Tony Duell wrote: Let's see: - awkward editing - the display does not show you the instruction, you have to list on the printer to see what you've done - key-sequence instructions rather than expressive language statements - very limited data-handling - limited to no structure to the language - etc., etc., etc., etc., etc., etc., etc., etc., etc., etc., etc., etc. Or how about this: pretty much every lesson learned and development made in languages and programming of the last 4 to 5 decades is missing. I have never been fond of 'key-sequence' programmable calculators, the 'language' is not very expressive or comprehensible. In short: tortured programming. BASIC is tortured too, but it's miles ahead of the above. I love my HP-21, a straight-forward non-programmable 4-level-stack RPN pocket calculator which I've had since 1976. But even back then I had no great desire to have an HP-25 (essentially the same machine but programmable). Additional complexity that's not needed in this scenario. The 9815 interface is essentially just what Chuck did (IIUC) and I have suggested: the 8 HPIB data lines and 3 data-handshake lines are brought across to the controlling machine. It's HPIB minus everything but the data-transfer. (I do have the manual with schematic. The calculator and plotter showed up at a university surplus a few years ago and it was one of those rare instances where it was the just about the complete package: cables, manuals, tapes, warranty card, and enough residual evidence to show it's lineage: used by a researcher in the Dept. of Pathology. The pens for the plotter were missing.) I noticed this in the HPIB interface manual and was thinking about doing that, or building a flash-memory storage device. It could have a little control panel to select a 'file' in the flash-memory. (Probably not going to happen, as if I don't have enough projects to work on). However, using HPIB between another machine and the 9815 is pretty much the same problem as using HPIB between another machine and the plotter. It would be nice as storage for the 9815 but as far as exercising the plotter goes, the 9815 just becomes an annoying and limiting waypoint between the newer machine and the plotter.

AFAIK the plotter in question (HP9872) only exists as an HPIB model. Anyway, the OP has the HPIB plotter. I don;t think he wants to buy another machine.... -tony

I believe the RS-232 version is the 7220 series. At least they look the same, I have both. Bob Tony Duell wrote:

Considering the machine is 3 decades old, that would partly explain that... Actually, if you used a larger, more modern, machine as a mass-storage unit for the 915, you could possibly figure out the structure of the program storage and eb able to write programs on said other machine and download them to the 9815. At least that would get round the listing and editing problems. I assume this plotter ahs a 'listen only' mode, where it's selected at power-on and doesn't need ot be addressed. Many HP plotters do. Of course you lose some functionality if you do that. Most HP plotters are HPIB talkers too, they can report the current plot size and pen position -- you could move the pen around with the buttons on the plotter and use the machine as a primitive digitiser (put a trace from some other instrument on the plotter bed, then move the carriage to points on that trace and read them into the computer). You might well not need this though. -tony

You _will_ lose the digitising functionality if you use the HPIB connector and set the machien to 'listen only' (if you can do that, most HP HPIB plotters do have that feature). In that mode there is no way the plotter can output data. I don't know the 9872 that well. I have soem experience of the internals of the 7245, which from what I've seen has somewhat similar electronics (the motor driver interpolator is almost identical circuitry, for exampl). Now on the 7245 the HPIB interface is very hardware-intensive -- even the HPIB bus commands (serial poll enable/disable, etc) are decoded in hardwre. Of course the handshanke and addressing are handles in hardware too. All the processor does is get a signal that says 'read a byte from the HPIB data register' and the processor does so (and interpets the bytes read as HPGL commands, of course). If the 9872 is like that, it should be very easily possible to work out what the 9815 interface is doing, since presumanbly it's also mostly handled in hardware, possbily even much the same hardware. From what you've said, it must be more than just a listen-only (unaddressable) HPIB interface, since the plotter can send data back to the 9815 -tony

