Wolfensten 3D requires VGA. While you can add a VGA card to the 1000 SL, it wouldn't be a good benchmark at all.

root42 wrote:

Who needs an Amiga when you can have a 16MHz V30! 😁

Pre-emtive multi-tasking?

This was a joke … the Amiga will of course have lots of advantages due to its custom chip set.

aazard wrote:

i 2nd that, im really interested in how this can go in terms of "usability"? wolf3d for 8086 and maybe a lo-tech 2mb ems card? i […]
Show full quote

Cyberdyne wrote:

Run Wolfensten 3D with it 😁

i 2nd that, im really interested in how this can go in terms of "usability"?
wolf3d for 8086 and maybe a lo-tech 2mb ems card? i think that wolf3d is a good usability test, darklands a torture test?

edit: when i say "usability", i mean beyond reasonable expectation for a typical or even standard upgraded tandy 1000 sl

Humm, a 2 seperate bus's.. 1 for cpu/memory and 1 for isa/onboard stuff? is that correct... that might make the pcs86 OC'ing an issue, unless its the same, like the Juko Nest N3 XT motherboard, and a few other juko motherboards from this era, do.

Has anyone ever done a DIP40 to "nec v33a" adapter?? its 8088/8086/80186/nec V-series instructions, NEC V33 & V33a series offers a method to expanding the memory address space to 16M bytes, It has two additional instructions BRKXA and RETXA to support extended addressing mode & is clock to clock equal in performance to a 80286... memory addressing might be useless but the additional speed at same clock is awesome
Or clock doubling circuit "on top of turbo mode"? so at could run 3 speeds like normal/turbo/turbo x2?
Like 4.77 / 9.54 /19.08 or 4.15 / 8.3 / 16.6 or 4.5 / 9 / 18 mhz
ALSO ANYONE GET AN 8087-1 10mhz to run @ 16.6mhz+ before, like even with active cooling
edit, spelling/spaces

I think 8087-1 is already overclocked. I think I once tried to track down a military grade version of it, but I don't think the -1 exists as a military grade part. Overclocking a regular one is futile. Those things are already stretched to their limits.

Has it been proven that Juko Nest N3 has a decoupled bus? The manual seems to suggest something along those lines, but the English in the manual is kind of poorly written, so unless somebody tests with a scope who knows.

Anonymous Coward wrote:

I think 8087-1 is already overclocked. I think I once tried to track down a military grade version of it, but I don't think the -1 exists as a military grade part. Overclocking a regular one is futile. Those things are already stretched to their limits.

Has it been proven that Juko Nest N3 has a decoupled bus? The manual seems to suggest something along those lines, but the English in the manual is kind of poorly written, so unless somebody tests with a scope who knows.

it indeed does say so, read this pdf, introduction section:

"able to run at much higher cpu clock speed, which the I/O bus speed is running at a safe 4.77mhz so that most expansion cards will function properly"
READ: change "which" to "with" and remove "is", these would be common translation errors, BUT thats an assumption, TO BE CLEAR

poorly written/translated, but fairly clear. a few olivetti units, juko units, and seemingly some tandy's used this technique

I want to see it on a scope. The numbers I'm getting from my SCSI adapter seem to imply the bus is running at full CPU speed. It's in the 700-800kbp/esc range, more than double a standard XT. Maybe there is a jumper to configure how the bus speed is derived?

Has anyone tried a 16MHz V30 in a NEST board?

aazard wrote:

Has anyone ever done a DIP40 to "nec v33a" adapter?? i

I'd be awesome if it was that easy, Likely some of the signals need conversion special logic in order to be usable in an XT. 🙁
Speaking of XTs, there also was an IBM XT 286. It had better RAM timings like the AT, even.
https://en.wikipedia.org/wiki/IBM_Personal_Co … r_XT#IBM_XT_286

For pure 808x XTs, there had been indeed 286 and 386 upgrades way back in the 1980s. however.
Often they included an ISA card and a ribbon cable with a 40pin adapter for "simulating" the presence of a 8088.
That way, real operating systems like Xenix, PC-MOS/386 and Concurrent DOS XM could run on "cheap" (non-)IBM PCs of the day.

IMHO, I wonder what use an expensive/complicated upgrade for a single XT machine has today, however.
I mean, the NEC V20/30 upgades made totally sense. They added basic 186/286 and 8080 binary compatbility and fixed certain shortcomings of the 808x,
that were actually noticable (DIR command was quicker, HDD interleaving could be optimized, games did less stutter etc).
Even today, the NECs are almost mandatory whenver modern DOS software is required to run on these dinosaurs.

