Thread
External Compact Mac CRT
So I bought one of the relatively cheap TTL CGA/EGA/RGB to VGA converters (GBS 8200). Was like $20 + shipping.
Would it be possible to run the signal directly from a compact Mac's motherboard to a VGA LCD using one of these converters?
Would it be possible to run the signal directly from a compact Mac's motherboard to a VGA LCD using one of these converters?
Although I'm not really an expert in electronics, I don't think it's really feasible. At least not employing the existing connection to the analog board, that is.
For that to work, the signal coming out of a compact Mac's logic board would have to be a CGA/EGA digital TTL signal, but as far as I know it's a really low level analog signal. I remember to have read somewhere that those logic boards actually drive the CRT directly, so I would guess that they must be generating something that ultimately drives the CRT gun without any fancy digital to analog conversion.
If someone could hack together a chip - say, an FPGA or something - that could sample the analog signal that's supposed to drive the CRT and process it on the fly into some more modern format, that would open up a lot of possibilities, but it's a pretty tough job. If I recall correctly there had been some attempts (possibly even here on these forums) to study the feasibility of such an adapter, but I think that nothing ever came out of them.
Another option would be to reverse engineer the Macintosh logic board and ROM/toolbox to find out exactly how the video memory is written, and try to hijack that information out of it. At this stage, the information to be displayed should reasonably be stored in some sort of "array" of pixels which could be "conveniently" read and converted by an external processing board to something actually drawable on a modern screen. Speaking as a programmer and software guy, this path would be a lot easier than the first one (not to say that's easy at all, but at least more feasible), especially since BMOW has already done extensive studies on the Macintosh toolbox ROM and the hardware architecture of the 512k/Plus - after all, his Plus Too project rendered everything via VGA!
That's just my two cents anyway, let's hear what the more experienced members say!
For that to work, the signal coming out of a compact Mac's logic board would have to be a CGA/EGA digital TTL signal, but as far as I know it's a really low level analog signal. I remember to have read somewhere that those logic boards actually drive the CRT directly, so I would guess that they must be generating something that ultimately drives the CRT gun without any fancy digital to analog conversion.
If someone could hack together a chip - say, an FPGA or something - that could sample the analog signal that's supposed to drive the CRT and process it on the fly into some more modern format, that would open up a lot of possibilities, but it's a pretty tough job. If I recall correctly there had been some attempts (possibly even here on these forums) to study the feasibility of such an adapter, but I think that nothing ever came out of them.
Another option would be to reverse engineer the Macintosh logic board and ROM/toolbox to find out exactly how the video memory is written, and try to hijack that information out of it. At this stage, the information to be displayed should reasonably be stored in some sort of "array" of pixels which could be "conveniently" read and converted by an external processing board to something actually drawable on a modern screen. Speaking as a programmer and software guy, this path would be a lot easier than the first one (not to say that's easy at all, but at least more feasible), especially since BMOW has already done extensive studies on the Macintosh toolbox ROM and the hardware architecture of the 512k/Plus - after all, his Plus Too project rendered everything via VGA!
That's just my two cents anyway, let's hear what the more experienced members say!
BMOW would know for sure. 
I have an SE in a 3rd party enclosure, and I wanted to see if it worked. The video card uses a proprietary connection, so there's no way for me to test it.
I have an SE in a 3rd party enclosure, and I wanted to see if it worked. The video card uses a proprietary connection, so there's no way for me to test it.
Just out of curiosity, did you try that "maybe it *is* TTL?" 9 pin to Multisync (via resistors) idea I floated over on the VCF for working with the add-on video card?
As for driving one of those scaler boards from a Mac motherboard, well, two issues.:
Per this page about driving a Mac monitor from a VGA card, one of the lines that comes out of a Mac motherboard is a "Horizontal Drive", not a standard "Horizontal Sync" signal. Roughly speaking, the issue is that the CRT inside a Mac isn't a "complete" monitor and the motherboard compensates for that, and as a result it doesn't output a standard Hsync signal. Someone better versed in television technology would have to come up with the solution for this; my guess is that a fairly trivial amount of circuitry would be required to reshape the pulse appropriately but you'll need to do that in order to have a different monitor work properly with it.
