Major havoc repro. Issues

Jamesv833

Well-known member

Donor 4 years: 2017-2018, 2023-2024
Joined
May 8, 2016
Messages
1,807
Reaction score
487
Location
Bellflower, California
Ok I recently finished my major havoc repro pcb and after I powered up in watchdog, I noticed I had cr8 installed backwards. I pulled and tested cr8(tested good) and replaced with a new 1n270 to be safe, still in watchdog. My 6502b was a cmos variant, and while waiting for one from china(still waiting...) I decided to install a 6502 from my tempest pcb, which brought me out of watchdog. Unfortunately I'm having issues with stray vectors, would this have anything to do with using a 6502 in place of a 6502b? Cr8 seems to be part of the linear scaling circuit and is tied to r70,71,72,73,74,75(which all tested good with a leg pulled from the board)c51 which is 10pf(doesn't register on my multimeter) and the dac-08. I've read through numerous major havoc repair threads and can't seem to pinpoint my issue(s). I did replace the resistor that a member found to be incorrect in the bom(I believe it needed to be 1k). Any help steering me in the right direction would be greatly appreciated. Here's a couple pics.
 

Attachments

  • IMG_1183.jpg
    IMG_1183.jpg
    423.7 KB · Views: 62
  • IMG_1184.jpg
    IMG_1184.jpg
    529.8 KB · Views: 49
What "stray vectors" are you referring to? :confused:
Are you using ZOUT to your scope and thus you wonder why the long trace lines?
 
What "stray vectors" are you referring to? :confused:
Are you using ZOUT to your scope and thus you wonder why the long trace lines?

Yes I'm referring to the long trace lines, im not using z-out on my scope, just X and Y. There is a port labeled "Ext Z Axis Input" on the back of my scope. Should I be hooking that up as well? Is this image to be expected without the use of the Z axis? I've been afraid to hook it up to my only working 6100.
 
Yes, you see the retrace lines originating from the center of the screen if simply using only the XOUT and YOUT.
 
My scope pic of MH playin' away ...

MH_scope.md.jpg
 
Phew, I plan on robbing the 6100 out of my tempest tomorrow. I know I'm not exactly in the clear yet, wish me luck. Thank you for all of your help.
 
I finally installed my working 6100. I'm having geometry issues that I hope can be adjusted out. It looks like I'm getting a few errors in self test. Test screen 3 doesn't seem to have a "shrinking white box". It's tough adjusting he pots correctly without having one of those 6100 extension cables, going to have a buddy be my eyes. Here's a pic of the errors I'm getting.
 

Attachments

  • IMG_1185.jpg
    IMG_1185.jpg
    539.5 KB · Views: 42
Flipping the pic correctly, we have -

mh111.jpg


Reviewing the manual, we have -

mh222.jpg


So, yes, ROM checksum errors? :(
NOTE TO SELF: "ignore E". LOL
 
Last edited:
It seems my issues are related to roms 1q,1m/n, and 1n/p. I pulled and tested them on romident, it doesn't seem to match any known mame romset. I'm using the return to vax roms, every other rom seems to check out ok. I noticed theirs a split pad between c192 and c193 that looks as though it had a little solder bridging the pads but after using a wick it seems that it's supposed to be bridged. Can anyone confirm? these roms are in close proximity to the alpha CPU, can this issue be related to the alpha being a 6502 instead of a 6502b? I know others are running return to vax, should these roms verify on romident or is it not part of mame?
 
9s and 1l show the source to be return to vax so I guess it's in their database. Strange how the three I'm having issues with don't come up as return to vax roms, just show checksums,crc32b... I tested the rom files from the return to vax set I downloaded and they show exactly the same thing. I downloaded them from Jess Askey's site. http://gamearchive.askey.org/Video_Games/Manufacturers/Atari/major_havoc.html

Says it's version 1.8

Speech seems to work.
 
I would assume the modified ROMs have checksum errors.
Interesting that the speech works. Not sure if anyone has confirmed that before ... (for that, I've been waiting for "confirmed" ROMs).
 
