PDP Recovery 11/04. LA30 Decwriter Terminal

Original author: Mattis Lind
  • Transfer
Continuation of the translation of an article on the restoration of an old interesting machine. In the first part, we set up the main unit board. In the second - the tape station TU60. A lot of heavy pictures. Italics are my comments.

LA30 Decwriter

Now I have begun the final part of my quest - to get a good terminal, which would be as vintage as the main system. I am fortunate that I have the LA30 in storage, which I am now about to start rebuilding. Our copy was made at the beginning of 1973. Although the printheads were apparently replaced in 1979. The

LA30 was apparently one of the first dot-matrix printers, which makes it interesting in itself, even in isolation from the PDP-11.

A print terminal with a really nice vintage look! It is also heavy. The body is cast (it seems to me) from aluminum, and the top cover is made of fiberglass reinforced with ordinary plastic.

The Decwriter on the right has a data rate switch, which is why I think it was the LA30S for the serial port, but the LA30P is written on the information plate on the back. In addition, it does not have two cards that are needed to connect via a serial port.

All printer control logic is transistor ( TTL ), with the exception of ROM for character decoding, a keyboard encoder and an optional UART.

After cleaning, the keyboard looks pretty decent. The layout is slightly different from modern keyboards. The microchip of a keyboard encoder is very beautiful, isn't it? A very early LSI created by SMC.

I started by fixing capacitors. Although all the capacitors were in quite good condition. Leakage current of 100 μA at the worst of them. Not bad, especially after 40 years. After that, I cleaned the printheads of old ink.

When (and if) I make it work, I will still have a serious problem: this is a parallel terminal that requires an M7910 board on the PDP-11. I don’t have one. Another option is to find two missing boards - M7731 and M7389, which, in my opinion, is even more difficult. The last option is to assemble a small adapter on Arudino from a parallel port to a serial port.


In the photo below, the print head of our LA30. The time stamp on it contains 1979. This means that the head has been replaced. This looks like the same head used on the LA36 and LA180.

I tested printheads by applying 15V and ~ 2A to each coil in an extremely short amount of time. Only 2 out of 7 needles reacted. Therefore, I bought a special recommended oil - Omega 636, and, by the way, it contains something called allspice. Liquid oil and a few days of rest brought life to three additional needles, and the remaining two need little pressure to release. I hope when I return the heads to the printer, these needles will work too.

Homemade Keyboard Encoder

SMC's KR2376, a pMOS keyboard encoding chip, is broken. Finding a suitable replacement would not be easy and very expensive. I made my own analogue on AVR.

Since the keyboard is capacitive, this means that everything and every key is a transistor. The collector of which is connected to the X-matrix, and the emitter to Y. The base, in turn, is connected to a variable capacitor, which is formed by the movable part of the key and the board itself. On the other hand, an oscillator is connected to this capacitor.

The original SMC chip is on the left and my replacement is on the right

Arudino has a pretty good keyboard library that only needs a few minor edits. In addition to the faulty encoder, there was a micro crack on the keyboard board, which I patched up with a small amount of solder. After that, the keyboard worked fine!

Software can be downloaded here .

Power Supply

The power supply contains a large transformer and four electrolytic capacitors. One of the transformer windings is connected to an AC capacitor to form a ferroresonant circuit. Thus, the power supply becomes more resistant to power glitches. It saves energy in the coil and capacitor. But this capacitor is already very old and most likely contains polychlorinated biphenyls. The nomenclature of the capacitor was recorded in the list of elements of the board, so I decided to replace it just in case, before continuing to work. I found a suitable capacitor from Cornell-Dublier

Old replacement capacitor

While I was waiting for delivery, I rechecked the electrolytic capacitors. Everything was fine with them, and as soon as the AC capacitor arrived, I was able to start the power supply under the test load without any problems. All specifications were excellent.

Enhanced Encoder Board

My makeshift keyboard encoder was a bit clumsy and didn't look very cool. For self-education, I decided to make a wiring and order a board in China. The result was wonderful!

Testing the entire terminal

Before starting the full test, I dismantled the printer in order to clean it and grease all the places where it was required. After reassembling and raising the machine to the “ON” position, of course, nothing happened. The first problem was that there were two lock keys that must be in the correct position in order for the block to be ready for operation. One of them checked that there was paper in the printer. When the printer is in the “READY” state, the “B INI L” signal is at a high level and the “READY” light on the keyboard lights up nicely. In addition, a strobe pulse is generated from the keyboard to the printer control logic. The LA30 version with parallel connection does not have a LINE / LOCAL switch, therefore this option is not user-controlled, but it turns out that the logic circuit with this feature is present on the M7712 board, however not used. When the “LCL ENA L” signal is supplied via pin AR2 on the M7712 board, it closes the printer to local mode of operation.

And voila, he does print something when we type on the keyboard, but since there is no ribbon in its place, it is very difficult to understand what it does print.

Looking through the store, I found three different ink ribbons in the original packaging, still in shrink and even in folded paper. It seems that the top one is for LA30, and the bottom one for LA36 / LA180. The average is not known exactly, but perhaps for the LS01 / Centronics 101. By the way, plain paper did not fit the LA30. It needs 250mm paper instead of 216mm.

Character Generation Problem

By inserting paper into the printer, I was able to check it, and it did print something, but this “something” was incorrect.

It seems that the last column of each character has disappeared and, in addition, despite the keys pressed, only a limited set of characters appears on paper. Referring to the ASCII table, you will notice that two bits are stuck. Having traced these bits, we see that the correct values ​​are applied to the input of the symbol generator. Since the column counter is also located inside this generator, it becomes the main suspect. First you need to check the triggers marked in red.

It turned out that these chips are not working properly. Replacing the broken 7474 made by National Semiconductor (again NS!) In 1972 with ceramic from Fairchild and Texas Instruments.

Visible last character

There was one more trouble. The "visible last character" function does not work the first time until the machine has warmed up. It begins to function some time after the start. This is a function that moves the print head two characters to the right after a timeout of 1 second so that the last character typed is visible to the printer. Great thing if you use the machine as a console terminal. I tracked the problem to a non-working 1 second timeout generator. Red area on the diagram.

Initially, it looked like a broken E14 (7440), but when I tested it using a cooling spray, it turned out that this counter E12 (7493) was behaving badly. NS chip again! Replacing it with a TI chip, released in 1969, I saw that the generator was working.


The last problem is related to the fact that the stepper motor stopped during the translation of the carriage on the filled line. Step pulses were just too fast. The LA30 has a separate unit that controls acceleration during carriage transfer. I simply lowered the maximum speed by tuning one of the potentiometers in the G936 module.

With this last edit, the printer worked as it should!

Only registered users can participate in the survey. Please come in.

Do you need this kind of article?

  • 96.8% Yes 591
  • 3.1% No 19

Also popular now: