:update-2015-09-22.txt


Any pointers, information and suggestions welcome.

regards

Richard
Re: KISS68030 + MF-PIC [message #4735 is a reply to message #4734] Thu, 31 May 2018 09:09 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
trick-1 wrote on Thu, 31 May 2018 05:54
Hi Folks

Working towards getting my KISS68030 running and have a couple of questions based on the fact that I can get nothing out of the serial port yet...Further, experimenting with various DRAM sometimes I get a solid green RUN light and I can see what looks like good activity on the bus monitor. Other times it wont RUN on power up or flips over to STOP within a few seconds and looks like it locks up.

**The KISS-68030 must be updated to version 1.1. The Wiki pages have the update info.
**The MF/PIC boards 1.0, 1.1, 3.0, 3.1 have one MANDATORY update; and one update required for the KISS-68030 board if you want to run Linux. Update to -008 suffix.


So the questions
1. Which ROM file should I look to use? kissbios.rom or the 68K firmware version 10.2 file ? Will the firmware come up even if there is no DRAM?

**The KISS instructions indicate that TEST15 should first be burned to ROM and run on a new board.
**After TEST15 passes, then the version 10.2 ROM (May 2016) should be used. Install the 68030 version, not the 68008 version.


2. I found the RAM Compatibility list https://www.retrobrewcomputers.org/doku.php?id=builderpages: tschaaps:kiss68030-ram-compatibility and reading this with the board page I take it that I can only use single sided 16MB, double sided 32MB, single 64MB or double sided 128MB 72pin, 60ns DRAM. The 4MB and 8MB DRAM that I have just wont work ? Any more detail here?

3. Is there jumper documentation for the KISS68030? Have setup as per the photos but I would like to understand that better if possible.

**Check the PDF schematics. The RAM size and 'sided-ness' must be set up correctly. ROM jumpers will need to be correct, too. The wait state jumpers should be set to conservative values, i.e., higher. I think the board itself is self-explanatory.

**What speed oscillators are you using? 8mhz determines the refresh timing, and may not be changed. The other two may be as high as 66/33, 64/32, 64/25, 50/25. Second is CPU timing; first is DRAM timing, which is usually about 2x CPU timing. Wait states matter, too.

4. I see reference to flash codes on the CPU board in various threads. Is there a more formal list or a video showing correct operation etc...?

**The README file in the Zipped distributions should have the info. The 10.2 ROM readme is complete; the TEST15 is lacking.

for reference my config is as follows
- 12 slot backplane (has been working nicely with the SBCV2 and a lot of other cards)
- ECB Bus Monitor
- MF-PIC
- KISS68030 with the modifications as per https://www.retrobrewcomputers.org/lib/exe/fetch.php?media=b oards:ecb:mf-pic:update-2015-09-22.txt

**The Bus Monitor may not be useful. The KISS puts addresses (32) on the bus, but not data; unless, it is accessing the MF/PIC board.

Any pointers, information and suggestions welcome

**I thought the MC68030 was overly sensitive to VCC. Check that VCC on the CPU board is within 5% tolerance, or better. The extra caps (elecrtolytic or tantalum) are a help. Extra electrolytic decoupling on the backplane is also recommended. I think you received the Backplane build notes.

**RS-232 problems are the !#@!@##-est to track down.

regards

Richard

Re: KISS68030 + MF-PIC [message #4736 is a reply to message #4734] Thu, 31 May 2018 09:20 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
A good source for memory on eBay is:

http://stores.ebay.com/18004Memory_W0QQ?_nkw=EDO

$10 or less; use Micron chips: I have memory in all sizes from these guys. 60ns, 72-pin, +5v. EDO or FPM. parity not necessary (and not used).

--John
Re: KISS68030 + MF-PIC [message #4737 is a reply to message #4736] Thu, 31 May 2018 19:07 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Thanks John

So found 2x 16MB SIMMS in my collection and have setup the board as per the information you have pointed towards.

