[N8VEM: 18048] XI debugging

[Jim Strickland]

I finally got my hands on the last couple capacitors to finish my Xi board
and backplane. (I don't have the 7 segment displays yet, but their drivers
are in) and I'm starting the process of debugging the xi. Right now, when
the power supply is turned on, the backplane shows standby power. When the
power button on the backplane is hit, the power LEDs on the backplane come
up, and the power LED on the xi board comes up. Unfortunately, that's where
the progress stops. The turbo light comes on at powerup, despite not being
jumpered to come on. Nothing else happens. No beeps, no lights on the
keyboard, no nothing. I get the same behavior minus the turbo light when j2
is set to 2,3 - software controlled. (the turbo light I am referring to is
the one on the xi card.)

The only differences that I'm aware of between my Xi and Sergey's in the
picture - I have a 29f002 (256kB) flash instead of the 29f010 (128kB)
called for (the empty space is filled with FF), my turbo oscillator is a
14.31818, which should mean I get 4.77mhz whether turbo is on or off. (my
8087 is only rated at 5mhz). I am using the DS12997+ clock chip, which
should mean I don't need x2 or the battery.

Jumpers: j1 (none set) j2 (none set) p3 (1,2) j3 (1,2) j4 (none set) j5
(not set)
Switches: sw2 (2,3,4 are on) sw3 (1 is on)
Cards: xi, known good VGA
Keyboard and monitor are connected.


Any suggestions before I go after this thing with the vom/scope/logic
analyzer? Is my flash setup viable?

-JRS

[Jim Strickland]

Followup on debugging:
Still no joy. The capacitor in the speaker circuit was wrong, and I had a
74LS02 in u30 where there was supposed to be a 74LS92. Replacing this has
not fixed my turbo-is-always-on unless jumpered to software control problem.

I also replaced the flash IC with the one specified (which I stole from my
cf card and reprogrammed.) No change.

I had it making static noises and what might have been a strangled beep for
a while, but after screwing with the trim capacitor in the clock crystal
circuit, it no longer does even that.

Questions:
1. Does anyone know what the variable capacitor in clock crystal circuit is
for and how to set it? I'm assuming it's to set the system clock to exactly
4.77mhz and if it's off slightly it won't stop the thing from booting.
2. When I ordered a 74ls688 for the decoding logic on the backplane,
Unicorn sent me a 25ls2521 and assured me it was compatible. Is it for this
application? The pinout looks reasonable. It seems unlikely this would
cause failure to boot, but it is listening on the address lines. of the bus.

-JRS

Post by Jim Strickland
I finally got my hands on the last couple capacitors to finish my Xi board
and backplane. (I don't have the 7 segment displays yet, but their drivers
are in) and I'm starting the process of debugging the xi. Right now, when
the power supply is turned on, the backplane shows standby power. When the
power button on the backplane is hit, the power LEDs on the backplane come
up, and the power LED on the xi board comes up. Unfortunately, that's where
the progress stops. The turbo light comes on at powerup, despite not being
jumpered to come on. Nothing else happens. No beeps, no lights on the
keyboard, no nothing. I get the same behavior minus the turbo light when j2
is set to 2,3 - software controlled. (the turbo light I am referring to is
the one on the xi card.)
The only differences that I'm aware of between my Xi and Sergey's in the
picture - I have a 29f002 (256kB) flash instead of the 29f010 (128kB)
called for (the empty space is filled with FF), my turbo oscillator is a
14.31818, which should mean I get 4.77mhz whether turbo is on or off. (my
8087 is only rated at 5mhz). I am using the DS12997+ clock chip, which
should mean I don't need x2 or the battery.
Jumpers: j1 (none set) j2 (none set) p3 (1,2) j3 (1,2) j4 (none set) j5
(not set)
Switches: sw2 (2,3,4 are on) sw3 (1 is on)
Cards: xi, known good VGA
Keyboard and monitor are connected.
Any suggestions before I go after this thing with the vom/scope/logic
analyzer? Is my flash setup viable?
-JRS

[Jim Strickland]

Fixed two unsoldered pins, now getting the 5 beep boot pattern. Probably means it's not seeing the video board. Progress!

