Topic: C128 80-column display on modern devices

[tokra]

I've posted this on the Lemon64-forum where a discussion about this emerged, but wanted to put this up in the "original" C128-forum as well. What do you think?

I have the feeling that a possibility to connect the 80-col signal of the C128 to modern hardware is sorely missed. The question pops up a lot on the Commodore-forums from time to time. The problem is that the C128 uses digital RGBI (RGB with intensity) which is basically a CGA-signal, while newer monitors and TVs use analog RGB, which is also used by the Amiga for example.

There are LOTS of solutions to display analog RGB on todays systems, and there even is a  converter from RGB to VGA available for around $40, however it still does not handle RGBI or full CGA 16-color mode:

http://www.jammaboards.com/store/cga/ega/yuv-to-vga-converter-pcb-gbs-8220/prod_291.html

So, the only problem remaining is converting the digital RGBI-signal to an analog RGB-signal. There are 3 ways to do this, with increasing difficulty:

1.) don't care: Just plug the RGB signales from RGBI into the corresponding analog RGB lines. Fairly easy to do, but you just get 8 of the 16 colors

2.) simple conversion: Use some small parts to integrate the intensity signal, like outlined here for example:

http://sites.google.com/site/h2obsession/CBM/C128/rgbi-s-video

Still doable with basic soldering skills I'd imagine, you get 16 colors, but one color is wrong (dark yellow instead of brown)

3.) conversion with brown-fix: the original CGA and C128 signal doesn't have a dark yellow, but instead a brown color. Read more about this here:

http://en.wikipedia.org/wiki/Color_Graphics_Adapter

A converter that takes care of this was posted by Wolfgang Moser in comp.sys.cbm in 2008, see this thread:

http://groups.google.com/group/comp.sys.cbm/browse_thread/thread/0116da09bf27f528

There is even a schematic and DIY layout linked in that article. This would be the holy grail of CGA (C128) display on modern devices. If I understand correctly you will need a small logic chip to implement the brown-fix, still looks doable if you know how to solder.

Sadly only a small number of enthusiasts will be able to do even option 1 and would probably gladly pay for option 3 to get the 80-column signal with full 16 colors on TV or with the added converter even on standard VGA-monitors. Option 2 seems like an easy enough solution to implement to test the waters on how much demand there is for such a device.

[Hydrophilic]

My memory may be failing again.  But wasn't the color on C= 1901/1902/1084 monitors dark yellow?  Did Commodore monitors muck with one of the colors like IBM CGA monitors?
 
I like your don't care solution     Just plug it up and hope nothing gets fried!  I don't know where to look for the manual to my TV... but I would be worried that it expects a 1V level, while the C128 puts out 5V level.  Shouldn't burn the house down or anything, but it would be a shame to ruin a nice TV...

[dabone]

I've got mine working on svideo. I still have to muck around with the resistors to get the colors right, but I have a stable display.
 
I'm using a rgb to svideo adapter from amigamaniac.
I've got a v2.0 he now makes version 3, I haven't tried the new version but it should be fine.
http://www.amigamaniac.com/RGB_to_PAL_NTSC_adapter.html
 
I've paired this up with the RGBI to RGB adapter from this page.
http://sites.google.com/site/h2obsession/CBM/C128/rgbi-s-video
 
When I get back around to it, I'm going to remake the board using 1.5k  pots so I can play with it and get the values for the best pic.
 
Later,
dabone

[RobertB]

I'm using a rgb to svideo adapter from amigamaniac.
I've got a v2.0 he now makes version 3, I haven't tried the new version but it should be fine.
http://www.amigamaniac.com/RGB_to_PAL_NTSC_adapter.html

     I keep forgetting to order one from him.

          Truly,
          Robert Bernardo
          Fresno Commodore User Group
          http://videocam.net.au/fcug
          July 23-24 Commodore Vegas Expo 2011 - http://www.portcommodore.com/commvex

[tokra]

My memory may be failing again.  But wasn't the color on C= 1901/1902/1084 monitors dark yellow?  Did Commodore monitors muck with one of the colors like IBM CGA monitors?

My Commodore 1901 definitely puts out brown, not dark yellow. Judging from MIRKOSOFT's picture of Tut Ench Amun, I'd say the 1084 he uses shows brown as well. So it looks like C128 VDC is the same as regular CGA, although the C128 Programmer's Reference Guide says "dark yellow". They may have screwed up the guide back then.

