Wrong colors on RGB-VGA and Indivision after installing ACA1211lc

Caution: Non registered users only see threads and messages in the currently selected language, which is determined by their browser. Please create an account and log in to see all content by default. This is a limitation of the forum software.


Also users that are not logged in can not create new threads. This is a, unfortunately needed, counter measure against spam. Please create an account and log in to start new threads.

Don't Panic. Please wash hands.
  • So today I got new ACA1211lc and Indivision MK3. Yay.

    I have installed the accelerator, powered my A1200 on and it worked, but colors on RGB (23pin to VGA adapter) are all wrong, grey is green, workbench is all weird.

    It looks correct on composite out. It is green both on the output from Indivision and on RGB.

    So I think something got fried or broken as I was installing the accelerator. It worked correctly before with Blizzard PPC.

    Does anyone have any idea what went wrong? It sort of looks like I see YUV instead of RGB?

  • If i read correct you are using VGA with an adapter on the RGB port of the amiga.

    Maybe you already did, but i would try another cable and or adapter if you have to see if its the same. It seems like "blue" isnt making it to the monitor.

  • I use both RGB-VGA adapter and Indivision AGA MK3 and both were showing green. Composite was OK.

    But after a while, now everthing works correctly. I guess the board warmed up and now it works? Bad solder somewhere on motherboard?

    Also, my Indivision works, but when I start the control program it tells me IndivisionAGAmk3 hardware not found. Continue?

    Strangely enough flashtool info works, and shows the overlay and so does flashtool flash firmware from rescue disk

  • This sounds like a bad connection between Indivision and the Lisa chip. Try to re-seat it. There are lots of connections, and a single one failing can cause funny effects.


    If you still observe a temperature-dependent behaviour after that, you might want to check if your PSU is up to the task. There are many in the market that are not suitable for an Amiga, despite being advertised as that.

  • Thanks.

    The PSU is rather strong, it is repurposed PC power supply, 400W or something thereabout) that I made a while ago with original cable from the previous PSU.

    I think I am getting problems with Lisa chip connections. I have HP Lisa with a SMD component next to it. It looks like it fits, but there is maybe tiny bit of space between PCB and the socket. I do not think the socket is resting on the SMD component and I have used considerable amount of force and tried to remove it and apply few times. Maybe it can still be pressed in deeper.

    FWIW Indivision does produce some display so it is getting some data off the chips. It does not always produce OSD overlay though.

  • The PSU is rather strong, it is repurposed PC power supply, 400W or something thereabout) that I made a while ago with original cable from the previous PSU.

    OK, so you know just enough about electronics to be dangerous. Sorry for the harsh words, but that's exactly what I've been criticising for years in terms of commercial known-bad PSU solutions. Please read our PSU FAQ.


    I think I am getting problems with Lisa chip connections. I have HP Lisa with a SMD component next to it. It looks like it fits, but there is maybe tiny bit of space between PCB and the socket. I do not think the socket is resting on the SMD component and I have used considerable amount of force and tried to remove it and apply few times. Maybe it can still be pressed in deeper.

    If you needed considerable force to remove the Indivision socket, it surely has been pushed on properly. You may have noticed that there is a small cut-out in the socket that gives the SMD component the required clearance. If there's a fraction of a mm of space, that's fine.


    It does not always produce OSD overlay though.

    That's expected behaviour; the OSD only comes up for a settable period of time (see documentation/configtool), and only re-appears if a new screen mode has been detected.


    So if the false colours problem doesn't return, the only thing you should update is your PSU, as that's plain dangerous to your precious retro gear. If colours are wrong "every now and then", your Lisa chip may need cleaning, which can be done with an old toothbrush and alcohol (please avoid contact sprays, as these often contain lubricants that you don't want in a computer).

  • Well, there certainly has been some interference between the SMD component and the socket and I tried to sand little bit of plastic off the socket to be able to push it deeper. While that was accomplished, it did not help with the problem. The settings program still does not recognize Indivision but couple times it managed to read EDID. flashtool does work though.

    The socket is seated as deep as I could push it. I can push double folded Post-It (paper) under it, except for bottom left corner which seems a bit more space, enough for 4 Post-Its.

