A very complicated, vintage, 'etch-a-sketch' a-like

[TonyDuell]

I've just finished restoring a 1958 Bryans 1806 'X-Y coordinate plotting table'. In other words an analogue XY plotter. I hope this is not too far off-topic for the forum, it's nothing to do with domestic radio, but I guess it comes under 'vintage lab equipment' And it certainly uses the sort of technology discussed here.

The chassis is a solid aluminium alloy casting. The top carries the paper. Under it is a shaded pole motor with a fan on each end. The top fan draws air though holes in the top plate of the chassis, holding the paper down. The bottom fan is for cooling.

At each side is a motor. A wound-field DC type of motor. Each drives a drum which carries steel cables to move the pen. One for the X direction the other for the Y direction. Owing to the layout of the cables and the fact there is no interacton between them (driving the X motor does not affect the Y position of the pen and vice versa), the Y axis drum is twice the diameter of the X one, which results in the same motion of the pen for a given angle of movement of the motor armature. The pen position is detected by a pair of the traditional slidewires (think of long,, wirewound pots).

Also in the chassis are 3 mains transformers (basically HT for the X amplifier, HT for the Y amplifier and lower voltages, with the HT for the timebase and the -ve HT rail taken off the amplifier transformers). A metal bridge rectifier that supplies DC to the motor armatures (straight off the mains). And the resistor networks for the input to the X and Y amplifiers. The latter operate the motor if the input voltage is non-zero, the input is a weighted (by resistors) sum of the input voltage, the voltage from the position control (+ve) and the voltage from the slidewire (-ve).

Plugged into the main chassis are 6 modules :

The PSU which contains metal rectifiers for the preamplifier supplies, a pair of EZ81s for the main amplifier supplies and 4 gas discharge stabiliser tubes to provide the + and - rails for the timebase, the position controls, and the slidewires. A relay keeps the amplifiers cut off until things have warmed up (see later).

The timebase, using an EF86 and an ECC81. This is controlled by a relay, triggered by a button on the panel or an external input. It supplies a ramp votlage to the X amplifier until the pen carriage operates a microswitch in the chassis.

I've attached my reverse-engineered circuits for the main chassis, the PSU and the timebase.

[to be continued]

[TonyDuell]

The remaining 4 modules are the preamplifier and amplifier for each channel. The former is transistorised on a PCB. The latter is valved, an ECC81 and a pair of PL82s, those drive the centre-tapped field winding on the motor.

Now for the restoration.

It clearly needed a mechanical overhaul as the safety clutches in the cable drums for each axis were slipping. So I had to unstring the cables, remove the pen carriage assembly, and dismantle the drive mechanism. Truely Philips tuning drives are a lot simpler!. The only good thing was that when I came to put it all back together I could see where the cable had been kinked round the pin in the drive drum. So I could fit that, wind one part of the cable on by turnign the drum, then wind the other bit by hand, round the pulleys and anchor it. It was a lot easier than I thought.

I then powered the chassis up with no modules fitted, intending to check the transformers. I turned it off quickly as the metal rectifiers in the armature supply were breaking down, smoking and sparking. I dismantled them, fitted fibre washers to turn them into fancy-looking tagstrips (!) and fitted a 1000V 3A silicon diode in series with a 47 ohm resistor for each original diode section. That seems to work fine.

The power supply smoothing capacitors were somewhat leaky, so I replaced the lot (including the one for the relay supply in the timebase module). For capacitors on the tagboards I just soldered new ones in place. For the cans on the bottom of the PSU module I cut the top of the cap off, extracted the guts, and fitted modern capacitors inside. Not really restuffing (I didn't attempt to close the can up again, the way they are mounted puts that against the chassis anyway), just the easiest way to mount them.

Now with the PSU and timebase modules fitted (but no amplifiers or preamps) I had an odd fault. One of the 85A2 valves in the PSU would only strike if the machine had been sitting unpowered for several hours. This took a lot of thought ...

It turns out that if the 85A2 is open-circuit, the voltage across it (due to the supply resistor and the load (position controls, etc) is around 100V. That is not enough to strike all 85A2s That's one reason for the relay in the PSU. It lets the 85A2s strike and then connects the loads after things have warmed up.

That relay was often firing as soon as I turned on the mains. It's operated by a thermal relay under the control panel. The idea is that that the thermal relay operates after a couple of minutes, the PSU relay pulls in and latches via one of its contacts. It also disconnects the supply to the thermal relay heater so that can cool down. Well, the thermal relay was not releasing quickly probably due the bimetallic strip and springs being distorted with age. The are 4 adjusting screws in the beast. The make and break contacts, the toggle spring, and an offset to the bimetallic strip. I fiddled with them and eventually got it to work properly.

Time to add the amplifiers. And be greeted by smoke from the cathode resistor on one of them. You guessed it. 'That capacitor'. Although the amplifiers are DC coupled there are capacitors across the grid resistors to improve the HF response (I guess). And one was very leaky. For obvious reasons I changed all 4.

And it works. I can move the pen with the position controls, lift and lower it with the switch on the panel, etc.

