Indivision AGA Mk3 / CD32 - default output is beyond monitor sync range

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.


Please understand that you need to create an account to be able to post, guest posting was disabled as an anti spam measure.

  • I have an NTSC CD32 and just received my Indivision AGA Mk3 today. Aside from the cable issue that I posted about in another thread, I'm having a hard time getting output on HDMI connection. My monitor sends a notification that the signal is out of scan range and tells me to set the output mode to 1920x1200@60Hz.


    Is this because the Indivision outputs at 50Hz by default? If so, is there an easy way to switch it to NTSC refresh rates? If so, where would I find the software for the CD32/4000D version? It looks like the downloads are all geared toward the A1200 version.

  • The default output is a 60Hz async mode that every single monitor should be able to display. We do NOT switch on VSync by default because not all monitors support it. The goal of Indivision AGA has always been to be plug&play.


    Did you buy directly from us, or from a reseller?


    Jens

  • The default output is a 60Hz async mode that every single monitor should be able to display. We do NOT switch on VSync by default because not all monitors support it. The goal of Indivision AGA has always been to be plug&play.


    Did you buy directly from us, or from a reseller?


    Jens

    I bought it from a reseller - Amiga of Rochester here in the US.

  • I'm sure he didn't have any time to run the config tool and save a different config, as he got our shipment just before he went to a show.


    Do you get the same "out of range" message on the VGA output? Can you run a WB and the config tool on your CD32?


    Jens

  • Hi Jens. Thank you for looking into this. As it turns out, the crystal oscillator was mounted in such a way that it prevented the socket from fully seating against the Lisa chip. With encouragement from Thomas, I was able to move the oscillator just slightly to the right, which allowed the Indivision to sit flush with the board. It is working as I write this.


    The only issue at the moment is that on HDMI, the screen is offset to the right by a fair bit on my Dell U2410. It is operating at 800x600@60Hz while the native resolution is something like 1920x1200@60Hz -- so I think this is just a monitor issue, or perhaps there is a better resolution to set it to.


    I did run the recovery disc for CD32 and it flashed the unit. However, the software on that disc (Indivision Prefs) crashes when launched from the AmigaOS version that shipped with my TF330. I am continuing to experiment with this, but I am confident that the board is good from a hardware perspective.


    Again, thank you so much for your support!

  • I'm still having a bunch of odd behavior with this thing and I'm not sure what the deal is, but it seems to reset to some kind of bad state when the power has been off for more than a few minutes. Here's the scenario...


    1. Cold boot the system = no display

    2. Switch monitor input to RGB - display is up, but not provided by Indivision

    3. Open Indivision.prefs and it warns "IndivisionAGAmk3 hardware not found. Continue?".

    4. Switch the system off, boot from recovery CD. Flash happens automatically and Indivision displays as expected, but output are offset to the right.

    5. Boot into Workbench, run games, whatever - everything is fine, but offset to the right.

    6. Power down system, walk away for 10 minutes, turn back on. No video until I use the recovery disc again.


    It's almost like when it's powered down it reverts to some mode the monitor doesn't like. Is it possible the NV memory is corrupt?

  • Also, when trying to save firmware to the device, I get "The configuration is not saved to the device", but no explanation as to why.

  • Between myself and Thomas from Amiga of Rochester, we figured out the issue - it's PSU related.


    Long story short, I needed a slight voltage bump on the 5V rail. Clues to the resolution were found in this thread:


    https://forum.icomp.de/index.php?thread%2F2135-individision-ava-mk3-4000d-cd32-no-signal-at-all%2F=&postID=19291#post19291


    I just wanted to take a moment once more to thank you for all you do for the retrocomputing community. I appreciate your time and guidance.

  • Between myself and Thomas from Amiga of Rochester, we figured out the issue - it's PSU related.

    That's a common theme in the retro computing world. I'll make a version of the CA-PSU for the CD32 later this year, then this problem will be solved.


    Jens

  • That's a common theme in the retro computing world. I'll make a version of the CA-PSU for the CD32 later this year, then this problem will be solved.


    Jens

    Yeah, I'm discovering that not just with my Amigas, but also my C64.


    Both of my Amiga PSUs are based on the Meanwell PT65b. One was built by Ray Carlsen (somewhat famous here in the US as a power supply builder) and the other is retrofitted into a stock Commodore PSU case. Both PSUs use the same square DIN connector required for the A500/600/1200, but I have a Keelog adapter for the CD32's 4-pin round DIN.


