Faulty Indivison MK3 - A1200

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.

    Purchased an Indvision MK3 for A1200 2022-02-16 from a local store here in Sweden and last week I started to get issues. HDMI stopped working and when using the Indivison in my A1200 it started to hang and get random crashes + having random gfx artifacts.


    It works with the VGA out but still has the random crashes.


    I have tried the following without result

    - Different types of PSU, even a re-capped original PSU + LAB PSU

    - Issue is move with the Indivision when I move it to another A1200

    - Still exist with non expanded A1200

    - Cleaned Alice + connector with IPA (while being ESD protected ofc).


    When it is removed the Amiga runs without any issues and if I connect another scandoubler it works without problem. What would a checkup + repair cost for this or is it warranty?


    Thanks in advance

    First of all, the flickerfixer is installed in Lisa, not on Alice - you'll break it if you put it on Alice.


    The flicker fixer is not connected to any part of the CPU/RAM subsystem of the computer, so technically, it does not have any means of making the CPU crash. We therefore need to dig deeper to find the root cause of your problem.


    If the computer crashes, it may *seemingly* be caused by the flickerfixer, but it may still be a power problem. Even if you have the same problem with different power supplies, you still cannot say if it's the PSUs or the flicker fixer, as there are practically only non-suitable PSUs on the market as of late - at least for expanded Amiga computers. They all work fine with light loads, but they start to fail if loads are getting higher.


    Indivision AGA MK3 is a lot more powerful than any other flickerfixer for the Amiga has ever been, and that takes it's toll in terms of power consumption. We've had numerous cases where a bad power supply was the root cause of the problem (mainly Meanwell-based PSUs like A-Power and the Polish "Elektroware" units, but also practically everything offered on Amazon/eBay).


    Most Lab PSUs that are in the hobby price range have a ripple problem: They are built to react quickly to an over-current situation (as you can usually set a current limit and a voltage limit on such a PSU), so they can't work with high-capacity, low-ESR output caps. This is a common problem due to the intended use of a lab PSU, so that's not a good comparison either.


    If the original re-capped unit is a 4.5A type from an A500, that would be great. However, before we jump to conclusions, it may be better if you post a picture of the gfx corruptions -I may be able to tell where things go wrong and solve this remotely, without any cost involved.



    Jens

    Hi,


    Thanks for the reply Jens. I was a bit tired after being up until 3am and them woken up at 7:30am again when my 2 y:o daughter thought it was time to get up :) Of course it is Lisa and not Alice, a tired human mistake :P


    Anyhow, the PSU's I have is the light 4.5A A500 PSU that is recapped with automobile graded capacitors. The laboratory PSU I own is a Rigol DP822A and is a bit above hobby and amazingly OK regarding the ripple.


    The third PSU that I have is a Meanwell, but not the standard ones. I'm using one that is industrial graded similar to the RT65B that is used in Fiber EDFA's which needs to be ripple and noise free. Honestly, I do not know what Meanwell has done the last years. IN the early 2000 they did pretty high end even for commercial products, but now days they are far from it.


    Anyhow, I did some more tests yesterday where I also removed the A1200 PCB from the housing so it would not budge when pushing on the Indivision MK3 to Lisa ( :P ). Still it does hang after a while and in some cases I have the graphic artifacts.


    I will try to get a photo of them when I do more tests. Easily described, it looks like garbage bytes thrown into the chip memorry where the bitplane pointers are located. Perhaps about 20-50 of them each time. The artifacts are static so they are not moving around.


    Claes

    Quote from ClaesB

    Honestly, I do not know what Meanwell has done the last years.

    They're good power supplies, but they are not suitable for the Amiga. They regulate 5V at their output, and they don't know that the load is "distant" on a cable. The cable is an inductor and a resistor, causing drop and a delay in current on transients. None of this is compensated for in a PSU chassis that doesn't know about the load being "distant". So no matter what grade they are, they can't possibly be up to the task, as they were built for a completely different scenario.