Tried the following
1. TEST15 - I get 4,3,5 flashes with a pause between each set of flashes, then it sits at RUN. Nothing on the Serial Port although if I type enough I will get garbage echoed back.
2. 10.2 ROM - I get 1 flash and then it sits at RUN. If i take the MC-PIC board out I get 5 flashes so as per the README that is expected and I think it means that the 030 board is working.

Open to suggestions but am thinking there is something not quite right with my MF-PIC board. Might replace the MAX232 first and then maybe time to get the scope out.

Re: KISS68030 + MF-PIC [message #4738 is a reply to message #4737] Thu, 31 May 2018 19:10 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
forgot to mention had already done the updates per the erata for both MF-PIC and 030 board.
Re: KISS68030 + MF-PIC [message #4739 is a reply to message #4737] Fri, 01 June 2018 12:49 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
The 68030 board looks good. Otherwise, you would see no LED flashes. Read the entire README.txt file in the TEST15 bin or src directories. The SRAM must be functioning, else you would get 2+8 flashes. There is no indication that the MF/PIC board is not found at 0x4?, with the UART at 0x48.

Has the MF/PIC board been used with any other system. If it has not, then the problem is most likely between the UART chip and the terminal you are using. I use a Wyse 30, and I have to provide ALL the modem signals to the terminal to get it to work. This will happen if you have used the MAX235 chip, but will not happen with the MAX232.

Also, how do you have the terminal/modem jumpers set? (3.0 board). If the MF/PIC is set to look like a terminal, then a null modem cable will have to be used between it and your terminal. If you have a straight-thru cable, then the MF/PIC board will have to be set to look like a modem. Two RS-232 devices may never be connected terminal-to-terminal. The connection must always be terminal-to-modem.

The MAX23x chip is most likely okay. Check the voltage levels on the RS-232 signal lines. My best guess is that the problem is here.

--John
Re: KISS68030 + MF-PIC [message #4740 is a reply to message #4737] Fri, 01 June 2018 13:10 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
The LED flash pattern is telling: 4, 3, 5

the 5 indicates that the UART is not being found. If it had, the sequence would be 4, 3, 2, then output @9600 bps.

--John
Re: KISS68030 + MF-PIC [message #4741 is a reply to message #4734] Fri, 01 June 2018 13:32 Go to previous messageGo to next message
adx is currently offline  adx
Messages: 21
Registered: October 2015
Junior Member
When I had UART problems with my m68k + mfpic it ended up being a bad buffer chip on the mfpic board.
Re: KISS68030 + MF-PIC [message #4755 is a reply to message #4741] Sat, 09 June 2018 17:15 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi folks

So still working on this. Found a bent pin on one of the 74F244s on the 030 board so that certainly wouldn't be helping my cause. Not had time to do any additional testing this week but a quick powerup test didn't show any new result. Maybe get some time later today.....

Will keep you all posted

Richard
Re: KISS68030 + MF-PIC [message #6445 is a reply to message #4755] Wed, 14 August 2019 00:52 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi everyone, so it has been over a year but I have finally got back to this issue. I realised that I had put the SIMM sockets in the wrong way round....removing them destroyed the board so I built a new one, paying attention to the SIMM socket orientation.

I am basically still at the point of getting the LED to flash 4,3,5

So i have the logic analyser on the MF-PIC board. I see /BOARD_SEL go LOW when A4-A7 are the same as jumpers (on,off,on,on), /RD and A0-A3 transition around this time also. So I suspect that for some reason the 16C750 is not responding.

Writing this I will go check the oscillator is working just to make sure....

open to ideas/suggestions, what else should .I check here?
Re: KISS68030 + MF-PIC [message #6446 is a reply to message #6445] Wed, 14 August 2019 09:42 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
The communication from the CPU by way of the CPU board LEDs was meant to help with situations like this: the problem of bringing up two boards simultaneously.