Would really like to have the best solution for showing C128 80-column on a modern TFT. I'm wondering if this sort of cable or PCB (if you choose the brown-fix solution) would be something for Jim Brain to offer in his online store. Which C128 wouldn't want that?

[Hydrophilic]

Well at least C128 Programmer's Reference Guide and my memory agree   As I recall, on the 1902, the dark yellow looked really like a dark version of the normal yellow, and not a true brown like I would get from the VIC.  Unfortunately my 1902 died many years ago.
 
So if your 1901 and Mirkosoft's 1084 show real brown, then either my memory is defective or my 1902 did not have the CGA "fix".  I might understand why there would be difference between the 1084, geared for the Amiga, and the 1901/1902 geared for the C128... but don't know why there would be a difference between 1901 and 1902...
 
Wish my 1902 was still working to be sure if it is the monitor or my memory that has the "problem"     Of course if it was working, I guess I never would have made that RGBI circuit... The worst part is my 1902 failed one month after the double-extended warranty expired (circa 1999)     I tossed the defective monitor long ago; I wonder if that warranty is somewhere... sad condolense if does exist...
 
Does anybody have a color picture of C128 and monitor (in 80-column mode obviously) released by Commodore?  Would be nice to know how they designed it to appear...
 
From my understanding, Commodore built their computers and their disk drives (using Asian drive mechs), but the monitors were branded 3rd party equipment.  I'm thinking the monitors were made by Phillips / Magnavox... at least in the USA... Of course my "understanding" on the subject could be as bad as my memory 

[tokra]

Ok, I've checked with Robert Hurst who I know from the VIC-20 forums and who popped up on a Google-search for the Commodore 1902 as well :-)

He confirms that the color it is in fact displayed as BROWN on his 1902 as well like on CGA and sent me a photo of his test as well. So at least it's the same all over the world, and to correctly reproduce a C128 signal on a modern device you would need the "brown-fix" (solution 3) device. Now, if only someone could produce these and offer them for sale...

[Hydrophilic]

I guess you are right and my memory is defective (or I had a non-standard monitor).  Here is a quote from 1901 service manual

An additional feature of the module enables brown colour wherever the computer produces red and green signals only.  This is done by IC02 SN7420 (Dual 4 input NAND gates) and transistor TI02.  The task is to reduce the green output level by RI21 and DI06 when the correct combination appears at the input lines.

I wonder if that is completely accurate, as that implies it would effect both normal (bright) yellow and dark yellow (brown).  Anyway, I've attached a clip of the relevant schematic.  I don't completely understand it... I would need to cross-reference their schematic with pin-out of 7407 and 7420 to have a good understanding.
 
Hopefully this will help somehow...
 
 

[tokra]

Phew, you loose me when it comes to hardware schematics. Maybe these two from the above mentioned thread on comp.sys.cbm on the brown fix by will help?

[17] http://d81.de/shared/C=128-VGA/C=128-VGA-IBM-Colorfix-sch.png
[18] http://d81.de/shared/C=128-VGA/C=128-VGA-IBM-Colorfix-brd.png

[Hydrophilic]

I don't really understand that last schematice I posted either!  You know I was wondering if there is a difference between 1902 and 1902A.  I said I had a 1902 but it was really a 1902A.  I think the difference was the 1902A supported analog RGB so it could be used with Amiga too.  However, it only came with C128 cables.  Also it had only one speaker, unlike the 1084s (which I believe has two).
 
Besides the 4 video formats it supported (aRGB, RGBI, Composite, Luma/Chroma), the thing also had "night vision mode" as I like to call it.  It turned everything into shades of green  I guess it was suppose to emulate monochrome monitors...
 
Anyway I found schematics of 1902A over at Ray Carlsen's website.  He's got lots of schematics on this page.  Anyway, looking at the schematic, I see it also has the "brown fix."  So I guess my memory is defective after all!  I've attached a relevant clip from the schematic.
 
It is about as complicated as the last one I posted, but I actually understand this one!  The RGBI lines are first inverted.  Then a wire-or circuit is used to active the "brown fix".  The wire-or uses diodes from inverted B and I lines (thus they must be high) and it uses open-collector inverters from the already inverted R and B lines.  Long story short, when the RGBI = 1100 then it grounds a resitor tied to the G line.
 