    Quote from ClaesB

    Still it does hang after a while and in some cases I have the graphic artifacts.

    Again, please take pictures of these artefacts, so I can see if that's happening on the Amiga chipset side, or the flicker fixer's side.


    Jens

    I will see if I get the issue with artifacts later tonight. I've ordered a new MK3 from GGS Data in Sweden to see do a quick check if it does work or still have an issue.

    Alright, I've tried to get the artifacts tonight and was successful one time but it rebooted very short after so I could not get a picture. The artifacts are not seen as often as before and tonight I have seen the following issues:


    #1 Hangs -> Reboot -> Green Screen

    #2 Hangs -> reboot -> Yellow screen

    #3 Hangs -> No sync to monitor

    #4 Hangs -> Forced CTRL+A+A -> Works a short time then graphic problems (https://photos.app.goo.gl/P1C5MPyQn8oAHYVo9)


    The MK3 is correctly seated and has the latest firmware upgraded. When it is removed, everything works for hours and when it is installed it does not.


    I've tried with following scandoublers tonight:

    - Indivision MK1 AGA

    - Scanjuggler Lite

    - Warped Vision (60Hz only)

    - Amiga 520D HDMI external


    All of the work without issues.


    Is there some test points I can check on the MK3? I will test the new Indivision AGA MK3 (purple version so they wont be mixed) when I get it and if it works, well :) I would say that something is not right with the old one I have.

    What else is in the system? I'd like to eliminate other stuff first. I see a huge cooler near the floppy drive in the fraction of a second where it's in the picture. The system should be as lean as possible if you want to isolate a fault.


    Jens

    This was the one closes to my office computer so it had to be the test computer yesterday. But I have ran the test on a vanilla A1200MB only with the same errors. If you insist I can of course do the tests again on the vanilla one and post.

    What strikes me is that the chipset behaves as if some random values have been written, or the CPU is trying to execute code from the $dffxxx area. This means that a CPU code fetch from chip ram has gone wrong.


    This may be possible to trigger from Lisa, as the chipset data lines are all connected, and even used by Indivision AGA MK3. What you could do is to run a chip ram test for an extended period of time - ATK is the right tool for that. If this fails with Indivision, but doesn't without, we have an indication that the data bus drivers of Indivision AGA MK3 have been damaged.


    Jens

    Got my new purple (I like it!) MK3 now that I've installed in the same Amiga 1200 as I tested in yesterday and it works and has been working for an hour now without any issues.


    Any how, I will do some chip memory tests using MBRTest-2 tonight to see if I can reproduce it. If it is the data bus driver, how much would that cost to repair?


    Claes

    Sorry for a VERY delayed answer but life has kept me really busy. Anyhow, it seems that the error was located in the ribbon cable. When I swapped the cable with the one I got for my new MK3 A1200 it has worked ever since.


    I have checked the cable for any visible damage but could not find any and will do a connectivity test when I have a minute or two free.

    I have no idea how such a damage could result in chip ram read errors. Are you sure that you didn't change more things on the setup?


    Technically, the blue and purple versions are identical: Same design files, same assembly, same test procedure. The test procedure includes testing with the exact cable that the unit is shipped with. As you may have noticed, they come assembled in the box, and that's the exact way that they have been tested here. We actually take pictures of each and every unit as it passes QC, and the cable is never separated from the unit after it got it's warranty ID sticker. So the most likely thing you'Re looking for is a cut in the cable that may have happened between the floppy drive and the shield.


    Jens

    Neither do I and I am 100% sure that I did not change anything in the setup. The problem disappeared on both Amiga's when I changed the cable + adapter.


    Before that it worked for around 2 years. Anyhow, I will do the test this weekend to try locating the issue. I'm happy that the mainboard of the MK3 wasn't broken since now I have two MK3 working :)