Translate

2020-12-08

[Tech] Nostalgia Attacks!

I'm weeding out my basement store of old computer  paraphernalia. Finally! 

Once I thought I'd lay may hands on an old PC, a coax ethernet card and coax cables so I could set up a Novell network, the way it was in my heyday. I'll leave them in my store even now, you never know, BUT I suspect this is not going to happen, there are always other projects with a higher priority to finish. Some never seem to finish.

I decided to scrap some disks but since I just can't force me to throw things away without recycling them in the futile hope they will be used in some later (nutty) project that never will be realized. My problem.
Well, digging around I came across my old Psion LZ64. Yay!
Since I had a new 9V battery I installed it to see if it (Psion) still worked. I did!

When I worked for Data General (1978 - 1994) it was my favorite piece of equipment. It had a calendar! I forget meetings (because I hate them) but here was a friend that could beep, in time for me to go to the damned meetings. It saved me a lot of problems! Besides, I had a complete list of all spare numbers in it. Very useful when ordering stuff from the main office for the next day. It was very usual to go to a customer to search for the reason of failure, order spares to the next day, getting them from the railroad express outlet in the early morning and then go back to the customer and fix the problem before lunch, fingers crossed.
It was programmable (OPL, Organizer Programming Language), I liked that. 

So here I am, my Psion working and I decided to look for information of it on the Internet, just for the fun of it. I found it.

In the wayback machine:
https://web.archive.org/web/19990225155521/http://www.org2.com/

In wikipedia:
https://en.wikipedia.org/wiki/Psion_Organiser

The most interesting one, Jaap Scherphuis homepage for Psion, You can read about him in the Wikipedia post:
https://www.jaapsch.net/psion/manxp.htm

This was a fun trip, way back!

OK, back to the scrap yard. I disassembled some old ATA disks, saving their head assemblies, spindle motors and Niobium magnets. I have no rational reason for this behavior, it seems to be a personality disorder. 

Especially the motors interested me. I had no clear idea about how they work. I thought this could be a tough thing to search for, things like that are sometimes buried inside the corporations for good. I did find some interesting info though and, not surprisingly, I found in a post in Russia! The were trying to use old disks as emery machines for tools! Hilarious invention! The guy doing the research for how this could be realized, is of the right mettle though. He knew what he was doing:
https://gtshina.ru/en/avtozvuk/kak-podklyuchit-motor-ot-vinchestera-pomogi-s-zapuskom-dvigatelya-ot-zhestkogo/

The motors are build as three phase, Y- or D-connected (tree or four wires) and if you feed them three square waves, shifted 120° it will work. Nice and interesting.

As a final of this post, an image showing the size of the read/write heads and actuator assembly second  on hard drive I worked on in 1978 (the first one aas a Diablo 2.4Mb :) compared to what they looked like 1995 or thereabout. The old disk was a 10Mb dual platter (one fixed [meaning you had to have a screwdriver on your toolbox], the other swappable), DG model number 6045.

 I have lost count of the number of times i repaired these. They were good drives but we had so many of them in the field. Sometimes they landed with their heads on the platter. The platter was destroyed, naturally (the surface looked like the magnetic material on recorder tape at the time, rust-colored), but after changing the platter, normally you could scrape off the deposit from the platter off the heads with a pen knife. Then it was "go" and no problems.  Cheaper for the customer! Rugged machine!



The has been a lot of development. And more was to come!:

https://www.nobelprize.org/uploads/2018/06/grunberg_lecture.pdf
This Nobel price guy wasn't the first (or only) at the time. There was another guy that invented the same thing. One of these two got his patent earlier and won the money for it (the terabyte disks arrived as a consequence of this), The other guy ( I think it was the Nobel Prize winner) had to be content with the honor.

Seeing this hefty magnetic actuator for the 10 Mb disks, just imagine the one I have in my garage that comes from the 96 and 192Mb drives. I hesitate when I have to lift it. It's enormously heavy. Hitherto my only use have been using it to weight down parquet boards in our living room after a leak from a water filled radiator.  Good enough, I was happy to have it at that time. 

A fun thing with these is that if you short circuit the coil on these (white to black lead) and then manually push it in either direction it will generate enough EMF ( https://en.wikipedia.org/wiki/Electromotive_force ) to continue driving by itself in in the same direction. It gives you a funny feeling, like a long, soft  and damped stop instead of just a immediate stop, caused by friction.

2020-12-05

Making Windows. Patience. That's what you need most.