    With my A1200, I had to bump the +5V rail in order to ensure the PiStorm/rPi had enough voltage to operate at full speed. I assumed the same would be required for the Ray Carlsen supply, which I was using with the CD32 -- but that was not the case after I wrote here last night. Switching back to the original CD32 PSU pretty much solved this problem (there are still some occasional no video conditions, but they are often solved with a toggle of the power switch). I'm a little hesitant to use the original PSU as it looks identical to the old C64 PSU that I know is infamous for failure-related damage to connected devices.


    In the thread I cited earlier, I see that you provided a quick utility to wake the Indivision at boot. I am going to integrate that into my startup-sequence - once the Indi is awake, it's good until the next cold boot regardless of the PSU used, so I think that solves my problem.

  • Both of my Amiga PSUs are based on the Meanwell PT65b.

    That's known-bad. Not because the Meanwell chassis is bad (it's in the medium-quality range, definitely OK for consumer applications with low-spec rippe requirements), but because the Meanwell chassis does not know that the load is "at a distance". It will regulate 5V at it's own output, and does not know how much voltage is dropped on the cable to the target, in the connector and in the input filter of the main board.


    The technical details are all summarized in our PSU FAQ. The TL;DR would be: The CA-PSU is currently the only one on the market that complies with all three key specifications that Commodore has put out for A500/600/1200 PSUs. All the offerings from Elektroware, A-Power, Keelog and the like are not suitable for an Amiga and therefore fall under the "false advertising" category. I'd send the unit(s) back for false advertising and a full refund.


    One was built by Ray Carlsen

    He makes the same mistake, but he has so many fans that I'm the bad guy for pointing that out.


    Your observation with the 5V "bump" is what the CA-PSU does automatically: It measures the current, calculates the voltage drop on the three loss-making paths (cable, connector, filter) and compensates for that at the source - constantly, in real time, so it's compatible with dynamic load changes like a variable-speed accelerator.


    The advantage is that you don't have to remember that the PSU is sending out a too-high voltage if it isn't loaded that high. Technically, yours out of spec now, as you can easily give the A1200 excess voltage just by clocking down the accelerator (or even by removing it). If you forget about that in maybe a year from now, you may end up with a fried Amiga (not knowing how much you had to crank it up).


    Jens

  • Also, when trying to save firmware to the device, I get "The configuration is not saved to the device", but no explanation as to why.

    Hi SeaWall07,


    Are you saying you were able to solve the "The Configuration is not Saved to the Device" by using another Power-Supply?

  • Hi SeaWall07,


    Are you saying you were able to solve the "The Configuration is not Saved to the Device" by using another Power-Supply?

    No. I was under the mistaken assumption that my Indivision was not displaying because a valid configuration wasn't being saved to the non-volatile memory. In actuality, it was saved and valid, but the Indivision wasn't able to be properly initialized because of inconsistent power on a cold boot. Once I popped the CD in, evidently there is an executable that forces the card to wake. I still have not found a better solution for the PSU issue - though it seems more stable with the stock CD32 powersupply. I do keep the CD handy in case I get stuck and have to wake the card again.

  • No. [...] it was saved and valid, but the Indivision wasn't able to be properly initialized because of inconsistent power on a cold boot. [...] I still have not found a better solution for the PSU issue - though it seems more stable with the stock CD32 powersupply. [...]

    Too bad, I'm suffering from it not Saving and stating that "The Configuration is not Saved to the Device". Then once it boots the display is all wonky like it's displaying hi-res screen with lo-res graphics. Once I open my config file it's all good, even upon full powerdown.


    Little changes to screen dimension and position are however NOT saved/corrected this way upon boot-up, so now I'm stuck always loading my config post boot-up. This can't be normal.


    I've even created a boot-disk which initiates a config write to Firmware in its startup sequence, but then it's also going into that chunky-pixel mode from which I can only save it by loading my saved config.

  • Too bad, I'm suffering from it not Saving and stating that "The Configuration is not Saved to the Device". Then once it boots the display is all wonky like it's displaying hi-res screen with lo-res graphics. Once I open my config file it's all good, even upon full powerdown.


    Little changes to screen dimension and position are however NOT saved/corrected this way upon boot-up, so now I'm stuck always loading my config post boot-up. This can't be normal.


    I've even created a boot-disk which initiates a config write to Firmware in its startup sequence, but then it's also going into that chunky-pixel mode from which I can only save it by loading my saved config.

    Yeah, hopefully Jens chimes in here. He was extremely helpful in troubleshooting - I'm sure he can narrow down the root cause.

  • If there is no connection between the cases, please open a separate thread and describe a short way to reproduce a bug or force unwanted behaviour. TBH, I don't get the actual problem here, it's not clear when error messages appear, what OS is used - it may just be anything, including a missing assign that causes the config tool to not find the last saved config.


    Jens

  • Jens

    Closed the thread.