(If you have access to one of those "Build your own Mac and save a bundle!" books describing how to build your own Hackintosh it's possible the requisite circuitry might be in there?... *googlegoogle*... No, annoyingly not. The book covers a number of different solutions, the one that involves directly connecting your Mac motherboard's video output to an external monitor is a "Power R" video adapter which has the necessary electronics on it; the book doesn't include the schematic. Hacking instructions for DIY'ing apparently originally appeared in a 1989 series of"Hackintosh" articles in Computer Shopper magazine. And finding the full text of those articles seems to be roughly as easy as finding the Holy Grail.)
The other thing I wonder about, reading the PDF instructions for the particular scaler you have, is if it actually supports "EGA", as in, "The frequency range that's roughly in the ballpark of a compact Mac's monitor.". The PDF says it supports signals in the range of "14.5-16.5K, 23.5-25.5k, and 30.5-32.5k", and none of those ranges *quite* cover the scan range of the Compact Mac's monitor. (Nor to they *quite* cover 350-line EGA.) If that's not a typo and it really doesn't support 21-2k-ish frequencies than it's not going to work with a compact Mac motherboard's output. (Is "23.5-25.5k" an Arcade board standard, maybe?)
As for driving one of those scaler boards from a Mac motherboard, well, two issues.:
Per this page about driving a Mac monitor from a VGA card, one of the lines that comes out of a Mac motherboard is a "Horizontal Drive", not a standard "Horizontal Sync" signal. Roughly speaking, the issue is that the CRT inside a Mac isn't a "complete" monitor and the motherboard compensates for that, and as a result it doesn't output a standard Hsync signal. Someone better versed in television technology would have to come up with the solution for this; my guess is that a fairly trivial amount of circuitry would be required to reshape the pulse appropriately but you'll need to do that in order to have a different monitor work properly with it.
(If you have access to one of those "Build your own Mac and save a bundle!" books describing how to build your own Hackintosh it's possible the requisite circuitry might be in there?... *googlegoogle*... No, annoyingly not. The book covers a number of different solutions, the one that involves directly connecting your Mac motherboard's video output to an external monitor is a "Power R" video adapter which has the necessary electronics on it; the book doesn't include the schematic. Hacking instructions for DIY'ing apparently originally appeared in a 1989 series of"Hackintosh" articles in Computer Shopper magazine. And finding the full text of those articles seems to be roughly as easy as finding the Holy Grail.)
The other thing I wonder about, reading the PDF instructions for the particular scaler you have, is if it actually supports "EGA", as in, "The frequency range that's roughly in the ballpark of a compact Mac's monitor.". The PDF says it supports signals in the range of "14.5-16.5K, 23.5-25.5k, and 30.5-32.5k", and none of those ranges *quite* cover the scan range of the Compact Mac's monitor. (Nor to they *quite* cover 350-line EGA.) If that's not a typo and it really doesn't support 21-2k-ish frequencies than it's not going to work with a compact Mac motherboard's output. (Is "23.5-25.5k" an Arcade board standard, maybe?)
http://www.vintage-computer.com/vcforum/showthread.php?44362-Full-Page-Display-signal-on-a-non-FPD-monitor&p=339699#post339699
Again, this is based on a hunch that because pin 6 and 7 on the jumper for the 9 pin connector are bridged together that the system was meant to work with a TTL monitor with the same pinout as IBM MDA, but not necessarily the same frequency. (As I said in the thread, the logic goes like this: Intensity+Video == bright white/green/whatever on MDA, if you were working with a system that didn't do grayscale you might just tie intensity and video together so instead of 2 bit grayscale you get one bit B/W from a single TTL output.)
(EDIT: And yes, IBM MDA is, strictly speaking, two bit grayscale not "monochrome".)
Again, this is based on a hunch that because pin 6 and 7 on the jumper for the 9 pin connector are bridged together that the system was meant to work with a TTL monitor with the same pinout as IBM MDA, but not necessarily the same frequency. (As I said in the thread, the logic goes like this: Intensity+Video == bright white/green/whatever on MDA, if you were working with a system that didn't do grayscale you might just tie intensity and video together so instead of 2 bit grayscale you get one bit B/W from a single TTL output.)
(EDIT: And yes, IBM MDA is, strictly speaking, two bit grayscale not "monochrome".)