That inspired me to draw up some simple circuit diagrams to implement the "brown fix".  Note you would still need RGBI -> analog RGB circuit.  Anyway, I've attached 3 different versions.  The first is the simplest and uses 7405 Inverter.  Then next 2 are just a bit more complicated and use either 7433 NOR, or a 7438 NAND.
 
I have not built any of these, so I'm not sure what resitance values to use.  I imagine the pull-up (on the left) should work fine with 2.2k or 3.3k ohms.  The one on the right (the pull-down attached to Green) would really require some experimentation.  I imagine I would start off with something like 470 ohms and increase/decrease to get a nice color...  It also might be a good idea to put small (68 ohm or so) resitor on each of the 4 input lines; otherwise the Green pull-down might not have the desired effect.
 
Just some ideas that may or may not be usefull...
 
Edit I was to lazy to draw the connection to ground of the ICs, which is definately needed.  I also was lazy and omitted the connection from Vcc to +5V in the first one (7405 inverter); that is also required.

[Hydrophilic]

I got bored and drew up a semi-passive circuit for the brown fix.  This doesn't require a power supply, Vcc... it draws most of the power from the B and I lines... which means Blue would be a little brighter than it should be, making brown a bit more washed out / less saturated (i.e., more gray and less pure color).  It also means the I-line would be a bit higher than normal, so brown wouldn't be as dark it should be.
 
A tiny amount of current is drawn from the Red line, but for producing yellow/brown, the important thing is the relative amounts of Red versus Green.  Not much you can do about the current need from Red to drive the transistor, however, you have great flexibility with the resistor on the Green line.  So you can setup the red/green ratio to suit your fancy.
 
The resistor on the Green line should be very small, around 100 ohms.  The other 3 resistors should all be the same size, assuming you used matching transistors.  And they should be considerably larger than the Green one... I would guess in the 2 to 10 k-Ohm range.
 
There are 3 transistors.  The top one attached to green/red, is an NPN (I suggest a 2N3904) and the other two are PNP transistors (I suggest those be 2N3906).  There are 3 diodes.  Just about any small-signal diode should work, like a 1N914 or 1N4148.
 
I haven't built this, so I have no idea if it would work.  If I can find another 9-pin plug lying around somewhere, I might actually give some of these circuits a try and post some photos.  But don't hold your breath!

[tokra]

Fascinating! If you ever get around to this, please let us know the results. I'm still dreaming for a ready-to-buy solution for the RGBI to RGBA-conversion (with or without the brown fix). I've dropped Jim Brain a mail regarding this, but I realize this is a question of how much demand there would be for such a device and how expensive it would be to build. I'd be willing to pay up to 40 EUR ($50) for such a device given it would be a small number run (or 20 EUR/$25 without the brown-fix). Good thing the RGBA->VGA adapter itself that goes after this is just $40, but this is mass-produced. A solution under $100 to make the C128 "future-compatible" would still be a valid investment for a retro-enthusiast I'd think.

[dabone]

There are a few 15k analog to 31.5k converters available, what I would love to see is a rgbi to analog rgb board that could fit in a db9 hood. Then you could add them to a product like this one.
 
http://www.miba51.com/CoCo_VGA_Adpater.html
 
This is a nice, stable 16bit analog 15k upscaler, It would work with my amiga 500, coco3, Snes, Genesis, Jamma Arcade boards, etc. And with a rgbi adapter then we could use it with the c128 and ibm cga systems.
 
Just my 2 cents.
 
Later,
dabone
 

[tokra]

Hmm, I've tried solution 1 today and got the GBS-8220 converter:

http://www.jammaboards.com/store/cga/ega/yuv-to-vga-converter-pcb-gbs-8220/prod_291.html

As the converter only has a 15-pin input, I needed a converter cable from 9-pin and used this one:

http://www.ambery.com/db9todbadca.html

I got myself a 5V 2A-power supply with the correct plug, connected everthing and... it doesn't work 
99% of the time I get no picture at all. If I enter 1 or lines of text I sometimes get a jumping, mis-aligned picture, when using the "wrong" colors, sync is lost completely again, same if I type too many lines. I got as far as trying the colors and could confirm only 8 colors got through (the bright ones). However as the picture won't sync correctly, this isn't really a working solution. Maybe I'm doing something wrong? According to the docs the converter should process:

14.5kHz - 16.5kHz
23.5kHz - 25.5kHz
30.5kHz - 32.5kHz    

Is the C128 standard (PAL) output not within this range? I thought it was 15.625 khZ

[Hydrophilic]

For both PAL and NTSC, the horizontal frequency is in the range 15.625 to 15.75 kHz, so I would think it should work with your converter. 
 
