Log of 2009 events
Back to Home Page
July 8, 2009
Aug 26, 2009
Sept 30, 2009
Oct 14, 2009
Oct 21, 2009
Dec 9, 2009
July 8, 2009
John Best, Dick Oswald, Dave Bennet and Joe Feng have designed and made a black box to operate the RAMAC mechanicals. (They havn't powered it up yet, "the smoke test" ;-)) " Left to right, Joe Feng, John Best, Dave Bennet, Dick Oswald. John Best has promised a better picture of the electronics :-))
The Black Box uses a microprocessor to read the analog position data and operate Pulse Width Modulation (PWM) to the electromagnetic clutches. (PWM is a more efficient - less power - method of driving the clutches than the original vacuum tube AM method.) There is an empty area on the mother board for another daughter board to enable reading of existing data. The group has not been given permission to write data. (The previously written data is still very readable (no apparent degradation, and has been recorded, to other media, by Joe Feng.)
August 26, 2009
John Best wants to make sure the current RAMAC team is credited with the current adventure. These are: Al Hoagland, John Best, Dave Bennet, Joe Feng, Dick Oswald, Mason Williams.
Dave Bennet (left) & John Best soldering a wire to connect an electronic chassis to a plug on the on the black box - that connects to cables from the RAMAC. Working further on the interconnect wiring, wire stripping, soldering connectors, ... Drawing diagrams is relatively easy, the details of the physical world take a lot of Tender Loving Care - just like in government ;-))
Photos from John Best's camera
This card outputs sensor and switch voltages - to ends of potentiomenter and to contact switches, and accepts inputs from the RAMAC, such as potentiometer voltages, crash stop contact closures, ... The back side of the card. Part of the purpose is to protect the control card from accidents in the real world. This is the card to drive the 5 air valves controlling functions on the RAMAC with 40 volts when commanded by the controller. The 2 large heat sinks and transistors operate the magnetic clutches which control the arm up/down in/out. The inductors help reduce the noise of the PWM control. All interface with the RAMAC and outside world go through these sockets in the back of the "Black Box". Here are the three main cards, including controller card with PIC microcomputer, not included is the power supply card, 40 volts for the air solenoids and about 180 volts for the clutches. Almost complete controller black box. Note the power supply card, complete with fuse ;-)) John's home workshop, with - with - everything including Diet Coke Almost complete controller black box with development debug probes.
Sept 30, 2009
Ed Thelen visited the RAMAC project. Left to right, John Best, Joe Feng, and Dave Bennet were present. John Best was using his laptop to control his RAMAC controller (see pictures above) which was controlling the RAMAC arm (with the read/write heads).
We took a 10 second movie of the RAMAC arm in motion, converted to .wmv format in two resolutions
OK resolution 990 KB, and Low resolution 184 KB
Here is some engineering data (50 years ago it would have been IBM Confidential)
- The top curve is acceleration, from the DC-Generator/tachometer. In ADC units, 0 volts is 2048 counts.
- The bottom curve is the position, from the position pot.
Oct 14, 2009
Oct 21, 2009
October 14, 2009
John Best reports that the RAMAC clutches:
delivered insufficient torque for fast operation Torque (for the same current) differed by more than 30 % with respect to each other
On October 7th, the RAMAC folks present (Dick Oswald, Joe Feng and John Best) had visited Allen Palmer of the 1401 project about Allen's experiences repairing and rebuilding the related 729 tape drive electromagnetic clutches. This work had been done several years ago at Grant Savier's workshop. Grant had made useful jigs and a special punch for making the smooth cuts for the felt washers. Allen also showed them pages 50 & 51 from the 729 CE Manual - and other material, including a disassembled clutch he was working on. There was considerable discussion of the original and new (non-IBM) magnetic powders. Allen had several pounds of each.
Back to the RAMAC Restoration Room
The clutches, motor, tachometer had been removed from the RAMAC, and were on a work bench.
clutches and tach
This image shows a clutch electrical contact slip rings, connected to the solenoid coil inside These are the fixed contact to feed the slip rings,
This is the driven end of the clutch, it rotates freely UNTIL a current, maybe 1 amp, is driven through the slip rings (next picture) then it torques the other part of the clutch, causing pull on the drive cable (above) John Best is trying to measure resistance (320 ohms) - often contacts epoxy.
