Indivision ECS v3 joystick stuck up

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.

    TL;DR version of my problem - A500 with Indivision ECS v3 (no joystick attached in CN2) constantly moves up. Works with standard Denise.


    Any idea what could be so close to failure that an original Denise ignores it, but the ECS triggers?


    At first I didn't even think about swapping back to the original Denise, since everything on the forums pointed me towards the multiplexer at U15.

    Multimeter and logic probe (I'm no expert) indicated that it was working, but I socketed U15 anyway. It tested OK with a TL866-II programmer, but I ordered

    replacements from China anyway. They also tested OK, and work with the standard Denise. The ECS v3 is flashed with the latest firmware (20201120) and I have the latest tools (v2.7).


    The A500 is usually maxed out (Gotek, ACA500+, ACA1232, XSURF500) with a CA-PSU and everything is stable. Now I'm testing with only Gotek and Indivision ECS v3, not even extra RAM.


    I should test the Indivision in one of my spare A500s, but this one's my favorite and I would like it to keep working.

    There's multiple cases like this already in this forum: The joystick multiplexer chip is defective and doesn't output a high-enough voltage to reach the minimum required for an LVCMOS input. The chip is located very close to the mouse/joystick ports and is normally a 74LS157. You can replace it with a 74HC157 or 74HCT157 if you don't find the LS version.


    Jens

    Yeah - I found some similar threads. The chips I ordered from China (for U15) were labeled SN74LS157N and validated as 74157.

    I could try ordering from a more reputable source.


    I got another Amiga 500 (also revision 6) out of storage and put the ECS in that. Guess what? Same behaviour...


    This one had a real floppy, so it's completely stock. I ran Amiga Test Kit on it. Controller port with original Denise is OK, but the ECS triggers up and left.


    Do you reckon faulty multiplexers in two Amigas?


    Edited to say I ordered some 74HCT157 from ebay.

    I have two ECS V3s. I use them in two of my A500++ replica builds. No issues. I have 3 more replica boards and alot of original A500 motherboards. Don't think I tried my two ECS V3s in my other A500 motherboards. I have a bunch of Rev.6A and Rev.5 motherboards. And the 5 A500++ replica boards based on the Rev.8A. They are all PAL. I think I have all kinds of spare chips, probably including 74LS157.


    Let me know if you want me to test the ECS V3 in a certain setup.

    Quote from turrican9

    Let me know if you want me to test the ECS V3 in a certain setup.

    Thanks. Well, this setup has worked great a couple of years. Now I have two Amiga 500 Rev6 with 74LS157 multiplexers that are are stuck in up-left joystick movement with ECS but OK with Denise.


    That said, if you found a problematic Rev6 I wouldn't be alone ;)

    This is not about a certain revision - it's about ageing 74xx157 chips that don't comply with their own specs any more. The output voltages are way too low if the chip in question displays the defect I mentioned.


    Jens

    I got some 74HCT157 multiplexers today from a reputable source, but unfortunately that didn't fix the issue either.


    The multiplexers work with Denise, but ECS is stuck with joystick port 2 up and left.


    I tested in two more of my A500 and they all show the same behavior. Thus I have tested the ECS in four of my A500s, all with the same behavior.


    Amiga Test Kit reports joystick port 2 stuck in up left (nothing attached) with ECS v3 and reports a working port (joystick attached) with Denise.


    Is it even possible that there's an issue with the ECS or do I just have bad luck with my Amigas?

    Quote from gaithcif

    Is it even possible that there's an issue with the ECS or do I just have bad luck with my Amigas?

    yes, it is. I now see that you've omitted important information in your first post that I only found now:

    Quote from gaithcif

    Controller port with original Denise is OK, but the ECS triggers up and left.

    Previously, you've only mentioned "up", and I had to assume that all other directions work. This new information indicates that a single pin of the CPLD on Indivision ECS can be pointed to: Pin#5. If you confirm that all other functions of Indivision ECS V3 work (stable picture, config tool doesn't report any error on startup), you could check continuity of Xilinx-pin#5 to Denise-pin #38. Also, since pin#37 of Denise is GND, there may be a whisker between the pins, causing the malfunction. Further, pin#4 of the CPLD is also GND, and a solder-whisker may be present there. Any of the two mentiones spots may explain the behaviour you're observing.


