Indivision ECS V2 ghosting

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.
  • I did send a patched version of the flicker fixer to Christian, but that obviously didn't improve it "enough" for him. Trouble with this patch is that I would have to do a new CE measurement (as the patch uses RF transistors in the output stage), which is out of proportion cost-wise.


    I have a suspicion that it's not even the output stage, but just the cable, which is from a different manufacturer. The pinout is identical from V1 to V2. Can you verify if that makes a difference?

  • I look forward to hearing what the result is after swapping them. Perhaps I will also construct my own cable to test also.


    **Update**


    Built a totally new cable and no appreciable change. :/ Oh well.


    **Update 2**


    Okay I need to try another real ECS Denise. I usually run Amiga RGB through OSSC because of the ghosting but two nights ago I was able to see the same issue but very VERY faintly with the RGB output too on the same Dell monitor on OSSC on HDMI.


    If that doesn't prove what component it is I am wondering if this could be from a slightly less than healthy Video Hybrid component?

  • I have quite bad ghosting on BenQ BL702A and BL912 when using ECS V2 VGA output. It get's a little bit better when using scanline emulation. It seems like the ghosting on the ECS V2 is even worse than on the ECS V1. ECS V1 has quite mild ghosting. Have no ghosting if using RGB output instead of VGA. Tried it on A500s, A500 Plus and A600s.

  • Updating once more:

    I pulled the 8373R4 Denise from one of my other A500s and tested it in my main A500 that has the IndivisionESC V2 in it. Here is what I found:

    1. The ghosting is still present on Amiga RGB with high contrast white / black transitions. My first place to look for it is in the menus on PPaint which uses a white menu background and black text by default. A faint dark bar is able to be seen extending to the right of the selected item just like when I use the VGA output from the Indivision, though the Indivision makes it much more visible. I don't know at what point the Indivision takes gets the necessary data to generate an image but it seems to me the Amiga's Video Hybrid component is maybe the next place to look but I don't know where I could get a new one to try and Kipper2k doesn't seem to sell the replacement anymore and I do not have the skills to build one from the BOM if I got PCBs made. :/

    2. I don't know how or if it is related but since pulling the Indivision I have not been able to duplicate the slight graphic corruption I sent my ACE2 back in to be checked out for. Once I got it back from repair the rate at which is appeared had become less frequent. However, with the real ECS Denise I have been totally unable to duplicate it even when I try to make it happen. (AmigaTestKit, AMOS Pro, and Workbench Install just resizing windows were sure triggers to make the glitch appear though it never seemed to hurt stability.)

  • Update.


    Have been testing between my ECS V1 and ECS V2. Turns out ECS V1 has no noticeable ghosting when using a good quality VGA cable. Also has a very sharp image on my BenQ BL 702A and BL 912. Both with 62Hz and 50Hz profiles.


    I have tried the same VGA connector cables that comes with the ECS V1 on my ECS V2. ECS V2 will still have bad ghosting. Of course I also tried the same VGA cable that goes to the monitor. ECS V2 seems to have ghosting no matter what profile or custom profile I do.


    I even created a profile to make the ECS V2 replicate ECS V1 62Hz profile. ECS V1 62Hz profile will make my BenQ screens switch to 1024x600@62Hz. I made the ECS V2 switch to the exact same resolution and everything by making a custom profile. Still it has the bad ghosting. And in general ECS V1 seems to have a sharper image VS ECS V2.


    This is definately something going on with the ECS V2. I have both original Amiga PSUs which I have recapped and more modern PSUs. Makes no difference on the ECS 2. I even tried a smoothing cap between VCC and ground in Denise socket. No difference.


    I do quite a bit of soldering in general. Is there any possible solutions using smoothing caps or resistors somewhere?