[Jim Strickland]

Okay. I'm getting the 5 tone beep reliably. No video. I still don't have my
7 segment LED display for the backplane, so I'm flying blind.
My video card is a generic wd90c00 - probably a WA-V77, but it's got no
markings to say conclusively. They appear to be common as dirt, and
everything with this chipset uses the same DIP settings, but the jumpers
are still something of a mystery. It does work. I put the card in my
celeron 500 machine with isa slots as the only video card and it performs
exactly as you'd expect.

Does anyone have a wd90c00 based card working with the Xi, or should I just
wait for the new run of Trident cards and build that? (I have one ordered.)

-JRS

[Jim Strickland]

Okay, something is desperately wrong with my turbo circuitry. With turbo
jumpered off (no jumper on the turbo pins) the turbo light is on, and the
system doesn't even try to boot, which is exactly the same behavior it
gives with turbo jumpered on. With turbo jumpered to software control, I
get the 5 note boot sound, which suggests the bios is loading. So two flaws
going on: 1. Turbo is on when it shouldn't be, and 2. The system doesn't
work in any fashion with Turbo enabled. I have a 14.31818mhz oscillator in
my turbo oscillator slot, so I shouldn't be overclocking anything.

Any suggestions?

-JRS

[Sergey]

Well, this is weird.
The turbo LED is controlled almost directly by the turbo jumper. Signal
goes go through a few gates: U43A, U40E, and U40C (on-board LED) or U40D
(turbo LED header).
If there is no turbo jumper installed, it will be LOW signal (provided by
R3 pull-down resistor) on pin 2 of U43A NAND gate, as the result U43A
output will be HIGH, U40E output will be LOW, and U40C/U40D output will be
HIGH, and the turbo LED will off (as it is connected between U40C/U40
output and VCC).
If the turbo jumper is installed, pin 2 of U43A is connected to VCC (logic
HIGH). But this alone doesn't automatically turns on turbo mode - it also
depends on input on pin 1 of U43A, which is connected to bus arbitration /
DMA logic (which turns off turbo during DMA operation).
Anyway there are few ways to debug this situation:
- Double check that components are installed correctly. Check resistor
values too :-)
- Try replacing related components.
- Get a logic probe, and try to check signals on pins described above. Look
at Xi 8088 schematics.
- Debug with an oscilloscope or a frequency meter (if you have access to
one)

BTW, the glitchless turbo switching circuit requires both clocks (turbo and
regular) to run in order to operate properly. Check your turbo oscillator,
and 8284 and 14.31818 MHz crystal. If you have a frequency meter that can
handle tens of MHz, you can try measuring frequency on pin 8 of oscillator,
pin 12 (OSC) of 8284, and also frequency on pin 14 (EFI) of 8284, and other
important clock frequencies. (If you have a frequency meter with lower
limit, say 4 MHz, you can try dividing frequency using a counter / divider
IC, e.g. 74*93)

Hope this helps,
Sergey

Okay, something is desperately wrong with my turbo circuitry. With turbo

jumpered off (no jumper on the turbo pins) the turbo light is on, and the
system doesn't even try to boot, which is exactly the same behavior it
gives with turbo jumpered on. With turbo jumpered to software control, I
get the 5 note boot sound, which suggests the bios is loading. So two flaws
going on: 1. Turbo is on when it shouldn't be, and 2. The system doesn't
work in any fashion with Turbo enabled. I have a 14.31818mhz oscillator in
my turbo oscillator slot, so I shouldn't be overclocking anything.

Any suggestions?

-JRS

[Jim Strickland]

On much deeper poking into the Xi, I can confirm the following things:

Powered down, pin 2 of j2 (the turbo jumper) has about 9.5k resistance to
ground. Powered up, the pin floats at about 1.25 volts above ground. When
grounded, the turbo light goes off and stays off, so it looks like my turbo
control problem is right in the front end of the turbo system.

The 74LS00 (u43, first in line in the turbo enable chain of logic) has a
minimum trigger voltage of 2 for a logic high. 1.25v shouldn't switch it.
I'm wondering if my 74LS00 is defective. I have another. I'll try that.

