Vintage chip Programmer : " Micropross ROM 3000U "

[Vince]

Hello people,

I am trying to diversify my collection of old stuff. I decided that having 40/45 old scopes was nice, but opening my mind to other types of TE would bring diversity and spice in my collection.
So since I like the early days of programmable chips, EPROMs, PROMs, GAL/PAL etc.... I thought it would be fitting to get a vintage / period programmer to go with these chips, if an opportunity arose.
Such an opportunity just arose the other day, on my local ad site Leboncoin.fr.  A big ass programmer for only 20 Euros + as much for shipping... it's a boat anchor.

Was very filthy, gave it a good clean last night and took it apart, played with it.

- It's branded " Micropross " ... made in France here, says the plaque at the back.

- Model " ROM 3000U "

- Made in early 1987 from the date codes I could find on chips. Some chips are dated 1984....

- It's got a removable programming fixture. The one that came with my unit as you can see is a 40 pin DIL ZIF socket.

- The black rectangle above the fixture is a piece of ESD foam, nice touch. Somehow the foam is still presentable up close, and not 100% turned to dust. As long as I don't touch it, it's fine...

- It's got two serial ports at the back , DB25, for I/Os.

- Has a floppy drive  ! Makes it look expensive ! 

It's got a built-in computer and tries to boot from the floppy at power up (it was supplied with a disk labeled "System Disk", albeit written in French) , so could technically be stand alone I guess... but it can't because there is no user interface on this thing at all. Well there is a buzzer and a PowerON LED, but other than that.... no display or keypad to speak of, at all.
The floppy looks like it's a backup copy, not the original, as you can see the label looks nice/computer generated, but it's applied on some generic Verbatim floppy. I tried to read the floppy on my vintage MS-DOS computer, but it says the disk is not formatted, or at least not formatted for a PC...
which I guess makes sense ? I mean the computer in this programmer is very low power, MIPS and ROM size wise, so I guess it can make sense that the designers came with a custom formatting and file format, bare bone, that would require very low CPU grunt and code size.
So I think I could do some more work... maybe I could try to do a binary dump of the disk and see if I can see "patterns" in it, hell maybe decode some ASCII text in there ! That would be fun.

So I guess it definitely requires to be connected via serial port to some desktop computer running some software, in order to operate.
I can't find any information about this programmer on Google, never mind software, so it's basically a brick, a big door stop, a boat anchor.
Still, it was so cheap and cute that I bought it anyway just to play with it and tear it down, and the joy of owning it. So that's what I did. I am not disappointed at all in this regard, I very much am happy I bought it. Result below with pics to be spread over multiple message for clarity.

I am writing this thread to document this thing. Given how hard it is to find any info on it, it makes it exotic/rare so worth being exposed I thought, especially since it's quite a substantial piece of gear...

Also I am hoping that people who actually know about this thing, might give me some info about it. Any info is better than none at all....
My hope is that it's not French, but rather Canadian. I mean French stuff is confidential and low volume, there is rarely any information about them, never mind a comprehensive service or user manual or anything really. If it were Canadian, hence more North American, I might caress the hope that like most high-end gear from that part of the world in the '80's, like Tek or HP.. there is decent documentation about it somewhere... somewhere.... but where ?

Thread could also be used to append stuff if I ever try go further with this thing. Like say try to repair the cosmetic damage, do some reverse engineering, or ever find some piece of info about it. I could put everything here, it would then be a one stop shop for other people as well, should anyone on this planet happen to have another one of these...



TEAR DOWN

Anyway, now I will post all the pics, sorted by category / theme, with some text to explain what's going on.

First round of pics for this introductory post, let's start with the exterior pics here, and cosmetic damage.
So as I said it's big, really big, and really heavy as well.
Back panel is a still plate. Rest of the enclosure is two halves top and bottom. Very thick, chunky plastic, which I suspect might be reinforced with fiber glass or something. Surface finish is grainy/textured.
At the back we have as I said not much but two DB25 serial port, and a cooling fan, equipped with an air filter luckily, which probably explains why the inside of this thing is kinda spot less, lovely.
It was very dirty and dusty when I got it. Spent 30 minutes cleaning it and rubbing the top with a wet scotch brite pad to remove most of the minor scratches and marks.

