5 Jul
2005
5 Jul
'05
7:08 p.m.
I finally allocated another day to spend on the 11/45. Unfortunately, I must
say I'm not sure if it's sick or not, hopefully someone can help.
First, none of the blinking lights programs I've seen blink the lights. I
can store and retrieve the program correctly, but when I put 1000 on the
switches and hit load address then run, it goes into run mode but no lights
blink (with any display switch setting). Halt makes the processor halt.
Here's what I did:
1000 005000 start: clr r0
1002 005200 inc r0
1004 006100 loop: rol r0
1006 000005 reset
1010 000775 br loop
On this program I put 1000 in the switches, hit load address, then run.
Address lights constantly show 1010, Data lights are blank, and the Run
light is on. Data Paths setting shows 0, bus register setting shows 5, uaddr
shows 177774.
Another test... if I store a 5007 (clr pc) in location 0 and start the
processor at location 0, the address lights show all 0's, the data lights
show 2, and the following lights are lit - pause, master, kernel. I also
tried this with location 0 set to 777 (br .) and get the exact same results.
More info... just after power up, the address lights show 20, the data
lights show 21043 (data paths). Master & Kernel are on. The uaddress setting
shows 177570.
So... I decided to go back to basics and check all the power points at the
pins on the backplane. Here's what I found:
A02A2 0
A06A2 4.931
A10A1 4.931
A26A2 4.931
A19A2 0
A16A2 0
A21A2 0
E02B2 -15
A17V2 0
A22V2 0
A17U2 0
A22U2 0
F17C1 0
E15A1 16
E01B1 9
Since my machine only has regulators at B, C, D, E, and F, I suspect the 0
volt readings above are ok. All the 5 volt test points have a very clear .2v
sawtooth pattern (which looks fairly smooth at 5v). Since one manual says
the ripple limit is .15, and another manual says the ripple limit is .20, I
suspect I'm ok there. The ripple on -15, 16, and 9 I can't see any
noticeable.
And by the way... the cpu cardset was borrowed from another listmember and
tested before I received it so I don't think it's something wrong with the
cpu cardset itself.
Any thoughts?
Jay West
5 Jul
5 Jul
8:27 p.m.
Are you sure that the lights aren't blinking too fast to see? Perhaps you
can hang a logic probe (with a pulse extender) on one of the lights to be
sure.
Joe
At 06:08 PM 7/5/05 -0500, you wrote:
I finally allocated another day to spend on the 11/45.
Unfortunately, I must
say I'm not sure if it's sick or not, hopefully someone can help.
First, none of the blinking lights programs I've seen blink the lights. I
can store and retrieve the program correctly, but when I put 1000 on the
switches and hit load address then run, it goes into run mode but no lights
blink (with any display switch setting). Halt makes the processor halt.
Here's what I did:
1000 005000 start: clr r0
1002 005200 inc r0
1004 006100 loop: rol r0
1006 000005 reset
1010 000775 br loop
On this program I put 1000 in the switches, hit load address, then run.
Address lights constantly show 1010, Data lights are blank, and the Run
light is on. Data Paths setting shows 0, bus register setting shows 5, uaddr
shows 177774.
Another test... if I store a 5007 (clr pc) in location 0 and start the
processor at location 0, the address lights show all 0's, the data lights
show 2, and the following lights are lit - pause, master, kernel. I also
tried this with location 0 set to 777 (br .) and get the exact same results.
More info... just after power up, the address lights show 20, the data
lights show 21043 (data paths). Master & Kernel are on. The uaddress setting
shows 177570.
