Presented in the order of oldest at the top. (The background comes first :-))
Please note: not all e-mails are presented -
E-mails saying "Thank you" or "I agree" or similar have been deleted
Date: Thu, October 12, 2017 9:00 pm
Subject: IBM 701 SpeedCode
From: "Sancho Villa, Roberto"
< rsancho@indra.es< mailto:rsancho@indra.es >>
Date: Thu, October 12, 2017 9:00 pm
To: "ed@ed-thelen.org"
Hi again Ed
I’ve been investigating on IBM 701. It is simulated by Rich Cornwell fork from SimH, but there is no sw at all nowhere.
I found several docs floating around that has comments on Speedcode, but just general descriptions. No traces of what it looks like.
But looking harder, I found at bitsavers the file
pdf/ibm/650/28-4024_FltDecIntrpSys.pdf that seems a port of SpeedCode
to 650. Source included! Only this is a very good reason to write
SimH simulator for 650 (added to my ToDo list ?)
The language is very interesting as it is on the timeline of Fortan, and it is a very clever usage of very limited machine resources.
So I googled again and found at CHM:
IBM Speedcoding System for the type 701 Electronic Data Processing
Machines Catalog Number: 102678975
Publication number 24-6059-0
Unfortunately this IBM manual is not scanned nor available anywhere.
And now here is my question: it is possible to check if source is included (was common in the era). And if so, it is
possible to ask for a scanned copy?
Should be a very nice addition to bitsavers, and a must to simulate and document the roots that leads to FORTRAN.
Many thanks in advance for your help
Best regards
Roberto
|
Oct 13, 2017, at 2:12 AM
On Oct 13, 2017, at 2:12 AM, ed@ed-thelen.org wrote:
> I’ve been investigating on IBM 701. It is simulated by Rich Cornwell fork from SimH, but there is no sw at all nowhere.
Roberto :-))
You seem to love challenges ;-))
I am CCing
Dag Spicer, who seems into these things
and
Al Kossow, who does BitSavers, and also works for CHM
If volunteer labor is useful,
I am willing/eager to do the scanning and/or be "gofer".
-Ed Thelen
PS Paul McJones did a massive accession of FORTRAN development documents for CHM.
The others have surprising energies/interests/talents.
Unfortunately, Bob Erickson, who used to maintain 701s, is no longer with us.
http://ibm-1401.info/TeamBios.html#Erickson
|
On Oct 13, 2017, at 9:12 AM
On Oct 13, 2017, at 9:12 AM, Paul McJones > wrote:
Dag and Al,
I wonder if someone over at the Shustek Center could check the 701 Speedcoding manual
(http://www.computerhistory.org/collections/catalog/102678975)
that Robert identified to see if it contains source code.
If it does, I would be willing to go over the Center and scan it.
Paul
|
Oct 13, 2017, at 9:17 PM
On Oct 13, 2017, at 9:17 PM, Dag Spicer wrote:
Hi Paul,
I asked Chris to pull the manual this afternoon and have a look.
It is a user's manual, with no source code listings or other internal info.
Please let me know if we may be of further assitance.
Best wishes,
Dag
Dag Spicer
Senior Curator
Computer History Museum
1401 N Shoreline Blvd
Mountain View, CA 94043
|
...
On 10/14/2017 6:35 AM
On 10/14/2017 6:35 AM, Sancho Villa, Roberto wrote:
Hi Dag, Paul
Thanks for reply.
I've found are another second candidate with near the same name:
IBM 701 speedcoding system
Catalog Number102632810
PublisherInternational Business Machines Corporation (IBM)
Extent60
CategoryManual
Lot NumberX3021.2005
Maybe this is the good one.
Anyway, do you have plans to scan the manuals? I think it is very interesting to see how it is defined.
Besr regards and thanks again
Roberto
|
14 de octubre de 2017 20:28
-----Mensaje original-----
De: Paul McJones [mailto:paul@mcjones.org]
Enviado el: sábado, 14 de octubre de 2017 20:28
Para: Sancho Villa, Roberto ; Dag Spicer
CC: Al Kossow ; Ed Thelen ; Robert Garner ; Marc Verdiell ; research
Asunto: Re: IBM 701 SpeedCode
Roberto,
Perhaps Dag can ask Chris or someone else to take a look at that manual too.
