SEB-3

[billdrom]

The SEB-3 is basically working. I kept trying to compare the video from the 540B card with the SEB-3 video and couldn't figure out why it was so different looking. It turns out that the SEB-3 video is dark on a white background. Also found that the transistor was installed backward. I was very careful to put it back the same way it was when I replaced it. After playing with resistor values trying to get a display and more closely checking the partial schematic I corrected the transistor orientation. Still could not get a display but did get an occasional flash of video. After switching to an old crt type monitor I was able to get it to synchronize. I guess the flat panel display wasn't good enough.
It still has all the jumpers installed. Maybe I should try putting it back to original anyway.
With the current boot disk it displays 71 characters and the cursor.

SEB-3 Display on monitor.jpg (366.4 KiB) Viewed 5468 times

[bxdanny]

Interesting.

Yes, the board produces "reverse video". I was going to mention that, but I figured it was clear from the image in the ad that my first message in this thread linked to.

So now i have seen it running on a green screen (the one I used with the board I was given) and an amber screen (yours). Not on a white one. (The converted TV that I used with my 540 board did not have a high enough resolution to display the SEB-3's output clearly.) Does your monitor have a control for horizontal width that would let you set wider "guardbands"? While 71 characters happens to be the size of OSI BASIC's input buffer, obviously the monitor is also limiting the number of characters that can be displayed.

Anyway, I was looking at the 6845 data sheet, and I realized that I was confused about what configuring it for 540 emulation would look like. I had thought that the characters would be wider in 64-character mode, filling the same screen width as the 80 in 80-character mode, but that is just wrong. The dot clock is fixed, so I guess the character size also would be, and 64 characters will simply display with wider guardbands. I also realized that two of the values I was writing to the 6845 (those for registers R3 and R8) had extra bits that did nothing. I think I put them in just to verify that they had no effect, and then didn't bother to take them out. Since they don't hurt anything, I think I'll still leave them.

Anyway, based on my current understanding of the 6845, I created a second set of parameters to put the board into 64-character (540 emulation) mode. I'm attaching a revised disk image that has programs to set both 80-character and 64-character mode. With it, you should be able to switch back and forth by running either BEXEC* for 80 characters or SEB364 for 64 characters. Let me know if that works.

P.S. The BEXEC* erases itself when done in order to get rid of the reserved space in front of the BASIC workspace. You can prevent that when booting the revised disk by holding down the left shift key during boot. In 64-character mode, the first two rows (128 bytes) will not be displayed (much like on the 540) and the first displayed character will be the one at $D080. In 80-column mode, displayed characters start at $D000.

[billdrom]

My friend responded to my email. He said he has two copies of the manual from different sources with different hand written notes. He scanned one of them and said he will scan the other this weekend. The one he sent is missing the page you sent me with the board layout.
The other thing he sent were some notes on the disk driver floppy disk which he is also sending this weekend.
He thought the notes were sent to him from Earl Morris and that may be true, but I think they may be your original notes Dan in your own handwriting because it is signed Dan S.

ohio scientific seb3 orig.pdf

Orion SEB-3 Manual

(1.87 MiB) Downloaded 429 times

ohio scientific floppy tech notes.pdf

(363.79 KiB) Downloaded 419 times

I was able to boot your new SEB-8v2 disk and test 80 column and 64 column modes. They both work OK. I tried different monitors which seemed to work but after I touched the horizontal controls the both lost sync and I could not get it to re-sync.
I'll show a few pictures of when they were working.

SEB-3 80 column on color monitor.jpg (653.72 KiB) Viewed 5447 times

SEB-3 in 80 column mode.jpg (670.58 KiB) Viewed 5447 times

[bxdanny]

WOW! Yes, those are my handwritten notes. I didn't remember writing that up or giving the board to an "Allen". I thought the guy from Grafix who gave me the board took it back. Could Allen's last name be Lipper? Allen Lipper was a member of the OSUNY users group (I think his first name was spelled that way). So the board you have may be the very same one that I once had.