Damage :

1)  Now looks much better only the very deep scratches are left, and there is one that's huge, very deep and very long... I would rather qualify it as a gouge, or even a trench  it's so deep the it dug into the core of the material, which as yuo can see is brown, which makes the gouge visible from a mile away 


2) At the front there is a "grill" that runs all along the edge. Most of it is fake, except where it is not : near the left corner it's an actual grill, provides an exhaust port, the PSU is right behind it. Two "bars" of the grill have been broken/missing. By sheer luck one of them actually ended up inside the device and I stumbled upon it when I tore down the thing. The piece is intact so I could potentially super glue it back into place. I would still be missing the other/second bar, but it looks much better with one missing bar than it does with two.. so better than nothing me think.

3) Seller did pack it properly so it got damaged during transport : the front lower right corner got a big hit, it cracked, starting from the bottom/hidden area, but then propagating towards the front and top of the case, going through the floppy drive bezel then ending up in the bottom left corner of the fixture area. There, the crack ended and led to a chunck of the case being  broken, ejected... but I don't have this missing piece sadly. So again one can see the brown material exposed, looks crap. The crack could be super glued I guess. The missing material I do'nt know... either I could fill it with epoxy or something, then paint it, trying to mix a custom colour to match the existing finish as best I can. The texture/grain would be lost but it would still be much improved.
Other possibility is to steal a small chunck of plastic from an inconspicuous area at the bottom of the case, and use that to craft a replacement piece that would fill the hole. This would require good skills to shape the piece just right, but would have the advantage of providing the same colour and surface texture as original.

[Vince]

Details of the main 3 damages : gouge on top, broken grill at the front left, big crack at the front right.

[Vince]

Now cracking the thing open !! 

Very nice construction inside. Thick black anodized aluminium plates, to begin with.
The floppy drive is secured to two angles using... overkill gorgeous stainless steel BTR screws, lovely.
On the left we have the power supply, which looks like an off the shelf industrial grade module. Only the black aluminium cover looks custom.

On the right, the heart of the system, that big black box...  which hosts 5 boards stacked horizontally. The back plane connecting them all together is therefore vertically mounted. that one is visible, it's the one you can see on the pic. Said board also supports a buzzer, and the output cable coming from the power supply connects straight to it.

Pic of the system disk as well.

[Vince]

Now the power supply.

All the outputs are available on the edge of the board which is kinda exposed. The terminals are labeled, to some degree.

so we have +/- 5V with sense wires, and +/- Vdc ... "Vdc", it's all it says ! I measured them and it's +/12V.

All rails are good and low ripple, so the PSU is fine and the filter caps in this thing are somehow still good ! 

I then popped the cover to see what it looked like inside.

It's a switching supply, 100 watts it says on the read panel.
Primary has two switching transistors, topology looks similar to the primary of the supply in my old Tektronix 22XX series scopes of the same era : no controller IC, but a discrete / home made oscillator. Primary of the transformer has a middle tap going to the supply, then the two transistors switch one half of the primary winding, in turn.

The supply with the most grunt is of course the +5V rail, to feed the bucket load of current hungry TTL chips you expect to find in such a device of this era.
This rail appears to be the "master" rail, regulated by the primary. On the secondary you din no regulation for it : just a beefy double diode mounted on a heat sink to full wave rectify the secondary winding from the transformer, then filter caps and that's all. I measured it at about 5.15V (under load/normal operation). As usual in those days they pump up the 5V TTL rail to make sure it can deliver in action...

The other 3 rails, "secondary" ones shall we say, do have their own dedicated regulation on the secondary side.

The +/- 12 V rails are done with a 7812 and 7912 in za TO3 package. The -5V supply looks like is done with an LM337 in a TO220 package.

[Vince]

Now the stack of 5 boards constituting the heart of the system, inside that big black box.

Removed the top plate so I could have a better look. removed two plates on either side, that secure the locking tabs at the back of the boards, just in case the boards decided to unplug themselves...

Then I could pull the boards one to have a look. Here it is.

Board #1  (starting from the top)

It's the CPU board. It runs with an 8 bit CPU, a 6502, running at 8MHz if I believe the crystal that's located..... two miles away, at the very opposite side of the board ! 

The EPROM for the code next to the CPU is D27128A, so only 16KB. Not much...
There is another EPROM on that board, but far away from the CPU so I don't think it contains code... maybe some lookup table that's part of the glue logic, who knows.