Since the Museum has moved the archives from headquarters (5 minutes from my house) to the other side of the Bay (25 minutes, if no traffic), I don't go there as often. But I expect to be spending a lot of time there later this year (late November or more likely December), so I could probably scan the SpeedCoding manuals then.
Paul
|
Sat, Oct 14, 2017 12:30 pm
Hello friends,
Chris and I will be over at Shustek on Tuesday a.m. so can inspect this second documet then.
Will that be OK?
Best wishes,
Dag
—
Dag Spicer
|
Sun October 15, 2017 at 4:28:26 AM PDT
From: "Sancho Villa, Roberto"
Subject: RE: IBM 701 SpeedCode
Date: October 15, 2017 at 4:28:26 AM PDT
To: Paul McJones , Dag Spicer
Cc: Al Kossow , Ed Thelen , Robert Garner , Marc Verdiell , research
Hi Paul
Nov/Dec is a great timeframe. I think there are invaluable gems buriend in CHM storages waiting to be re-discovered. SimH needs sw!
BTW, I created a GitHub account to hold my work, with a ready to run demo of a full loaded IBM 7094 datacenter
https://github.com/rsanchovilla/SimH_cpanel/blob/master/test_run/I7000_cpanel_test.zip
Includes CPU, data channels, tapes with vacuum columns simulation, card reader, and 2302 Disk simulation. All under SimH.
A side question: do you know who manages http://simh.trailing-edge.com/software.html?
I've reconstructed with Dave Bryan the first version of RTE OS for HP2100. It would be a nice addition to sw kits
Regards
Roberto
|
...
Sun, Oct 15, 2017 1:23 pm
Subject: Fwd: IBM 701 SpeedCode
From: Robert Garner
Date: Sun, Oct 15, 2017 1:23 pm
To: Paul Pierce , Bob Supnik
Cc: Robert Garner , Ed Thelen , Dag Spicer , Van Snyder
, Paul McJones , "Sancho Villa, Roberto" , Al Kossow
, Marc Verdiell , Luca Severini , Ron Mak
, David Brock , Hansen Hsu
Paul, Bob,
fyi .. some amazing emulation projects by Roberto Sancho Villa:
729 unit emulator, COBOL on 1401, Monitor Sys 1 on 1620, RTE OS for HP 2100,
and now (I believe) targeting an entire 701 and 7090….
Roberto,
> A side question: do you know who manages http://simh.trailing-edge.com/software.html?
SIMH’s author, Bob Supnik, copied, should know… ;-)
- Robert
|
Sun, Oct 15, 2017 2:43 pm
Subject: IBM 1401 Operating System
From: "Sancho Villa, Roberto" < rsancho@indra.es>
Hi Folks at 1401 team
I remember reading 1401 has no OS.
Well, I’ve found at BitSavers an interesting doc from Univ of Milwaukee. No sources, but … sounds like a Batch OS similar to IBSYS.
Any of you were aware of this promising piece of SW?
Regards
Roberto
|
Sun, Oct 15, 2017 5:36 pm
Subject: Re: Fwd: IBM 701 SpeedCode
From: Bob Supnik
In fact, I run the simh.trailing-edge.com web site, so the message has
come to the right place.
/Bob
|
Sun, Oct 15, 2017 5:37 pm
Subject: Re: IBM 1401 Operating System
From: Bob Supnik
I've not heard of it before. I'm copying Van Snyder, who has been SimH's
expert on 1401 software.
|
Sun, Oct 15, 2017 5:50 pm
Subject: SimH SW Kits - HP2100
From: "Sancho Villa, Roberto"
Hi Bob
Great! And very pleased to contact you. Not all days I met someone who has microcoded cpus
So ... I propose you to add to sw kits
- RTE version 1: first version of long lived Real Time OS for HP2100. Self hosting source included.
- MTS: Magnetic Tape System: MagTape monitor for HP2100. Only 5 (five) commands, who says less?
