VidCard Rosetta Stones? DA-15 <-> DE-9 <-> HD-15Translations?

[Gorgonops]

In the meantime, the pinout/traces are pretty clear in the enlarged section of the listing pic in the second post.

All right. If we operate under the assumption that that DE-9 port is anything "PC compatible" then it has to be either MDA or CGA, it can't be EGA. Reason: It looks like pins #1 and 2 are tied to ground, and pins 6-7 are tied together. Pin #2 is defined as a second ground on MDA and CGA, but is one of the color lines for EGA running in the 350 line mode, which rules out driving an EGA monitor. Meanwhile, tying 6 and 7 together would make sense if you wanted to use an MDA monitor to display only 1-bit video at the high-intensity brightness. The only question remaining would be why they bothered wiring the CGA color lines (3 through 5) if it's only a monochrome card that wouldn't likely support a 200 line video mode. Perhaps they're there for multisync TTL monitors expecting 3 or 4 line RGB (and they're all actually tied together on the host card) (*), or perhaps there's some other type of monitor supported by that port that uses those pins. (Always a possibility) But, yes, the one thing we can definitively say is it's not EGA.

(*) Edit: This conjecture is actually supported by the pinouts on that document NJRoadFan attached, IE, look at the "Pin assignments for other computers" section, 8 and 16 color TTL modes.

 

[Trash80toHP_Mini]

I unbagged the NuBus "Mac II Dual Page" Card for the photo session. I've got great pics to run through GraphicConverter before posting them. After loading the roughly culled pics on a fresh Zip250 to sneakernet them over to the QS'02, I decided it was time to fire up the pet IIfx. It has all my DeclROM spelunking utilities ready to rock and roll. It was a real kick to see the big Radius PrecisionView 2150 in action again after what must have been at least a couple of years to get the pet IIfx back into action! [] ]'&gt;

Anyhoo, here are the results:

Lapis Dual Page Video ©1990

Rev: 400cOEM1024x768only

The etched ©1991 Lapis Technologies Inc. verbiage is so tiny I didn't notice it before I put it under the magnifier lamp.

So now I've got a second Lapis Card with a DE-9 connector, this time with HD-15 thruholes. This one is NuBus so I can actually test it. I've also found a couple of solder cup terminated HD-15 connectors in the hoard. That rocks because I can locate one anywhere I can find an appropriate spot on any backplane plate that matches up with the soldered DE-9 connector and another port. Older NICs come to mind as possible donors.

 

[Trash80toHP_Mini]

OK, here are the promised piccies:

Solder side mirrored to match to component side for comparison





Conventional view of solder side.









Tha all important pinout details:





edit: I really should have blown the dust off the cards. :

 

[Gorgonops]

I'm ignoring the Nubus card, because you can't actually see which pins any of the traces run to because they're obscured by the connector itself on the component side.

As for the SE board, I don't have the patience at the moment to chase the schematic through the interlink between the output board and the card itself and from there try to chase signals the various components on it (which looks to be something of a fool's errand anyway, because most of it is PALs and the Xilinx ASIC, which we have *no idea* what they're set up to do), but if you look solely at the tracework on the output board you'll see that the only common pins between the DE-9 and the DB-15 are some grounds. The one pin on the DB-15 that does have a signal is pin 5, which is driven through a resistor and a diode from its own line on the interlink connector. In my opinion this tells us all we need to know about the DB-15: Pin 5 is "green video" on the regular Mac connector, which can also carry a composite sync signal, and therefore is capable when combined with a ground of driving a BNC-plugged Radius FPD monitor via the appropriate cable. The fact that no other pins on the DB-15 get a signal indicates that this card can drive that sort of monitor and *only* that sort of monitor with that port.

As for what it says about the DE-9 port, the unfortunate answer is "nothing". Just eyeballing it it looks like some, if not all, the lines going to it are buffered through a 74LS244 and I don't see *any* resistors or anything else that would act like a DAC, so I would bet that it only supports TTL monitors. Beyond that, well, I think it's either RTFM or hook it to an ancient TTL-capable multisync with cable compatible with CGA and MDA and see what happens.

 

[ScutBoy]

OK - below are the pics of my card. Note that my connector board is silk-screened as "SE Output Board - VGA" for whatever that's worth. Let me know if any more pictures will be helpful. I did these quick on the top of my desk, so apologies for the bad lighting.

 

[Trash80toHP_Mini]

In my opinion this tells us all we need to know about the DB-15: Pin 5 is "green video" on the regular Mac connector, which can also carry a composite sync signal, and therefore is capable when combined with a ground of driving a BNC-plugged Radius FPD monitor via the appropriate cable. The fact that no other pins on the DB-15 get a signal indicates that this card can drive that sort of monitor and *only* that sort of monitor with that port.

Actually, I've got 5-BNC connections on the Radius PrecisionView 2150 (which won't be used for this kinda crap) and on the 21" MAG Innovision CRT with the cute little MHz (refresh rate) readout panel. Both these MultiSyncs handle just about whatever I throw at them, so we'll see! I'm particularly looking forward to receiving the NuBus Radius FPD Card. If it gets here as quickly as the Lapis card did from the same seller, I could have it by Friday. With that info on MNS, it should be fun to hook it up and see where it goes!

 

[Trash80toHP_Mini]

Sweet pics, ScutBoy! Now it's time to start playing in Illustrator. I think I'll do a comparison of your card's VGA pinout/traces with those on my Lapis DVP NuBus card. Luckily your card's traces are on the solder side and those on mine are on the component side for a connector that's conveniently unimplemented! Nice! [] ]'&gt;

 

[Trash80toHP_Mini]

 

[Gorgonops]

Trash, I hate to say it but I *really* don't think looking at those SE cards is going to help you at all in figuring out your Nubus card. The Nubus card has the signal lines from the blank port outline routed to an area full of unpopulated jumpers, and there's a clear section adjacent to it with soldered jumpers for selecting the monitor type the card is supposed to go with. I kind of suspect converting it from whatever it's for now to another monitor type will require component changes elsewhere on the board, not just slapping a VGA plug on there. I also kind of suspect based on the "ECL" markings on the board that the 9-pin connector isn't anything like the one on the SE board. There is a floating around a pinout for a DE-9 ECL monitor plug, it could well be for that, which has *nothing* in common whatsoever with an MDA/EGA/multisync pinout.

On the topic of the two different SE output boards, well, they're sort of interesting. First off, the one signal pin that went to the DB-15 on the non-VGA one isn't connected to anything; that supports the speculation that it's a special composite-video/sync output specific to a monochrome FPD and has completely a different output chain on the main board. Secondly, it appears that the MDA/CGA/EGA sync lines are connected to the same pins on the VGA connector; that makes perfect sense since VGA uses TTL compatible sync. And thirdly, the three lines that would be the CGA RGB lines appear to be connected to the VGA RGB lines via a set of resistors, *AND* it also looks as if all those lines are connected to each other. That suggests that earlier speculation that the three lines on the DE-9 port are there to enable driving an RGB TTL multisync is correct, and the same lines are simply pulled over to the VGA port and converted to analog levels via resistors. (Since the lines are interconnected that means the card will of course only output "white" on a color monitor, which makes sense; it is after all a 1-bit card.)

What I'm sort of curious about is how the card "knows" what kind of monitor is connected. IE, is there any sort of autodetect (not sure how that would work), or if it's up to the user to specify it to the driver. There don't appear to be any jumpers to hard-force a specific mode.

 

[Trash80toHP_Mini]

Yep, I figured I'd be needing to fake up the resistor set for completing the VGA traces from the setup I see on ScutBoy's card. Dunno if anything is connected to the other end of the resistor connections, but it's at least worth a look. :-/

Here's the pic from avadondragon's thread:



It looks like he has distinct header feeds for three different connectors. I'm dying to see a full set of pics of card and breakout board. I saw him posting recently, methinks it's time for a PM.

edit: whoa! take a close look at the header portion of that card, eudi!

 

[Trash80toHP_Mini]

 

[Gorgonops]

edit: whoa! take a close look at the header portion of that card, eudi!

I guess someone made a mistake laying out the production board there. Sadly not that uncommon a thing.

Anyway... going *all* the way back around the horn, I still don't really see how looking at these SE video cards are going to help you figure out how to retrofit your Nubus card from what I'm now 99.44% sure a DE-9 port fitted out to drive an ECL monitor to a VGA port. (One of the additional headers on Avondragon's card is clearly labeled "ECL", which doubly (triple-y?) confirms the suspicion that the ECL output circuity is completely separate from the VGA/TTL driver circuitry.) The jumper area on the Nubus card sort of *implies* that they might have sold a version that supported a TTL monitor instead on that port, but without documentation as to what monitor types were actually available you're just shooting in the dark. Back in the day full/two page displays were often sold as a matched set with a display card, so while a given manufacturer might have used the same display card PCB to support different sorts of monitors there's no reason whatsoever they would necessarily include all the components necessary to support *all* possible monitors on the boards. Different versions may well have differently programmed ASICs or PALs and use different ROMs.

