• Please review our updated Terms and Rules here

New Sol-20 Owner

snuci said:
I tried to reseat the memory but other than that, I don't have memory to try to replace. I went from 9 lines of 9090 to 16 lines, to 1 line to 9 lines and am back there.
Click to expand...

With working power (new power switch, C14 entry module and a few new wires and heat sink hardware), I'm now at about the same point you are (looking at power-on video displays). The 9090... is a good sign. Here's how to proceed:

1) Pull the personality module and any external RAM. For now, we don't want any system RAM.

2) All video RAM should be installed (U14-U21)

3) After power up, use a jumper wire to momentarily short pin 1 of U2 to ground. This forces the starting line counter on the screen to line zero so you'll get a full 16 lines of display data. The random power up state of this register is why you saw differing number of video lines displayed.

4) Use a jumper wire to momentarily short pin 2 of U75 to ground. This ensures the video display circuits are all enabled.

At this point, if your screen is all 9090909... then your video section (and video RAM) is all good. This pattern results from the CPU fetching 0xff's from the non-existent personality RAM area. This is a RST 7 instruction which pushes the return address onto the stack and jumps to address 0x38. There the CPU finds another RST 7 instruction, pushes the return address (0x00, 0x39) onto the stack and jumps back to 0x38 again. As you can see, this quickly ends up writing 0x00, 0x39 through memory (including the video RAM) since the 8080 SP is decrementing with every RST 7 instruction. In the character generator ROM, 0x00 is the squarish "0" you see and 0x39 is, of course, an ASCII "9".

Mike
 
Progress! When running the immediately preceding test, my Sol displayed garbage in the first 32 characters of all lines, but had the desired "9090..." pattern in the right half of all lines. This sounded like A5 from the CPU was stuck high since only locations in video RAM with A5 set were being initialized with 0x39, 0x00. Sure enough, I pulled the 8T97 ( U68 ) that buffers A5 from the CPU and the corresponding output pin had "rotted" off. I put in a new 8T97 and was rewarded with a full screen of 909090..."

I then installed the personality module and was greeted with a cleared screen, the SOLOS prompt and a cursor! I powered down, connected the keyboard, but as expected, got nothing from the keyboard. I unscrewed the keyboard mechanism from the keyboard PCB and rotted foam fell out everywhere. I have, however, been able to type a few monitor commands by "typing" directly on the PCB pads with my fingers. So, time to find or make keyboard pads I guess!

Mike
 
Last edited:
Mike, that's fantastic!

I haven't had a chance to try out the instruction you posted yet. I will either try tomorrow or wait for my switch to arrive on Monday and then try. I'm on my last box of fuses but I'll give it a shot if I don't get called into work. I've tried to follow the Sol-20 manual with the progressive tests but the tests are in the process of putting it together and doesn't really tell you what you need to remove from a fully assembled unit to try that test. I get the "9090" with my Solos module attached so I may be in a bit more trouble.

Maybe I'll get the keyboard ready and replace the foam pads until the switch and memory come. I already have them made up and I assume they are the same diameter as the Apple Lisa keyboard.

Thanks again. I will report. Sounds like your some foam pads away from success.
 
By the way, for the video tests I mentioned earlier, you don't need to remove the data RAM on the main board (i.e., U3-U10), just make sure there is no external S-100 RAM plugged in.

Mike
 
I did the test as described and it worked fine. I also swapped video for system memory and it also worked so I'm guessing my memory is okay. Getting down to the Personality Module test is where I have an issue. I tried the Chapter IV Step 11 test and whether it's in or not, it acts like it's not there. I see the same pattern of "9090" even with personality module plugged in. The module look sokay and I cleaned the contacts just in case. It looks like this:

IMG_20140406_072355.jpg

I think I'll have to go to Chapter IV Step 9 and check voltages with the 9216 and 74LS08 removed.
 
Just bought a Sun type 4 keyboard off eBay. Hopefully the pads are better in the old Sun keyboard than in the even older Sol keyboard!

Mike
 
snuci said:
Corey, would you mind taking a picture of the screws next to a ruler? I also want them for completeness and the quest for stuff like this is part of the fun.

I would really appreciate this (while I wait for my memory and power switch).
Click to expand...

Sorry I just saw this. I was very sick all last week and even missed VCF. I'll try to take that picture tomorrow and post it. I have my screws put away since I open the covers on my sol-20s way too much...

Cheers,
Corey
 
After some testing, I'm getting about 85 ohms between B14 and B15 of the J5 personality module edge connector. It should be 10k or more or a short is present. I'll have to print off some schematics and take a look further.
 
snuci said:
After some testing, I'm getting about 85 ohms between B14 and B15 of the J5 personality module edge connector. It should be 10k or more or a short is present. I'll have to print off some schematics and take a look further.
Click to expand...
Actually, that sounds about right. You're looking across the main board 5v supply to ground -- not just the personality module.