Don't get me wrong, if there was a trouble-free 286 upgrade (daughter card etc) that fits in a 808x socket, it would be awesome I think. 😎
Especially if it had a bit of extra RAM included, say 64KiB to 2MiB, so programs could roll out their data a bit..
That way, XMS, VCPI extenders, DOS=HIGH and Windows 3.1 would work (Windows 3.0 in Real-Mode had little software support).

However, what purpose would it have to install an extremely powerful, but likely troublesome CPU such as V33, in a system full of bottlenecks. 😕
Going by logic, wouldn't it be more sensible to fix the bottlenecks of the PC XT platform first, like the internal 8-Bit bus ?
Memory access on a 8088 is much slower than on an 8086. So why not just combine efforts and make a proper,
"fixed" XT machine design with 16-Bit I/O, NEC V30/8086 and a true, modern XT BIOS - instead of creating a lowend 286-wannabe each time ?
The replica , say, from Sergey, is awesome (honestly!), but also adds AT specific stuff (in BIOS) that might lower XT compatibility softwarewise (few apps could get confused).
Don't get me wrong this is no complain, I do just think about it. I guess you can turn an old plane (piper cup or cessna)
into a car by chopping off its wings just fine and add a steering wheel and mechanism. But why, except for curiosity ? 😉
In short, if the identity of something valuable has to be sacrified in the name of progress (higher, faster, farer), then what's left of it ?

I've always disliked" "system-on-a-card" upgrades... its just not the same machine its been "replaced", direct cpu upgrades, socket adapters, add-on cards like ems... that im fine with, even using parts made/invented yesterday.... as it preserves that system as "upgraded" not replaced.

WHY?: The reason i'm after an XT with the nec v30HL @ 16mhz, 1024kb base memory, with 32mb of EMS memory (onboard like the pcs86 would be my goal) is because i know its possible 🤣 (ps, i confirmed at my co-workers yesterday that a normal v30 will see 32mb on a ramquest EMS board, but he no longer has the original 8086 to see if it would see it also, but i dont see why it wouldnt)

i want to load pc-dos, a real-mode "modded" copy of win-fwg 3.11 and opera 3.62 and post to the net from an XT

PS found this for all the XT/AT users.. i need DR-DOS versions tests & Free-DOS versions tests (did i miss a "useful" dos?) note: i did not make this, here is what the original poster of this info said:

"I just went through all my boot disks, and ran CHKDSK on each one to display the amount of free RAM (out of 640K) on each one, using a totally clean boot (no CONFIG.SYS or AUTOEXEC.BAT):

PC DOS 2.00 ... 630,672 bytes
PC DOS 2.10 ... 630,672 *
PC DOS 3.10 ... 616,432
MS-DOS 3.10 ... 616,432
PC DOS 3.21 ... 609,392
PC DOS 3.30 ... 600,528 ***
MS-DOS 3.30 ... 600,368
IBM DOS 5.00 ... 593,328
MS-DOS 5.00 ... 593,328
MS-DOS 6.00 ... 592,256
IBM DOS 6.10 ... 593,056
MS-DOS 6.22 ... 592,256
PC DOS 6.30 ... 593,024 **
PC DOS 7.00 Revision 0 ... 593,840 ***
PC DOS 2000 (7.00 Revision 1) ... 593,760 *** (Y2K, euro symbol(?), etc fixed, but no FAT32, so still runs on 8086 class << thats my note, aazard

I also have an MS-DOS 4.01 boot disk, but not a copy of CHKDSK which will work with it, so I can't give it an accurate free RAM amount, but judging by what other utilities report, it's somewhere in the ballpark of 590,000 bytes. So IBM claimed that PC DOS 7 offered the most free RAM of any version of DOS since 3.3, and that does appear to be true!"

If "hot-rodding" an XT. I have 8-bit isa question... and you guys seem the defacto XT "testers"

is the "1024kb DRAM Cirrus Logic CL-GD5420", the card that wont work with a 8086 but will with a v30, faster and/or more compatible than the "1024kn DRAM ATi VGAWonder XL"?

They seem to be the 2 best video cards that will run in a 8-bit isa bus, according to various posters.... anyone here used both for comparison?
all i can seem to find on ati vgawonder xl is that its max res is: 800x600 15-bit colour depth @60hz
vs the Cirrus Logic CL-GD5420 @ 1024x768 15-bit colour depth @60Hz (this i have the manual for, so i know alot about it)

I dont mean to seem crazy, 🤣. I'm just cherry picking the best known parts i can for my projects. Socket 3 was easy, DIP-40 XT''s are fun and challenging

aazard wrote:

PS found this for all the XT/AT users.. i need DR-DOS versions tests & Free-DOS versions tests (did i miss a "useful" dos?) note […]
Show full quote