Obviously your *best* chance would be to find pictures of the *same card* as yours outfitted with a VGA port and compare the component differences. Lacking that, I think your only chance would be to use a DVM and trace the VGA lines (start with the sync lines, then the video), and try to work out how they *would* have been connected through any missing resistors, etc. (If this is a strictly black-and-white card or only supports a very limited number of grays then you'll likely find the video lines were coupled through a simple resistor network to pull the levels down from TTL instead of a proper RAMDAC.) I imagine that eventually the lines you're interested in are either going to disappear into the ASIC, or they'll end up at that terminal block area where you may get a clue if the "same" output is used to drive both ECL and TTL/VGA monitors (presumably said output would be TTL level with a block of external circuitry to covert it to ECL, but that's not a sure thing) or, again, it ultimately looks like it comes out two different sets of drive wires. At that point, well, my best suggestion would be to try to get the thing running, probe it with an oscilloscope, and see if you get something that looks like it could be sync and video signals at that point in the circuitry. If you do then you can try populating downstream with a best guess at appropriate component values and see where that gets you. (If the signals look to be TTL level you can reference the various monitor conversion docs you find on the web for some ideas as what might be appropriate.)

 

[Trash80toHP_Mini]

I never gave that resistor patch a second glance, look at the header labels:

J2 - Video Out

J3 - SPD

J4 - ECL2

J5 - ECL1

Got no SE to get the one card running, got no sillyscope to probe the NuBus card that I can get running. Is there any testing I can do with an autoranging digital multimeter?

Yep. mine's a single bit card from all I've read, so that simplifies things. Converting TTL to VGA sounds like a neat idea for these kinds of cards in general.

The NuBus Radius FPD card will be fun to poke around with when it gets here.

edit: just tracked it, looks like it should be here by Friday!

 

[techfury90]

More likely to me is that there's two ECL headers/pinouts because they bought monitors en masse from multiple OEMs. By putting two headers on the board, you could just plug the output board into the right one to compensate for whichever of the two ECL monitor pinouts were shipped with that particular unit without having to stock two different output boards for the different pinouts. 

Here's a simple TTL-analog/VGA circuit: https://www.repairfaq.org/sam/vidconv.htm#nvctlvg

Simply connect the H and V sync outputs to your VGA monitor as-is, and make sure to connect the video intensity signal to all 3 RGB pins, unless you just want red, green, or blue. 

Note that you will need to convert your board back to TTL if it's already on ECL, I assume the jumpers just need to be moved. Gorgonops' analysis about the HD15 connector is spot-on. I think that NuBus card also may have provisions for TV out via the HD15 connector, hence some of the mysterious unpopulated components. On that card, the signal is generated as TTL and then converted to ECL via the Motorola MC10H124L TTL-ECL shifter at the top right corner. ECL chips usually came from Motorola and had part numbers like MC10XYYYZ, where X and Z are letters, and Y are numbers. You must make sure that none of them are connected to your output, because it's possible that the negative logic voltages used by ECL outputs could cause damage to non-ECL inputs. Conversely, you can identify if a monitor is ECL by opening it up and looking for the opposite, often found on a small board separate from the monitor chassis.

 

[Trash80toHP_Mini]

That the analysis gibes rather nicely with this newsgroup post:

Lapis made a host of PDS video cards for the SE/30, part of which may havebeen dedicated to use with a specific monitor, but there were also the"DisplayServer" cards that had Lapis waxing about their general usefulness.I cite from the original packaging of a "DisplayServer SE/30" DSPDS/30multi-purpose video card:"With its support of most commercial videodisplay standards, it becomes the ideal traveling companion for yourSE/30...The DisplayServer SE/30 lets you choose from a large selection ofmonitors from both the Macintosh and PC environments: EGA, VGA, Multiscan.Full Page Display, Dual Page Display, Apple Mono, RGB, Portrait and DualPage Displays, Overhead Projection (LCD) Panels and TVs.Copyright 1990 Lapis Technologies, Inc.!"Note that this card, too, on thecomponent side carries the P/N# 10006Rev#3, while the S/N# field is leftblank. It has a sticker saying "Lapis PDS/30 Retail 9110 49781" on thesolder side, and what I think is the (socketed!) ROM chip has a stickerwith

"�1990Lapis Technologies, Inc.All Rights ReservedVer SE 30 R30G."

Interesting indeed!

edit: my reformatting of that text isn't showing up for some reason or other. :-/

 

[avadondragon]

Here are slightly better quality close ups of the front and back of the headers on my PCB:

 

[Trash80toHP_Mini]

Yawn! Before I woke up/had coffee I made the mistake of deleting a few the annoying eBay temptation emails and noticed a BIN listing for an SE MoBo from my buddy blairdus. Hip shot that sucker before even checking to see if I needed to transfer funds to cover it, that'sthe depth of semi-consciousness I was in that half-lidded stupor. :

Are you listening to this techfury?