Board #2

This one is the floppy drive controller. The left half of the board contains all the chip that make up the controller, and the right half is an array of RAM chips dedicated to the operation of the controller.

Board #3

This board is I would say the "master" I/O board. It is connected to the two remaining board below it, via a couple ribbon cables as you can see.
It also carries the connector for the extra mega super wide huge ribbon cable that goes to the programming fixture / ZIF socket.

The left half of this master IO board (c), has mostly some diverse analogy stuff:

- A local voltage regulator, LM317 in TO220 package.
- Op-amps : x2 LM358
- Comparators : x2 LM339
- x3  AD558 = 8 bit DAC
- The obligatory, lone discrete transistor... a 2N2222A of course, what else ! 

The right half on the other hand has a clear structure / pattern to it. Definitely pure I/O stuff.
It has some TTL chips, lots of ULN series chips so must be Darlington arrays I guess.
Lots of diodes as well, and lots of LM339 comparators.

Board #4

Two more I/O sections similar to the one described above.

Board #5

This board, the last one, is analog. Looks like it's sole purpose is to generate various programming voltages.
All along the top edge of the board, we have a series of x5 LM317 voltage regulators mounted on the individual heat sinks, with nearby, a cluster of x5 multi-turn trimmer pots, with a dab of glue on their screw.

The third of the board near the bottom / back plane, is filled with TTL chips, I guess to interface with the backplane/ CPU board.

Then in the remaining space on the board, middle left and bottom right, we have two spots of analog stuff :

a local voltage regulator LM337, some op-amps again, LF351 and LF358, some ULN interface chips again, and a LF398 op-amp no, looked thatone up and it's something completely different : a sample and hold chip ! So I guess these programming voltages are not just "open-loop"... they must actually be monitoring what's going on as well, to be 100% sure ! 

[srb1954]

It is odd that there is a CRTC display controller chip (MC6845) on the CPU board but I don't see a socket for a video monitor on the rear panel.

Is there an internal connection you can hook up a video monitor to for debugging purposes?

The extra EPROM that you observed not connected to the main CPU bus might be a character generator ROM for the video display.

[Paceguy]

Hi Vince,
 I would highly suggest that you contact Matthieu Benoit at matthieu.benoit.free.fr. He has much info on older eprom programmers and could be of help to you. I too have picked up a couple of vintage programmers and Matthieu has been of great help. I believe that yours is not a stand alone one and has to be linked to a computer. Good luck with your project!

[pcprogrammer]

It is odd that there is a CRTC display controller chip (MC6845) on the CPU board but I don't see a socket for a video monitor on the rear panel.

Good catch

Maybe the board has been used in other devices too, that did have a display. By the looks of it there are a couple of chips (dip6 and dip8) not mounted, that might be for the video output.

[Vince]

Oh forgot one last round of pics sorry ! It's the programming fixture, which I took apart as well, of course ! 

Inside, lots of goodies : about 20 DIP reed relays, as many dipped tantalum caps, and as many lovely wire wound resistors. Right there I have my expense covered ! 

The bottom of the fixture case is a thick metal plate which adds a lot of weight to the fixture and makes it feel "expensive" 

Pic of the two sockets  on the programmer, once you pull the fixture off.

Pic of the board on the inside of the programmer, on the other side.. as you can see it holds a LOT more stuff than just two sockets !
Lots of electronic stuff on that board...  TTL chips, ULN interface chips, a few DIP reed relays again, all branded "Celduc", whoever they are... including a fancy one in a metal housing !
There is also a 4MHz crystal oscillator in there... strange.

THAT'S ALL FOLKS


That's all I can say for now... following the quick tear down I did last night on this thing.

Again if you have any information, links, material about this programmer or maybe at least this company/manufacturer, or similar products in this range, better than nothing... then please speak up and share ! 

[Vince]

It is odd that there is a CRTC display controller chip (MC6845) on the CPU board but I don't see a socket for a video monitor on the rear panel.

Is there an internal connection you can hook up a video monitor to for debugging purposes?

The extra EPROM that you observed not connected to the main CPU bus might be a character generator ROM for the video display.

Oh wow, thanks for that !!! 