I can offer these suggestions:

1. Default data rate, before running Setup, is 9600. Be sure this is what you are using.
2. Serial communications are notorious. I suggest investing in one of the multi-LED line testers that can be put between your terminal and the UART.
3. Make sure the MF/PIC output is "modem", not "terminal." This way you do not need a null modem cable. If using a null modem cable, then the MF/PIC UART output should be set to "terminal."
4. Check to see what line status your terminal requires. I have a terminal that will not operate unless ALL modem control lines are at the proper levels. If using the serial port on your PC with a terminal program, generally, the modem control signals are ignored.

5. Do the usual check of all solder joints on the MF/PIC; and check that all chips are properly seated. I would assume that you have done this already, if you are at the stage of putting a probe on signals on the MF/PIC board.

--John
Re: KISS68030 + MF-PIC [message #6447 is a reply to message #6446] Wed, 14 August 2019 14:38 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi John

Thanks for the response.

So the LED's on the CPU board are certainly helping and at least it is indicating that the CPU board is functioning Smile

1. Yes I have Tera Term setup to 9600 (I can get a console on my SBCv2, CPU280 and MiniMax8085, Sun Machines) so I know that part is definitely working.
2. Yes serial can be a pain. I have an HP4951C protocol analyzer on the bench for this reason. I just don't think it's getting that far but will double check. The way I understand the final LED flashes is that it is trying to "connect" to the UART. What sort of test is it? Loopback or something else? I have also double checked the cable to make sure that the right signals are mapped to the right pins.
3. Set to terminal like all my gear. I have a known good null modem cable.
4. Yes serial port on a PC as per 1 known working
5. Yes have done so.

I have also pulled all the majority of the 74x chips and tested them in a TL866, all checked OK. I have also swapped over the NS32202 and 82C55 with no change.

Have also pulled all the 74x chips on the CPU board and tested them just to be certain. Particularly the bus interface chips. They are OK to.

I have tried 2 different backplanes. no change.

I am at the point of re-checking each socket to make sure that I haven't bent a pin.

If that doesn't come up with anything today I am going to build another MF-PIC board which I have.

thanks again

Richard
Re: KISS68030 + MF-PIC [message #6449 is a reply to message #6447] Wed, 14 August 2019 22:20 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi john and the rest of the folks out there

One word: success!

I spent a few hours going over the original board, checking joints under the magnifying glass and sockets for continuity.
Couldn't find anything that would change the out come.

Have therefore gone and built up the second board using machine sockets..

I plugged it in and turned it on...I was greeted with 4,3,2 LED flashes and the following on the console.

Have then put kiss bios 10.2 in the rom and it boots to the following screen

I would like to express my appreciation for all the help and suggestions. Truely appreciated

Richard

[Updated on: Wed, 14 August 2019 23:16]

Report message to a moderator

Re: KISS68030 + MF-PIC [message #6452 is a reply to message #6449] Thu, 15 August 2019 08:17 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
TEST4.BIN should be putting out an endless number of lines. It is not.

The final 'A' means the test is of Aligned memory (i.e., 32-bit word is accessed with one memory cycle). Subsequent tests may be flagged 'U', for Unaligned (i.e., 32-bit word address is not a multiple of 4).

The SRAM is good, and is used during the DRAM testing. It looks like there may be a DRAM problem.

With the 10.2 BIOS, boot, while the boiler-plate text is coming out hit Reset. This is a convenient way to get to the memory test that is part of the BIOS. It is essentially TEST4.

With DOUBLE = 1-2, the auto-select should be working. 64M ON is correct.

*** Which revision of the board are you using? There is an update to the earliest board which amounts to a connection to pin 21 of U405 (DRAM9). 1.1-006 is the highest revision of the board.

What CPU/DRAM oscillator combination are you using? 25/64 or 25/50 would be suggested.

Try DRAM is slot A only. I trust your DRAM is 60ns or faster.