Update: If you want to know more about what machinery I used see:
https://a-zproj.blogspot.com/2020/08/so-you-have-rotten-window-mullioned-to.html


This blogpost is mostly my note so I don’t forget how to do what’s described below. I’ve adapted is slightly to outside viewers. I’m an autodidact so I may give you misleading advice, it’s up to the viewer/reader to determine if this is fake news. I take no responsibility for my bad communication skills.

The procedure below may seem cumbersome and it is. This, however, is the price you pay when buying a table saw for $1000, weight about 30 kg instead of buying a saw for $4000, weight 300kg (which is the one you’d really would like to have). It’s also a question of movability, the cheaper one fits in my car's luggage compartment. The number of available square meters in you shop is another deciding factor. Something that speaks for owning a table saw at all, are the many odd jobs you can do in five minutes with perfectly usable precision, like a 2m wooden strip or a 80 cm long shelf etc. where using a universal handsaw sucks royally. As usual, reality corners you.

Better start with some nomenclature. Stolen from ” https://www.thespruce.com/window-muntins-and-mullions-1822920 ” is this piece of information. Just for the fun of it I’ve decided that my windows consist of top and bottom rails, sash stiles and sash bars, both vertical and horizontal.

When i speak of "windows", it’s only the part carrying the glass that matters (to me), the part of a window fixed to the wall (the window frame) is not a part of my project. Yet.

The image below is stolen from https://www.salisburyjoinery.com/hs-fs/hubfs/Blog%20images/sj-diagram-sash-window-terminology.png?width=744&height=595&name=sj-diagram-sash-window-terminology.png where you possibly can learn more about modern windows


And for my example this would mean:



I start with sawing the rails and stiles to the correct measures, 32.2X45.0mm, cross 

sectional and 548.0mm, 515.0mm in length for stiles and rails, respectively. I try to 

be as exact as physically possible, using my table saw, which means I try to get

the dimensions as close  to 1/10 of a mm as  possible. This will *not *be possible 

in real life, but always striving for it will limit errors adding up when assembling the 

different parts in the end. Please use calipers when measuring! :0)


I end up with this:



Admittedly, not very exciting, but (relatively) exact.


Then it’s time to make the grooves carrying the glass (outside) and the decoration 

(on the inside). The choice of decoration is guided (an approximation) by what's 

already is used in the house where the finished window will be installed. 

I use a Makita router, fixed to a table to make these. For the decoration I use a 

router bit from Cobolt, 217-095, R=9,5 mm

(see table in picture). It looks like this (not the roller bearing in the small end):




For the square groove i use a straight bit that leaves a square groove.

In my case this groove is 10x10mm and experience shows, this is too much to route

in a single action so I made an insert (clamped to the vertical support) that limits 

the depth of the first cut to 5mm. After removing it, I make a second cut to get the final

10mm depth.


The first thing I made after buying the router was a router table. It looks like 

this in my application:




Changing the router bit to the Cobolt one, the depth of the lower groove (image below) 

is going to be sideways constrained by the width of the roller bearing (at the small end

of the router bit). Adjusting the height of the bit from the table I have to match the other 

groove to it so they will look equal. 

This is not possible to discern in the image though.




The windows I will replace had a problem. Water, when raining, was collected under
 the lower rail, sucked into the wood and rotting started. These windows was made 
with a sort of forking  joint (the one to the left):


























I use the simpler joint to the right.


[Comment: I want to move the joint sucking water nearer the inside of the window and

I want to make the joint in one single point, avoiding the outer, the one

more exposed to moisture. This will lessen the strength of the joint but may

delay critical sucking up water.  

Let’s say it still does rot and I would like try to forestall the total disintegration of the 

window, then I could use a wooden plug from the inside but not extending to the  

outside. If glued, this could prohibit the collapse when my joints become totally rotten, 

near end of life. 

Just a thought. We’ll see after some years of use if this is going

to be necessary.  I use polyurethane glue which is very water resistant so 

that ought to work out.

As long as the window doesn’t scrape the sill when opening it, there will be

very little stress on these joints. The stability of a finished window is more

dependent on the glass and putty than on the corner joints.]


To make the A cuts I use a home made contraption to assure the cut is 90°. The cuts

are the like the A cut but not quite as deep. Fine adjustments are made later.





























Then I do final adjustments until the two peaces fit exactly together. This

takes some time. I try the fit, adjust some and retry until they fit exactly.

I mark the ends with a number so I know which rail end fits perfectly to which

stile end. Since everything is handmade it’s necessary (at least when I’m 

doing it:). There *will* be differences between cuts, unfortunately.