Mike
 
Hrmmm... Maybe I should start looking at the Solos-1 9216 ROM chip then just to be sure it's operational. I did see a thread started by glitch where Nama was able to test out his 9216 ROM by recreating it on a 2716 and comparing it here. I will also check those voltages in the thread. Unfortunately, the link to the ROM image in that thread has changed.

If Nama is watching, can you tell me if "/Sol 20/ROM images/SOLOS.bin" is the correct image? The image "/Sol 20/ROM images/ROM images/solos.rom" is bigger. I assume the first one is correct? I'd like to try to read my 9216 ROM and compare.

Thanks for the continued help. I'm still learning and appreciate it.
 
snuci said:
How do you dump it? I tried to read it as a D2716 and got all FFs for the first 1k and 00s for the last 1k
Click to expand...
I cheated and read it out of a running SOL! The 9216 is a hybrid of the 2708 and the 2716. Pin 21 and 19 are the problems.

Pin 21 is A10 on the 9216, Vpp (normally +5v) on the 2716, and -5v on a 2708.
Pin 19 is +12v on the 9216, A10 on the 2716 and +12 on the 2708.

It's highly unlikely that the ROM is the failure point, I wouldn't bother trying to make an adapter to read it. Instead, I'd start looking for chip select issues and for problems with the "internal" data bus path. The 8080 reads data from both the 1K RAM and the personality module on the internal bus instead of the normal DIO bus. A set of four mux's, and of course, the logic associated with generating the select and enable lines for these muxes, control data coming into the 8080.

Mike
 
Thanks for that but I did manage to make it work!


I couldn't read the 9216 properly so I just figured I'd "go for it" and programmed the 2716 and made the adapter from the other thread. When I started the Sol-20 up, I saw garbage on the screen. This was new! I swapped out the memory one by one and nothing changed until I got to the last 2102. When I swapped it, presto!
Sol 20 - 1.jpg

I had to then retry my 9216 Solos-1 chip and it also worked. I really don't understand why. It sounds hauntingly familiar to Nama's fix here.

Once I put everything together, I then did the little sample in the manual (I had already redone my keyboard over the weekend) and all looks good.
Sol 20 - 2.jpg

Here it is in one piece.
Sol 20 - 3.jpg

I have one last issue. It keeps blowing fuses. I went through about 7 of them even after changing my switch. I'm not sure why but I do see some sparking in the "commoning block" between the switch and the fan. I thought it was the switch but I'm not sure what that is. I was not able to get slo-blo fuses and used normal 1A fuses. Would this matter?

In any case, I'm still happy. Thanks for the help.
 
Last edited:
Check your commoning block may be setup wrong. Some wires are supposed to touch/connect and others not. I'd sort that out before you burn out something on the motherboard.
 
I already checked that via the manual. All three wires in both commoning blocks are in the right spot and all wires within each block is connected to one another. Unless they are not supposed to be, seems okay. I'll double check the manual.
 
snuci said:
If Nama is watching, can you tell me if "/Sol 20/ROM images/SOLOS.bin" is the correct image? The image "/Sol 20/ROM images/ROM images/solos.rom" is bigger. I assume the first one is correct? I'd like to try to read my 9216 ROM and compare.
Click to expand...

Sorry for the late reply, I've been following this thread on and off. Seems that you have sorted everything out now. Great stuff!!!!
If you can get your hands on a Northstar disk controller board and a 5.25" drive, then it's a very easy step to get both CP/M and and Northstar DOS up and running too:

http://www.neoncluster.com/projects-sol20/sol20-NSDOS.html

http://www.neoncluster.com/projects-sol20/sol20-CPM.html

In fact there is an even simpler way to do it now. I need to find time to document this also.

Phil
 
I'm still waiting on the Sun Type 4 keyboard in hopes its pads are still good, but in the meantime, I found that I've got problems with the UART used in the cassette interface. SAVE operations to cassette from SOLOS finish in 4-5 seconds now matter how much data I'm saving. No error is displayed.

I wrote a test program and determined the transmit buffer empty signal (TBRE) from the UART is permanently asserted, so all the data transmits in less than 100ms. Using a meter, I didn't find continuity from the TDRE pin on the UART to the line it connects to on the input bus. This explains why the signal always appears asserted. I pulled the UART out of its socket, cleaned up numerous ugly pins, reinserted it, and then the TDRE pin showed continuity to its input bus line. Thought I had an easy fix... but now when I run a tape save or my test routine, it hangs forever! I'll have to drag out the scope and dig into that tomorrow night.

Mike
 
You must log in or register to reply here.
Back
Top