IIsiColorPivotII_PDS_Card_HackProject™

[zuiko21]

Still quite busy... but here's a bit of news about the subject -- not as good as I'd like, though

I made a crude, ugly connector for this card, but so far I connected the several SENSE and SYNC lines plus the associated grounds only. No coaxial cables were soldered thus no picture would be possible; but at least I could check if my versatile Acer monitor syncs up properly with the card's (partial) output. Some evidence of this outrage:

Now, the testing: I used a IIsi with 33 MB and System 7.5.5. An Apple 12" RGB was connected to the internal video output (to be able to see what was going on) while the Acer was hooked to the ColorPivot thru a typical, known good VGA adapter (and my cable).

Upon power-up, during the RAM test (33 MB do take a while on the IIsi!) both outputs were muted, as expected... but as soon as the internal video arose, the Acer detected some signal too! No picture of course, but from its menu there was a 68 kHz / 75 Hz signal, like that of a Portrait display -- confirming my previous probing with the oscilloscope. However, this was always the same any way I went with the VGA adapter's switches, plus it never showed up on Monitors control panel although Gauge series's Slot Info was detecting it properly every time (slot $A). Warm restarts make the sync signals to the Acer disappear for a moment and then back again, as expected.

On the other hand, I recently got (from macmetex too) a Radius Pivot NuBus Interface (632-0070). Sharing the same DRAM (8x M514252A-80, just different package), the same big custom chips (297-0308 & 297-0309), the RAMDAC (Bt478KPJ50), the famous 14.3181 MHz crystal can and regulators (LM385 & LM393) plus most discrete components, I don't think it's crazy to assume both cards would have similar capabilities... even the ROM (DIP vs. PLCC) is labeled quite much the same: (C) 1992 Radius Inc. COLOR PIVOT V2.6 256K U5-0039-04-A (the PLCC version reads CLR/PVT V 2.6 0048-2A, there's much less space on it)

This NuBus card has the usual DA-15 connector, so I could easily check that it supports several display modes:

• 512x384, 60 Hz (12" RGB)
• 640x480, 66 Hz (13", not VGA)
• 832x624, 75 Hz (16")
• 640x870, 75 Hz (Portrait)

But whenever I had any other mode set on the adapter's switches, it kept displaying the usual grey screen with the Radius logo at the bottom at the "default" portrait mode: 640x870, 68 kHz, 75 Hz; looking very stretched on the wide Acer monitor -- not the safest thing to do if connected to a fixed-frequency incompatible monitor Of course, when on a compatible mode the Radius logo appeared for a moment, then the Mac cleared the screen and started booting, as expected.

Could it be that somehow the Pivot IIsi is not detecting a compatible monitor and thus disabling itself? Maybe when I get the coaxials soldered it could show the Radius logo, as the 'scope-detected activity on the RGB lines suggests. BTW, the sense lines' continuity on my crude cable checks out OK...

I hope to be able to complete the task in a few days, and come back with more info. I can't conclude anything so far

[Trash80toHP_Mini]

Take a screen shot to see what's going on in the RCPII/IIsi's VRAM.

[zuiko21]

According to my experience, screen shots (Cmd-Shift-3) will only capture whatever screens show up on the Monitors control panel. I have however re-checked this particular case, and it captures only the 512x384 image of the 12" RGB I had at the IIsi built-in video port regardless of the sense code applied to the Pivot's connector -- either from the VGA adapter's switches, or from real Apple monitors, in particular the Portrait Display.

However, a funny effect was observed... I tried to do the same with no monitor at the built-in port, pressing Cmd-Shift-3 blindly and then Keyboard power-enter (for a proper shutdown). In order to avoid the long RAM test (33 MB), I was going to remove the 12" connector on the fly (yeah, I've done that many times without ill effects) and then power-R to reboot cleanly...

...but as soon as I disconnected (while running!) the 12" from the built-in port, the Pivot's output was muted -- the Acer no longer received sync signals! Then, I issued (blindly) the restart command and... got Chimes of Death right after the normal bong If cold-started this way, the IIsi bongs, does the RAM test and then the Death Chime Plugging back a monitor into the built-in port makes everything go back to the usual way...

This is quite strange because, if the Radius Pivot card is not installed, the IIsi can certainly run completely headless: I've used it as a LocalTalk server that way, and a Cmd-Shift-3 does make a fancy 32x32 image with the "Menu bar" occupied by the small Finder icon (of the Application menu) and garbage in the lower half, with the arrow somewhere in the middle

EDIT: just tried again, but putting a cruder jumper between one of the card's video outputs (pin 2, 5 or 9) to one of the monitor's DA15 video input lines -- for instance, pin 5 to pin 5 (green signal). The Acer display shows, clearly, a streched (it's portrait mode on a wide display!) screen with the usual Radius logo at the bottom, all green of course. Separate SYNC (13" won't do)

Unlike its NuBus sibling, this particular card rejects any sense code I could apply, considering it an unsupported monitor. Same when using the APD (only pin 9 gets the input, this time with the right aspect ratio)

Booting from 7.1 (instead of 7.5.5) changes nothing.

