rainbow needs help

[RetroHacker_]

Before you break out the soldering iron, try "piggybacking" known good RAM over the existing RAM. Basically, you just bend in the pins of the new RAM very slightly, and stick it on top of the old RAM, so it makes contact with all the pins - the tiny amount of spring tension of the legs will hold them in place. Then try it, and see if it passes the test. If it does, then simply remove all but one, and keep trying it, moving the one chip to the next in line until you find the one bad chip.

Note that this method doesn't always work - if the old RAM has a shorted cell, or something, piggybacking RAM won't change anything. But it does work fairly often, and it's a good troubleshooting technique, and has, in my experience, really helped narrow down a bad chip when the onboard diagnostics don't make it easy.

-Ian

 

[harry]

Before you break out the soldering iron, try "piggybacking" known good RAM over the existing RAM. Basically, you just bend in the pins of the new RAM very slightly, and stick it on top of the old RAM, so it makes contact with all the pins - the tiny amount of spring tension of the legs will hold them in place. Then try it, and see if it passes the test. If it does, then simply remove all but one, and keep trying it, moving the one chip to the next in line until you find the one bad chip.


Note that this method doesn't always work - if the old RAM has a shorted cell, or something, piggybacking RAM won't change anything. But it does work fairly often, and it's a good troubleshooting technique, and has, in my experience, really helped narrow down a bad chip when the onboard diagnostics don't make it easy.

-Ian

Hi Ian, thanks for reply, I have never had to try out checking ram with the "piggy back" method, I have often read of people trying it , so this afternoon I tried it out , I installed six new 4164 ic's on top of the first bank of six, no change,I then moved them over to the next bank of six, no change, I moved 4 of them to the remaining row, hoping something would change,... but nothing changes on screen . still confronted with "message 19 main board." it was worth trying , and I
was hoping for some sort of reaction. I am waiting for some psu headers to arrive, I can then try and see what is happening, with the board out of the case.
thanks again.......harry

 

[Lou - N2MIY]

Harry,

You mentioned 16 of the memory dips. In the print for the motherboard, there were eight memories in two rows of four in one corner of the motherboard. Also, in NEXT's photo of his motherboard, I see the same eight I saw in the print. Where are the 16? Are they on another board, or are there eight more on the motherboard that I missed?

Lou

 

[harry]

Harry,

You mentioned 16 of the memory dips. In the print for the motherboard, there were eight memories in two rows of four in one corner of the motherboard. Also, in NEXT's photo of his motherboard, I see the same eight I saw in the print. Where are the 16? Are they on another board, or are there eight more on the motherboard that I missed?

Lou

Hi lou, if you look at Next's photo the ram chips are in two rows of 6 starting from the top left, but only 4 in the final third row, his board appears identical to mine. .....harry

 

[Lou - N2MIY]

Harry,

Off thread, my confusion has been explained to me. You have a 100B (aka 100+) motherboard. The print set I have is for the 100A. Best of luck with all your desoldering.

Lou

 

[harry]

I've done a bit more work on the rainbow,and made up the extension cables for the psu to m/b, and done some further testing with the m/board out ,( before surgery begins) one problem it seems, is what apears to be a short between the positive side of one of the electro caps, and psu neutral ( ground) ? I snipped one end of the cap to isolate it ,and this shows there is a short on the board, unfortunately I cannot trace the circuits as is appears to be between layers? I am really needing a link or info to get some shematics of this 100B board,I have of course been searching on line, and will continue, but any help would be great. .....many thanks Harry

 

[harry]

Harry,

I would bet lunch that you will find a bad bit in one of those 4164s. I have fixed enough machines to have seen a few this happen a few times now. 4116s seem to be even worse though.


Lou

Lou, I felt I must let you know, you were correct in thinking it was bad ram stopping the rainbow from booting, I decided to go for it and remove all 16 ram chips,
and fit new turned pin sockets in place, I started by removing the bank of four to see how things would go, I managed to damage one track, but learned ! I have put up photos which show the repair I've done with the yellow wire. I tried booting after replacing each bank, but as per sods law! it was not until I had replaced the final bank the on screen TESTING carried on and completed , It was great to see the rainbow 100 screen as shown, you won't believe how relieved I was after over three hours work! 4164 15ns chips were fitted, and with the aditional ram board the rainbow shows 256k memory.
I am now needing to make or obtain a suitable boot disk, there is a hard drive fitted but although it spins, I have no responce from the keyboard when I choose W or any of the boot options, infact the only responce is via the coursor keys which when pressed starts/ stops the floppy drive, do I need to boot from disk to configure something to be able to use the keyboard,I imagine the keyboard should be working.from when the first screen appears so one can choose an option? or not so?
many thanks for your help, I am now a little closer to a working machine, ...............harry

 

[Lou - N2MIY]

Harry,

That looks like a nice bit of work. Did you have a desoldering iron, or buy one at Maplin?

It appears that the shotgun approach did the trick. Now that it's all in sockets, it'll be a bit easier to repair next time there's a bad bit in memory. It also looks like you recovered fine from the destroyed trace.

So now you need to make a system disk. If you go here you will find the famous Imagedisk utility: http://www.classiccmp.org/dunfield/img/ At the bottom of the page you will find a link labeled "System / Install disks" . Follow that link and you will find Rainbow 100+ system disk images. You will need to download and install the Imagedisk software on a PC fitted with a 1.2MB 5-1/4" floppy drive.

Let us know how it goes.