  • I noticed something interesting that I can't explain easily. I tested the V2 that Jens updated in several A2000s (A2000A, A2000B rev 4.5, A2000B rev 6.2) and the ghosting was still present. All of these have well working power supplies (LiteOn - 2 of them recapped with rather expensive Panasonic caps, voltages around 5.02-5.07V)

    I did not yet try switching the cable between the V1 and the V2 as the V1 is in my Video Toaster A2000 and that is a major disassembly effort. But I did test exactly the same V2 using the A604n on an A600. With one of the monitors and cables that had ghosting on the A2000. And I could barely make out any ghosting - almost had to imagine it.

    About the only thing that I can imagine is a grounding issue on the A2000s. I don't remember my V2 coming with the grounding cable - I will look in the box again - even though that is shown in the wiki and recommended. My V1 in the Video Toaster A2000 is properly grounded with the cable.


    Christian

  • I noticed something interesting that I can't explain easily. I tested the V2 that Jens updated in several A2000s (A2000A, A2000B rev 4.5, A2000B rev 6.2) and the ghosting was still present. All of these have well working power supplies (LiteOn - 2 of them recapped with rather expensive Panasonic caps, voltages around 5.02-5.07V)

    I did not yet try switching the cable between the V1 and the V2 as the V1 is in my Video Toaster A2000 and that is a major disassembly effort. But I did test exactly the same V2 using the A604n on an A600. With one of the monitors and cables that had ghosting on the A2000. And I could barely make out any ghosting - almost had to imagine it.

    About the only thing that I can imagine is a grounding issue on the A2000s. I don't remember my V2 coming with the grounding cable - I will look in the box again - even though that is shown in the wiki and recommended. My V1 in the Video Toaster A2000 is properly grounded with the cable.


    Christian

    I have had mine in my A500 grounded with the supplied cable as noted in the manual. The V2 has been in and out a couple times for me to move my A500 into my Checkmate case and now twice after that to test the both my original DENISE and a full ECS version. Each time the ground cable was reconnected to the recommended screw. I have also made a new cable from scratch. No change.

    As you describe "almost had to imagine" the ghosting... that is how the ghosting appeared through the Amiga RGB going through a component converter to a Panasonic CRT TV. I now can see it with ease with the RGB out to the OSSC and my Dell monitor as I described earlier, V2 installed or not, it doesn't matter. The V2 just seems to make the ghosting much more visible which leads me to believe it is something to do with the Amiga. My next step is to get one of my spare A500s setup and test it there.


    This post got me thinking... I would actually buy a V3 if there is ever to be such a thing if it mimicked the way the OSSC works and had a built in HDMI passthru and output via HDMI... It could be a great internal, all in one, highly configurable option for highly expanded systems with one unified video output to modern displays.

  • I have had mine in my A500 grounded with the supplied cable as noted in the manual. The V2 has been in and out a couple times for me to move my A500 into my Checkmate case and now twice after that to test the both my original DENISE and a full ECS version. Each time the ground cable was reconnected to the recommended screw. I have also made a new cable from scratch. No change.

    As you describe "almost had to imagine" the ghosting... that is how the ghosting appeared through the Amiga RGB going through a component converter to a Panasonic CRT TV. I now can see it with ease with the RGB out to the OSSC and my Dell monitor as I described earlier, V2 installed or not, it doesn't matter. The V2 just seems to make the ghosting much more visible which leads me to believe it is something to do with the Amiga. My next step is to get one of my spare A500s setup and test it there.


    This post got me thinking... I would actually buy a V3 if there is ever to be such a thing if it mimicked the way the OSSC works and had a built in HDMI passthru and output via HDMI... It could be a great internal, all in one, highly configurable option for highly expanded systems with one unified video output to modern displays.

    I've now moved the ECS V2 to a spare A500 have I have setup and tested it on two more monitors. This A500 is an NTSC 6A running 2.0 ROM that hasn't yet been recapped and is using a late 90's "Big Foot" converted PC power supply. Same ghosting regardless of the resolution settings in the configuration program. Cable doesn't matter (cheap with no ferrite cores and 2m long or short 1m long with huge cores and heavy thick cable.)

    I give up. I am sure most of the ECS V2s are fine but I don't believe this is right. I will probably sell it as used / marginal condition heavily discounted because it simply doesn't work right and then buy a ScanPlus ECS.