So... I decided to go back to basics and check all the power points at the
pins on the backplane. Here's what I found:
A02A2 0
A06A2 4.931
A10A1 4.931
A26A2 4.931
A19A2 0
A16A2 0
A21A2 0
E02B2 -15
A17V2 0
A22V2 0
A17U2 0
A22U2 0
F17C1 0
E15A1 16
E01B1 9
Since my machine only has regulators at B, C, D, E, and F, I suspect the 0
volt readings above are ok. All the 5 volt test points have a very clear .2v
sawtooth pattern (which looks fairly smooth at 5v). Since one manual says
the ripple limit is .15, and another manual says the ripple limit is .20, I
suspect I'm ok there. The ripple on -15, 16, and 9 I can't see any
noticeable.
And by the way... the cpu cardset was borrowed from another listmember and
tested before I received it so I don't think it's something wrong with the
cpu cardset itself.
Any thoughts?
Jay West
8:45 p.m.
Joe wrote...
Are you sure that the lights aren't blinking too
fast to see? Perhaps you
can hang a logic probe (with a pulse extender) on one of the lights to be
sure.
I'll try that tomorrow... but I wouldn't think that is what is happening.
Also, pardon my lack of knowledge, but... shouldn't I be seeing the .2v
ripple as an AC component (ie. a sine wave)? It's a very clear perfect
sawtooth. And it's exactly the same on all the 5V regulators, which strikes
me as odd.
Jay
11:31 p.m.
On Jul 5, 2005, at 6:45 PM, Jay West wrote:
Joe wrote...
The sawtooth is correct for h744's. 0.2v is a bit much, but shouldn't
be a show-stopper. The 11/70 maint. manual has a
decent description of the h744 on pages 4-59 thru 4-62 (EK-11070-
MM-002_May79.pdf on bitsavers). It says the voltage should
range from 4.95v - 5.05v. (The description in the 11/45 manual isn't
as good.)
Also check for high frequency spikes at the valleys of the sawtooth,
for example:
5v_ /\ /\ ... If these are more than 2 volts (go below
3v), you will have "issues".
/ \/ \/ Welcome to the wonderful world of aged DEC
switching supplies =:O
| |
3v_ | |
Ascii art, feh. Put up some pics of the `scope display. :-)
Mike
Are you sure that the lights aren't blinking
too fast to see?
Perhaps you
can hang a logic probe (with a pulse extender) on one of the
lights to be
sure.
I'll try that tomorrow... but I wouldn't think that is what is
happening. Also, pardon my lack of knowledge, but... shouldn't I be
seeing the .2v ripple as an AC component (ie. a sine wave)? It's a
very clear perfect sawtooth. And it's exactly the same on all the
5V regulators, which strikes me as odd.
Jay
6 Jul
6 Jul
6:55 p.m.
Joe wrote...
Have you tried single-stepping the program from the frontpanel?
Be warned, though, that I had a particularly nasty fault on my 11/45. It
would mess up branch instructions (and some others IIRC) but when I
single-stepped it at the microcode level (I have a KM11 card), it all
seemed fine. It turned out to be a chip that was 'going slow'....
Are you sure that the lights aren't blinking
too fast to see? Perhaps you
can hang a logic probe (with a pulse extender) on one of the lights to be
sure.
I'll try that tomorrow... but I wouldn't think that is what is happening.
Also, pardon my lack of knowledge, but...
shouldn't I be seeing the .2v
ripple as an AC component (ie. a sine wave)? It's a very clear perfect
sawtooth. And it's exactly the same on all the 5V regulators, which strikes
me as odd.
No, that sounds right. Even on a linear supply, the ripple is not
sinusoidal -- it's an almost-linear ramp down as the smoothing capacitor
discharges follwed by a little bit of a sinusoid as the capacitor is
recharged. On a switching regulator (which this is), the ripple is likely
to look even more like a sawtooth.
You might like to replace the output capacitors in your regulator bricks,
but I really don't think this is the problem.
-tony
9:08 a.m.
>>>> "Joe" == Joe R <rigdonj
at cfl.rr.com> writes:
Joe> Are you sure that the lights aren't blinking too fast to see?
Joe> Perhaps you can hang a logic probe (with a pulse extender) on
Joe> one of the lights to be sure.