  • If the configtool does not find the unit even after cleaning Lisa with alcohol and a toothbrush, there may still be a bad connection - not between Lisa and the socket, but between the main board and Lisa. If oyu have very good lighting and a magnitying glass (or better: Microscope), you may be able to see something. If you have the tools and skills, you could just re-solder every pin just to be sure.


    I doubt that the width of 4 pieces of paper can cause the contacts, which are "from the side", to go bad.

  • I did wash Lisa with contact cleaner and isopropyl alcohol. Will try scraping a bit with a toothbrush.

    I have a microscope but it has a base so can't use it on a big board. Will try to find my old 2MP Celestron.

    My soldering is horrible (I am a software engineer), but I can try touching up the pins with TS100 as the last resort.

    Indivision sort of works, the live config tool (Ctrl-Shift-Tilde) works, and the display is sharp and nice, flashing firmware works, so I can live with it as it is. Not being able to use the config tool is more of a psychological problem ;-)

    Also, the config tool does show some stuff, for instance it shows that the display mode is 253 lines, which is probably wrong. (I have a PAL Amiga in US).

    If the config tool does not see Indivision I guess Indivision is not piggybacking on data/address lines properly but it still has power, clocks and color data lines?

  • Not being able to use the config tool is more of a psychological problem

    The tool lets you set different output modes, especially the higher-resolution ones. You're missing out on quie some functionality, and I'm all for making that work for you.


    If the config tool does not see Indivision I guess Indivision is not piggybacking on data/address lines properly but it still has power, clocks and color data lines?

    Yes, one of the data lines is probably not making contact. All other lines such as clocks, RGB data and even the register address lines are required for the rest to work properly, so your search area is limited to the upper 16 data lines.

  • OK, so you know just enough about electronics to be dangerous.

    So true ;-)

    Yeah, I get your point about PSU, but I really do not have many options here in US.

    As for clearance between the socket and SMD component, I did take off some more plastic there, and pushed the socket, the SMD component subsequently made an indentation in the plastic, so I suspect on my board it is mounted very close to Lisa chip.

    I will probe all the socket pins on top of the Indivision PCB with DSO Mini to see if any of them is not connected.

    Is there any databook/PDF that shows mapping of the socket through hole pins to chip pins?

  • I believe I have seen a PDF for Lisa somewhere, but that won't tell you which socket pin is which signal - you don't know the socket layout, and you would have to account for the 90-degrees rotation plus inverted count that results from "going on top of the chip".


    This one should solve it pretty easily:

    The upper pins that didn't fit the screen are RGB signals only. Pins with only a number in them are not used by Indivision.

  • I guess the PC PSU does not have any adjustment? You might want to try reducing the cable length. After all, it's the drop on the cable and in the input filter that causes the drop below the magic 4.85V that's required for most 3.3V regulators. That's the whole reason why I have developed the cable drop compensation circuit.


    Some signals in the "lower" area didn't look very digital in that video - lots of noise, but I could not figure out if that is caused by the probe not applied properly. There is "some" movement, so yes, it appears to make contact. If it does not work with the minimum 4.85V (better: 4.95V), we'll have to take a look at the unit here.

  • Well, I guess I need a better PSU. Can't do much with my ATX ;(

    So I wonder, which ebay 110V PSU is the least bad? ;-)

    elektronika4u (link removed)

    electroware (link removed)

    or this? (link removed)

    Or A-Power (link removed)


    Also, since you develop power supplies, maybe you could consider something like just the converter box that takes any off the shelf 19.5/20V laptop power supply with VRMs for 12, 5 and -12V and regulation hardware with a short cable to connect to Amiga, this way it wouldn't need special versions for US and EU (and UK) or other regions., just run off regulated 20V provided by off the shelf PSU?

    Otherwise, what is the timeframe for your US version of Amiga PSU?

  • All the products you've linked to are based on MeanWell-chassis. These are not designed for a load on a long cable; they will cause even more problems, as the cross-regulation problem is even bigger on such small chassis. On the legal side, it's plain false advertising: Just because someone put an Amiga-connector on the cable does not mean that the PSU is suitable for an Amiga. Also, these sellers are sitting on a barrel of dry powder, as none of them is covered by a product liability insurance (at least not to my knowledge).


