Restoring a NorthStar Horizon, problems with SRAM board

Hi,

This is a really long post and I am sorry about it but this is a long
story so if any one can help, I am providing as many relevant details
as I can think of.

I recently purchased a NorthStar Horizon with the intent of restoring
this old machine. After it arrived, I inspected it for damage and
followed the procedures for restarting an old computer.

http://s100-manuals.com/Repairs.htm

When I bought the machine, I knew it had been unused for many years it
and was going to be a challenge to get running, so this is no surprise
but still I could use a hand in figuring out why it won't boot.

I stripped out the cards, labled, photographed, and removed everything
I could from the power supply and reformed it over a day with the
variac. The power supply seems to be in good working order, as does
the fan and power indicator LED.

One by one, I added the cards and disk drives by themselves to see if
any rectifiers or tantalum capacitors burned up and none did. I
reconnected all the pieces and tried to boot but so far, no luck. I
have a whole collection of 5.25" disks including at least one boot
disk. Also, I get absolutely no drive activity on start up which is
rather suspicious.

The Horizon contains a NorthStar ZPB-A3 CPU board, a Tanner (DRC) 64K
Static RAM board, a NorthStar MDS-AD3 floppy controller, a Hayes
MicroModem 100, and a Vector Graphics Flashwriter (clearly not a II so
its probably a I) video board.

The motherboard is wired to support a serial printer on the right
serial port, the left serial port is disabled, the parallel output
port does not appear to be configured, the parallel input port is
attached to a custom keyboard.

The ZPB-A3 CPU board contains the 2708 PROM boot option which makes
sense since this computer has been highly modified and does not rely
on a serial terminal. Rather it relies on the parallel keyboard and
the video board for user IO. Unfortunately, the 2708 PROM is very old
and I do not have the tools to read it since it requires 3 different
voltages. My EPROM programmer can read and program 2716 but they are
too different from the 2708 to be useful.

On cold boot, the video board does display a screen full of static
garbage characters which the same pattern with each cold boot. As far
as I can see the video has never changed.

Based on some observation, I believe the problem lies in the SRAM
board:

http://www.digibarn.com/collections/parts/s-100-boards/Image18.jpg

http://www.smallscalerailway.com/images/56KRAM.jpg

http://www.hartetechnologies.com/manuals/Digital%20Research/Digital%20Research%2064K%20SRAM%20Manual.pdf

The SRAM board has a single LED in the lower right hand corner (see
the pictures at the links) which I believe is an "active" indicator.
When I place it in the Horizon with any or all of the four other S-100
boards, it refuses to turn on, ever, regardless what I do or how long
I wait.

My first question: can someone confirm the single SRAM board LED is a
"active" indicator? In other words, the LED turns on when the board is
ready to serve SRAM to memory request on the S-100 bus? If so, it
should be on almost all the time, right? There is a schematic in the
manual but the documentation does not say what its purpose is.

When I tested the SRAM board separately on my other S-100 chassis
(looking for the burning and exploding tantalum capacitors) I could
see the LED flicker at start up and shut down. The test S-100 chassis
has no peripherals or cards in it at all. Also if I left the SRAM
board for a while and pressed the RESET button the LED would light up
brightly like everything was working fine. However, as soon as I put
in another S-100 card like the CPU board or modem board the LED
refuses to light no matter what.

I thought maybe it is a voltage regulation issue since when a board is
alone in an S-100 chassis with unregulated +8v rails, the voltage
tends to be higher than when the power supply is loaded with several
other cards drawing current. So I take my VOM to measure the +8v rail
and if I so much as touch the PS +8v rail with a VOM lead, ground or
hot, the LED blinks out and stays out until I RESET the system and
remove the lead. Now that is strange behaviour. Touching a power
supply rail with a voltage meter probe should not affect whether the
card works or not!

Something just isn't right with the SRAM board. Also, it has two 7805
voltage regulators with heatsinks. Neither of them even gets warm to
the touch which is highly suspicious considering this board contains
28 2k x 8 SRAM chips plus buffers and miscellaneous glue logic. I
think those heatsinks should be at least hot to the touch with all the
current the SCRs have to supply. What is even more suspicious is that
the 7805's have two tantalum capacitors on each of them.

However, when I check the pin 3 of both of the 7805's, I see it they
are both producing normal +5V. I also checked the tantalum capacitors
(in circuit) for low impedance or shorts but none of those were
obviously bad or getting hot on the test S-100 chassis.

I pulled and reseated all the logic chips on the SRAM board. I
cleaned off some suspicious crusty stuff off the back of the board and
pressed in all the SRAM chips. I reviewed the documentation and all
the S1 switches (1-8 OFF for no extended addressing) are off per
manual. Also, the upper (E000-FFFF) 8K of SRAM are disabled on S2,
with extended addressing, full and half phantom lines turned off (S2
1-4 ON, 5-8 OFF).

Since the Horizon will not boot at all, the power supply seems to
check out almost certainly the problem is either the motherboard
itself, the CPU, SRAM, or disk controller. However, the only board
exhibiting clearly strange behaviour is the SRAM board. The rest, I
do not know.

If anyone is willing to help out, I certainly would appreciate it.
Thanks in advance!

Andrew Lynch

.