I think it could be done, but it looks hard. Chapter 9 of Larry Pina's Macintosh Repair & Upgrade Secrets describes how to build an adapter to interface a compact Mac with a "PC-style TTL external monitor". It involves some soldering and building a simple circuit with some inverters. Then you could connect the output of the circuit to the TTL-to-VGA adapter you bought. But even then it might not work - as Gorgonops said, the horizontal refresh rate of the Mac video signal is not quite the same as typical PC TTL video. According to Pina's book, some TTL monitors worked better than others with a Mac video signal, and some didn't work at all.
Some months back, I proposed a different method of getting a compact Mac to VGA adapter and posted about it here. My idea was to use an FPGA or other similar hardware to grab the Mac's video data from the logic board connector, then buffer it, and "replay" it at a different horizontal refresh rate and with appropriate letter boxing etc. to make a standard VGA video signal. It should be possible in theory, maybe I'll try it someday.
Bbraun proposed a different technique using a fast microcontroller connected to the 68000 busses, and snoop for CPU writes to the frame buffer memory. The frame buffer is just a specific address range in normal RAM for the compact Mac. Then you'd have the Mac's screen image bitmap data, and could do anything you want with it: output it as VGA, save it to disk, send it over the network, etc.
Let me know if you decide to try Pina's circuit - I'd love to see it tested in the real world.
Some months back, I proposed a different method of getting a compact Mac to VGA adapter and posted about it here. My idea was to use an FPGA or other similar hardware to grab the Mac's video data from the logic board connector, then buffer it, and "replay" it at a different horizontal refresh rate and with appropriate letter boxing etc. to make a standard VGA video signal. It should be possible in theory, maybe I'll try it someday.
Bbraun proposed a different technique using a fast microcontroller connected to the 68000 busses, and snoop for CPU writes to the frame buffer memory. The frame buffer is just a specific address range in normal RAM for the compact Mac. Then you'd have the Mac's screen image bitmap data, and could do anything you want with it: output it as VGA, save it to disk, send it over the network, etc.
Let me know if you decide to try Pina's circuit - I'd love to see it tested in the real world.
I was wondering about the possibilities for d0ing a digital conversion of the frame buffer data of any given video output. The wedge would consist of tapping the data as it enters the pins of the RAMDAC or equivalent.
The internal Video Card for the PowerBook 100 was likely wedged at the Ram Expansion Card slot as it was reportebly RAM based.
Such is probably not particularly Compact applicable, but mirroring oddball video output via DVI conversion seems to be the most flexible approach to this VidNoob. Analog rocks, but doesn't play well with modern LCD technology. Heading it off at the digital pass seems appropriate.
Up-scaling ought to be only slightly non-trivial at that point I would think.
The internal Video Card for the PowerBook 100 was likely wedged at the Ram Expansion Card slot as it was reportebly RAM based.
Such is probably not particularly Compact applicable, but mirroring oddball video output via DVI conversion seems to be the most flexible approach to this VidNoob. Analog rocks, but doesn't play well with modern LCD technology. Heading it off at the digital pass seems appropriate.
Up-scaling ought to be only slightly non-trivial at that point I would think.
How about video output to an external monitor from a Macintosh Portable? Would that be any help in solving this?
Well, at that point, I'm wondering if it would be feasible to build a replacement "analog board" to completely bypass the original AB.
My reasoning is that a large part of the original analog board is pretty much video circuitry employed to generate the appropriate output to drive the CRT (refresh rates and so forth). If it's possible to hijack the signals coming from the logic board to connect an external monitor of some kind, then it would in theory be possible to build a more compact and durable analog board that only provides power, much like an ATX power supply would do.
I know that this would pretty much ruin the historical value of a compact mac, but on the other hand it would provide a way to make perfectly working logic boards shine again (and maybe it would also open up the possibility of some crazy modding experiments, for those who are interested, like LCD upgrades and so on).
My reasoning is that a large part of the original analog board is pretty much video circuitry employed to generate the appropriate output to drive the CRT (refresh rates and so forth). If it's possible to hijack the signals coming from the logic board to connect an external monitor of some kind, then it would in theory be possible to build a more compact and durable analog board that only provides power, much like an ATX power supply would do.
I know that this would pretty much ruin the historical value of a compact mac, but on the other hand it would provide a way to make perfectly working logic boards shine again (and maybe it would also open up the possibility of some crazy modding experiments, for those who are interested, like LCD upgrades and so on).
I fail to see how putting that #%>$^%|@$ A/B away in a box for safekeeping/preservation could be considered ruination.
That sucker and the MoBo always needed a fan or a MacChimney to work very long after the convection cooling design's airflow got borked anyway.