Also, having changed the turbo oscillator, I now get to the multi-tone
powerup sound (as listed in the bios comments) no matter what mode turbo is
in, so that's progress. Given your comments about requiring the turbo
oscillator even when not in turbo mode, I suspect my original turbo
oscillator was not making good connection or was just bad. Turbo osc is now
16mhz, overclocking my 8087 by 300khz and underclocking my 8mhz 8088.

Question: Will booting proceed if no valid video card is present? I know my
video card works - as a 16 bit card - and while it has the jumpers and
switches to enable an 8 bit bios, because it's a generic card, it's
possible there is no 8 bit bios in there.

I do have a frequency meter that goes into the hundreds of mhz, and I'll
have to use that, since my scope tops out at 10mhz. I also have a logic
shrimp (the 4 pin version of the logic sump).

Thanks for the suggestions. That's pretty much what I've been doing. :)

-JRS

Post by Sergey
Well, this is weird.
The turbo LED is controlled almost directly by the turbo jumper. Signal
goes go through a few gates: U43A, U40E, and U40C (on-board LED) or U40D
(turbo LED header).
If there is no turbo jumper installed, it will be LOW signal (provided by
R3 pull-down resistor) on pin 2 of U43A NAND gate, as the result U43A
output will be HIGH, U40E output will be LOW, and U40C/U40D output will be
HIGH, and the turbo LED will off (as it is connected between U40C/U40
output and VCC).
If the turbo jumper is installed, pin 2 of U43A is connected to VCC (logic
HIGH). But this alone doesn't automatically turns on turbo mode - it also
depends on input on pin 1 of U43A, which is connected to bus arbitration /
DMA logic (which turns off turbo during DMA operation).
- Double check that components are installed correctly. Check resistor
values too :-)
- Try replacing related components.
- Get a logic probe, and try to check signals on pins described above.
Look at Xi 8088 schematics.
- Debug with an oscilloscope or a frequency meter (if you have access to
one)
BTW, the glitchless turbo switching circuit requires both clocks (turbo
and regular) to run in order to operate properly. Check your turbo
oscillator, and 8284 and 14.31818 MHz crystal. If you have a frequency
meter that can handle tens of MHz, you can try measuring frequency on pin 8
of oscillator, pin 12 (OSC) of 8284, and also frequency on pin 14 (EFI) of
8284, and other important clock frequencies. (If you have a frequency meter
with lower limit, say 4 MHz, you can try dividing frequency using a counter
/ divider IC, e.g. 74*93)
Hope this helps,
Sergey

Okay, something is desperately wrong with my turbo circuitry. With turbo

jumpered off (no jumper on the turbo pins) the turbo light is on, and the
system doesn't even try to boot, which is exactly the same behavior it
gives with turbo jumpered on. With turbo jumpered to software control, I
get the 5 note boot sound, which suggests the bios is loading. So two flaws
going on: 1. Turbo is on when it shouldn't be, and 2. The system doesn't
work in any fashion with Turbo enabled. I have a 14.31818mhz oscillator in
my turbo oscillator slot, so I shouldn't be overclocking anything.

Any suggestions?

-JRS

[Sergey]

1.25V is way to high for pin 2 of turbo switch (given that it is not
jumpered anywhere). It should be less than 0.8V - valid LOW level for TTL
gates. Not sure why you get so high voltage. Perhaps your 74LS00 sinks too
much current? You can try using 4.7K R3 resistor instead of 10K (or connect
another 10K in parallel with existing one). Or replace 74LS00 with another
IC (maybe use 74ALS or 74HCT/AHCT instead, they have much lower input
current).