It is interesting, these guys are antique car buffs, taking 'em apart to fix 'em. They each seem to have a garage full of tools, used frequently. They are comparing parts of this RAMAC to parts of antique cars.
And where to get bearings (the bearings of one of the clutches seems really noisy and rough), and how much force it will take to press the bearings out and in. If the current bearings are not the correct size, what work arounds are available? - The air is full of techie/mechanical talk -
Dick Oswald had the hots to open the clutch, John Best was more cautious - Its Open! Dave Bennet (left) looking at the driven plate (the conical section (would have been under the bottom) has been removed.) Near the tips of his fingers is the felt seal to keep the iron powder in the clutch. The part near Dick Oswald's fingers contains the iron rotor and coil.
Oct 21, 2009 at Grant Saviers' work shop on Mt. Hamilton Road
We met at Grant Saviers' workshop with the intent of repairing the two clutches used in the IBM RAMAC 350 rotating storage device. OK, I'm the paparazzi - This might be called the heart of Grant's work shop, "rather well equipped" ;-)) - better picked up than many? and the visitor's area
Christmas comes in many forms - here Grant Saviers (left) receives a RAMAC dual clutch assembly for play for a day from Dave Bennett and John Best (right) Ah Look - a damaged shaft to be rebuilt. Some frozen bearing had removed maybe 0.003 inches. How lucky can you get? ;-))
OOPS - a distraction, what is that long red thing? A two stage rocket that went 42,000 ft straight up.
The red shaft is the top stage. Grant is holding the "engine", with rocket propellant, pressure container, and nozzle. Here is the instrumentation/control package. Battery, GPS unit, 3 axis accelerometer, a 460 MHz transmitter for telemetry and tracking, and a black powder charge for popping the parachute. Aren't integrated circuits great !!
We then got into a discussion of amateur rocket propellants
Grant inserting the nose cone. His left hand is just below an area of heat bubbled red paint that got too hot at Mach 2.3. Here Grant holds a brown cylinder which is the allowed volume for various student rocket projects. These projects include autonomous vehicles to get from a launched rocket back to the launch area. How about that for miniaturization? From Grant
From Grant about the payloads:
- Amateur rocket folks successfully sued the BATFE to prevent them from "regulating" rocket propellants as explosives. When I was peripherally involved, an assembled rocket motor was considered (the B word), and if larger than a certain size could not be transported on highways - you had to cast 'em on site - a major bother. Now you can cast 'em in a factory or other controlled area and truck 'em to a launch site
- Amateurs are in general using manufactured Composite Propellants of the ammonium perchlorate based (APCP) type.
- Current techniques use multiple "grains" of say 6" long, 3" diameter, with a 3/4" center hole. Relatively uniform gas production (pressure at the nozzle) of about 800 psi is obtained because the burning at the grain boundaries including the inter-grain boundaries gives a rather even burning surface and other magic.
The "official" web site is http://www.arliss.org/ The clubs web site is http://www.aeropac.org/ Several types in the 6" dia x 11" format - examples but not constrained except that return from 12k agl ft ejection to specified lat-lon must be autonomous: powered & not powered gliders/airplanes moon rovers (to earth by parachute) retrievers - pick up an immobile beacon and take it to the L-L point guided parachutes, parafoils, etc. Smaller rockets fly one "cansat" and 3 "cansats" fit in the payload container I am holding - limit 1lb and about the size of a coke can plus parachute. The experiment is anything except vertebrae animals.
Ah - back to "business"
John Best and Dave Bennett had taken some torque vs current measurement of the two clutches while mounted in the RAMAC - a rather difficult environment. Maybe we should do it again in a better lab environment, isolated from many other variables. "We" will do static torque measurements from this side and apply current to the magnetic coil from this side.