Those last two pictures make me feel good, especially the next-to-last one. You did get that amber monitor to display the full output of the board, horizontally and vertically, whereas it was cutting off parts in both directions before. So congratulations. Too bad you didn't leave it be at that point. I'm still wondering why the LCD monitor can't display it. I have heard that they can be quite picky about what signals they will accept, but if it accepts the 540, I don't know why it doesn't accept the SEB-3. Maybe the reverse video? Anyway, is the crystal on your board 12.68 MHz, like the manual says? I was expecting it to be 800 times the horizontal frequency, which you say is 15,809 Hz, which would be 12.6472 MHz. And the vertical frequency should be 1/263 of the horizontal, which would be 60.11 Hz. Is that what you see (if you are able to make such a measurement)? What are the corresponding frequencies (crystal, horizontal, and vertical) on the 540?

The BASIC and OS on the disk images I posted here are the stock standard 3.2 versions, but I have posted my Enhanced Basic elsewhere in the Files section. BASIC occupies tracks 2, 3, and 4 on an 8-inch diskette, PLUS track 8, sector 3 was also modified, and needs to be copied as part of copying "Enhanced Basic". Oh, and the "cursor up" code (CHR$(11)) is not implemented on the SEB-3 driver I posted. All the other control codes and escape sequences listed on that handwritten note to "Allen" are.

[bxdanny]

I don't know why i said the cursor up code wasn't implemented. I just double checked, and indeed it is.

[billdrom]

Its a small world. I am glad you got at least some of your notes back.
He said the handwritten notes were forwarded to him by someone else. I'll ask him if he remembers where he got the board.
His name is Allen Cohen.
He said he asked for help in a letter to the editor of Peek65 and a reader sent him a copy of the manual.
I don't know who made the changes on the back of the card.

The bottom two pictures were made using a different amber monitor. Its a Magnavox "Computer Monitor 80" .

SEB-3 Bexec.jpg (613.52 KiB) Viewed 5428 times

[billdrom]

To answer your question, the crystal on the SEB-3 card is marked 12.658, with my scope it measures as 12.6461 and with a cheap counter its 12.6508 .
The OSI 540b card crystal is marked 12.083915, with my scope it reads 12.0757 and with the counter it reads 12.0786

I both cards again with the Magnavox monitor and they both work with out adjusting the monitor.
I also tried the flat panel which is a Curtis 12 inch TV set a lot of input types.
It worked when first hooked up but then slowly fades out the video brightness.

Bill

[bxdanny]

Well, it's certainly a small OSI-related world. But let me get this straight: The notes I wrote were addressed to someone named Allen. You got them from someone named Allen. But he says he is not the same Allen that they were addressed to? It sounds rather unlikely, but I guess stranger things have happened.

Is there any chance that Allen might actually have the diskette that those notes were meant to accompany? I suppose it's quite unlikely that the disk would have survived all these years, even if he ever had it, and he may not have ever even had it. But in the unlikely event that he DOES have it, I would certainly LOVE to get a copy (as an image file, of course).

[bxdanny]

Bill,

I was looking more closely at your photo of the back of the board, to see if i could figure out what those jumpers actually did. And at some point - when I saw how pins 5, 6, and 7 of that added chip were tied together, and HAD been tied to pin 8 as well, but then carefully cut away from it - I remembered. I'M THE ONE WHO ADDED THOSE BLUE JUMPERS AND THAT EXTRA CHIP. Two chips, actually. Yes, that is definitely the very same board I had had. I think the circuit has to do with "transparent access" to the video RAM. I was NOT using the SEB-3 to generate the CPU clock, but I think I found some way to make the access transparent after all. Is that possible? The DE (Display Enable, pin 18) signal from the CRTC chip is involved. Possibly one of the other gates on the added chip (74LS04) was used for something involving reading the keyboard (KBD signal).

You can see if video memory access is "transparent" by running a tight loop like

10 I=PEEK(53248):GOTO 10

and seeing if it causes any black "hash" on the screen as the CPU accesses video RAM. I don't think it will. (Or would it be white/amber hash, only visible when the screen was full of text? I still don't think it will.)

[dave]

This has been a very interesting thread to follow. I had seen ads for those boards, but had never seen one in the wild. I suspect if OSI had thought of introducing a SB2/C1P with a 6847 chip, dynamic RAM, and used the saved space to add an on-board disk interface, they might have had a decent competitor to the tandy CoCo.

Does anyone have a copy of the full schematic of the SEB-3?