1ST1
Veteran Member
I have the lotech xt-cf low profile.
- VCF South West - June 14 - 16, Davidson-Gundy Alumni Center at University of Texas at Dallas
- VCF West - Aug 2 - 3, Computer History Museum, Mountain View, CA
- VCF Midwest - Sept 7 - 8 2024, Schaumburg, IL
- VCF SoCal - Mid February 2025, Location TBD, Southern CA
- VCF East - April 2025, Infoage Museum, Wall NJ
-
Please review our updated Terms and Rules here
Paradise PVC2 questions on Olivetti M19
- Thread starter Zare
- Start date
1ST1 if you want to try something out before I get the XTIDE working again, I think I have an OK procedure.
86Box will produce a CHS-correct "card image". You just use a raw HDD image with CHS values that are a closest fit for your CF capacity. Prep/install everything using 86Box, then dump the image to CF.
I did a 32MB SanDisk using 614/6/17 yielding a 32,065,536 bytes image file.
I did not try autodetection on Pentium's CF-IDE but with manual setup entry the card boots normally.
I wonder whether XTIDE is able to autodetect this CHS. XUB can do manual values, as a last resort. But autodetection would be nice.
With DOS 3.31 we can do 512MB partitions. I also ordered a SanDisk 512MB, because it's important to know whether CHS formatting just works, or CHS formatting works only on cards that cannot be physically addressed via LBA. If CHS universally works we can format a CHS layouted 4x512MB MBR on DOS 3.31 (CHS limit is 8G) and have that on M19 boot. The Transcend 4G/8G cards that I've mentioned could be used.
P.S. I think there's merit to renaming this thread to "general M19 talk" or something alike
86Box will produce a CHS-correct "card image". You just use a raw HDD image with CHS values that are a closest fit for your CF capacity. Prep/install everything using 86Box, then dump the image to CF.
I did a 32MB SanDisk using 614/6/17 yielding a 32,065,536 bytes image file.
I did not try autodetection on Pentium's CF-IDE but with manual setup entry the card boots normally.
I wonder whether XTIDE is able to autodetect this CHS. XUB can do manual values, as a last resort. But autodetection would be nice.
With DOS 3.31 we can do 512MB partitions. I also ordered a SanDisk 512MB, because it's important to know whether CHS formatting just works, or CHS formatting works only on cards that cannot be physically addressed via LBA. If CHS universally works we can format a CHS layouted 4x512MB MBR on DOS 3.31 (CHS limit is 8G) and have that on M19 boot. The Transcend 4G/8G cards that I've mentioned could be used.
P.S. I think there's merit to renaming this thread to "general M19 talk" or something alike
Have ordered NEC V20 and NEC SV9000. Unfortunately the TVGA9000i-2 doesn't seem to work. They'll take some time to arrive meanwhile here are benchmarks on 4.77 MHz system(s).
- 512kb Rev 3.62 with Fujitsu 4.77MHz CPU, onboard graphics, disk controller in slot 1
- 640kb+parity Rev 3.58 with Siemens 4.77MHz CPU, TVGA8900C in slot 2, disk controller in slot 1
- 86box 512kb Rev 3.62 with Intel 4.77MHz CPU, integrated graphics profile, XT-IDE disk controller
Naturally, the 86box does not emulate this machine's graphics chips and connections but merely adds Plantronics modes to basic CGA emulator.
The real hardware is a bit faster in text ops. But with VGA the speed increase is significant. 285 vs 297 vs 536 chars/sec.
This is also felt in practice, the menus in TUI programs work much faster, the redraw rate feels almost instant.
At this point I really hope SV9000 works because I need to find a low profile card to fit the case, and if the performance is similar to 8900C, with V20 in the text rate could get really high for XT machine.
Had tweaked around 86Box and made config with some generic 8MHz XT profile and a Tseng card that I believe was in real world big and expensive. That config reaches about 750chars/sec.
- 512kb Rev 3.62 with Fujitsu 4.77MHz CPU, onboard graphics, disk controller in slot 1
- 640kb+parity Rev 3.58 with Siemens 4.77MHz CPU, TVGA8900C in slot 2, disk controller in slot 1
- 86box 512kb Rev 3.62 with Intel 4.77MHz CPU, integrated graphics profile, XT-IDE disk controller
Naturally, the 86box does not emulate this machine's graphics chips and connections but merely adds Plantronics modes to basic CGA emulator.
The real hardware is a bit faster in text ops. But with VGA the speed increase is significant. 285 vs 297 vs 536 chars/sec.
This is also felt in practice, the menus in TUI programs work much faster, the redraw rate feels almost instant.
At this point I really hope SV9000 works because I need to find a low profile card to fit the case, and if the performance is similar to 8900C, with V20 in the text rate could get really high for XT machine.