[Trash80toHP_Mini]

It sounds like your Mac Video Adapter might not be up to the task to me. IIfx had problems getting anything but 640x480 out of the Apple Display Card 24AC I sent him until he got a good adapter, now it's running fine.

I did all my RCPII/IIsi tests with the Enhance "Liberty" Adapter, the best of the best in my estimation. Which one are you using?

As for what shows up in the Monitors Control Panel:


Full House 4: PixelTown Config.

The next post shows what happens when they don't play nicely together in Full House 5

I couldn't actually follow the details of what you were reporting, I'm really tired ATM. What monitors are you using?

[Trash80toHP_Mini]

I forgot to mention the Video Adapter Docs Scan Dump.
Let me know which one you're having trouble with and, if it's in the pic or on my list, maybe I can help you out with some documentation.

In your PM, you said you were trying to use your RCPII/IIsi with an LCD. Check the previous page for the output range of the card and check to make sure your LCD is a MultiSync.

[uniserver]

so i can buy this video card, and use it as long as i make my own cable?

sounds like a pain in the butt

http://www.ebay.com/itm/Radius-Color-Pivot-IIsi-Video-Interface-Card-Apple-Mac-/150933000786?pt=US_Vintage_Computing_Parts_Accessories&hash=item23244ed252

and again does this card get some kind of spectacular resolutions?

[tt]

Not sure about resolution, but I think it is 8-bit color, same as the SE/30 version. There is a Radius mirror somewhere that has specs...

[Trash80toHP_Mini]

I don't think I'd call it spectacular, but 16" @ 8bit Color is bigger, better and doesn't screw around with Bank A like the Vampire Video setup on the MoBo. That bit of nastiness only gives you 12" RGB or 640 x 480 in 8bit and Portrait at I'm not sure how many bits, but I'll bet it's not 8. If no monitor is sensed on the MoBo Video Connector at startup, none of the Meg of RAM in Bank A gets buffered for use as makeshift VRAM.

p.s. check out the first page, it lists all the resolutions tested.

[volvo242gt]

JT:

640x870 on a IIsi or IIci will give 4-bit. I remembered that from my freshman year in college, when taking a college chemistry class - small computer lab next to the main chemistry labs - contained an Apple IIgs ROM 3 that was pretty loaded, a Mac 512, a IIsi with Sony CPD-1304S hooked up, and another IIsi with a regular Color Pivot hooked up to the VampireVid port.

Interesting thread, btw.

-J

[Trash80toHP_Mini]

THX, J. I figured as much with the tiny amount of blood-sucked DRAM available as ersatz VRAM.

I went back to the IP and had to fix some of the connector specs that I'd fumbled while reporting the initial test results.

[volvo242gt]

Cool. Not a problem.

OBTW: with respect to the SuperVideo control panel, didn't notice much improvement in video acceleration. Seems the same. Main issue seems to be the drawing of windows, etc. While playing Super Tetris or Apeiron, it seems function at a normal rate for a IIci. There is a separate Accelerator INIT used with the 2.07 control panel, which was rolled into the control panel for version 2.7... For now, I've upgraded back to 2.7.5.

-J

[Trash80toHP_Mini]

Most of the MoBo Video systems are as, or faster than, NuBus Video for most purposes, but limited to lower resolutions. Cards like the RCPII/IIsi run at the system bus speed as opposed to the 10MHz NuBus. Higher resolutions a/o DSP acceleration for doing transformations in pro level software like CAD Systems and filters in PhotoShop would be the realm of NuBus VidCards. Back in the Day, it was the standard DTP config to have a 13" RGB running off the IIci's MoBo Video and a grayscale TPD or FPD running off a NuBus Card.

For gaming, you're probably not going to see any improvement in speed because gaming resolutions are typically in the low range where MoBo video shined in that rime period. A NuBus Card with DSP Support would have been the winner hands down if gaming had ever taken off on the Mac.

As an interesting side note: by the time Dual Voodoo IIs were the cat's meow for supporting gaming engines in the PC world, the Mac had lost any chance at competing in the gaming market. My friends who are serious (read as obsessed) gamers are impatiently waiting for the development of gaming engines for Linux, Windows having become a serious limitation for them in terms of reliability in today's world.

The RCPII/IIsi, along with its competitors, are a special case. Overcoming one of the IIsi's purposely implemented system bottlenecks and giving a bit smaller bump in resolution/Color Depth/adding second screen support for FPDs and Pivots being their strong suit.

The IIsi was the combination of an upgraded/lamed SE/30 and a lamed IIci, having most of the features of both, but intentionally limited to keep it at the third tier of the Mac II line. Requiring a NuBus Chipset and Math CoPro to take up the expansion slot as well as soldering a severely limited amount of RAM in Bank A made it a pro-sumer IIci. Adding color/GS support, 4 MHz of bus speed and support for NuBus improved upon the SE/30 spec. Removing the L2 cache, support for more than a single NuBus Card, reducing FPD g/s bit depth and shaving 5MHz off the IIci's system bus knocked it down a peg in the Mac_II lineup. Limiting system RAM kept it from competing with either system for the Professional Graphic Art/DTP/CAD Workstation market segment.