Yes, system should boot without video card connected, but it will be
difficult to get any indication that system is working or debug it if
something goes wrong :-)

Thanks,
Sergey

Post by Jim Strickland
Powered down, pin 2 of j2 (the turbo jumper) has about 9.5k resistance to
ground. Powered up, the pin floats at about 1.25 volts above ground. When
grounded, the turbo light goes off and stays off, so it looks like my turbo
control problem is right in the front end of the turbo system.
The 74LS00 (u43, first in line in the turbo enable chain of logic) has a
minimum trigger voltage of 2 for a logic high. 1.25v shouldn't switch it.
I'm wondering if my 74LS00 is defective. I have another. I'll try that.
Also, having changed the turbo oscillator, I now get to the multi-tone
powerup sound (as listed in the bios comments) no matter what mode turbo is
in, so that's progress. Given your comments about requiring the turbo
oscillator even when not in turbo mode, I suspect my original turbo
oscillator was not making good connection or was just bad. Turbo osc is now
16mhz, overclocking my 8087 by 300khz and underclocking my 8mhz 8088.
Question: Will booting proceed if no valid video card is present? I know
my video card works - as a 16 bit card - and while it has the jumpers and
switches to enable an 8 bit bios, because it's a generic card, it's
possible there is no 8 bit bios in there.
I do have a frequency meter that goes into the hundreds of mhz, and I'll
have to use that, since my scope tops out at 10mhz. I also have a logic
shrimp (the 4 pin version of the logic sump).
Thanks for the suggestions. That's pretty much what I've been doing. :)
-JRS

Post by Sergey
Well, this is weird.
The turbo LED is controlled almost directly by the turbo jumper. Signal
goes go through a few gates: U43A, U40E, and U40C (on-board LED) or U40D
(turbo LED header).
If there is no turbo jumper installed, it will be LOW signal (provided by
R3 pull-down resistor) on pin 2 of U43A NAND gate, as the result U43A
output will be HIGH, U40E output will be LOW, and U40C/U40D output will be
HIGH, and the turbo LED will off (as it is connected between U40C/U40
output and VCC).
If the turbo jumper is installed, pin 2 of U43A is connected to VCC
(logic HIGH). But this alone doesn't automatically turns on turbo mode - it
also depends on input on pin 1 of U43A, which is connected to bus
arbitration / DMA logic (which turns off turbo during DMA operation).
- Double check that components are installed correctly. Check resistor
values too :-)
- Try replacing related components.
- Get a logic probe, and try to check signals on pins described above.
Look at Xi 8088 schematics.
- Debug with an oscilloscope or a frequency meter (if you have access to
one)
BTW, the glitchless turbo switching circuit requires both clocks (turbo
and regular) to run in order to operate properly. Check your turbo
oscillator, and 8284 and 14.31818 MHz crystal. If you have a frequency
meter that can handle tens of MHz, you can try measuring frequency on pin 8
of oscillator, pin 12 (OSC) of 8284, and also frequency on pin 14 (EFI) of
8284, and other important clock frequencies. (If you have a frequency meter
with lower limit, say 4 MHz, you can try dividing frequency using a counter
/ divider IC, e.g. 74*93)
Hope this helps,
Sergey

Okay, something is desperately wrong with my turbo circuitry. With turbo

jumpered off (no jumper on the turbo pins) the turbo light is on, and the
system doesn't even try to boot, which is exactly the same behavior it
gives with turbo jumpered on. With turbo jumpered to software control, I
get the 5 note boot sound, which suggests the bios is loading. So two flaws
going on: 1. Turbo is on when it shouldn't be, and 2. The system doesn't
work in any fashion with Turbo enabled. I have a 14.31818mhz oscillator in
my turbo oscillator slot, so I shouldn't be overclocking anything.

Any suggestions?

-JRS

[Jim Strickland]

Okay, I'll try the lower resistance pulldown. Meantime I managed to short
out the ISA bus and fry one of the TTL chips in bus arbitration, so I'm out
of action for a while.Before I had my little meltdown, I did try measuring
the voltage on the turbo select pin with no ttl installed in u43, and got
no voltage, so I suspect you're spot on that the combination of 10k and
74LS00 aren't compatible. The voltage is coming from the IC itself, not a
short under the socket or something (which I thought was the case.) I'll
try changing the resistor once I have the board showing signs of life
again. Moral of the story: when you use steel wool to clean IC legs, don't
leave it lying on your workbench. :P