PS found this for all the XT/AT users.. i need DR-DOS versions tests & Free-DOS versions tests (did i miss a "useful" dos?) note: i did not make this, here is what the original poster of this info said:

"I just went through all my boot disks, and ran CHKDSK on each one to display the amount of free RAM (out of 640K) on each one, using a totally clean boot (no CONFIG.SYS or AUTOEXEC.BAT):

PC DOS 2.00 ... 630,672 bytes
PC DOS 2.10 ... 630,672 *
PC DOS 3.10 ... 616,432
MS-DOS 3.10 ... 616,432
PC DOS 3.21 ... 609,392
PC DOS 3.30 ... 600,528 ***
MS-DOS 3.30 ... 600,368
IBM DOS 5.00 ... 593,328
MS-DOS 5.00 ... 593,328
MS-DOS 6.00 ... 592,256
IBM DOS 6.10 ... 593,056
MS-DOS 6.22 ... 592,256
PC DOS 6.30 ... 593,024 **
PC DOS 7.00 Revision 0 ... 593,840 ***
PC DOS 2000 (7.00 Revision 1) ... 593,760 *** (Y2K, euro symbol(?), etc fixed, but no FAT32, so still runs on 8086 class << thats my note, aazard

That was originally posted at vcfed.org by vwestlifein a discussion about Dos versions. Here's the thread http://www.vcfed.org/forum/showthread.php?296 … f-x86-DOS/page3 I posted the same thing on to vogons a few years ago. It's interesting alright. At the time I tested out the current FreeDos. It was well down on the list compared to classic Dos vasriants by at least 50k without any tweeking. DR Dos 7 iirc was around the 586k mark iirc. PTS Dos 2000 was similar to MS Dos 6 and it was back-would compatible with XT class hardware as well.

Of course when you start using memory managers things even out a bit.

aazard wrote:

i want to load pc-dos, a real-mode "modded" copy of win-fwg 3.11 and opera 3.62 and post to the net from an XT

Even if you could mod a copy of WFW311, you're still not going to have enough memory to do anything. First of all, some of the programs you want to run use 32-bit protected mode, and secondly even if you could get those to run in real mode, they're not designed to make use of EMS. Even Windows 3.0 has almost no software for real mode with EMS support. EMS 3.2 probably wouldn't cut it either. It would have to be 4.x with large page frame mode. And even then, if you had patched software to go with the patched operating system, Opera would still be too slow to do anything useful. Just get DR-DOS with GEOS and be happy.

I had this idea for a while now, that it might be possible to get 16-bit memory transfers from Orchid Ramquest 8/16 on my Juko NEST motherboard by tapping the data lines on the edge connector, and wiring them directly up to the pins on the CPU. Any reason why that wouldn't work?

Anonymous Coward wrote:

I had this idea for a while now, that it might be possible to get 16-bit memory transfers from Orchid Ramquest 8/16 on my Juko NEST motherboard by tapping the data lines on the edge connector, and wiring them directly up to the pins on the CPU. Any reason why that wouldn't work?

i'll be honest, i know hardware/software very well from working years as a sys admin with too much time & freedom plus an interest in the "classic era". But knowing much about how this is truly working "under the hood", i have tid-bits of info, much of it not done by me personally (aka "o yea, young tech, we had a vax-xxx and there was a slide ruler in a glass case with a hammer as a joke")

I think there are a few late "turbo XT" style units that had 1024kb+ to 8mb "ems x.x" built onto the board, and i know of mentions of others, or 1 i've seen...like the pcs86 i always mention... that had EMS 4.0 support built in. Most had sockets on the board itself past 640kb-1024kb norm, some used proprietary modules (i think ibm xt-at or pc-convertible) a few had simm slots.

BUT, that level of Frankenstein-ing is past my limit (and honestly probabley my solo abilities), i might go as far as, like on a biostar mb-8433uud-a, connecting a "missing" path (as in the mentioned case of 8433 "missing" the trace to allow 1024kb cache) or to change an oscillator/switch for a oc'ed multiplier or frequency.

I think this might be the best place to ask the question though 😀

aazard wrote on 2019-09-26, 22:29:

Humm, a 2 seperate bus's.. 1 for cpu/memory and 1 for isa/onboard stuff? is that correct... that might make the pcs86 OC'ing an […]
Show full quote

aazard wrote:

i 2nd that, im really interested in how this can go in terms of "usability"? wolf3d for 8086 and maybe a lo-tech 2mb ems card? i […]
Show full quote

Cyberdyne wrote:

Run Wolfensten 3D with it 😁

i 2nd that, im really interested in how this can go in terms of "usability"?
wolf3d for 8086 and maybe a lo-tech 2mb ems card? i think that wolf3d is a good usability test, darklands a torture test?

edit: when i say "usability", i mean beyond reasonable expectation for a typical or even standard upgraded tandy 1000 sl

Humm, a 2 seperate bus's.. 1 for cpu/memory and 1 for isa/onboard stuff? is that correct... that might make the pcs86 OC'ing an issue, unless its the same, like the Juko Nest N3 XT motherboard, and a few other juko motherboards from this era, do.

