We have a new EPROM programmer. This one is from a different company than
Data I/O so it does things a bit different than what you might be used to.
But fear not, it is quite a bit better. Although it doesn't program EPROM's
any faster than the DIO, it does everything else substantially faster.
It has 512k bytes of internal ram so you can program sets with up to
8 27512's (or 16 27256's). The maximum word size allowed is 128 bits
which ought to cover anything we'd do in the near future. It does not,
however, allow you to patch data from the keyboard (too bad, Ed and Bob).

I've setup a new command procedure to run it and you use it from any of
the terminals in the DIO room. The commands are the same as the one's
for the DIO except you start it up with:

	$ IMI			!instead of DIO

You only need to turn the unit on (switch is in back, upper right corner)
to activate it for computer (no Sel F1 Start required). After it does its
self-test, the message IMI 3016 appears in the window. If its already on,
and the START button is lit, then press it. If the button remains lit,
cycle the power on the unit.

There are no manufacturer codes required in this unit, so the DEVICE command
needs only the generic part name and I assume it means you can program mixed
vendor parts simultaneously. But you still need to watch out for 27128's and
27128A's! 

Each socket has an LED over it and this LED can either be off, glow red,
glow yellow or glow green depending on what the IMI is doing or has done.
Basically RED means the EPROM didn't program or verify properly or in the
case of set programming, indicates a missing part. Green means the EPROM
programmed or verified correctly and yellow or off means the socket is
not being used.

Set programming is done the same as with the DIO except the word size is
handled by the VAX not by the IMI. This means that the IMI always thinks
that the word size is 8, so if you peek into memory from the keyboard you
might not see what you think you ought to. The VAX scrambles the data words
during the download into sections the size of a single EPROM and loads them
into the IMI in the order that they appear in the input file. For example
assume a word size of 24 bits and an EPROM size of 1000 (hex):

	input file offset:	+0 +1 +2 +3 +4 ... +0FFC +0FFD +0FFE
	data appears at socket	 1  2  3  1  2         1     2     3
		"		 4  5  6  4  5         4     5     6
				      etc.

This is the same as what the DIO would do.

The IMI program does what is called a LOAD TEST before beginning programming or
verifying. This test checks that all EPROMS are properly loaded into their
sockets and if set programming, also checks that there are enough EPROMS to make
up at least one set. If either of these tests fail, then the program will
display what is needed on the terminal, light the LED's over the sockets to
which you must do something, turn on the START button and wait for you fix
whatever is wrong and press the START button. The tests are repeated each time
you press the lighted START button. 

If you are programming, the IMI program checks that all the EPROMS that are
loaded into the unit are blank. If one or more is not blank, then the red
LED over the non-blank part is illuminated, the START button is turned on
and the system waits until you load a blank EPROM and press the START button.
The blank test is repeated until is succeeds.

For the really advanced EPROM user, you can also dynamically change the
programming pulse time with a /PULSE=m[.n] command on the FILE or PROG command.
The maximum value is 50.9 milliseconds and the minimum time is 0.1 milliseconds.
The typical value for the older EPROMs (2732 and smaller) is upwards of 40
milliseconds. The 27256's and co. use a pulse time of about 4 milliseconds. If
you change the programming pulse, DO SO AT YOUR OWN RISK! The data may be
programmed into the part so it'll verify but shortening the pulse time will also
shorten the lifetime of the data in your EPROM. That is to say, your data may
"erase" itself over a course of a few days. This may be acceptable for
development work where you re-program the EPROM's daily or weekly, but not so
hot for parts destined for a field test game or some such. 

Report errors to this address,
ds 

The people at IMI sent me some new program Eproms for the IMI with fixes for
some of the bugs that I told them about. I, in turn, fixed the VAX program
to compensate for their bug fixes and in the process, changed it a little.

1) It will now allow you to program partial sets.

2) Data downloads will go faster if you specify a fill character (default
   character is 00) that matches MOST of the data in the EPROM. The program
   pre-fills the RAM in the IMI with this fill character and won't download
   streams of 10 or more of those characters found in your input file.

ds