5t4nd1ng @ th3 3dg3 0f 5p3ct4t10n!

Welcome to Walt Perko's Brainless.org-anizm


My Vintage Altair 8800 Computer SN:222820K
First, you can learn more about MITS and the Altair 8800 with simple searches online.


After exploring inside of my new computer and imaging the S-100 circuit boards and other guts, I found this computer once belonged to Fremont School. Now I'm working on tracking it's original provenance.

I found the computer on CraigsList.org and bought the computer from a guy in Mill Valley, CA. that got the computer from a neighbor that died.


.
Front Panel is well worn!

.

Looking inside


.

Not an original MITS Power Supply

.


Maybe in the future I'll pull this power supply and reaplce it with Mean-Wells

In the meantime, after testing the power supply to be usable, I was able to add a longer bolt to allow DB25 connectors in the back panel.

It's nice to have the slots further apart so one board doesn't heat the next board.
One more 88-2SIOJP behind the current board and maybe the last light green slot can be a parallel board.
.

A closer look at the boards
.

Nice Looking MITS 8080 CPU Board


.

The Console 88-2SIO board addresses 10h/11h and 12h/13h and setup for TTY rather than RS-232

.

The 2nd 88-2SIO board addresses 00h/01h and 02h/03h and setup for TTY rather than RS-232


.

Came with the Solid State Music 4K Static RAM board

This board is For Sale or Trade

.

Also came with a Tarbell Cassette I/O board

This board is For Sale or Trade
.


.


Replacements & Improvements
.


Replacing the SSM 4K RAM with a FDC+ board from FarmTek

The best solution for 64K RAM and PC Serial Drive Server serving four Floppy Drives to the computer!

.


Replacing the two 88-2SIO boards with one Martin Eberhard 88-2SIOJP board

Smarter and easier to use/configure 88-2SIO design
.


Front Panel Cable B4 Fixed

Wires soldered to the back of the board

After Front Panel Cable Repairs

.


The MITS 8080 CPU board

LM320K-5.0 Voltage Regulator in lower left corner
No clock fix on the board
.


Initial Tests after Assembly

Per the manual checking power supply voltages on the buss without any boards in any slots
PIN 1 = +8V - PIN 2 = +20V - PIN 51 = +8V - PIN 52 = -20V - PIN 53 = GND

Testing with an oscilloscope the ripple or 60Hz hum on the d.c. voltages I cannot see any ripple, this is a sign of good capacitors

More Board INFO

Next test was a continuity test between the solder points on the front of the front panel to the solder points on the back of the CPU board to insure the connections are all good.

After turn the computer ON, I STOP → RESET → set all the address switches DOWN → DEPOSIT → DEPOSIT NEXT
several times to about address F → then RESET → EXAMINE → EXAMINE NEXT →

Today it's D8 that 's ON most of the time → and D0 so far is ON all the time.

When I EXAMINE up to address 8 the D8 LED turns OFF until address 11 when it turned ON again.

I'm hoping those symptoms might help to narrow down the fault.

Then I swapped the original MITS 8080 CPU board out for the Re-Production board in my Altair 8800c computer ...
Now the Altair 8800c computer has the faults and it looks like the vintage Altair 8800 still has some faults too.

Swapping out the NEC D8080A-2 (White/Gold) chip with a NEC 8080AFC plastic chip, the CPU seems to work a little better, but still has some faults.

.


First fix was replacing the front panel connectors
Second fix was swapping out the CPU chip with a new CPU chip
Third fix was to install almost all the CPU board ICs into sockets

as of Earth Date: 20230704
.


A few questions about the CPU board

Why is there a second 5V regulator kluged on the board? Are both in parallel or is the one on the top disconnected?

Oh wait, the second is a negative 5V, 1.5A regulator. It is a 320 not a 340. What uses so much current at -5V?

Just looked at the schematic and the -5V is provided by a small zener. The board should not draw over 1ma worst case, 0.01ma typical draw on the 8080. Is the -5V going somewhere else off board?

Walt showed a few close-up photos earlier of this dodgy "upgrade" to the -5V regulator performed by the previous owner. The -5V is not going anywhere else other than the CPU board. The original resistor and Zener that provided -5V regulation was removed when the LM320-5 was installed. Yes, using a regulator in a TO3 package is overkill in this application - my guess is that the previous owner may have used it just because he happened to have one available. As to "why" he did this - my guess is because of concerns that the home-made power supply he built to replace the original Altair power supply puts out -20V which is a bit more than the original.

I know what the part is - I was just curious as to why such a high current regulator was kludged onto the board. It looks like a recipe for sparks. If I had that card, I'd remove it and replace it with the correct zener diode. Perhaps the former owner didn't have any 7905s lying around...

If the voltage was higher, all that is needed is a larger value resistor before the zener. No big deal, but I am just curious why people do things like this.


This makes me wonder if I should pop off that kludge and install the proper zener diode which would also give me the space to install IC sockets for the last three chips.

I would be worried about shorts as well, it has a large case and the case attaches directly to the -16V input power from the backplane. BTW, a 7905, being a half to one amp regulator would be smaller but just about as over kill as the LM320. The schematic values are correct at 220 ohms but the resistor could be changed to 390 to allow for up to 25V in.

Now I'm having 2nd thoughts about using the CPU board in my computer. I cannot even see where the zener diode goes on the board.

I would put in the correct zener and make sure the 220 ohm (or 270 if the 220 is not handy) is in place. The 8080 voltage should be withing 5% of 5V so pick a 1% or 2% zener such as the 1N5231CTR.

This sounds like a better fix:
I would consider a 79L05 in a TO92 package as an improvement over the Zener diode.
The pins are not EBC, but rather "IN", "OUT" and "GND" in some order (you'll have to look it up).
The "IN" pin should be connected to the incoming power wire. You should be able to pick that up at one of the pads for the resistor.
The "GND" pin should be connected to ground, which should be at one of the pads for the Zener.
The "OUT" pin should connect to the 8080, which will be at the other pad for the Zener.

But restoring the MITS 8800 CPU board to the original design sounds best!
.


For Testing Purposes I will start with a Reproduction MITS 8800 CPU board

It takes about $100 and a couple of hours to assemble

Next step, test the new CPU board in my Altair 8800c computer

.


?????????????

??????????????

.


???????????

??????????????
.


?????????????

??????????????

.


???????????

??????????????
.


Little Programs for Testing 8080 CPU Computers
.


This program sends "A" to "Z" to the serial port

; TEST PROGRAM TO CONTINUALLY SEND A TO Z CHARACTERS TO THE SERIAL PORT AT 10H,11H
; THIS PROGRAM ASSUMES THAT THE SERIAL PORT IS CONFIGURED AND THE BAUDRATE IS SET
; BY HARDWARE JUMPERS ON THE SERIAL CARD.
; BUILD USING ASM.COM WITH CPM COMMAND: ASM SERTST

; DEFINE PORT CONFIGURATION FOR SERIAL PORT 0011 = CONOUT EQU 11H ; SERIAL PORT DATA OUTPUT PORT
0010 = CONSTAT EQU 10H ; SERIAL PORT STATUS INPUT PORT
0002 = CONTBR EQU 2H ; SERIAL PORT TRANSMIT BUFFER READY STATUS BIT
. . . . . . . . . . . . . . . . . . . . . . ; ASSUMES BIT IS HIGH WHEN THE UART IS READY TO SEND A CHAR

0000 ORG 0

START:
0000 0E41 MVI C,'A' ; LOAD REG C WITH ASCII A ( 016 101 )
LOOP:
0002 DB10 IN CONSTAT ; READ CONSOLE STATUS PORT ( 333 020 )
0004 E602 ANI CONTBR ; CHECK TO SEE IF UART IS READY TO SEND A CHARACTER ( 346 002 )
0006 CA0200 JZ LOOP ; LOOP UNTIL UART IS READY TO SEND NEXT CHAR ( 312 002 000 )
. . . . . . . . . . . . . . . . . . . . . . ; ASSUMES BIT IS HIGH WHEN THE UART IS READY TO SEND A CHAR

0009 79 MOV A,C ; ACC = TEST CHARACTER ( 171 )
000A D311 OUT CONOUT ; OUTPUT CHARACTER IN ACC TO CONSOLE ( 323 021 )
000C FE5A CPI 'Z' ; IS LAST CHARACTER A Z ( 376 132 )
000E CA0000 JZ START ; YES, RESTART SEQUENCE AT A ( 312 000 000 )
0011 0C INR C ; NO, NEXT TEST CHARACTER ( 014 )
0012 C30200 JMP LOOP ; REPEAT WITH NEW CHARACTER ( 303 002 000 )



; OCTAL CONVERSION SUMMARY AS SHOWN ABOVE IN COMMENTS
; THERE DOESNT APPEAR TO BE A WAY TO GET ASM TO OUTPUT OCTAL LISTINGS
; SO THE CONVERSION WAS DONE A BYTE AT A TIME FROM THE HEX IN THE LISTING FILE.
; LOAD THE FOLLOWING BYTES STARTING AT ADDRESS 0, THEN START EXECUTION AT 0
; OR BUY AN IMSAI SO YOU CAN USE HEX DIRECTLY FROM THE LISTINGS!
;( 016 101 )
;( 333 020 )
;( 346 002 )
;( 312 002 000 )
;( 171 )
;( 323 021 )

;( 376 132 )
;( 312 000 000 )
;( 014 )
;( 303 002 000 )

0015 END START ; INCLUDE START ADDRESS IN HEX FILE


016
101
333
020
346
002
312
002
000
171
323
021
376
132
312
000
000
014
303
002
000