Had tweaked around 86Box and made config with some generic 8MHz XT profile and a Tseng card that I believe was in real world big and expensive. That config reaches about 750chars/sec.
Attachments
Last edited:
1ST1
Veteran Member
Interesting, I expected an MFM drive, but it is the same simple SCSI (!) controller like it is used in the ETV 260 (which is also based on M19).
That one's from "my" machine with a Rodime disk, the newly acquired one has a MFM+Seagate.
The controller and riser/adapter are quite interesting. Here's a bit from wikipedia
I think this is the adapter, or a clone of it. On third picture you can see how it looks from the top of the machine, the ISA slot is beneath.
NDC5127-50 is in there. It did come out in 1986 but it's a highly integrated pricey controller.
The machine sees the option ROM and the drive but reports interrupt error. I'm not able to benchmark this thing yet.
Btw curious to see construction only to hold a small fan. I have one mounted on the rear side of the HDD via some smaller construction. It has been turned off since forever without any negative impact. And it did contribute a lot to the noise coming out of the box.
Especially the other M19 with this cooling apparatus, no dB measurements but it's close to level of a standard AT build with crammed case with multiple vents.
I am trying to build a more powerful XT out of a M19 because it's small and produces 0 dB with a CF adapter replacing the disk. Curious to see some machines out there are built in very noisy fashion.
The controller and riser/adapter are quite interesting. Here's a bit from wikipedia
I think this is the adapter, or a clone of it. On third picture you can see how it looks from the top of the machine, the ISA slot is beneath.
NDC5127-50 is in there. It did come out in 1986 but it's a highly integrated pricey controller.
The machine sees the option ROM and the drive but reports interrupt error. I'm not able to benchmark this thing yet.
Btw curious to see construction only to hold a small fan. I have one mounted on the rear side of the HDD via some smaller construction. It has been turned off since forever without any negative impact. And it did contribute a lot to the noise coming out of the box.
Especially the other M19 with this cooling apparatus, no dB measurements but it's close to level of a standard AT build with crammed case with multiple vents.
I am trying to build a more powerful XT out of a M19 because it's small and produces 0 dB with a CF adapter replacing the disk. Curious to see some machines out there are built in very noisy fashion.
Attachments
Is there any additional info like the model name available?
Also if I may ask you historical question about evolution of these products. In the beginning there was M24. Then M28 was released as successor but in the same year M19 and ETV260 were released. Is M19 lineage in desktops or typewritters, e.g. was it planned to release two PCs as 2nd gen, both expensive and budget variant, and then create a typewritter system from the budget variant, or was M19 derived backwards from the typewriter design?
1ST1
Veteran Member
M19 was little sister of M24, a kind of "low cost" model (Olivetti never was low cost) and it was made for such environements like class rooms in schools etc. The ETV 260 came much later in 1987 and it was renewd in 1989 as the ETV 3000. The M19 is PC but the ETV 260 (and the ETV 500 which is just a relabeled M19 with a typewriter attached) wass marked as a wordprocessor system which als can run MS-DOS software as a benefit. The real successor of M24 was M240 which has the same power as M24SP. M28 and M280 were more like an upgrade to M24 and M240.
So M19 is indeed a second original x86 Olivetti desktop, really interesting that it had no real successor. You'd also expect a lot more online resources about it. All the others you have mentioned have them, M24 is fully documented as far as design and service manuals go too, disk images are available, etc.
I believe it was still sold brand new in then West Germany in 1989/90 for a fraction of original 1986 price.
Another fact I know it was running in year 2000 or 2001 on a lobby computer in a hotel in Copenhagen.
I believe it was still sold brand new in then West Germany in 1989/90 for a fraction of original 1986 price.
Another fact I know it was running in year 2000 or 2001 on a lobby computer in a hotel in Copenhagen.
1ST1
Veteran Member
Still in 2000/2001 in usage? Wow!
The real successor of M19 was M200 and PCS-86. And yes, there is almost no documentation about M19.
The real successor of M19 was M200 and PCS-86. And yes, there is almost no documentation about M19.
1ST1
Veteran Member
Most M19 were only monochrome CGA as well. M19 color monitor is VERY rare.
Indeed makes sense for mono-only machine to be a successor then.
In that manual a lot depictions were about color monitor w/expansion box. Never seen one in real life. From the description it holds "support for color monitor and a full size ISA card", and judging by the size of it, can't fit more than a PSU and 1 slot card.
I think M19 with composite out is more rare than this.
One more unexplained thing is the sound port hole in the rear. I don't have much experience with computers of this vintage but I haven't seen one with a PC speaker potentiometer that M19 has. Who knows, maybe they wanted to terminate beeper at rear so you can plug speakers in.