  • No, I don't expect an answer or solution after so many times, but want to comment it. ;) Have the same problem as

    DownshiftDX , how to sell this "smearing thing"? What is it worth in this condition?

  • I have the same issue when using it in my PAL A600. I never noticed it in my NTSC A2000 when using the same unit, but it's really bad on my A600, both on my Acer monitor and my Elgato HD60 Pro capture device. I attached a screenshot to help visualize. The bleeding is at its worst when there is a longer horizontal line of white, such as the top edges of the icons.


  • No, I don't expect an answer or solution after so many times, but want to comment it. ;) Have the same problem as

    DownshiftDX , how to sell this "smearing thing"? What is it worth in this condition?

    It is clear this isn't going to be resolved after this long.

    I don't know what it may be worth and if I do list mine for sale I am going to be clear about the issues and probably start at $50 US plus shipping.

    If I had a PC CRT monitor to test on it might be acceptable but I can't justify buying a CRT just to find out it is still a poor quality image. For now the ECS V2 is just sitting in my spare A500 on the bench while my real day-to-day A500 has a real ECS Denise in it with an OSSC.

  • It is clear this isn't going to be resolved after this long.

    I don't know what it may be worth and if I do list mine for sale I am going to be clear about the issues and probably start at $50 US plus shipping.

    If I had a PC CRT monitor to test on it might be acceptable but I can't justify buying a CRT just to find out it is still a poor quality image. For now the ECS V2 is just sitting in my spare A500 on the bench while my real day-to-day A500 has a real ECS Denise in it with an OSSC.

    The ghosting is quite bad in workbench. But with scanlines on and gaming it's not too bad.

  • I did investigate this last week with the Samsung Syncmaster 214, which shows the issue here when using high output pixel clocks (=high output resolutions like 1600x1200). The change involves quite a few SMD components to be soldered, and if you go "all the way", it's 15 SMD components to be changed/added, as what I came up with is a new output amplifier stage that has almost nothing in common with the original Indivision ECS V1.


    Samsung's input stage is quite piccy, and in essence, it appears to be developed against high-frequency-DACs of modern-day graphics cards. In order to mimic the conversion process of those, I had to change the output driver transistor with a much faster one with 2GHz transition frequency. To properly control the base of this transistor, one base resistor needs to be changed, and another one added, otherwise the output voltage levels are not right. Next, the analogue voltage supply of the RGB output stages needs more support with a higher capacitance after the supply filter.


    I'll try to optimize the mod, so fewer components and bodge wires are required. Otherwise, I fear that too many Indivision ECS V2 will die in the attempt of applying this mod.

  • Don't get me wrong - this is not a "fix" - it's an upgrade, so it's a paid service. I will publish everything I've found, but I encourage everyone who experiences this problem with his monitor to do the fix on his/her own.

  • I did investigate this last week with the Samsung Syncmaster 214, which shows the issue here when using high output pixel clocks (=high output resolutions like 1600x1200). The change involves quite a few SMD components to be soldered, and if you go "all the way", it's 15 SMD components to be changed/added, as what I came up with is a new output amplifier stage that has almost nothing in common with the original Indivision ECS V1.


    Samsung's input stage is quite piccy, and in essence, it appears to be developed against high-frequency-DACs of modern-day graphics cards. In order to mimic the conversion process of those, I had to change the output driver transistor with a much faster one with 2GHz transition frequency. To properly control the base of this transistor, one base resistor needs to be changed, and another one added, otherwise the output voltage levels are not right. Next, the analogue voltage supply of the RGB output stages needs more support with a higher capacitance after the supply filter.


    I'll try to optimize the mod, so fewer components and bodge wires are required. Otherwise, I fear that too many Indivision ECS V2 will die in the attempt of applying this mod.

    The follow up is appreciated Jens. However, on every single monitor,Amiga, and now video out cable I have tried the resolution has not made one bit of difference. Since I first realized the issue my ECSv2 has been restored with the recovery tool and latest firmware and I run it with the default 800x600 settings for an NTSC system.