You have just given me a rabbit hole to explore, sounds really cool interesting fun and exciting all at the same time !!! 

I will now take it apart again and go that route ! Well, maybe not right now... have so many other projects I need to do first.

[Vince]

Hi Vince,
 I would highly suggest that you contact Matthieu Benoit at matthieu.benoit.free.fr. He has much info on older eprom programmers and could be of help to you. I too have picked up a couple of vintage programmers and Matthieu has been of great help. I believe that yours is not a stand alone one and has to be linked to a computer. Good luck with your project!

Thanks a lot, will contact him for sure !! 

[Vince]

It is odd that there is a CRTC display controller chip (MC6845) on the CPU board but I don't see a socket for a video monitor on the rear panel.

Good catch

Maybe the board has been used in other devices too, that did have a display. By the looks of it there are a couple of chips (dip6 and dip8) not mounted, that might be for the video output.

Yes I think generally speaking there must have been other versions of this programmer, because some stuff is unused or unpopulated.

- On the CPU board I notice an empty foot print right next to the code EPROM, same size. So maybe provision for an extra EPROM, more code to implement more features.
- Still on the CPU board, right next to the CPU there is an equally large foot print that's unpopulated. Looking the pads / traces pertaining to it, it's clear that it's meant to host a large backup battery eh ?
- on the back plane PCB, there are two ribbon cables plugged there, + the PSU connector.  But I also see 3 unused connectors : a small 2 pin one next to the one that feeds the PowerON LED at the front.
Then there is a 2 row header that most likely meant for a ribbon cables. Then a third header, single row this one, for God knows what.  Maybe for a video output !! 

[Vince]

OK, Googled a bit more. Still can't find anything about the programmer or any programmer they might have made, but at least I found some info about the Manufacturer itself, I am less in the dark.

The bad news : it IS indeed French ! So all hopes of finding documentation or anything, has just vanished into the aether !
Will contact Benoit anyway of course, but I see on his Website that despite him being French hence local to Micropross like me, and despite his very comprehensive collection of devices... he doesn't have a single bit of kit from Micropross, so not looking good ! 

So, what Google found me, is an old article from a local/French business oriented journal, " L'usine Nouvelle ".

https://www.usine-digitale.fr/article/le-lillois-micropross-s-adosse-a-national-instruments-pour-se-developper-dans-le-test-de-paiement-mobile.N359819

Article explains that Micropross is a little French company, based in Lille (big city in the far North of France, near the Belgian border), that was founded in 1979 when I was 2 years old. The subject of the article, which is dated 2015, is to say that in 2014 Micropross, who gave up on chip programming H/W and moved on to the design of other types of test equipment, sold itself / partnered with a much bigger... and American.. company. None other than NI / National Instruments !  In 2014 they were staffed with only 47 people, made 16 Millions Euros Gross revenue and 92% of that was made abroad !
Honestly I don't quite see how a 47 company could have designed a programmer that complex, money wise, and in a time frame that made sense... so I suspect that in the '80s when they developed this programmer, that had a few more people around...

Also found their website. It's got its own "identity" but as you can see from the domain name it's indeed part of NI....
Here is the link to their "About us" page which explains them joining NI, along with a pic of their HQ in France:

https://micropross.ni.com/company/


OK so that's "something" I guess. Ni being a big company, if they care one bit about "customer" service, maybe they can help me out here.... nope, not a chance. At best they will redirect me to Micropross which, again at best, will be kind enough to reply to me, only to say "sorry obsolete product we have no documentation left paper or digital, it's all long gone, goo luck and thanks for your interest in our products ! "......

Will try anyway, on a "nothing to lose" basis ! 

[Zoli]

...
- It's got two serial ports at the back , DB25, for I/Os.
...

My suspicion: one serial is for computer communication, one for terminal communication; so it would be worthwhile to check for Tx activity on the ports during startup.

[Paceguy]

Matthieu Benoit may not have anything on his site concerning Micropross, but he may know someone who could be of help to you. I believe that he has a fairly wide network of contacts with people who are into programmers.

[Vince]

...
- It's got two serial ports at the back , DB25, for I/Os.
...

My suspicion: one serial is for computer communication, one for terminal communication; so it would be worthwhile to check for Tx activity on the ports during startup.