    I can also check our QC pictures if you give me the warranty ID of your unit.


    Jens

    Thanks for your response, I'll do my best to check everything carefully. I sent the warranty id via email.


    It was Amiga Test Kit that detected that up and left are active. The Flashfloppy HxC AUTOBOOT program where I first noticed infinitely scrolls up.


    If Xilink pin 4 & 5 are counted from the lower left corner (the dot) I got continuity to pin 37 and 38 on Denise. No shorts between those pins.

    I cleaned with IPA and cotton swabs and blew some compressed air as well.


    Other than the joystick issue the ECS is working normally.


    The config tool reports no error when I flash the firmware but the command flashtool flash firmware fails.

    Don't know if that's supposed to work with my configuration. Tried from the Gotek and via the ACA500.


    Could be that the commandline flash tool fails because some memory mapper inhibits direct register accesses. Did you actually boot from the rescue disk?


    Continuity to pin #37 and #38 of Denise means that the pin is shorted to GND. This could be either physical (solder whisker), or it's in the CPLD. If you don't see it, we probably need to check it here. Read this article for info about how to send stuff back to us.


    Jens

    This actually gives the clue that something with the CPLD is wrong. If the unit still starts and you can flash firmware from the config tool, the FPGA and flash chip are fine.


    Does the VGA output actually drop out (for a short time) when the rescue disk attempts to boot?


    BTW, I'm not sure if the rescue disk even works under Kick1.3. The whole product is designes for KS/WB 2.0 or higher.



    Jens

    Quote from Jens

    Does the VGA output actually drop out (for a short time) when the rescue disk attempts to boot?

    I can see no dropout - both SCART and VGA show the same thing and the ECS "watermark" is visible while booting.

    Quote from Jens

    BTW, I'm not sure if the rescue disk even works under Kick1.3. The whole product is designes for KS/WB 2.0 or higher.

    The wiki says "Command line flash and config tool requirements: Kickstart 1.3 (v34)".


    With the ACA500+ I can select Kickstarter 3.1. Oddly enough the signature seems to vary every time.


    An ACA1233, attached but disabled in replies #11 & #13. Enabled in reply #9.


    I have a simple ON/OFF switch attached to the jumper to enable/disable the A1200 accelerator depending on what I want to do.

    OK, then the accelerator/software is not the cause. We've got two indications that point to "CPLD bad" - please send the unit, we'll take care of it.


    Jens

    Things are not going well :(


    I sent the ECSv3 back for service and swapped to my A1200 instead. Then I noticed that the sound was terrible...


    After some troubleshooting I noticed that my CA-PSU is not outputting -12V.


    I can easily verify this by measuring pins 4 (+12V) and 11 (-12V) on the LF347 OpAmp on the A1200. +12 is spot on, but -12 shows 5.2V.


    I got one of my spare original PSU:s and started the A1200 with that. Perfect sound and +12, -13.5 on the OpAmp. This A1200 is in pristine condition and I recapped it in 2019.


    How did I not notice this while running the Amiga Test Kit about a million times while troubleshooting the ECS?


    I tested on my A500 and the first 5 seconds or so of the test tune sound OK before the distortion takes effect...


    So... Since I did the ECSv3 troubleshooting with missing -12V from the CA-PSU will it then be fully working in your workshop?


    How do I then proceed with the CA-PSU?

    Quote from gaithcif

    So... Since I did the ECSv3 troubleshooting with missing -12V from the CA-PSU will it then be fully working in your workshop?

    Indivision ECS V3 does not neet the +12V or -12V rail - only the 5V rail.


    Quote from gaithcif

    How do I then proceed with the CA-PSU?

    Just send it back for service. And it would be good if you checked for any expansion that (over-)loads the -12V. It's *very* rare that the -12V regulator fails.


    Jens

    Sure, but so far, I haven't been informed that this combination even arrived. So it's either you've included a nice error description and there were no questions, in which case you'll get an eMail when everything is completed.


    Or there's no error description at all, in which case my employees has strict instructions to not do any unprompted/pro-active action, but instead to wait for clear instructions from the customer.


    Jens