    Also, since you develop power supplies, maybe you could consider something like just the converter box that takes any off the shelf 19.5/20V laptop power supply with VRMs for 12, 5 and -12V and regulation hardware with a short cable to connect to Amiga, this way it wouldn't need special versions for US and EU (and UK) or other regions., just run off regulated 20V provided by off the shelf PSU?

    The "converter box" approach is essentially what the CA-PSU is: It uses a 12V PSU with good regulation on the 12V rail and generates everything else from that. Further, it takes care of cross-regulation, so if load on the 12V rail rises (for example when a floppy or CD drive spins up), the 5V rail won't drop. However, I only designed that in because I've had a spare op-amp in the circuit; it's a nice-to-have thing and only solves a practical problem in corner cases.


    Running off a "regulated" 19 or 20V input won't be required; the input may be wobbly to a certain degree, as modern DC-DC converters are monitoring voltage/current on a cycle-by-cycle basis. However, I won't be able to expand the existing design in that way, as I'd have to go through all the tech review with the insurance again. I'd rather find a supplier of 110V-compatible 12V/5A PSUs that complies with all required standards and use the existing design, which already went through yet another revision for adding an adjustment potentiometer (luckily, I was able to demonstrate that all key safety measures are not affected by this change, so the cost for the review was OK).


    So yes, I'm on the CA-PSU, but still need to wait for the key DC-DC converter to come in. That chip shortage thing really hurts here; I can't switch to a different chip, as that would void all the insurance reviews we've already passed.

  • Getting UL and/or CE listed PSU in US seems to be impossible for mere mortals. Amazon is full of burn-your-house down stuff when you search for 12V PSUs for LEDs. That is why I think 20V laptop power supplies are better choice, While there are many fakes, it is not that hard to buy something directly from Dell or whatnot. I do not know if your PSU-CA needs to be certified together with AC-DC PSU, but I think maybe better option would be to sell just the converter box and let people provide their own 12V PSU?


    I am tempted to get couple adjustable buck converters (although ones with -12V are hard to find) and cobble something together to power my A1200 in meantime ;-) They are adjustable so I could compensate for drop on the cable,

  • I do not know if your PSU-CA needs to be certified together with AC-DC PSU, but I think maybe better option would be to sell just the converter box and let people provide their own 12V PSU?

    At this point, I'm doing measurements on the 12V PSU as well, as I am responsible for the whole thing, from the wall plug to the target computer. Since the 12V go into the target computer with only minor checking, it's my responsibility to verify startup behaviour, overload shutdown and transient response. These are all things that the average customer can't measure, especially since measuring on the primary side requires expensive gear that only a fraction of companies even have. I bought a Rohde&Schwarz HMC8015 power analyzer just for this project - including the HZC815 adapter.


    I am tempted to get couple adjustable buck converters (although ones with -12V are hard to find) and cobble something together to power my A1200 in meantime ;-) They are adjustable so I could compensate for drop on the cable,

    The adjustment would have to be automatic, depending on the load on 5V and 12V. Since you're dealing with Ohm's law, it's all linear, so no complicated curves to create. Since the negative 12V only need a few hundred mA, the circuit for that is pretty simple - if you know how the buck converter works in the first place...

  • Yeah, about that. i do not believe in electrons ;-) I think the bits are carried around the boards by little dwarves or maybe ants. It is the curse of being software guy. At my electronics class all I could do was to make blue smoke ;-)

    So I was thinking about using off the shelf DROK converter for 5V, which should keep the voltage constant at the source, but obviously cannot compensate for increasing drop on the cable as load increases ;( and a "TPS5430 Positive Negative Dual Output Module Regulator Power Supply Module with Switching 12V Output" and just hook them up to some 20V PSU I have laying around.

    It is probably bad idea. Plus, I do not know if their startup times are the same and if there is any sequencing for power rails in Amiga power supply.

    OK, enough bad ideas ;-)

  • The last reply was more than 365 days ago, this thread is most likely obsolete. It is recommended to create a new thread instead.