Stupid me, thanks for pointing the elephant in the room !   I guess the bigger the elephant the harder  it is to see.....
OK will do some probing soon, stay tuned ! 

[Vince]

Matthieu Benoit may not have anything on his site concerning Micropross, but he may know someone who could be of help to you. I believe that he has a fairly wide network of contacts with people who are into programmers.

Yeah, probably !  We will soon find out, as I already e-mailed him a couple hours ago, and pointed him to this thread. 

[PaulAm]

At that era serial display terminals were quite common.  Today you would just use a comm program on a pc, of course, but they were pretty rare in those days.  Looks like it was probably a pretty pricey item when new.

[capt bullshot]

Hi Vince,
 I would highly suggest that you contact Matthieu Benoit at matthieu.benoit.free.fr. He has much info on older eprom programmers and could be of help to you. I too have picked up a couple of vintage programmers and Matthieu has been of great help. I believe that yours is not a stand alone one and has to be linked to a computer. Good luck with your project!

Well, that name rang some bell ...
Yes, it's the place where I used to find some useful information about my ALL-03 porgrammer:



BTW - Vince: I enjoyed your teardown pictures. Afair, the 6502 doesn't have an internal Xtal oscillator, so IMO there's a dedicated oscillator (might be as simple as some '04 gate) for the 8MHz and the 1.8432MHz baudrate crystals. No need to place the crystals near the 6551 and 6502 chips.

[Vince]

Good point Cap'tan about the oscillator.....

I already got a reply from Matthieu Benoït, that was quick  !


Sadly not much info. He says :

1) used to know someone who knew these programmers well, but the e-mail address he had for him doesn't appear to be valid/working any longer. Also he added that this person was really really really old... so chances he passed away sadly, taking valuable information with him for eternity, bummer !!

2) He does remember these particular programmer very well himself. He can vividly remember a particular shop in Paris,  back then, that had two of these very programmers, along with documentation and disquettes. He chose not to buy one because he found them too bulky... even back then ! 

He gave me the name of the shop and the town, which is near Paris. The name of the shop rings a bell, think it's quite well known, probably still exists today, will check and try to contact them  about it !
You never know your luck... maybe they have some stuff related to it still in an old box hiding somewhere under a desk, or someone remember something about it...
I bet whoever worked in that shop back then, is long retired or dead, but maybe before doing so, they passed some knowledge to the next generation of staff, who knows... worth a shot !

But as he says, given that these were French / local, chances are there are still some French people alive today that live somewhere, and could help. The challenge is of course to FIND these people ! 

[Vince]

...
- It's got two serial ports at the back , DB25, for I/Os.
...

My suspicion: one serial is for computer communication, one for terminal communication; so it would be worthwhile to check for Tx activity on the ports during startup.

OK I played in the RS232 department. Here are the results before I fall asleep in a couple minutes probably (it's only 21H40 but well... I don't understand...).

So, I scoped both serial port during power up / boot from the floppy, for activity on the TxD pin.
Zero, nada. Well, I get -10.9Volts. Then I scoped each one of the 25 pins, and some of them have voltage... 4 of them to be exact, same result on both ports :

#4 =  +12V
#5 = +6V
#6 = +6V
#20 = +12V

So at least I know these port run +/-12V signals, that's a beginning... Sure it was to be expected since the PSU gives 12V rails, and it's meant to be connected to a computer which also usually run their RS232 on 12V. But well, nothing prevented them from doing some voltage upping to reach the 30V of the RS232 standard.

So what we also know that clearly the UART uses some of the control signals rather than just RxD / TxD. I guess it was to be expected for such an old device.
I removed the PCB that holds these connectors, from the rear panel, so I could get a closer look at the tracks going to the DB25 connectors.
Indeed many track are routed to them, and they are connected to the backplane PCB via a 20 way ribbon cable. A good part of that is used by the 6 way DIP switches used to configure each port.

Anyway. This means the programmer probably is not set up to send data on its own, at power up, unsolicited. The terminal connected to it, must be expected to ping/interrogate the programmer first, then only the programmer will respond.
Since I don't know the command set, nor what these DIP switches do, how to configure them, what control signals are used and how... it's a bit hopeless trying to communicate.... 
I mean it's not IMPOSSIBLE... but would require some painful reverse engineering to draw the schematic of that board, what the DIP switches do exactly, then lots of trial and error, trying every baud rate under the sun, combined with ever possible configuration of hardware control and start and stop bits configurations. Oh my god.... maybe one day... but not today.