It would do me no good to re-read the repliece. I know what you've said, but _I_ have a problem with it. What you're describing is, I beleive, technically impossible. I believe you are claiming a bidirectional data transfer over an HPIB (IEEE-488) line with no addressing of devices. Now, from all I've read (indluding the original IEEE488-1975 standard) and all I've ever worked with, addressing is what determines the direction of data transfer. The device addresed as a talker outputs data, the device(s) addressed as listener(s) input said data. I _has_ to be that way, becvause there is no particular device that has to be invovled in every transfer. It may be convnetional to always have the controller (computer) taking part in every data transfer, but it certainly is not required. The controller could address one device (say a voltmeter) as a talker, another device (a printer) as a listent and then 'sit back' while the readings from the voltmeter were printed out. So there io no 'read/write' line or equivlaent on the HPIB interface. I've looked at hpmuseum.net. Alas the 98130 (olotter interface) manual is not there, so I can't see shcemaitcs of that unit. And the schematics of the 9872 are quite hard to read -- I hate reading scanned schematics on a screen anyway (and no, this is not because I am trying to read them on my text-only display :-)). But anyway... Actually, I thought it was only 7 data lines (DIO8 is not wired to the 9815 connector). That makes sense, HPGL is sent in ASCII, which is a 7 bit code. Looking at the scheamtics, the 9815 interfaces uses the same hardware as the HPIB interface (the appropriate pins on the 2 connectors are wired in parallel). So the handshake must be the nroaml HPIB one, since all that is handled in hardware anyway. My guess, and it's only a guess, is that it works something like this : At power-up, if the 9815-mode line is asseeted, the plotter goes into listen mode (the HPIB talk/listen modes can be controlled from the plotter's microcontroller, this is also used for self-tests on the HPIB interface, as it is on the 7245...). The 9815 talks and sends it data. If the poltter gets a command that should output data (say a digitise command), it goes into talk mode automatically and sends said data. The 9815 side of things knows it's just sent such a command and goes into listen mode to recieve said data. -tony

Is this docuemtned anywhere? I can't see it in the 9872 operating/service manual, but as I said I've not seen the 98130 (9815 plotter interface) manual. HPIB-style, but not HPIB!. It may use the HPIB lavels, it may use the same 3 wire handshake, but it's not HPIB if it changes direction like that without any commands fromt he controller. That's what confused me. You kept on insisting it was HPIB, I still claim that, strictly, it can't be. -tony

[HP9815 interface conenctor on the HP9872 plotter, and the fact that it's bidirectional] OK, do you know this for certain, for example by doing tests on the system, disassembling the ROM in the 98130 interface or whjatever, or are you guessing it like I did? You claimed it was an HPIB port! It isn't. Ah, but it's not/ There's a trivial corrrection that i's only 7 data lines, and a more major one that it's darn well not HPIB. If it was truely HPIN, then without the ability to receive commands, it can't be bidirectional. Read the IIEEE-488 standard if you don't believe me (I have read it). OK, it uses the HPIB 3-wire handshake. I think that much is clear from the schematics. I was (and still am) interested in knowing how it revereses the transfer direction. I've _guessed_ it was done based on the 2 devices (9815 and plotter) knowing what commands would cause the plotter to send back data and knowing how to determine the end of that data. But is that known _for sure_? > >I know what you've said, but _I_ have a problem with it. > >What you're describing is, I beleive, technically impossible. > > Apparently if one comes to a conclusion before you, it's technically impossible. Not at all... Noe I said 'I believe'. Not 'it is'. In other words, I am not sure. I still believe (and will continut to beleive until somebody shows me how it can be done) that a trune HPIB system without addressing fuctionality can't be bidirectional. A device that changes from being a listener to being a talker without receiving addressing commands is not an HPIB device. Hang on a second. Last time I checked this was a list for discussing classic computers and their peripherals. It is not a list to provide you (or anyone else) with your wants (in this case HPGL files). If you post about a machin, its quite possible you'll get a thread started that goes off in direction you'd never anticipated. If you don't like that, well, tough! I make no secret of the fact that I like HP desktop calculators. I am interested in them and their peripherals. And I openly admit to not knowing everything about them. I also said from the start that I'd neve seen a 9872 plotter or its interface. The reason I started asking about it was to _learn something_. This might be an alien concept for you, but some of us don't know everything. Oh come on... I did ask questions. You could also have tried correcting my comments _with verifyable facts_ -- that is, stuff from the manuals, stuff from your own measurements. Or even just guesses ('Well, it can't be true HPIB, but the way it changes direction could well be...'). Instead you flame me. Pot. Kettle. Black. -toyn