And here we have the heart of the torque measuring. Grant made a rod to clamp perpendicularly onto the center shaft, marked with dimples at 2", 5" and 10" radii. John Best has set a current which Dave Bennett is applying to the slip rings. John is pushing a force measuring instrument at the 5" dimple, measuring the force to cause the clutch to slip. And the pen and paper. The measurements were repeatable and in line with the measurements made previously on the intact RAMAC
Enough of this Physics Lab stuff, time for fun - take it apart! Hmmm - is this torque safe??? Will I break a screw? Ah - look at that :-)) The whole goal of the clutch is to control torque to that slotted low inertal mass rotor. (The rotors of the 729 clutches were not slotted.) Visible at the center of the rotor is the felt washer to keep the black magnetic powder where it belonged -
All cleaned up, screws in one plastic cup, magnetic powder in another. Visible is the deep ring grove that the rotor spins in. One side of the grove will have one magnetic polarity, the other side the other polarity, from the current in clutch coil winding. The magnetic powder stretching from pole to the other, with the rotor in the middle. and the logbook
Grant removes the inner pole piece to get at the bearings supporting the clutch on the main shaft. Ah - look at that :-)) and the inner felt
Grant making a nice supporting structure to support the clutch while pressing out the bearings. This is not a casual operation - serious first class. Worthy of a holy relic -er- irreplaceable artifact. and a nice little grove with different radius when we turn the clutch over
And a proper pusher thingie - OK, so I ain't no machinist :-(( You give it a name !! Kinda pretty ain't it? And sometime a pulling impact is desired. Here Grant is sliding the shiny slug back along the shiny rod until it hits a stop - and out comes the bearing :-))
OOPS - Hello there - the second clutch had NO felt washer to:
- keep the magnetic powder in
- protect the bearing from the powder. - - This bearing was REALLY bad.
Grant says punching felt to make washers properly concentric is "interesting". It took 3 days to make this jig for the 729 felt washers, (a different size)
Here a felt has been punched for outside diameter. The die here is placed as concentrically as possible by eyeball. Unfortunately, if you strike it with your brass hammer - it seems to shift. But gentle vertical pressure in an arbor press does just fine :-))
Grant says he has a magnetometer - lets measure the magnetic flux in the gap with different currents at different positions.
Grant says that a method of making this round iron powder is to boil iron in a really hot flame or arc (probably in a neutral or reducing atmosphere). Here are Internet sources 1, 2
Let's look at the powder through Grant's microscope. It was stored in cups like this during disassembly - really not much per clutch, on the order of 20 grams, about the weight of 20 dimes. OK - This image was made through the microscope eyepiece onto my little pocket Canon image stablized camera. With normal 35 mm focal length, the lit image was a small part of the view - someone suggested optical telephoto, much better but rather dark. I've lightened it up in post processing.
Well, the new bearings (Timken) had been pre-ordered, and delivered, we now have good felt washers. Time to put things back together. There are rude jokes about what to do about left over parts :-(( Poor taste at the end of a long tense day? I believe Grant is pressing in a bearing, using 2 of the jigs he just made, and his arbor press.
Would you believe we didn't have any parts left over? ;-))
and didn't appear to have lost/broken/mangled any !!! :-)) and every moving part seems to move smoothly :-))
and this is not government accounting -
Dec 9, 2009
Your correspondent has been busy, then abroad :-)) then sick :-((
But back now -
Present were Dave Bennet, John Best, Joe Feng, Dick Oswald, and Mason Williams. Visitors were George Ahearn, Stan Paddock, Ed Thelen of the 1401 Restoration Project -
Dick Oswald said that he was taught a course by Fran Underwood, architect of the 1401 who spoke at the 50th Anniversary of the 1401 Announcement held at CHM November 10th. 2 hour presentation
Dave Bennet examining the track detent switch We haven't discussed this before. It is a three position switch - closed one way, open, closed the other way. This verifies that the air operated track detent is either IN or OUT. Unfortunately it is misaligned. Needed some shimming. John Best had made shims with proper holes punched. Installation of those shims, properly, was an interesting exercize.
This is a similar switch that verifies that the disc detent is either fully IN or fully OUT.
For a quick refresher,
Here is John Best explaining some point to Mason Williams, a new member of the RAMAC Restoration Team Mason Williams was introduced as a Cal Tech physicist, and seemed very familiar with magnetic clutches, oersteds, B-H curves, and other easy to forget physics things. http://en.wikipedia.org/wiki/Oersted http://bhcurve.com/
Mason Williams showed me a movie he is making to demonstrate principles of the RAMAC. This is a few seconds of the result of his 1st week's effort - very preliminary - about 750 K bytes, .wmv format click here