Ok I reprogrammed the problematic roms and although they now verify on romident, I'm still having errors in self test. While it's the same roms giving me issues, some of the letters/numbers on the right have changed. Here's a new pic.
 

Attachments

  • image.jpg
    image.jpg
    431.7 KB · Views: 30
I would assume the modified ROMs have checksum errors.
Interesting that the speech works. Not sure if anyone has confirmed that before ... (for that, I've been waiting for "confirmed" ROMs).

Are you saying it's expected to see these errors using the "modified roms"? Everything seems to be ok except the image pot adjustments I need to make. I was able to get the shrinking white box to display by increasing the screen pot on the high voltage transformer.
 
Yes, mod'd ROMS likely have errors in self test.
Fix the display and play some games. If it does not crash, then OK.

As for ROMIDENT, not sure what's going on there as I've been checking other ROMs too (for another game) and get no confirmed results. I assumed it was the ROMs ... but maybe not?
 
Ok cool, even the messed up screen geometry hasn't stopped me from playing.lol.
Thank you for all of your help!
 
Ok, I received a r6502b today from China. Although I wasn't holding my breath that this would fix my issues, I wanted to have the right component populated. Swapped out a standard 6502 that I was using temporarily, and I'm still having some issues. While the game is very much playable, I would like to resolve these issues. During attract I don't seem to get any of the scrolling info that gets displayed under the Major Havoc title(storyline,enemies...) A lot of the text seems to be misshaped, and I can see a slight "jitter" within the text as well. I'd love to play return to vax, but I'm thinking I should be using one of the standard rom sets for troubleshooting. What set does everybody seem to be running? Can anyone give me any ideas based on the image/description where I should be looking within the circuit to help diagnose these issues? Here's a pic of the title screen about 10 seconds after it finished drawing the incomplete? Title.
 

Attachments

  • IMG_1186.jpg
    IMG_1186.jpg
    402.9 KB · Views: 33
Notice that the scrolling text that's missing from the 6100 does show up on the scope... which tells you that it's a problem with the Z circuit. The 6100 respects the Z and blanks those vectors, but the scope shows everything coming out of X and Y.

But if Z is broken why do you see anything? Because it's a problem with the Window circuit, whose job it is to blank only certain (programmer-defined) areas of the screen. it was used in Battlezone, for instance, to keep the battlefield out of the score and radar area at the top of the screen. Easier to do it by setting a window boundary than by trying to calculate and truncate each vector that might be going too far. It is used in MH for the credits crawl as well as the scrolling maze phase of the game... which is fascinating to watch on the scope. On the scope it's jerky and frenetic as extra segments are drawn past the top. With the Z controlled by the window circuitry, you get a nice stable boundary which really improves the perception of smooth scrolling.

The Window circuit is on sheet 8A of the schematics, so make sure all those components are correct and all pins are soldered. To troubleshoot you'll have to probe through it while attract is running and catch it while it's being used. When you have a signal on VYD11, make it makes it all the way into pin 16 of 12D (that's when the boundary is being set by the programmer). It is supposed to turn on the switch at 12D and connect the Y Window signal (from the Y Output stage) so it gets amplified by 12C and 9K and sent to the RGB Output circuit to create the ZBLANK signal coming out of 8K. That is used to clear (or not) the flip flop at 13J which is the final arbiter (before the buffer at 13K) for the Z levels for all three colors (sheet 9A).

As for CR8 going in backwards, and possibly contributing to malformed and/or shaky vectors, I'd replace Q7 and perhaps Q6, and try a new DAC-08. It's all in the circuit called Linear Scaling but keep in mind it used for all scaling factors including one (in other words, it's used all the time).

PS. Yes the speed of the 6502 can wreak havoc on the game if it cannot run fast enough. The 6502 will run at up to 1MHz, the 6502A up to 2MHz, and the 6502B up to 3MHz. You may get lucky with an CPU that can exceed its rated speed, but it's not guaranteed.
 
Sweet, thank you for the very detailed post. Ill place an order on those possibly jeopardized components and post back with results. Thank you for the run down of the window circuit, can't wait to get this game in the house. You guys are awesome!
 
Back
Top Bottom