In that manual a lot depictions were about color monitor w/expansion box. Never seen one in real life. From the description it holds "support for color monitor and a full size ISA card", and judging by the size of it, can't fit more than a PSU and 1 slot card.
I think M19 with composite out is more rare than this.
One more unexplained thing is the sound port hole in the rear. I don't have much experience with computers of this vintage but I haven't seen one with a PC speaker potentiometer that M19 has. Who knows, maybe they wanted to terminate beeper at rear so you can plug speakers in.
1ST1
Veteran Member
There was no sound option for M19. Soundcards for PC only appeared later and all of the early samples were full height, no chance to install it in the M19. The expansion box has only one ISA slot, maybe that one is too valueable for just a soundcard. M19, ETV 260 and ETV 500 also did not have composite video out option. Only ETV 3000 has composite out (at uncommon videofrequency) for the monochrome monitor, for that the mainboard has been modified.
The thing with color monitor and harddisk is a bit strange as normally the monitor has to supply the computer with power, there is no power supply inside the computer. The monochrome monitor normally has not enough power a harddisk, that is what known literature about M19 says but anyhow such examples of M19 and ETV 500 (really rare!) could be found from time to time. Instead the literature says that for harddisk the expansion box is required as it contains a power supply and the expansion box is the only way to have M19 in color as the color monitor does not supply the computer, only itself.
In total, M19 is a beautyful PC, but a strange one as well.
The thing with color monitor and harddisk is a bit strange as normally the monitor has to supply the computer with power, there is no power supply inside the computer. The monochrome monitor normally has not enough power a harddisk, that is what known literature about M19 says but anyhow such examples of M19 and ETV 500 (really rare!) could be found from time to time. Instead the literature says that for harddisk the expansion box is required as it contains a power supply and the expansion box is the only way to have M19 in color as the color monitor does not supply the computer, only itself.
In total, M19 is a beautyful PC, but a strange one as well.
Indeed strange, or is it just lack of primary sources that can explain a thing or two in detail.
Attached the brochure. Next to the manuals this is the only original publication I know of.
In the section expansions and options, on 4th page right, you can see listed "TV adapter, for transfer of video signal via SCART". It is impossible to do this via ISA expansion because its DMA doesn't allow device to device transfers, any sort of video option must be present in the circuitry of the card. Furthermore, Paradise PVC4 is documented to support composite out. And on the mainboard, there is an empty chip socket right on the edge where video connector is. I've done only quick visual circuitry "discovery" so no proof but this socket lies "before" IC latches and buffers from the video port point of view so it can't be involved in video processing itself.
Also didn't know about the power and disk bit. I would like to know more if possible about that literature, because in brochure there is nothing that would imply this.
I've ran mine with that SCSI controller, hard drive, FDD, XTIDE and Adlib from mono monitor without any problem.
Color monitor doesn't have enough space for integrated PSU, that space needs to be at the bottom rear because the PSU is heavy which shifts the monitor center mass. The mono screen is designed for it, slanted to the back and standless. If Olivetti made that monitor, the expansion box would be totally redundant apart for the full size ISA slot.
And since power is just standard AT over 9pin molex outside supply can be easily connected even back in the day.
Problem and strangeness is a video signal. If it were of normal frequency M19 is a CGA+ machine as-is and would be viewed as such, but it was tied to a specific color monitor which was tied to a half-redundant expansion box making everything nearly unobtainable. That's why it ended up as just a monochromatic machine, which is sad, because just take a look how much board estate the graphics part uses, about 40%.
For the sound stuff, I agree with you but just saying the facts...there's a hole back there, and a note icon engraved on plastic above it. This only means one thing. So I presume what it could be, agreeing with you that soundchip is highly unlikely, and the next guess is PC speaker connection to a jack in that place. So you can connect external speakers. After all, M19 has physical volume control, again unlike any non-soundcard machine I've ever seen.
Attached the brochure. Next to the manuals this is the only original publication I know of.
In the section expansions and options, on 4th page right, you can see listed "TV adapter, for transfer of video signal via SCART". It is impossible to do this via ISA expansion because its DMA doesn't allow device to device transfers, any sort of video option must be present in the circuitry of the card. Furthermore, Paradise PVC4 is documented to support composite out. And on the mainboard, there is an empty chip socket right on the edge where video connector is. I've done only quick visual circuitry "discovery" so no proof but this socket lies "before" IC latches and buffers from the video port point of view so it can't be involved in video processing itself.
Also didn't know about the power and disk bit. I would like to know more if possible about that literature, because in brochure there is nothing that would imply this.