An easy way to check that out: replace the "rol r0" by an "inc r0".
That way r0 will cycle through all 65536 values, and you should easily
see that in the lights.
If that doesn't help, try replacing the reset with a "mov r0,
@#177570" (010037 177570, and of course the branch becomes 000774).
Then the incrementing pattern should show when the Display register is
selected.
paul
10:19 a.m.
Jay,
Have you single stepped through the program to see if it is doing what
you expect it to?
--tom
At 06:08 PM 7/5/2005 -0500, you wrote:
I finally allocated another day to spend on the 11/45.
Unfortunately, I
must say I'm not sure if it's sick or not, hopefully someone can help.
First, none of the blinking lights programs I've seen blink the lights. I
can store and retrieve the program correctly, but when I put 1000 on the
switches and hit load address then run, it goes into run mode but no
lights blink (with any display switch setting). Halt makes the processor
halt. Here's what I did:
1000 005000 start: clr r0
1002 005200 inc r0
1004 006100 loop: rol r0
1006 000005 reset
1010 000775 br loop
On this program I put 1000 in the switches, hit load address, then run.
Address lights constantly show 1010, Data lights are blank, and the Run
light is on. Data Paths setting shows 0, bus register setting shows 5,
uaddr shows 177774.
Another test... if I store a 5007 (clr pc) in location 0 and start the
processor at location 0, the address lights show all 0's, the data lights
show 2, and the following lights are lit - pause, master, kernel. I also
tried this with location 0 set to 777 (br .) and get the exact same results.
More info... just after power up, the address lights show 20, the data
lights show 21043 (data paths). Master & Kernel are on. The uaddress
setting shows 177570.
So... I decided to go back to basics and check all the power points at the
pins on the backplane. Here's what I found:
A02A2 0
A06A2 4.931
A10A1 4.931
A26A2 4.931
A19A2 0
A16A2 0
A21A2 0
E02B2 -15
A17V2 0
A22V2 0
A17U2 0
A22U2 0
F17C1 0
E15A1 16
E01B1 9
Since my machine only has regulators at B, C, D, E, and F, I suspect the 0
volt readings above are ok. All the 5 volt test points have a very clear
.2v sawtooth pattern (which looks fairly smooth at 5v). Since one manual
says the ripple limit is .15, and another manual says the ripple limit is
.20, I suspect I'm ok there. The ripple on -15, 16, and 9 I can't see any
noticeable.
And by the way... the cpu cardset was borrowed from another listmember and
tested before I received it so I don't think it's something wrong with the
cpu cardset itself.
Any thoughts?
Jay West
25 Jul
25 Jul
6:18 p.m.
Are you sure that the lights aren't blinking too fast to see? Perhaps you
can hang a logic probe (with a pulse extender) on one of the lights to be
sure.
Joe
At 06:08 PM 7/5/05 -0500, you wrote:
I finally allocated another day to spend on the 11/45.
Unfortunately, I must
say I'm not sure if it's sick or not, hopefully someone can help.
First, none of the blinking lights programs I've seen blink the lights. I
can store and retrieve the program correctly, but when I put 1000 on the
switches and hit load address then run, it goes into run mode but no lights
blink (with any display switch setting). Halt makes the processor halt.
Here's what I did:
1000 005000 start: clr r0
1002 005200 inc r0
1004 006100 loop: rol r0
1006 000005 reset
1010 000775 br loop
On this program I put 1000 in the switches, hit load address, then run.
Address lights constantly show 1010, Data lights are blank, and the Run
light is on. Data Paths setting shows 0, bus register setting shows 5, uaddr
shows 177774.
Another test... if I store a 5007 (clr pc) in location 0 and start the
processor at location 0, the address lights show all 0's, the data lights
show 2, and the following lights are lit - pause, master, kernel. I also
tried this with location 0 set to 777 (br .) and get the exact same results.