Maybe it is the voltage levels?  If you are not using any intensity mod (RGBI to RGBa) then the levels going into your converter should be near +5V.  The RGBI to RGBa should drop the voltage levels, but depending on which circuit, these may or may not drop down to +1V.  May be the GBS-8220 requires voltages around 1V level... I don't know...
 
Even though it is actually AC, you can get a near DC output for testing with a simple voltage meter by clearing the VDC screen (ye old print shift+home) and setting the background color for testing.  For example COLOR 6,3 should make the full screen RED.  Then you can test the voltages going into your converter.   In this example, G, B, I should measure near 0V while R should measure 1~5 V depending if you used a RGBa conversion (16 colors) or a straight connection (only 8 colors).
 
Other than that... check your wiring to be sure you don't have HSYNC and VSYNC mixed up.  Well it looks like you bought a cable and from your link, the connections on the C128 look correct (the VGA side looks correct too).
 
You might need to check that your converter uses seperate H and V syncs and not a composite sync in the mode you want to use.  I looked at the web page you linked it is really skimpy on the details.  It says RGBHV is supported only in the 30kHz range, but the CGA/EGA will support the 15kHz range... I'm not sure what differences your converter has between these two modes (other than frequency and the Pxx setting).
 
I didn't start using IBM PC until VGA was well-established, so I don't know the details about CGA/EGA... but from what I understand one or both of them may use an inverted level for HSYNC or VSYNC.  So another thing you can try is inverting these signals...
 
Before wiring up inverters in hardware, you can test in software by changing bits 6,7 of VDC register 37 with command

Code: [Select]

H=1:V=1:BANK15:SYS DEC("CDCC"),H*128+V*64,37
You can try different combinations of H and V (set equal to 1 or 0) to see if any work.  That is based on info from the C128 Programmer's Reference Guide on page 294 (register map).  However, in the register-by-register description, register 37 is completely undocummented (should be on page 334)!  So this may not work on all revisions of 8563/8568...
 
Anyway, good luck!

[tokra]

Thanks for your input. I did some searching and the register 37 is only available on the newer VDC-chips:

http://www.go64.de/english/online_e/06_00_2e.htm

and apparantly only of use if you have an EGA-monitor.

However I found this post:

http://forum.arcadecontrols.com/index.php?topic=98692.0

which also suggests that the RGBHV connection may not work. Using the RGBS connection however would mean to connect the HSYNC and VSYNC cables, which *may* ruin the source - so I'm not really looking into doing this...

Maybe I should try this device instead:

http://www.ambery.com/rgbcgatovgac.html

This ia a lot more expensive though, the GBS-8220 was just $37 delivered...

[tokra]

Searched around some more. Looks like all those converters are just for analogue RGB and when they say "CGA" they falsely mean 15 kHz analogue RGB. Even the one from Ambery say it needs:

Input RGB Signals:

    RGB: 0.7 Vp-p 75 ohm
    H sync/ V sync: 2 Vp-p 75 ohm

But the TTL RGB of the C128 and "true" CGA would only output 0V or 5V, not 0.7 max.

I've found this (old) cable solution online:

http://www.commodore128.de/64kabel9.htm

This uses 470 Ohm and 1 kOhm resistors apparantly. Would this produce signals within the limits needed for the converter? And what about the limits of H sync/V sync? Are these in the correct range from the C128?

[Hydrophilic]

All the outputs from C128 VDC connector are TTL (0~5V).  I think it's weird you found something with different voltage levels for RGB as compared to H/V syncs... they're usually all TTL (0~5V) or all VGA-ish (0~1V).
 
As worried the seperate H and V syncs might cause trouble.  As you noted, you can connect them for composite sync... just don't wire them directly together!  You can put a small cap (less than 1 uF) between the two and connect one to the composite sync line.  It sounds like you aren't willing to try that just yet.
 