I've ran mine with that SCSI controller, hard drive, FDD, XTIDE and Adlib from mono monitor without any problem.
Color monitor doesn't have enough space for integrated PSU, that space needs to be at the bottom rear because the PSU is heavy which shifts the monitor center mass. The mono screen is designed for it, slanted to the back and standless. If Olivetti made that monitor, the expansion box would be totally redundant apart for the full size ISA slot.
And since power is just standard AT over 9pin molex outside supply can be easily connected even back in the day.
Problem and strangeness is a video signal. If it were of normal frequency M19 is a CGA+ machine as-is and would be viewed as such, but it was tied to a specific color monitor which was tied to a half-redundant expansion box making everything nearly unobtainable. That's why it ended up as just a monochromatic machine, which is sad, because just take a look how much board estate the graphics part uses, about 40%.
For the sound stuff, I agree with you but just saying the facts...there's a hole back there, and a note icon engraved on plastic above it. This only means one thing. So I presume what it could be, agreeing with you that soundchip is highly unlikely, and the next guess is PC speaker connection to a jack in that place. So you can connect external speakers. After all, M19 has physical volume control, again unlike any non-soundcard machine I've ever seen.
TVGA9000B just doesn't work, shame since NEC version is smallest it can be.
There is another adapter Everex that's of similar size. But I do not intend to buy and test random cards.
At this point developing a case mod becomes a moderately involved thing to do as opposed to being very easy. Usage of this machine would be a testbed for a longer running 8 bit diy hobby project, naturally didn't want to have a computer in parts. It's not end in itself so I won't waste time on it. Going to fit 8900 on a panel where disks used to be an use a cable to connect it to bus and fix a VGA cable through the rear. picoPSU will be in there but adapter is going out and power button will just float somewhere. Crude, but enough to close the computer case with everything safely connected.
The light blue wiring that's different in between produced machines is actually around latches and converges to a bus mux/demux. So yeah, this is a part of "configuration" and in between different BIOS version and a very picky choice of VGA cards, I say this "solution" may not be universal.
Additionally the M19 is equipped with 4xUPD41416-15 which indeed counts for 32kB of RAM. The PVC2/4 cards have 64kB so they can support Hercules. This is a primary reason why not all Paradise configuration ports are open. Again, given enough time and dedication one could make a Hercules mod by replacing the chips and rearranging the blue signal cables but is as far as I'm concerned a herculean task. The only bit of info left I want to find about it is what exactly is controlling the horizontal sync rate.
There is another adapter Everex that's of similar size. But I do not intend to buy and test random cards.
At this point developing a case mod becomes a moderately involved thing to do as opposed to being very easy. Usage of this machine would be a testbed for a longer running 8 bit diy hobby project, naturally didn't want to have a computer in parts. It's not end in itself so I won't waste time on it. Going to fit 8900 on a panel where disks used to be an use a cable to connect it to bus and fix a VGA cable through the rear. picoPSU will be in there but adapter is going out and power button will just float somewhere. Crude, but enough to close the computer case with everything safely connected.
The light blue wiring that's different in between produced machines is actually around latches and converges to a bus mux/demux. So yeah, this is a part of "configuration" and in between different BIOS version and a very picky choice of VGA cards, I say this "solution" may not be universal.
Additionally the M19 is equipped with 4xUPD41416-15 which indeed counts for 32kB of RAM. The PVC2/4 cards have 64kB so they can support Hercules. This is a primary reason why not all Paradise configuration ports are open. Again, given enough time and dedication one could make a Hercules mod by replacing the chips and rearranging the blue signal cables but is as far as I'm concerned a herculean task. The only bit of info left I want to find about it is what exactly is controlling the horizontal sync rate.
With NEC V20 the 8MHz settings do not work. The KB blinks and that's all. 4.77MHz settings do work.
There are a couple of undocumented jumpers around the board, some of them paired with what seem to be oscillators. I'll try something in that area as a last resort.
Btw, here are pictures of the sound 'feature'. It's absolutely no doubt something was planned. "JSOUND" and "PSOUND" are connected via PCB traces underneath, the former is a soldier point for the outside jack that breaks through that hole on first pic, the latter is the point of connection to something inside the case. What that something is unknown.
There are a couple of undocumented jumpers around the board, some of them paired with what seem to be oscillators. I'll try something in that area as a last resort.
Btw, here are pictures of the sound 'feature'. It's absolutely no doubt something was planned. "JSOUND" and "PSOUND" are connected via PCB traces underneath, the former is a soldier point for the outside jack that breaks through that hole on first pic, the latter is the point of connection to something inside the case. What that something is unknown.
Attachments
1ST1
Veteran Member
Is it an 8 MHz V20 version? Does an original 8 Mhz version of i8080 start?