--John
Re: KISS68030 + MF-PIC [message #6453 is a reply to message #6452] Fri, 16 August 2019 00:29 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi John

Thank you for the information and the help so far.

So I am running 16/64 with a 25Mhz 030. Reading the parts list I understood this to be OK...but maybe not? Thoughts?

I am running the following hardware



    030 board - KISS 1.1-006 no errata
    MF-PIC board rev 3.1-007 with mandatory update wire RP2:8 to U3:28 ONLY at this stage.
    ECB Mini Backplane - 3 Slot
    PC Powersupply
With regards to the memory I have a small pile of 50ns and 60ns memory to chose from. Also, I did purchase 2x 64Mb 50ns EDO SIMS per your link earlier in the thread. I have cycled through all of these today and I get output similar to the following each time. Thanks for the tip about accessing the mem test using the reset.

Welcome to the KISS-68030 System BIOS 10.2

Copyright (C) 2011-2016 John R. Coffman <johninsd@gmail.com>
Copyright (C) 2015 William R. Sowerbutts <will@sowerbutts.com>


RESET exception trap requests a full memory test.

04000000 <--HMA1 unequal to HMA2--> 0C000000
Using HMA = 0C000000


The memory tests use a 33-bit rotating bit pattern; i.e., each successive
memory word32 is written with a pattern rotated left one bit from the previous.
The 33-bit length is used to pick up shorted or open address traces. The seed
value for each of the test lines is shown on the left, long+Xbit.


33-Bit Address Aligned or
Pattern+X Start End+1 Unaligned

0635F570 1 00000000 00010000 A pass B pass
632A4163 1 00010000 01000000 A fail
BA2CF0DD 0 00000002 00FFFFFE U


I also tried BIOS 10.1 and although it got further I had no response on the keyboard. If I sent a break sequence then I could step through all the way to the boot prompt.

I have therefore gone back to test15 - test4.bin to make testing/diag easier. It too is showing failure.

TEST4.BIN 12-Mar-2016 John R Coffman. Licensed for hobbyist use only.

MF/PIC UART is 16750
SRAM present
SRAM Btest pass

00000000 04000000 A fail
00000001 01000001 U fail
01000002 02000002 U fail
02000003 03FFFFFF U fail
33251231 7AED3431 U fail

Also loaded up TEST3.BIN as I see it also deals with 64Mb RAM. Removed the jumper and gave it a go. Similar result.


I have gone over the board again, checked I have used the right chip version LS,AS, F etc per the file KISS-68030-parts-list.pdf.

Also gone over with the soldering iron just to make sure....still getting memory test failure.

In case there is something visually obvious on the board I have attached front of board photos with memory in and memory out showing pin 1 of the 72pin SIM closest to the board edge per the silkscreen.


Richard
  • Attachment: IMG_0712.JPG
    (Size: 2.56MB, Downloaded 6 times)
  • Attachment: IMG_0711.JPG
    (Size: 2.74MB, Downloaded 9 times)

[Updated on: Fri, 16 August 2019 00:55]

Report message to a moderator

Re: KISS68030 + MF-PIC [message #6454 is a reply to message #6453] Fri, 16 August 2019 07:18 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
Does your memory support CBR refresh? It must. (CBR = CAS before RAS)

I note that the first test of 64K of memory passed; then with longer tests, failures.

This board has had issues with:
1. Voltage sensitivity: check that board voltage is as near as possible to +5v, say, within 2%. 68030 is finicky, and a power hog.
2. Decoupling. High frequency decoupling (0.1 uf ceramic/monolithic caps) is okay. However, more electrolytic decoupling may be in order. I used some very old electrolytic caps on the first board I built up, and it took hours of burn-in before the caps re-formed; then it behaved.

[Refresh transient on some memories is 4A!! Caps on the memory cards should be handling this.]

16/64 is probably okay. Some 68030's are rated for a lowest frequency, too. I think 16mhz is okay on a 25mhz chip, but check the spec sheet. Do you know which mask revision you are using?