Thanks for the help, and I'll keep you posted
-JRS

Post by Sergey
1.25V is way to high for pin 2 of turbo switch (given that it is not
jumpered anywhere). It should be less than 0.8V - valid LOW level for TTL
gates. Not sure why you get so high voltage. Perhaps your 74LS00 sinks too
much current? You can try using 4.7K R3 resistor instead of 10K (or connect
another 10K in parallel with existing one). Or replace 74LS00 with another
IC (maybe use 74ALS or 74HCT/AHCT instead, they have much lower input
current).
Yes, system should boot without video card connected, but it will be
difficult to get any indication that system is working or debug it if
something goes wrong :-)
Thanks,
Sergey

Post by Jim Strickland
Powered down, pin 2 of j2 (the turbo jumper) has about 9.5k resistance to
ground. Powered up, the pin floats at about 1.25 volts above ground. When
grounded, the turbo light goes off and stays off, so it looks like my turbo
control problem is right in the front end of the turbo system.
The 74LS00 (u43, first in line in the turbo enable chain of logic) has a
minimum trigger voltage of 2 for a logic high. 1.25v shouldn't switch it.
I'm wondering if my 74LS00 is defective. I have another. I'll try that.
Also, having changed the turbo oscillator, I now get to the multi-tone
powerup sound (as listed in the bios comments) no matter what mode turbo is
in, so that's progress. Given your comments about requiring the turbo
oscillator even when not in turbo mode, I suspect my original turbo
oscillator was not making good connection or was just bad. Turbo osc is now
16mhz, overclocking my 8087 by 300khz and underclocking my 8mhz 8088.
Question: Will booting proceed if no valid video card is present? I know
my video card works - as a 16 bit card - and while it has the jumpers and
switches to enable an 8 bit bios, because it's a generic card, it's
possible there is no 8 bit bios in there.
I do have a frequency meter that goes into the hundreds of mhz, and I'll
have to use that, since my scope tops out at 10mhz. I also have a logic
shrimp (the 4 pin version of the logic sump).
Thanks for the suggestions. That's pretty much what I've been doing. :)
-JRS

Post by Sergey
Well, this is weird.
The turbo LED is controlled almost directly by the turbo jumper. Signal
goes go through a few gates: U43A, U40E, and U40C (on-board LED) or U40D
(turbo LED header).
If there is no turbo jumper installed, it will be LOW signal (provided
by R3 pull-down resistor) on pin 2 of U43A NAND gate, as the result U43A
output will be HIGH, U40E output will be LOW, and U40C/U40D output will be
HIGH, and the turbo LED will off (as it is connected between U40C/U40
output and VCC).
If the turbo jumper is installed, pin 2 of U43A is connected to VCC
(logic HIGH). But this alone doesn't automatically turns on turbo mode - it
also depends on input on pin 1 of U43A, which is connected to bus
arbitration / DMA logic (which turns off turbo during DMA operation).
- Double check that components are installed correctly. Check resistor
values too :-)
- Try replacing related components.
- Get a logic probe, and try to check signals on pins described above.
Look at Xi 8088 schematics.
- Debug with an oscilloscope or a frequency meter (if you have access to
one)
BTW, the glitchless turbo switching circuit requires both clocks (turbo
and regular) to run in order to operate properly. Check your turbo
oscillator, and 8284 and 14.31818 MHz crystal. If you have a frequency
meter that can handle tens of MHz, you can try measuring frequency on pin 8
of oscillator, pin 12 (OSC) of 8284, and also frequency on pin 14 (EFI) of
8284, and other important clock frequencies. (If you have a frequency meter
with lower limit, say 4 MHz, you can try dividing frequency using a counter
/ divider IC, e.g. 74*93)
Hope this helps,
Sergey

Okay, something is desperately wrong with my turbo circuitry. With

turbo jumpered off (no jumper on the turbo pins) the turbo light is on, and
the system doesn't even try to boot, which is exactly the same behavior it
gives with turbo jumpered on. With turbo jumpered to software control, I
get the 5 note boot sound, which suggests the bios is loading. So two flaws
going on: 1. Turbo is on when it shouldn't be, and 2. The system doesn't
work in any fashion with Turbo enabled. I have a 14.31818mhz oscillator in
my turbo oscillator slot, so I shouldn't be overclocking anything.

Any suggestions?

-JRS