Regards
Roberto
|
|
Sent: Wednesday, August 16, 2017 3:01 AM
From: Sancho Villa, Roberto [mailto:rsancho@indra.es]
Sent: Wednesday, August 16, 2017 3:01 AM
To: ed@ed-thelen.org; Robert Garner
Cc: 1401 Software Team; Paul Pierce; Marc Verdiell; Ron Mak; Ken Shirriff
Subject: Call me crazy ... emulating 729 MagTape
Hi again Guys
I’m developing a control panel functionality for SimH. I’m working on old IBM irons, and gone to the point of simulating the 729/727 reel movements.
I would like to reproduce as close as possible the fascinating and always unexpected reel spinning. After some tries I realized the way is to model the medium into the vacuum columns.
I’ve seen many times the 729 debug videos on youtube, but … need help.
So I ask for your support on fully understanding the pressure sensor works and the hardwired logic that drives the reels motors.
Can anyone help me?
Best regards and Thanks in advance
Roberto
BTW: screenshots are not mockups. Tape reel spins on cpu commands, r/w head moves up and down on tape attach. You can see the medium thru reel fences going from one side to the other. And of course, computer computes. If you like to test drive my wip please let me know J
|
Enviado el: miércoles, 16 de agosto de 2017 19:38
De: Marc Verdiell [mailto:marc.verdiell@gmail.com]
Enviado el: miércoles, 16 de agosto de 2017 19:38
Para: Sancho Villa, Roberto ; ed@ed-thelen.org; 'Robert Garner'
CC: '1401 Software Team' <1401_software@computerhistory.org>; 'Paul Pierce' ; 'Ron Mak' ; 'Ken Shirriff' ; Carl Claunch ; 'Ignacio'
Asunto: RE: Call me crazy ... emulating 729 MagTape
Roberto,
You look a normal person to me, just more gifted than usual. Awesome project. We sure want to try it!
The vacuum column control is simple. If you follow the video it will be obvious. Two holes in the vacuum column (one towards the top, one towards the bottom) are connected to a tube which leads to a vacuum switch (the one I show behind the front door). The vacuum switch is actually operated by a bellow that gets inflated/deflated when the vacuum gets broken by the tape going lower than the hole Vacuum sucks from the bottom of the column, air enters on top, tape acts as a movable barrier in the middle. When tape is higher than hole, the switch see vacuum, bellow contracts. When tape is lower than the hole, the switch sees atmospheric pressure, bellow inflates. The tape reels tries to maintain the tape around the top or the bottom hole, depending on which direction the tape is going. More precisely, the take up side tries to maintain the tape around the high point and the feed side tries to maintain the tape at the low point. For that it simply activates the reel clutch if the tape goes past the hole (too high in the feed side, too low in the take up side). I believe it is a simple on/off relay driven thing, the vacuum switch directly activating the reel clutch through relays. There is only some intervening relay “logic” between switch and clutch that changes reel direction (in our case, meaning which clutch is selected), on/off versus off/on, and which hole is monitored to achieve the desired result depending on the direction of tape travel.
If you need a specific video let me know, but that should be much more straightforward than most things you are trying to model. You are going to need to know the position of the holes, physical size of the column and the reel speed to get a correct simulation. We can try to measure that today, unless Iggy knows it on top of his head.
Marc
|
Date: August 16, 2017 at 11:03:02 AM PDT
From: "Sancho Villa, Roberto"
Subject: RE: Call me crazy .. emulating 729 MagTape
Date: August 16, 2017 at 11:03:02 AM PDT
To: Marc Verdiell , "ed@ed-thelen.org" , 'Robert Garner'
Cc: '1401 Software Team' <1401_software@computerhistory.org>, 'Paul Pierce' , 'Ron Mak' , 'Ken Shirriff' , Carl Claunch , 'Ignacio'
Hi
Thanks Ed, Marc for your great explanations.
I’ve some doubts concerning border cases, please let me share them
- I’ve been told that on reading and backspacing short records, it is possible to have the reels rotating on opposite direction at same time (on short periods of course). Is this possible? If so, how?!?!?
- On mounting tape, 729 repair youtube video shows some kind of read movements (seems forwards). Is this because tape just mounted is read by cabinet for some non cpu visible data? To calibrate something? To position tape?