Trash80 makes a good point, replacing the A/B in this design would be more fixing a bad engineering decision in the first place than anything else.
Not too surprising there. IBM MDA/Hercules and the compact Mac are both roughly 350 line video standards but MDA runs at 50hz while the Mac is a hair over 60hz vertical refresh, with a corresponding increase in hsync. So roughly speaking you're "overclocking" an MDA monitor by about 20% if you connect it to a Mac motherboard. I'm sure some monitors stood up to that better than others.
As for adapting the Compact Mac's output for more modern monitors, probably the most straightforward approach is a simple scan doubler. Here's the schematic for one that's far more complicated than you'd need for this because it includes an NTSC decoder and handles color. Basically cut out the entire NTSC decoder, axe the A/D converters as well, use only one line buffer (instead of 3), and dumb down your D/A output converter to a simple resistor. (It would probably be pretty trivial to replicate this all onto an FPGA; since it's single-bit monochrome your line memory only needs to hold about 64 *bytes*. Of course, on the input side you'll need to adapt it to the Compact Mac's oddball hsync output, per a circuit like Pina's.) This would give you a double-scanned (electrically) VGA compatible signal at a (roughly) 43khz hsync. The result *will* be an oddball video mode and LCD panels which are strict about only supporting VESA-standard video modes won't like it but it's well within the capability range of more forgiving multisyncs. The HSYNC would be about the same as 640x480@85hz, or a little higher than 800x600@60hz and well under 1024x768@60hz.
I, for one, would be very glad to have another way of running a Macintosh Plus logic board without the original AB, since both my ABs have failed on meI fail to see how putting that #%>$^%|@$ A/B away in a box for safekeeping/preservation could be considered ruination.
That sucker and the MoBo always needed a fan or a MacChimney to work very long after the convection cooling design's airflow got borked anyway.
I was just considering that, from a collector's point of view, having a classic Mac without its original AB is a suboptimal solution, and one that would not employ the integrated CRT anymore. In order to replace the analog board but still use the original CRT, one would have to design a video circuit that mimics the functions of the original, but that is somewhat beyond the purpose of the "external video" hack (and would probably require a good amount of reverse engineering anyway).
Still, I'd be very interested if someone decided to go ahead and try to design such a board, even focusing on just having the external video output working for the time being! All the proposed concepts are very intriguing, but are very much beyond my electronic capabilities...
I would imagine one could make a far better, and smaller, A/B with modern components than what is in there.
I would imagine the logic board and A/B of a given Mac, while still being 100% compatible with all the other original hardware, could be easily re-done much better using modern components.
I would imagine the logic board and A/B of a given Mac, while still being 100% compatible with all the other original hardware, could be easily re-done much better using modern components.
I know that at a certain point, BMOW was looking into making a new logic board, completely compatible with the original Plus one, using "modern" components as part of his "Plus Too" project; however three years have passed, and I don't know whether he kept the research active or put the project on the shelf!
if you want to take the compact mac video signal, and convert it into something that could be displayed by any type of monitor, you need a scalar processor.
have to buffer the signal, or at least build a pixel clock. Write the data into dual-port RAM. the SE/30 this is real easy to do, you can tap off the clock that drives the 74LS166. SE I have no idea.
Then the frame data gets retrieved from the second port to scan a secondary display. Once this is designed, you could technically interface an LCD panel, or anything.
have to buffer the signal, or at least build a pixel clock. Write the data into dual-port RAM. the SE/30 this is real easy to do, you can tap off the clock that drives the 74LS166. SE I have no idea.
Then the frame data gets retrieved from the second port to scan a secondary display. Once this is designed, you could technically interface an LCD panel, or anything.
That's some interesting stuff! Seems like someone has preceded us by almost three decades in successfully having a 128k logic board running without its AB. The video circuit seems pretty crazy stuff - does it output a standard composite video signal? That would be pretty easy to convert into VGA!
Also the power supply seems a common ATX one of some sort, it's a pity that the image is not in high resolution and it's difficult to understand exactly what cables have been connected just by looking at it. The video circuit is neat, though - the schematics is there and ready for testing!
No. If you read the fine print he says he figured out a modification he could do to an Apple composite monitor that would up its line rate sufficiently to sync with the Mac's output. Remember, the Mac's output simply won't *fit* on a monitor using standard NTSC sync; it only allows for 262.5 vertical lines per non-interlaced frame (technically broadcast NTSC is 525 line *interlaced* using two half-frames) which of course the Mac outputs a picture 342 lines + some blanking lines tall.