Has anyone ever done a DIP40 to "nec v33a" adapter?? its 8088/8086/80186/nec V-series instructions, NEC V33 & V33a series offers a method to expanding the memory address space to 16M bytes, It has two additional instructions BRKXA and RETXA to support extended addressing mode & is clock to clock equal in performance to a 80286... memory addressing might be useless but the additional speed at same clock is awesome
Or clock doubling circuit "on top of turbo mode"? so at could run 3 speeds like normal/turbo/turbo x2?
Like 4.77 / 9.54 /19.08 or 4.15 / 8.3 / 16.6 or 4.5 / 9 / 18 mhz
ALSO ANYONE GET AN 8087-1 10mhz to run @ 16.6mhz+ before, like even with active cooling
edit, spelling/spaces

Sorry for the necro, but I just wanted to say if you can get an NEC D9008D co-processor, it was specifically made to work with the V30 at an 8087 and there is real world proof that it is perfectly happy working off of a 16 MHz clock. They are hard to find, but can be had even these days.

Paralel wrote on 2024-02-16, 06:12:

Sorry for the necro, but I just wanted to say if you can get an NEC D9008D co-processor, it was specifically made to work with the V30 at an 8087 and there is real world proof that it is perfectly happy working off of a 16 MHz clock. They are hard to find, but can be had even these days.

Fascinating...
Can it work with the V20 as well?

There's plenty of V20 @ 12 MHz boards, seemingly without proper FPU option.
There was only Microway, selling 10 MHz 8087 chips tested to run at 12 MHz.

i was unable to make the D9008D on my hotrod Tandy. If that FPU is present the system does not start POST.

did you manage to make it work ?

Grzyb wrote on 2024-02-16, 10:12:

Fascinating... Can it work with the V20 as well? […]
Show full quote

Paralel wrote on 2024-02-16, 06:12:

Sorry for the necro, but I just wanted to say if you can get an NEC D9008D co-processor, it was specifically made to work with the V30 at an 8087 and there is real world proof that it is perfectly happy working off of a 16 MHz clock. They are hard to find, but can be had even these days.

Fascinating...
Can it work with the V20 as well?

There's plenty of V20 @ 12 MHz boards, seemingly without proper FPU option.
There was only Microway, selling 10 MHz 8087 chips tested to run at 12 MHz.

Theoretically, it should work, since it was intended for the V20 & V30. However, I don't believe this combination has ever been tested, as far as I know.

pshipkov wrote on 2024-02-16, 10:20:

i was unable to make the D9008D on my hotrod Tandy. If that FPU is present the system does not start POST.

did you manage to make it work ?

Which type of V30 did you try it with, the OG V30 8 or 10 MHz or the V30HL 8 or 16 MHz?

Were you able to get any further information using a POST card?

A friend I know has it working as a coprocessor in a V30 @ 8 Mhz in an upgraded 1000 SL/2. It is reported as an 8087, and is seen as such in diagnostic software.

I don't know if he's ever tested it's actual functionality, as far as actual calculations and checking if the results are correct. From what I have read, it is possible to install them, POST just fine, have diagnostic software recognize it, but in some instances if people actually tested it, the results of calculations it was given were completely wrong. Sometimes it had something to do with other hardware causing trouble, but I know of a few instances where the exact issue remained unknown.

here is my Tandy story.
16MHz rated V30. System does not light-up at all with that FPU. There is of course a chance the FPU is dead and somehow drags the entire system with it ...

Good notes at the end of your post.
Do you have something on the internet in written form that we can latch on ?
Thanks.

pshipkov wrote on 2024-02-16, 21:34:

here is my Tandy story. 16MHz rated V30. System does not light-up at all with that FPU. There is of course a chance the FPU is d […]
Show full quote

here is my Tandy story.
16MHz rated V30. System does not light-up at all with that FPU. There is of course a chance the FPU is dead and somehow drags the entire system with it ...

Good notes at the end of your post.
Do you have something on the internet in written form that we can latch on ?
Thanks.

This is basically most of the information on the internet about this particular chip:

https://www.cpushack.com/2021/09/01/necs-forgotten-fpus/

https://www.cpu-world.com/forum/viewtopic.php?p=304639

It is fairly unknown. As far any Tandy 1000 specific information, that is just from my friends own experimentation.