Some good news though, well maybe : notice how there are two unpopulated connector footprints on this rear PCB ? One is 2x18 pins, could be for GPIB maybe. The other one is 2x8 0.1" header, so could be for a VGA output as far number of pins is concerned. 16 is more than 15.. so good enough. I guess one would need some adpater to connect a DB15 VGA to that header, but maybe the manufacturer provided that...

Now, the two ribbon cables for the serial ports, are connected to the front of the programmer, on the backplane PCB. On that PCB there is a 2x8 header... same as we found on the rear  I/O PCB hmm...; so logic would say that maybe, the video output from the CRT controller on the CPU board, is routed on the back plane to that 16 pin header... so all I have to do is scope that back plane,, easy enough, to see if I see some video action going on....

HMMMMMM !!!!!!


STAY TUNED !!!     

[Vince]

Oh boy I could not help I worke don the thing some more, it's addictive, a drug, pure and simple. I should have picked another hobby when I was a kid....

OK so, several things


Benoît Matthieu

He e-mailed me again, saying the shop in Paris is long gone, so no go !
However he did give me the name of that old chap that he knowe and who knew that programmer well. I can try finding him see if he is still with us today...

GPIB

OK so I think that 2x18 footprint on the rear / IO PCB is most likely not for GPIB. As you can see on the picture posted earlier of that PCB, all the traces from that connector are routed straight to the smaller 2x8 header footprint....
also, on the back plane, I have a matching 2x8 heaer but no 2x18 header anywhere.
So clearly, the back plane feeds only the 2x8 connector at the rear, and the 2x18 is just a duplicate that offers the same functionality, just using a different connector.

so this is bad news, it means that this 2x8 header mos tlikely is Not our much desired video output... because  then it would mean there was on the market some video cables using a 38 pin cable ?!    Not likely is it.... never heard of it anyway..
In order to confirm that, I scoped the matching 2x8 header on the back plane, and yeah, zero sign of activity on all the pins, nada.

UART

I looked the datasheet for all the big DIL chips on the CPU board, to see what's what.

- MC6845 = CRT controller as was already exposed.
- R6522 = VIA chip, in modern talk  : an I/O chip... contains 2x8 bit ports, 2x16 timers, and some shift registers.
- MB8416, 4 of them. They are 2KB static RAM chips. 2 are next to the CPU, and the other two next to the CRT controller, so probably a frame buffer or something.
- R6551 = the UART.

HOWEVER... what I don't get, is that the datasheet says this chip contains only ONE UART !!!  But we have TWO serial ports at the back ! 
so I guess there must be some multiplexing going on, and the CPU can use only one or the other, but not both at the same time... it must make a choice at some point... so that's quite an important / interesting finding here.

ALSO... I just got a crazy idea to figure out what is the configuration of the UART !!!  Really silly, but hey it's still an idea so it's gonna count for something eh....
Here goes : using a logic analyzer, I could spy the data bus and control signal at the UART chip, at power up, and see how the CPU is configuring the chip ! Spy every write the CPU makes, to see what registers are getting written, and with what value.
No, I don't have a logic analyser.... or DO I maybe !
Remember... the other day I got tha told 54645D HP scope, "Megazoom" ASIC, 2x100MHz .... MSO !! It has a built-in 16 way logic analyzer !!
16 lines is enough to look at th 8 data bus + some control signal ! I could watch the chip enable pin to trigger the scope, then watch the R/W pin and RS / register select  pins. Oh yeah, at last a real world interesting use for "new" HP scope !!!!!!!!!   

VIDEO

I kept the best for dessert of course !!

So if you have been following along, you now know there are two unused connectors / header on the backplane PCB. It's now clear that the 2x8 one goes to the rear IO PCB. But what about the other one ? 1x8 one ? This one no idea where it goes !

So I scoped it anxiously.... and yeah found LOT of activity on there !!! CRT controller is monochrome so the signal were expected to look "digitally". And that what I found !

From memory, out of  pins, 2 show no activity, probably ground, and one is at 12V solid, so probably the 12V supply. But the other 5 pins all have constant activity, trains of pulses,  5V " TTL" of varying content and timing, all perfectly repeatable / stable on screen.... looking GOOD ! Most likely some kind video signal !!!