- If tape is on high point on take side (low on feed side), and tape asked for backspace, does it have enough tape to go on while take reel accelerate in reverse position? Does it wait to reverse columns medium levels before going on?
- After backspacing command ends execution on tape, columns stays “reversed” until next command, or returns to normal levels (Left column full, right one empty) guessing that next command will be forward operation
(please forgive me if too many questions)
The measurement to make calculation will be very helpful ! Many thanks
Best regards
Roberto
|
Aug 16, 2017, at 1:06 PM
On Aug 16, 2017, at 1:06 PM, Bob Feretich wrote:
The movement of tape across the he read/write head is controlled by capstan rollers just to the left and right of the head. I believe that Tape commands only affect these rollers and that the vacuum columns operate completely independently. Even the two tape columns operate independently from each other. The reel motors can not respond as quickly as the capstans. So, a sudden forward/backward/forward sequence of tape movements can creates excess tape in both vacuum columns, causing both reels to try to remove tape from their columns.
The capstan rollers can move tape at 75 inches per second and reach full speed in a about 3 milliseconds.
The reel motors respond much more slowly, but are able to move tape at a faster rate, so even though there is a delay, they will catch up. I don't know the acceleration numbers for the reel motors.
There is a reflective metal tape strip at the beginning (and end) of each reel of tape. The tape at the beginning is called the "load point" marker. When a tape is mounted, the operator must manually wind the tape past this marker. When the operator presses the "Load Rewind" button, the capstans will engage and move the tape in the reverse direction until the drive senses the "load point" marker, then stop and is able to tell the 1401 that the drive is ready. (The "Start" button also needs to be pressed before it will tell the 1401 that it is ready.) This load movement is done autonomously by the 729, the 1401 is not involved.
The state of the tape in a vacuum column is not communicated to the 1401, except when the drive detects a tape break (or tape running off the end of the reel, usually caused by a missing piece of reflective tape). Then the 729 just becomes "not ready". The 1401 has no clue regarding why the 729 has become "not ready".
Regards,
Bob
|
Date: Wed, August 16, 2017 1:34 pm
Subject: Re: Call me crazy ... emulating 729 MagTape
From: Allen Palmer
Date: Wed, August 16, 2017 1:34 pm
To: Bob Feretich
Cc: Robert Garner , Iggy Menendez ,
Ed Thelen , Marc Verdiell ,
"Sancho Villa, Roberto"
Bob
When rebuilding the reel clutches using the much improved magnetic powder Grant and I spent time measuring the ‘acceleration curves’ of the clutches so as to duplicate the ‘old powder’.. those notes are in the tape logs. The amount of tape moved in any ‘time frame’ of course varied by the circumference of the tape on the reel. A one second of rotation on an empty reel vs on a full reel. Again another reason for the independence of the different part of the tape movement system in the drive.
Thanks for the excellent write up on capstan / prolay control
|
Date: Wed, August 16, 2017 2:27 pm
De: ed@ed-thelen.org [mailto:ed@ed-thelen.org]
Enviado el: miércoles, 16 de agosto de 2017 23:20
Para: Allen Palmer ; Bob Feretich
CC: Robert Garner ; Iggy Menendez ; Marc Verdiell ; Sancho Villa, Roberto
Asunto: RE: Call me crazy ... emulating 729 MagTape
The measurements on mag tape clutch torque
with various powders and currents is at
ibm-1401.info/729%20V%20clutch%20powder.xls
-Ed Thelen
Subject: RE: Call me crazy ... emulating 729 MagTape
From: "Sancho Villa, Roberto"
Date: Wed, August 16, 2017 2:27 pm
To: "ed@ed-thelen.org"
Interesting.
Ma = miili amperes? Top of scale = 360 ma? = 0,3ª. Such low comsumption has the motor?
On Aug 16, 2017, at 3:02 PM, wrote:
The magnetic clutch coils have an average resistance of
104 ohms
as per
http://ibm-1401.info/Sched2005October.html
in Electrical Measurements:
P = (I^2)*R = 13.5 watt
I really didn't get my hands into recently active clutches,
but no one complained of them being hot -
-Ed Thelen
|
|