BIOS, CP/M, mem tests -- all run in real mode. Try disabling the Cache and/or MMU by installing jumpers. MMU is only required by Linux. Especially check that the low memory test (64K) passes with the Cache disabled. Very important that it does pass.

Passing the first 64K memory test says that it should be possible to R/W all of DRAM memory.

So, check: CPU speed, voltage, decoupling, CBR refresh, no Cache, no MMU.

--John
Re: KISS68030 + MF-PIC [message #6455 is a reply to message #6454] Fri, 16 August 2019 16:36 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi john

Will take a look further at the CPU and it's tolerance. Also will find out if the memory is CBR..

Have increased electrolytics to 1000uF @ 100V - probably overkill but for the sake of the experiment.

Cache and MMU both have jumpers installed already.

Have left test4.bin running and currently failing...will let it run for a few hours. Will check voltage later.

Cheers

Richard

Re: KISS68030 + MF-PIC [message #6456 is a reply to message #6455] Fri, 16 August 2019 21:08 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
Leaving the power on will form the capacitor(s) on the memory strips, too -- in case they are old inventory.

Virtually all memory (32-bit, 72 pin, DRAM) is CBR refresh. I have had the best results with Micron chips, and Samsung chips.

C74N (ceramic) and F91C (plastic) are the most common MC68030 masks. I have both of these running at 25mhz and 32mhz, respectively. The list of masks is attached.

--John
Re: KISS68030 + MF-PIC [message #6458 is a reply to message #4734] Sat, 17 August 2019 01:34 Go to previous messageGo to next message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Hi John

OK so todays further activity
- Replace the PSU with a Mean Well 5V 14A fixed supply. That should be able to deliver plenty of current. Tweaked the supply so that it gave out 5V with the board powered on (5.03 powered off).
- With the new supply I ran the BIOS 10.2 Memory test and it managed to pass 2 lines before failing. It has gone back to failing after the first line now, so I suspect that it is timing or tolerance issue.
- Tried un-jumpering cache and mmu. but no improvement. Put the jumpers back in.
- The mask on my 030 is D66C which I note is a different fab from the 2 that you have referenced.
- I have confirmed that the 25Mhz 030 will operate at 16Mhz http://datasheets.chipdb.org/Motorola/68030/MC68030.pdf page 23 min: 12.5Mhz Max: 25Mhz for a 25Mhz part.
- I have ordered a 25Mhz oscillator, 68030RC16 and 68030RC33.
- might try and find some of the known working memory per the memory list that I have seen.

Unless you have any additional thoughts I will put this on the shelf until the parts come in.

thanks again

Richard

[Updated on: Sat, 17 August 2019 01:37]

Report message to a moderator

Re: KISS68030 + MF-PIC [message #6460 is a reply to message #6458] Sat, 17 August 2019 09:11 Go to previous messageGo to next message
jcoffman is currently offline  jcoffman
Messages: 259
Registered: October 2015
Senior Member
I did some experiments back when the KISS-68030 was first released with differing CPU/DRAM oscillators, the written results of which are long lost. As I recall, the good combinations were: 32/64, 25/64, 25/50. I do not remember the results using lower CPU speeds.

I am a little concerned about your current oscillator mix: 16/64 as I recall. I hope the boost to 25/64 will help.

--John

Re: KISS68030 + MF-PIC [message #6461 is a reply to message #6460] Sun, 18 August 2019 05:09 Go to previous message
trick-1 is currently offline  trick-1
Messages: 119
Registered: September 2016
Senior Member
Thanks john,

Once I get the parts in I will run some tests on the following combinations 16/64,25/64,32/64,25/50 and document the results, as appropriate, with the 16/25/33Mhz 030's.

stay tuned.

Richard
Previous Topic: Battery Power for SBC V2 and Backplane
Next Topic: I'm here.


Current Time: Sat Aug 24 10:59:56 PDT 2019

Total time taken to generate the page: 0.02106 seconds