Sorry can't show you anything, can't hold the probe on the pins with one hand and hold the camera with the other to capture the scope's  display....

It's 00H05 here, I am dying, so tired.

So tomorrow I will plug some old FDD ribbon calbe onto that header (male/female dupont wires would be best but I don't have that...) so that I can  hooke the scope probe ontot those signals and have both hands free to operate the scope, take notes and take pics to show you these signal and have your input.

Can't wait !!! 

Once we have the characteristics of all the signals, we can try to figure out what kind of video output standard that corresponds to, and from that we can start coming up with a plan to figure out how to display that signakl onto a monitor !!! 


See you tomorrow !!!

[syau]

Oh boy I could not help I worke don the thing some more, it's addictive, a drug, pure and simple. I should have picked another hobby when I was a kid....

OK so, several things


Benoît Matthieu

He e-mailed me again, saying the shop in Paris is long gone, so no go !
However he did give me the name of that old chap that he knowe and who knew that programmer well. I can try finding him see if he is still with us today...

GPIB

OK so I think that 2x18 footprint on the rear / IO PCB is most likely not for GPIB. As you can see on the picture posted earlier of that PCB, all the traces from that connector are routed straight to the smaller 2x8 header footprint....
also, on the back plane, I have a matching 2x8 heaer but no 2x18 header anywhere.
So clearly, the back plane feeds only the 2x8 connector at the rear, and the 2x18 is just a duplicate that offers the same functionality, just using a different connector.

so this is bad news, it means that this 2x8 header mos tlikely is Not our much desired video output... because  then it would mean there was on the market some video cables using a 38 pin cable ?!    Not likely is it.... never heard of it anyway..
In order to confirm that, I scoped the matching 2x8 header on the back plane, and yeah, zero sign of activity on all the pins, nada.

UART

I looked the datasheet for all the big DIL chips on the CPU board, to see what's what.

- MC6845 = CRT controller as was already exposed.
- R6522 = VIA chip, in modern talk  : an I/O chip... contains 2x8 bit ports, 2x16 timers, and some shift registers.
- MB8416, 4 of them. They are 2KB static RAM chips. 2 are next to the CPU, and the other two next to the CRT controller, so probably a frame buffer or something.
- R6551 = the UART.

HOWEVER... what I don't get, is that the datasheet says this chip contains only ONE UART !!!  But we have TWO serial ports at the back ! 
so I guess there must be some multiplexing going on, and the CPU can use only one or the other, but not both at the same time... it must make a choice at some point... so that's quite an important / interesting finding here.

ALSO... I just got a crazy idea to figure out what is the configuration of the UART !!!  Really silly, but hey it's still an idea so it's gonna count for something eh....
Here goes : using a logic analyzer, I could spy the data bus and control signal at the UART chip, at power up, and see how the CPU is configuring the chip ! Spy every write the CPU makes, to see what registers are getting written, and with what value.
No, I don't have a logic analyser.... or DO I maybe !
Remember... the other day I got tha told 54645D HP scope, "Megazoom" ASIC, 2x100MHz .... MSO !! It has a built-in 16 way logic analyzer !!
16 lines is enough to look at th 8 data bus + some control signal ! I could watch the chip enable pin to trigger the scope, then watch the R/W pin and RS / register select  pins. Oh yeah, at last a real world interesting use for "new" HP scope !!!!!!!!!   

VIDEO

I kept the best for dessert of course !!

So if you have been following along, you now know there are two unused connectors / header on the backplane PCB. It's now clear that the 2x8 one goes to the rear IO PCB. But what about the other one ? 1x8 one ? This one no idea where it goes !

So I scoped it anxiously.... and yeah found LOT of activity on there !!! CRT controller is monochrome so the signal were expected to look "digitally". And that what I found !

From memory, out of  pins, 2 show no activity, probably ground, and one is at 12V solid, so probably the 12V supply. But the other 5 pins all have constant activity, trains of pulses,  5V " TTL" of varying content and timing, all perfectly repeatable / stable on screen.... looking GOOD ! Most likely some kind video signal !!!

Sorry can't show you anything, can't hold the probe on the pins with one hand and hold the camera with the other to capture the scope's  display....