More info... just after power up, the address lights show 20, the data
lights show 21043 (data paths). Master & Kernel are on. The uaddress setting
shows 177570.
So... I decided to go back to basics and check all the power points at the
pins on the backplane. Here's what I found:
A02A2 0
A06A2 4.931
A10A1 4.931
A26A2 4.931
A19A2 0
A16A2 0
A21A2 0
E02B2 -15
A17V2 0
A22V2 0
A17U2 0
A22U2 0
F17C1 0
E15A1 16
E01B1 9
Since my machine only has regulators at B, C, D, E, and F, I suspect the 0
volt readings above are ok. All the 5 volt test points have a very clear .2v
sawtooth pattern (which looks fairly smooth at 5v). Since one manual says
the ripple limit is .15, and another manual says the ripple limit is .20, I
suspect I'm ok there. The ripple on -15, 16, and 9 I can't see any
noticeable.
And by the way... the cpu cardset was borrowed from another listmember and
tested before I received it so I don't think it's something wrong with the
cpu cardset itself.
Any thoughts?
Jay West
6:50 p.m.
What is this??? I sent this over THREE weeks ago!
Joe
At 06:18 PM 7/25/05 -0400, you wrote:
Are you sure that the lights aren't blinking too
fast to see? Perhaps you
can hang a logic probe (with a pulse extender) on one of the lights to be
sure.
Joe
At 06:08 PM 7/5/05 -0500, you wrote:
>I finally allocated another day to spend on the 11/45. Unfortunately, I
must
>say I'm not sure if it's sick or not,
hopefully someone can help.
>
>First, none of the blinking lights programs I've seen blink the lights. I
>can store and retrieve the program correctly, but when I put 1000 on the
>switches and hit load address then run, it goes into run mode but no lights
>blink (with any display switch setting). Halt makes the processor halt.
>Here's what I did:
>
>1000 005000 start: clr r0
>1002 005200 inc r0
>1004 006100 loop: rol r0
>1006 000005 reset
>1010 000775 br loop
>
>On this program I put 1000 in the switches, hit load address, then run.
>Address lights constantly show 1010, Data lights are blank, and the Run
>light is on. Data Paths setting shows 0, bus register setting shows 5,
uaddr
>shows 177774.
>
>Another test... if I store a 5007 (clr pc) in location 0 and start the
>processor at location 0, the address lights show all 0's, the data lights
>show 2, and the following lights are lit - pause, master, kernel. I also
>tried this with location 0 set to 777 (br .) and get the exact same results.
>
>More info... just after power up, the address lights show 20, the data
>lights show 21043 (data paths). Master & Kernel are on. The uaddress
setting
>shows 177570.
>
>So... I decided to go back to basics and check all the power points at the
>pins on the backplane. Here's what I found:
>A02A2 0
>A06A2 4.931
>A10A1 4.931
>A26A2 4.931
>A19A2 0
>A16A2 0
>A21A2 0
>E02B2 -15
>A17V2 0
>A22V2 0
>A17U2 0
>A22U2 0
>F17C1 0
>E15A1 16
>E01B1 9
>
>Since my machine only has regulators at B, C, D, E, and F, I suspect the 0
>volt readings above are ok. All the 5 volt test points have a very clear
.2v
sawtooth
pattern (which looks fairly smooth at 5v). Since one manual says
the ripple limit is .15, and another manual says the ripple limit is .20, I
suspect I'm ok there. The ripple on -15, 16, and 9 I can't see any
noticeable.
And by the way... the cpu cardset was borrowed from another listmember and
tested before I received it so I don't think it's something wrong with the
cpu cardset itself.
Any thoughts?
Jay West
6933
days inactive
6953
days old
8 comments
6 participants
participants (6)
-
ard＠p850ug1.demon.co.uk -
jwest＠classiccmp.org -
mcesari＠comcast.net -
pkoning＠equallogic.com -
rigdonj＠cfl.rr.com -
uban＠ubanproductions.com