The sync signals just stabilize the picture, and shouldn't interfere with the pixels / colors.  The way I figure, it would work or it would not.  If it did work, I don't think it would cause any quality issues.  I would give it a try before ordering another converter.  The only real concern should be the voltages...
 
As far as voltages too high, you can put some resistors on the line.  But this really requires knowing the voltage requirements and impedance of your converter.  Otherwise you would just have to use trial and error.
 
The converter cable with only resistors in your last link would give you RGBa type signal... the actual voltage seems like it would be 0V, 1.666V, 3.333V, or 5V depending on the color combination (or less depending on impedance of GBS-8220).   So if you need 0~0.7V (I'm still not sure exactly what your converter needs), then this would not work.  Also there are no resistors on the H / V syncs, so again it would not work if you need 0~2V.
 
Did you get any type of documentation with your converter that tells you what input voltage it needs for the lines?  I did not see that info in the web page you originally linked...
 
As far as register 37 goes, I thought because your converter supports CGA/EGA modes the software setting would be worth a try.  Of course if you have an older VDC chip then it won't help at all...

[tokra]

Thanks again for your input. I think I will have to ask a colleague some day who is more keen on  this sort of thing if he can maybe set up a little adapter-board. I also contacted Allan Bairstow who according to the Commodore Free magazine did some experiments on this converter board as well.

Really a shame calling these boards CGA to VGA converters, when their input requirements are in fact definitely NOT CGA. Sadly the model I got came completely without manual. Only the small PDF can be found online. The manufacturer's website http://www.gonbes.com is down more often than up, but I couldn't find a better manual there either.

At least now I have this thread for future reference 

[dabone]

I've got one of those boards also, and I've had no real luck with it. Either with my 128 or Amigas.
To me these boards are a complete waste of money.
 
BTW, CGA is a accepted term in the arcade business, and the old cga arcade monitors are 15k and 0 to +5 ANALOG input levels.
(I used to collect and repair arcade machines)
So in theory a arcade style cga to vga should work, just not these.
 
FWIW.. Most people don't remember CGA as a digital monitor signal, so whenever you see something listing cga compatibility assume Arcade style cga.

Here is a Brand new LCD display that should work just fine with the c128 video port.

http://www.provantage.com/miracle-business-lt15a-ttl~7MRCL02Y.htm


Later,
dabone
 

[tokra]

Browsing comp.sys.cbm I found someone who claimed he has successfully connected the Ambery-converter to a C128. I inquired and a few days later, he made a Youtube-video of the converter in action:

Commodore 128 LCD + 80 col..flv

So this means at least the $89 box produces some kind of stabke picture. Sadly he didn't check the color-output... guess I'll have to get one myself. After some digging I found that it's really a Cypress CM-397:

http://www.cypress.com.tw/english/display.asp?id=244

and it can be gotten all over the world. So I'll probably fork out the money and give this a shot.

On the other hand I thought some more about the Gonbes-8220 converter. From the manual it looks like it can only handle CGA with Composite Sync but the C128 puts out HSync, VSync. However the C128 also puts out a monochrome video signal on pin 7:

http://www.hardwarebook.info/C128_RGBI

and this includes a composite-sync signal at 0.5V apparantly. Would it be possible to just connect pin 7 to the composite sync in of the converter? I think I read somewhere that the extra video-signal wouldn't really matter if you just need the sync. This might be worth experimentation as well.

[dabone]

When I worked on connecting a original psx to a arcade monitor, I extracted composite sync from the composite video signal using a lm1881 circuit. It's a little hard to get a good clean sync on the scope using a perfboard, but should be good enough for a test. I'll try this when I can get to it, I think I have a couple left in the parts bin.
 
 
Later,
dabone
 

[frank128]

Hi,

I got myself a 5V 2A-power supply with the correct plug, connected everthing and... it doesn't work 
99% of the time I get no picture at all. If I enter 1 or lines of text I sometimes get a jumping, mis-aligned picture, when using the "wrong" colors, sync is lost completely again, same if I type too many lines. I got as far as trying the colors and could confirm only 8 colors got through (the bright ones). However as the picture won't sync correctly, this isn't really a working solution.

I run into the same symptoms with Gonbes GBS-8220
 

I've got one of those boards also, and I've had no real luck with it. Either with my 128 or Amigas.
To me these boards are a complete waste of money.

You are not the only one who has waste money. But the real bad thing is that there is no working converter solution available.