So why do I call it an 'Etch-a-Sketch'? The problem is that I don't (yet) have the mating connector for the input plug. I _think_ it's a Plessey Mk4 aluminium one, 12 pins. So I can't actually feed signals into it. All I can do is move the pen with the position knobs.

I've attached the circuits for the 2 sections of the amplifer along with my notes on assembling the drive cables.

[TonyDuell]

Here are a few photos of the unit.

The first shows the undeside of the chassis with the PSU and timebase units in place.

The second shows the top of the chassis with the steel cables strung round the pulleys

The third shows the main amplifier module

The fourth shows the preamplifier PCB

And the last one is the assembled unit.

[Electronpusher0]

Very nice job Tony, I have an X-Y plotter from a later generation and that's complicated enough for me, I don't think I would have had the courage to start!

Peter

[Red to black]

Fantastic stuff!

[TonyDuell]

Oh I've done later ones too. I reverse-engineered the HP9125 (The plotter for the HP9100 calculator) which has some very 'interesting' electronics. For that one I had to rewind the 2 motors (little permanent magnet DC things). I've even worked on plotters with modern things like ICs in them.

But as I said over in the 'state of repairs' thread I'll have a go at just about anything. Certainly anything electrical or electronic. Although I am not as experienced as many people here I feel (even if I say so myself) I have enough experience to get out of any problems that might occur.

I want to see a drawing made by it.

[TonyDuell]

My art skills are not up to doing it by hand...

Once I find a source of the input connector I intend to link it to my EAI1000 analogue computer. We'll see what that can produce.

[Electronpusher0]

Originally Posted by TonyDuell

I don't (yet) have the mating connector for the input plug. I _think_ it's a Plessey Mk4 aluminium one, 12 pins.

Have tried posting in the wanted section with a photo, you never know.

Peter

[McMurdo]

do you know who might have used these things? For example I associate plotters with hard copies of cad-cam projects or large engineering drawings. What would this plug into originally?

What would this plug into originally?

Two analogue signals, I wonder if there iis a pen up/down signal too?

[G6Tanuki]

Originally Posted by McMurdo

do you know who might have used these things? For example I associate plotters with hard copies of cad-cam projects or large engineering drawings. What would this plug into originally?

Engineers, cartographers, signwriters, anyone doing any science that involved graphs. I've even seen overhead-projector 'transparencies' made using them.

(I remember 'borrowing' a client's Hewlett-Packard 7475A X-Y plotter for a weekend to do some PCB masks)

Usually it would be connected to a minicomputer of some sort (back in the days when a "minicomputer" was something at least the size of a wardrobe).

[TonyDuell]

To answer some of the questions...

I have no idea what it was originally used with (and it may well originally have been built into a larger instrument, I am convinced the bottom cover and feet on the unit are not original). As it dates from 1958 it probably wasn't originally used with a digital computer, but rather with some analogue measuement or simulation/computer system.

As for pen lift, if you use the internal timebase, the pen drops when it is triggered and lifts at the end of the scan (contolled by the relay in the timebase chassis). There is a toggle switch on the panel to enable this (if 'off' then the pen is always up). If you use an external X input then said toggle switch just lifts and lowers the pen. There are pins on that 12 pin connector that will trigger the timebase (assuming you've selected it with the panel controls) and also lower the pen in external-input mode. You just need to short the appropriate pins together to lower the pen (I assume an external relay is the easiest way to do this).

The pinouts, etc, are in the chassis wiring circuit I posted.

I have a photo of the input connector somewhere, I'll dig it out and post it in the wanted section. At least somebody (I hope) can tell me what to look for.

As for 'minicomputers the size of a wardrobe' I have several. At least one has a quad DAC board in it with TTL outputs as well (the latter could be used to control the pen relay). But I also have smaller plotters with digital interfaces (mostly HPIB) that are easier to use with such machines.

[mark_in_manc]

When I started in Noise and Vibration consulting (early 90s) we used such GPIB (or HPIB, I can't remember) plotters evolved from this to get screen dumps off FFT analysers. We'd cut the pictures out, glue them into word-processed reports with gaps left for the figures, sort out any unwanted stuff with tipex, and photocopy the lot Such was the state of trying to 'suck' graphs into word, we carried on doing that for a surprisingly long time. And yes - even using them to plot OHPs for conference presentations and lectures, later at the university.

(The digital ones ran on plotter files which were text based and readable, and it was easy to go in and change your data, if you really needed to!)

[TonyDuell]

HPIB, GPIB and IEEE-488 are all essentially the same interface. I think HPIB specifies a few things that GPIB doesn't but you can generally plug instruments of any of the 3 interfaces together and it'll work. It certainly won't do any damage.

The text-based plotter files were probably HPGL (Hewlett-Packard Graphics Language). Very easy to create and modify. I've done it many times.

As well as a couple of HPIB plotters, I have an HP7470 Opt 003 here. That's the HPIL version for the handheld calculators (!). HPIL has the same logical structure as HPIB (there was an HP interface -- 82169 -- that converted between the 2) but is serial.

But this has little to do with my Bryans analogue plotter.