[Comment: As you can see, the rail in the groove that holds the contraption to 

the saw table is made of wood. The precision of the saw is good enough to 

produce these, I made several to other tools and a sled. In the right picture,

some wooden blocks in a pile sawed to ”exact” number of mm differing in 1 mm

steps, used to make a coarse setting of the width of the cut. They make setting

up the saw a bit quicker, fine adjustments and some testing on scrap bits to 

check the exact result have still to be done.

Since measuring and setting is time consuming, making all parts with the 

same dimensions in a row saves a *lot* of time.]




Then I make a new cut B, perpendicular to the one shown. Doing this, I also fine

adjust the A cut’s length. I also remove most of the part of the decoration.

See the image above.


The "decoration" side of rails and stiles is is on the window's inside and to join them

I have to make a 45° cut in all of them. They must end at the exact point where the

distance A ends, not one mm or part of a mm more, otherwise the joint would look

outright awful from the inside. Below is the picture of the finished ends of rails and

stiles:





[Comment: As you can see the rails have the material removed  from the side facing inwards, the stiles from the side facing outwards.

Do not repeat my mistake removing material  from the same sides, oblivious 

of my error due to lack of concentration on the subject at hand. I do this a lot, 

probably some mental deficiency. If you do, you will have to remake these parts.]


The next step is the making of the horizontal and vertical sash bars fit when mounted.

This is means cutting ends to 45° at the ends of both rails and stiles.

This a purely try, fit, and try again moment.


The picture above shows the ends of the stiles and rails when the 45° cut is made.

Anyone can see there are discrepancies from the one tenth of a mm rule. The cuts

shown after lots of iterations making the ends fit nearly seamlessly. Reality is what

it is.


Coming this far you have a nice framework (the window parts on top of it, is a

"don't care". The term is coming from the IT-market and means you should disregard it.):



















Note: the very useful band clamp. A must have when making windows!


The next step is to make the crossed sash bars. In my world these are 32.0mm 

high, and the witdh is 30.0 mm. That means they have a top and a bottom side. The

bottom side is defined by routing a grove on each side, 10mm by 10mm and the 

top is defined by routing decorations on each side. Then you get:




Note: the 45 ° cut must end by the red line in the image above. There is no room 

for error, sawing too much.

If you do, you will have to throw the piece on the scrap heap.


Next, the sawing of the cross of sash bars. Exciting times, if I miss I'll have to

make new ones. Not a very attractive alternative at this point, I assure you. I've 

been there. But before doing that, I have to check that they are exactly of the

correct lengths. For that I need the band clamp shown in the image at the top, 

showing a completed widow. When applied, I can lay the finished sash bars in 

their correct positions, checking that they end exactly at the groove (indicated 

by the little red circle in the image) at both ends. If not, go back and remake 

the sash bar.


When that's OK, I now saw the 45° cut on all the four ends of the bars. Next I 

have to find where the middle of the bar is and mark it. For the marking to be

possible to see when sawing I use a slip of paper, the kind I use for printing 

things from my computer. I cut a strip some 20mm wide and use as a ruler. 

Like this:



I use the straight side, not the one resulting from my scissor paperwork, tightening it 

hard checking the "ruler" side is aligned at the free end and draw a thin line after the 

papers edge.Time to saw. I iterate with the blade until I reach a point where it touches

the line marking the middle, then turn the sash bar 180 degrees and do it again.

It sounds straight forward but since after the turn I have no line left I will have to go

by the already made cut. This is best done after some training cuts. The result is like:




As you can see, I missed the mark slightly, that is not a problem, i can then saw to the
 left or right until the checking against the end of the other sash bar fits exactly, thus 
moving the V-shape compared to the mark (until the V becomes too wide! :). I can also miss and then the sash bar 
becomes fire wood.

When done, they will look like this:

Now there's only one thing left. To make them possible in a cross mount I will have to
 saw a little extra groove, leaving them looking like this:


So, now I test that they work. It should be a really tight fit, i may have to fiddle 
with them for a while pressing them together, and dismounting them several times
before they finally give in and fit. Perfectly.

Having a cross, I check it's square 90° (otherwise each pane has to be tailored to it's
place, which would be ridiculous). Then I test it in the frame and check that it fits
squarely in it. 90° degrees everywhere, you know the drill. Then I mark the exact 
midpoints of the bars endings on the frame. That's where the next sawing will take 
place. They will look like this when finished:

You're done! Mount the window and tighten the band clamp you already have bought 
and the result is like what is shown in the beginning of this HOWTO-rant.


Disclaimer: I don't recommend making windows my way, I just make them. Find a 
better way if you can. I just want to show that with a table saw and a hand router, 
you can make windows. I didn't even take a course to do it.