I hope this makes some kind of sense . . .

p.s. I also appreciate your feedback, J. It's encouraging, this project gets a lot of views, but very few replies.

[Trash80toHP_Mini]

. . . i can buy this video card, and use it as long as i make my own cable?

sounds like a pain in the butt

Actually it's not too bad, it's a fabulous deal, buy two in case you mess up the first one or we get it working in the SE/30.

Desolder the connector and then solder your wires directly into the thru-holes. I've got the cable so I haven't bothered to source thin co-ax like that used for stand-alone WiFi antennas.

Find that and a solder cup DA-19 connector and you'll be able to make a cable consistent with the original Radius spec. Dunno about RFI interference or possible crosstalk problems for video a/o the rest of the system if you don't use co-ax. The boffins in the 68kMLA Tech Support Division will need to chime in on that one, I'm just a knuckle draggin" hardware hacker in the SpecOps teams.

Eventually, I'd like to noodle out the wiring diagram to just throw 16" 8bit Color directly out of an HD-15 connector at the backplane.

So many hacks . . . so little time . . .

[zuiko21]

Phew... After a long time with tons of work, I'm back again

Glead to hear the discussion going on. I've had little time to keep twinkering on it, but here's an update/remainder of the state of my research:

•The card supplies a 640x870 (portrait) signal, no matter the sense code applied (from the limited selection I tried, that is) with the usual Radius logo on startup.
•The card doesn't show up on the Monitors control panel, thus the OS isn't supporting it Most likely I'm not supplying it the proper sense code, as you suggested. Anyway, the card keeps the Radius logo output, as if an unsupported monitor was detected
•My LCD monitor does sync OK with this default output (although being a wide screen, the card's portrait output look awfully stretched!)
•I have been able to put the card's "default" output on the real RFI genera... I mean, Apple's Portrait Display but with a sense code manually suplied thru my adapters -- same results, only the Radius startup screen, not detected by the OS.

About the adapters I used: here's a pic of all I have:

•The one at the top right has no switches -- came with a surprisingly short-lasting Sony 19" monitor and set ("modern") Macs to the highest-res Multiscan mode.
•The one at the top left seems pretty much the same as the one at bottom right on the picture of yours -- two rows of 8 DIPs (not shown). The original leaflet was lost long ago, but I found this which seems to be correct for it:

TTL UniMac 82D manual

•Bottom right is my very first Mac-to-VGA adapter. It had a small mode table printed, but it's now rubbed off However, I did try each one of its 64 combinations on the Quadra 700, and wrote the results somewhere
•Bottom left came as a freeby in another deal, and never had any instructions for it. However, I think I found its manual on the web:

10-DIP Mac-VGA adapter

But still haven't cycled thru all the settings in order to find a compatible mode for the IIsiColorPivot

Since you have the original cable for it... may I ask you to check continuity on the three sense lines in order to determine the full sense code expected by the card? You know: check each line against ground, and then between each pair of them, both ways -- some sense codes use diodes and thus are polarity-dependant. That way, at least we could try with one less variable...

All the best,

[Trash80toHP_Mini]

Sorry, forgot about your request, Z. I've got to sort out all the hacks in progress and put a few away so I can get to it . . .
. . . it looks like a computer recycling bin exploded in my apartment.

I think I'll dig through the Mac Video Drawer for the VGA adapters I have that match yours and test drive them.

Speaking of RFI Generation Units, I put a grounded, galvanized steel plate between my MPD and the PecisionView 2150.

[Trash80toHP_Mini]

Z, where have you been?

Meanwhile:

Found a very interesting page googling for info on my NuBus Pivot card.

http://maclassic.com/wiki/Radius_Pivot_030_Video_Card

It seems as if someone has already done all this work for us!

The bottom line is interesting, is that new info . . .

Sense Pins: Connect pin 7 to pin 10 for VGA. Leave pin 4 not connected.

. . . I'm wondering how they managed the 60 Hz timing conversion for VGA . . .
. . . forgot to mention that nice piccie . . .
. . . at least they linked to some connector possibilities!

http://www.digikey.com/scripts/DkSearch/dksus.dll?WT.z_header=search_go&lang=en&keywords=4-640620-5&x=0&y=0&cur=USD
http://www.digikey.com/scripts/DkSearch/dksus.dll?WT.z_header=search_go&lang=en&keywords=A30965-ND&x=0&y=0&cur=USD

Be that as it may: someone linked to an adapter pinout somewhere and there is one component to add between two of the lines, so there is progress to be made on this cable cloning project.

[Trash80toHP_Mini]

Found it:


http://pinouts.ru/VideoCables/MacVideoToVGA_pinout.shtml

I've gotta get my @$$ in gear and order a pile of hackin'crap from DigiKey!