It's 00H05 here, I am dying, so tired.

So tomorrow I will plug some old FDD ribbon calbe onto that header (male/female dupont wires would be best but I don't have that...) so that I can  hooke the scope probe ontot those signals and have both hands free to operate the scope, take notes and take pics to show you these signal and have your input.

Can't wait !!! 

Once we have the characteristics of all the signals, we can try to figure out what kind of video output standard that corresponds to, and from that we can start coming up with a plan to figure out how to display that signakl onto a monitor !!! 


See you tomorrow !!!

The floppy disk controller board has a TMS9914 which is a GPIB controller so likely one of the header is for GPIB

There is also 75160 and 75162 GPIB transceiver

[Vince]

The floppy disk controller board has a TMS9914 which is a GPIB controller so likely one of the header is for GPIB
There is also 75160 and 75162 GPIB transceiver

Oh, thanks for that, interesting indeed ! 

I will look into that more closely then, to confirm that this x16 pin header is indeed GPIB, that would be cool !
... another cool thing to investigate, great 

[Vince]

OK, Matthieu Benoît  e-mailed me again with more info !  He rocks indeed !! 

1) He was mistaken, the programmer he saw in this Paris shop back in the day was not a model 3000U but 5000D... his bigger brother which he remember had a MONITOR along with it !
So yeah, it's now a certainty, there was indeed a higher end model, and all these unused connector headers and foot prints here and there now are confirmed to make perfect sense.


2)  He steered me to a thread on this French forum : https://forum.system-cfg.com/viewtopic.php?t=5743

Thread dates back to 2014 but still... it's pretty much "recent", as far as this programmer is concerned, isn't it...
Lots of little pieces of interesting info in there !
Look there is even one of the engineers who worked on designing this programmer back in the day ! 

Documentation :

- Someone almost got some : found a web site which was selling this programmer. He asks them about documentation, replied they had some but would charge 100 bucks to photocopy and ship it !  So the guy didn't buy it 

- Someone contacted Micropross directly about documentation, and they did reply !  ... .only to say that they "lost"/ scrapped everything they had when they moved building. 

Floppy Disk :

- They confirm what I envisaged : if I can't read this floppy from my computer, it's not because it's bad/dead (maybe it is, we just can't know at this point !), but it's because it does indeed use a special/proprietary format !

- Someone said that usual floppy copy/imaging S/W on PC's can't handle this format, no joy.

- Someone else said that on some old computers like the Atari 520ST, you had utilities that could make an actual copy of a floppy, i.e disregard any formatting or data structure, and just copy bit for bit from end to end, blindly, whatever is recorded on the disk ! I need that !

- The Engineer from Micropross said that A) the only special thing about their in house disk format is that they use sectors of 256 bytes instead of 512.  and B) They wrote a PC utility in house, so that they could work on their disks using a regular PC.

- People confirm that the machine needs the floppy in order to boot. It's mandatory. You can't just make it work using the serial port connected to a PC. You first need to power it up, boot from the floppy, then connect to the PC.


Operation :

- Someone has a working 5000D unit, with a good boot floppy. He says the unit can be operated STANDALONE. It boots from the floppy and then puts stuff on the screen with menus. One of the options is to make a copy of the boot floppy. However he tried that and when the programmer tries to format the disk, it fails...




Conclusion :

All this good info found only on the FIRST page of this that thread, and there is apparently 5 pages total ! So will read all that and probably learn some more cool stuff !!! 


NEXT : looks like two members there, have a working 5000D ! So they could help me I am sure... for example they could maybe help in telling me details about the serial links, how it is configured exactly, H/W control scheme used, what to do with the DIP switches, etc etc...
If the 5000D unit is stand alone,  I guess the extra serial port is used to connect a keyboard ?
That would explain why there the UART chip has only one port inside : either you use the thing standalone hence the K/B connected to one of the serial ports.. or you use it remotely from a desktop computer, hence you can use the keyboard from that computer hence you don't need to use a local K/B. That could make sense...

So we would need to reverse engineer that local K/B interface to see how it works, so I can design an interface to use a regular K/B...

Oh it's fantastic, so many new paths to explore now ! Looks like this is one hell of a vast fun reverse engineering project with many ramifications and child/sub projects ! This bloody programmer will keep me busy for YEARS to come !