[pups] pdp11/05 questions.

[pups] pdp11/05 questions.

[Milo Velimirovic]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1180 bytes --]

hi all,

The recent flurry of activity in early pre-C UNIX for an 11 with a  
small memory got me back to working on my 11/05. So far I've  
identified two nasty problems with the data paths board, the M7260.  
One of the 8266 MUX chips looks like the plastic boiled and bubbled  
and circuit board is discolored underneath it. I'd welcome both  
sources for replacement chips and techniques for replacing it.

Additionally there's a lifted and broken trace on the non-component  
side of the module near the F edge connector. Any sugestions for  
repairing a damaged trace would be welcome.

Lastly, I'd just as soon use a DL11W in the 11/05 rather than go to  
the trouble of setting up an external clock to feed the on board UART.  
I can get both 9600 baud and RS232 from the DL11W instead of 2400 baud  
current loop from the built-in interface. I haven't yet found the  
jumpers to remove/install that would disable the built-in console  
interface. There's also the LTC.

TIA,
Milo

--
Milo Velimirović,  Unix Computer Network Administrator
University of Wisconsin - La Crosse
La Crosse, Wisconsin 54601 USA   43 48 48 N 91 13 53 W
--
Unix: Where /etc/init is job #1.

[pups] pdp11/05 questions.

[John Holden]

On 08/07/2008, at 7:35 AM, Milo Velimirovic wrote:

> hi all,
>
> The recent flurry of activity in early pre-C UNIX for an 11 with a
> small memory got me back to working on my 11/05. So far I've
> identified two nasty problems with the data paths board, the M7260.
> One of the 8266 MUX chips looks like the plastic boiled and bubbled
> and circuit board is discolored underneath it. I'd welcome both
> sources for replacement chips and techniques for replacing it.
>

Tricky. The 8266 is rather odd in that it inverts one of the inputs
and I know of no direct replacement. I'll check if I have a spare  
M7260 card.

As for replacing chips, particularly on old cards. The only safe way  
is to:-

1)	with a fine pair of cutters, cut every leg of the chip and remove  
the body
2)	with needle nose pliers, grab each leg in turn, heat on the other  
side
	with a soldering iron and pull out the leg (use a light force and  
make sure
	the solder is molten on the top side, particularly if it has a track)
3)	with a solder sucker (or desoldering station), remove the solder  
from the holes


> Additionally there's a lifted and broken trace on the non-component
> side of the module near the F edge connector. Any sugestions for
> repairing a damaged trace would be welcome.
>

If the chip was toasted, then this track probably took all the  
current. You can
either glue it down and bridge the break with solder, or cut the track  
where it's
good, scrape off the PCB lacquer, then carefully solder in some wire.  
If it is
a short run, use some stiff tinned copper, then hot glue in place.

> Lastly, I'd just as soon use a DL11W in the 11/05 rather than go to
> the trouble of setting up an external clock to feed the on board UART.
> I can get both 9600 baud and RS232 from the DL11W instead of 2400 baud
> current loop from the built-in interface. I haven't yet found the
> jumpers to remove/install that would disable the built-in console
> interface. There's also the LTC.

You can disable the serial interface by removing W1 from the control  
logic /
microcode board M7261. It is not possible to disable the LTC, simply  
don't
enable it on the DL11W. I have, in the distant past, modified a M7260  
for
9600 baud RS232, but it involves removing removing several components  
and
modifying the 9602 oscillator timing. Using a DL11 is a better choice.

John

[pups] pdp11/05 questions.

[Milo Velimirovic]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 3823 bytes --]


On Jul 7, 2008, at 10:27 PM, John Holden wrote:
> On 08/07/2008, at 7:35 AM, Milo Velimirovic wrote:
>
>> hi all,
>>
>> The recent flurry of activity in early pre-C UNIX for an 11 with a
>> small memory got me back to working on my 11/05. So far I've
>> identified two nasty problems with the data paths board, the M7260.
>> One of the 8266 MUX chips looks like the plastic boiled and bubbled
>> and circuit board is discolored underneath it. I'd welcome both
>> sources for replacement chips and techniques for replacing it.
>>
>
> Tricky. The 8266 is rather odd in that it inverts one of the inputs
> and I know of no direct replacement. I'll check if I have a spare  
> M7260 card.

I noticed that oddity while staring at the printset. One thought (and  
a really ugly one at that) was to build a two chip replacement. Of  
course the pinouts of an inverter and a "normal" quad mux not being  
anything at all like the 8266 would really complicate matters. A last  
resort.... Also would the two chip sol'n change the timing through the  
AMUX - is it in a critical path?
>
>
> As for replacing chips, particularly on old cards. The only safe way  
> is to:-
>
> 1)	with a fine pair of cutters, cut every leg of the chip and remove  
> the body
> 2)	with needle nose pliers, grab each leg in turn, heat on the other  
> side
> 	with a soldering iron and pull out the leg (use a light force and  
> make sure
> 	the solder is molten on the top side, particularly if it has a track)
> 3)	with a solder sucker (or desoldering station), remove the solder  
> from the holes

No surprises here. It's just painstaking, delicate work that I have  
little experience with.
>
>
>
>> Additionally there's a lifted and broken trace on the non-component
>> side of the module near the F edge connector. Any sugestions for
>> repairing a damaged trace would be welcome.
>>
>
> If the chip was toasted, then this track probably took all the  
> current. You can

That makes sense but it's a trace to the F connector (FK2 I think) and  
the 8266 is at position E10? above the A connector...
>
> either glue it down and bridge the break with solder, or cut the  
> track where it's
> good, scrape off the PCB lacquer, then carefully solder in some  
> wire. If it is
> a short run, use some stiff tinned copper, then hot glue in place.

Again no surprises. The broken trace is in an area densely packed with  
runs as all the signals from the component side feed through the  
module and are interleaved with the lines from the solder side, all  
feeding into an area filled with resistors and diodes.

>
>
>> Lastly, I'd just as soon use a DL11W in the 11/05 rather than go to
>> the trouble of setting up an external clock to feed the on board  
>> UART.
>> I can get both 9600 baud and RS232 from the DL11W instead of 2400  
>> baud
>> current loop from the built-in interface. I haven't yet found the
>> jumpers to remove/install that would disable the built-in console
>> interface. There's also the LTC.
>
> You can disable the serial interface by removing W1 from the control  
> logic /
> microcode board M7261.

I found the alleged location of W1 on the M7261 - it's not obvious  
what's a jumper and
what's necessary copper. back to the print set before wielding and  
Xacto.

> It is not possible to disable the LTC, simply don't
> enable it on the DL11W.

Easy enough to disable the LTC on a DL11W.

> I have, in the distant past, modified a M7260 for
> 9600 baud RS232, but it involves removing removing several  
> components and
> modifying the 9602 oscillator timing. Using a DL11 is a better choice.

I'd much prefer the DL11 choice to more board mods.

  - Milo
--
Milo Velimirović,  Unix Computer Network Administrator
University of Wisconsin - La Crosse
La Crosse, Wisconsin 54601 USA   43 48 48 N 91 13 53 W