There are industrial composite monitors that would be willing to accept that signal unmodified: here's an example, you can order it with a 21khz sync rate, that should work right off the shelf.
That would have been ISA, such was the world of the Cat Mac/Hackintosh, all inexpensive, Industry Standard HARDWARE hacking! Debasement of the time honored latter term for almost mere OS trickery is an abomination.
I'm sorry, I'm not really practical with all the standards that were used back then. I know that ISA was a standard card interface used by a lot of IBM clones in the mid-80s, but I'm not accustomed to power supply standards and so forth!
I've read about the IIGS monitor, but I haven't ever owned one, so I was not sure about the connector and sync/refresh rates it used, I assumed it must have been some sort of composite/antenna signal just by looking at the circuit output.
So basically, an industry specific monitor would be needed for that hack to work? Would it be possible to re-use it in some ways with more common CRT screens as found in TVs and old multiscan monitors? NTSC monitors won't like such a signal as you pointed out, but I'm wondering if it's possible to obtain something less-exotic than an industry grade monitor or if we are forced to do some post-processing to that signal.
I've read about the IIGS monitor, but I haven't ever owned one, so I was not sure about the connector and sync/refresh rates it used, I assumed it must have been some sort of composite/antenna signal just by looking at the circuit output.
So basically, an industry specific monitor would be needed for that hack to work? Would it be possible to re-use it in some ways with more common CRT screens as found in TVs and old multiscan monitors? NTSC monitors won't like such a signal as you pointed out, but I'm wondering if it's possible to obtain something less-exotic than an industry grade monitor or if we are forced to do some post-processing to that signal.
No worries, no criticism intended, I only get upset with upstarts who've hijacked a perfectly suited hardware hacking term for sleight of soft hands.
ISA would be the entire (IBM PC AT) architecture: I/O Cards, Box, MoBo, MoBo specific backplane openings (where implemented at all) PSU and those specific Power Connectors pictured that finally settled down into the ATX/PCI form factor after much ado.
Such passed for standardization in that 8/16bit AT day and on thru the VESA Local Bus, pre-PCI era of 32bitdom that developed by the Clonemakers (Gang of Nine) after some upstart in Cupertino went all 24/32bit over the entrenched PC industry. The gang went one way after a lot of false starts, while IBM MicroChannelled itself into . . .
. . . well . . . lets face it . . . Lenovo.
.
ISA would be the entire (IBM PC AT) architecture: I/O Cards, Box, MoBo, MoBo specific backplane openings (where implemented at all) PSU and those specific Power Connectors pictured that finally settled down into the ATX/PCI form factor after much ado.
Such passed for standardization in that 8/16bit AT day and on thru the VESA Local Bus, pre-PCI era of 32bitdom that developed by the Clonemakers (Gang of Nine) after some upstart in Cupertino went all 24/32bit over the entrenched PC industry. The gang went one way after a lot of false starts, while IBM MicroChannelled itself into . . .
. . . well . . . lets face it . . . Lenovo.
.
Well, the IIgs monitor *is* a standard NTSC composite monitor, but the author of that article hacked it to accept the higher scan rate. Undoubtedly it's possible to do that with other monitors, but, yes, you're unlikely to run into "just any" composite monitor that's willing to take it. There are some "multisync" composite monitors out there, mostly designed to handle both PAL and NTSC signals, but the Mac's output signal is going to be way out of spec for PAL as well so I doubt many of them will be forgiving enough to handle it.
Just to note, it's actually possible to get monochrome Composite Video out of a VGA card using just a few resistors; I built one of these adapters on a lark once. The thing is you have to use a driver to reprogram the VGA card to output an interlaced picture at NTSC scan rates. Alas there's no way to reprogram the Mac's video hardware (short of attacking it with a soldering iron) to be more friendly.
that 32-bits VESA was still a pain to work with. It wasn't very well designed to handle more than 1 card. I've had display get corrupted and hard drives not work because of capacitance and interference. It took me quite a while to find a 32-bits video card and 32-bits IO card that worked together for my then 486DX2. I'm glad they went PCI which behaved a whole lot better shortly after rather than try to keep the overly long and flaky 32-bits slot and added an extra 32-bits upgrade