Is there any converter that works? – I accept the simple 8 color solution below.

1.) don't care: Just plug the RGB signales from RGBI into the corresponding analog RGB lines. Fairly easy to do, but you just get 8 of the 16 colors

Frank

[Hydrophilic]

I love tokra's idea: don't care   
 
I always thought my TV only had composite and S-Video inputs, but recently I discovered it also has YCbCr inputs... I don't have any 'real' YCbCr cables, but I have many many unused sets of Composite/Audio_Left/Audio_Right cables.  Some of them have a very thick composite (video) line with thin left/right (audio) lines... but most of my extra cables have the same diameter for each of Composit/Left/Right, so I assumed they were made with the same specificiations... which is to say each should be able to carry a video signal.
 
Based on that assumption, I tried connecting my satellite receiver to my TV via the YCbCr inputs.  It works and the picture is very good.  Honestly I don't see any difference between YCbCr and SVideo... maybe with 'real' YCbCr cables there would be a small but noticable difference ?
 
ANYWAY, the fact that YCbCR input on my TV works made think of something that could help people trying to connect a C128 80-Column to a modern display device.  Attached is a circuit diagram I drew up...
 
NOTICE: I have not tested this so it may not work... in fact there are several variables so there is a good chance it will not work... BUT it is the cheapest video solution that is possible, I think.  There are only 6 resistors and 2 capacitors... that should cost less than $5
 
Once I get moved into my new place, I'll have a chance to try this.  I'll let you know my results, but in the meantime, see the attached circuit diagram.
 
You should notice H.Sync and V.Sync are connected to Y output via capacitors.  The value of 1uF is just a guess.  In fact, sync may not even go on the Y line, but I'm guessing that it does.
 
More importantly, each of R/G/B connects via a resistor network.  It is approximately a 1/5 output.  The idea is that the C128's 5V (TTL) outputs should be scaled down to 1V.  Of course this may reduced by the load of your particular TV/Monitor.
 
Also the Intensity (I) line is not used so you only get 8 colors (if it works at all).  If you try this and get a good 8-color display, then you should be able to apply my RGBI -> aRGB circuit fix.  I've released several of these; most without the 'brown fix' and few with...
 
Anyway, this is just another idea to try.  Also, the signals are meant for YCbCr input of a TV/Monitor, but it might also work with a multi-sync VGA monitor... so instead of having 3 seperate RCA-type plugs for Y/Cb/Cr, you could try running all the lines into a 15-pin VGA connector...  I have a VGA connector I can test, but as I said, I won't be able to test anything for a while...
 
Maybe this will help somebody... if not, I think it should still be a fun experiment!
 
EDIT
I just noticed the circuit is not symetrical: there is no resistor on the H.Sync / V.Sync lines.  So you may want to try adding a 390 Ohm resistor on those lines.

[Richard42]

Hi everyone.  Just wanted to pop in and give you some more info about video conversions.  I've also been working on this topic for my Amiga 1200, C128, and Coco 2/3 for a while.  I just bought a digital scope and now can actually troubleshoot and make things work, so I've learned a lot about these converters in the last few days.

I have a GBS-8220 and an Ambery AV-1 (which is also sold as a Cypress CM-397).  My AV1 appears to be dead and I've ordered another.  I got the GBS-8220 working with my A1200 and I've figured out it's limitations.  This is actually a very good board - it's cheap and the video quality is nice.  And it will work well with the C128 if you know how to use it.

Here are the limitations that I've found with the GBS-8220 and how you can work around them:

1. It only converts 60 hz.  This makes it pretty worthless for the Amiga, since most games and demos are 50hz.  This should be okay for my C128 though.
2. The sync signals must be in the 0-2v range, and the input RGB signals must be in the 0-0.7v range.  So you need to use a resistor divider network to go from the TTL 5v levels of the C128 to the analog levels of the converter.
3. Here's the kicker: the GBS-8220 only takes composite sync for 15khz video.  So, you must use a 74 series logic chip (NAND or XOR) to combine H and V sync from the C128, then a resistor divider to get it down to 2v before feeding it to the converter.  I don't know if it can handle either positive or negative polarity; the amiga has negative and I got it working with that.

I just put all this info together and finished up with my Amiga tonight.  I'm going to build a circuit and get it working with the C128 in a few days.  I can give you all more info if you want.