Lou

 

[tezza]

Well done Harry, you are getting there.

The Rainbow is an interesting machine. Wish I had one.

Tez

 

[tingo]

Yes, well done Harry! Nice work.

 

[harry]

lou, thanks for link, and info, I'll look at making a disk, although me, and disk images don't always work out? but I have had some success in the past.
I do have a desoldering station, which is a great asset, and without it I don't think I would have tackled this job.


Tez, I seem to remember you wanting a lisa not so long ago, and look what turned up! there is bound to be a rainbow somewhere with your name on it!

 

[harry]

[Q

So now you need to make a system disk. If you go here you will find the famous Imagedisk utility: http://www.classiccmp.org/dunfield/img/ At the bottom of the page you will find a link labeled "System / Install disks" . Follow that link and you will find Rainbow 100+ system disk images. You will need to download and install the Imagedisk software on a PC fitted with a 1.2MB 5-1/4" floppy drive.

Let us know how it goes.

Lou[/QUOTE]

Its been some time since I done the work on the rainbow, but this week I managed to make the cpm 86 disk and the dos disk, from the link Lou kindly gave me.( getting a 386 pc built up and configured has been another story!!) The rainbow will boot up to cpm or dos, no problem, but I now find myself confronted with no characters from any of the letters I type, only what appears to be a comma instead of any letter? The numeral keypad works fine, as does the setup key, although the setup page is not as per my manual.the usual options it seems are not given. seems to be stuck in some sort of terminal mode. I am using the keyboard from a vt 220 but it is an LK 201 so I assume it should be ok.
The machine will reset correctly ( ctrl/setup), and I am confronted with the normal boot screen, but typing any letter to boot from has no response, the only way I have found to boot the floppy is to use the coursor up key, this then loads the system disk. I have been through the setup screen many times, but am unable to get characters on screen? pressing shift and any letter will change the screen character to a small square, same response to any letter key. any help really appreciated...........Harry.

 

[Lou - N2MIY]

Harry,

I'm glad to hear that the disk creation went well.

I suppose the first order of business is to work out if the problem is at the keyboard end, or in the machine. I don't have the LK201 prints in front of me here, but it's worth looking at the keyboard's output on your scope. Is the keyboard really sending the same character every time you strike a letter key? If so, then the trouble is with the keyboard. It's worth opening the keyboard anyway to make sure some conductive junk isn't causing trouble with the keyboard encoder IC.

Lou

 

[harry]

Lou,

I dug out the vt 220 and checked out the keyboard, it appears to be working ok, types all characters to screen fine. Looks like I will have to check things out inside the rainbow, I was hoping it was going to be a set up glitch, maybe not so simple . .....Harry

 

[jdreichard]

I hope you don't mind me reviving an old thread, but I may have the same problem with my DEC Rainbow 100B as harry.

I am getting all seven diagnostic LEDs at power on. The technical reference says this is "Error 30 (rom crc, rom 0)".

Here's what I can verify:

1. PSU voltages are in range.
2. 8088 and Z80 both checked and working.
3. Other voltages on the system module seem to be okay.
4. I dumped all three PROMS and verified their contents to match ROM dumps I found online.

I fear, like harry, I've got bad RAM, which would explain why the ROMs aren't loading and the processor isn't executing code.

But, thinking about desoldering 16 chips x 16 pins each gives me 0001 0000 0000 reasons to be anxious!

Any other advice would be appreciated!

Thanks!

 

[DougIngraham]

I have seen old PROMs get slow. The external readers will present the address and then quite some time later (many microseconds) read the data. The data looks good when read by an external device but in the system where it is expected the data is stable in 500ns or less they dont read well. I suggest you try burning a set of PROMs and trying those. It might not be your problem but it is possible and an easy thing to try.

Good luck!

 

[jdreichard]

I have seen old PROMs get slow. The external readers will present the address and then quite some time later (many microseconds) read the data. The data looks good when read by an external device but in the system where it is expected the data is stable in 500ns or less they dont read well. I suggest you try burning a set of PROMs and trying those. It might not be your problem but it is possible and an easy thing to try.

Good luck!

Thanks Doug! I'll give that a try! I really appreciate the response!

 

[jdreichard]

I wanted to follow up on this thread. Miraculously, I have the DEC Rainbow running. Not RAM, not ROMs, but the oscillator! Upon checking with an oscilloscope, I discovered that the pin that should have been outputting 24.07342 MHz for the rest of the system was dead as a doornail. So here's how I brought my Rainbow back to life:

1. I used an Arduino and the Adafruit SI5351A clock generator chip to generate the right frequency (25 * (28 + (11101/12500))/30).
2. To make this easy, I soldered the SMA connector to the board and then hooked up my oscilloscope probe with a BNC coupler.
3. I hooked onto the clock's pin on the bottom side of the main board.
4. Booted up the Rainbow and... viola! No Massive Hardware Failure!
5. Keyboard error of course.
6. I hastily constructed the LK201 (of course, like everyone else, the LK201 I have is dead as a doornail) emulator on another Arduino Nano.
7. It kinda works --- and sometimes gets me past the system check to the main screen.
8. I can get to the Terminal, but the emulator incessantly and erratically issues key repeats. It may be a software issue.
9. Onward and upward.

But, the Rainbow lives!

I found a supplier of 24.073 MHz crystal oscillators. I will replace it when they arrive and update this (and cross threads) on the progress.

Thanks!

Josh

 

[tradde]

That sure is progress. Congrats. You'll have it fully working in no time.