What's new
By:

Pulstar and Ironclad conversions help

Ok it wasn't that bad, the dirty pins weren't making good contact to the burner on the reads. So far only C5 is bad for sure, and maybe C1(erased before double checking). I might as well pull all the V roms too while I'm at it. Basically redoing the whole cart. Definitely will verify after writes from now on. The monitor test cart and Ironclad went really well, so I was starting to trust the writes a little too much. That and I have do 8 banks per 322 eproms, so I just did "samples". I did convert the 3 DIP switches to change banks to a rotary DIP switch, so all I have to do is turn the dial instead of flipping DIP switches for every bank.
 
Well crap. I desoldered every eprom and tested them. They all checked out except for C5 and maybe C1. After burning C1 and C5 again and verifying, soldering them back in, it looks exactly the same. :( Sprites still look the same as before. Does Pulstar have encrypted roms? With all of the non-JEDEC EPROMs, I have to use the adapter on my TL866, (only M1 is JEDEC) and I have trouble getting the OE pin to detect. I continued with the reads and file comparison to the original all came back identical to the original. So I chucked it up to not needed OE for the burner, hopefully that's not the reason. Because most EPROMs showed that pin not detected.

Since the burn checked out, that means I have a bad soldering somewhere maybe? It'd be difficult to get the same bad soldering twice in a row?

So quick recap for better readability:

1. All jumpers are correct.
2. Initial NEO GEO logo UNIBIOS, jingle logo, and subsequent 3rd logo all garbled, but FIX layer is ok.
3. All sprites garbled, desoldered all EPROMs and read the contents out and compared to original MAME rom, only C1 and C5 were different, they were reburned and retested
4. Same exact symptoms are still present.

Maybe I try the binaries I read from the EPROMs in MAME? No point really, I checked the md5sum of the files that were read back, they're the same as MAME files.
 
Last edited:
The SNK customs on these boards do fail. Just to be sure, did this PCB work good on the original game before you started this conversion?

Make sure the edge connector on the CHA PCB is clean.

Inspect for any broken traces or lifted traces, especially around where you've been soldering and re-soldering chips. These MVS boards don't have plated through holes on the EPROM locations which makes de-soldering easy but it also makes damaging the traces very easy as well.

It's also possible you have a bad ROM despite it verifying ok. I've run into this a few times where a chip reads fine in the programmer but fails in-circuit, sometimes this is due to something obvious, like a weak pin on the precipice of breaking off other times it's a mystery (maybe a weak internal connection?).

The best thing to do here is swap out the C-ROM data 1 ROM at a time in MAME with a blank ROM file, do this until the bad sprites/tiles from the blank file in MAME matches the same bad sprites/tiles on your PCB... then replace that ROM with a fresh ROM.
 
Thank you twistedsymphony for your continued assistance on these conversions! I did indeed check the game for 30 min prior to desoldering, it worked fine without issue. 1 pad for the M1 rom did lift, but I patched it back with a magnet wire. During resoldering, I looked through all the areas for other issues, I didn't find any. But will double check to make sure.

Ahh, the just-barely-hanging-on rom issue. Now I remember from your other thread. I will do the blank ROM file check using MAME and see if I can isolate. If I find any, this will have to wait as I'm fresh out of 322's. The next shipment somehow ended up in Hawaii, and is being redirected.
 
I thought of another thing. did you change any of the jumper settings from the original game? if so it might not hurt to double check that the solider joints are good on the jumpers/no unintended bridges.
 
twistedsymphony said:
I thought of another thing. did you change any of the jumper settings from the original game? if so it might not hurt to double check that the solider joints are good on the jumpers/no unintended bridges.
Click to expand...
I did have to change quite a few. Initial continuity checks were good, but I didn't test beyond the soldering point. Perhaps I can find where the pad is coming from, and test from that to make sure the soldering is good and no bridges.

Are the game rom zip files just a simple zip? I'm having trouble getting mame to load even a unzipped then rezipped rom without any modification. Gotta read up more on how the roms are created/used. My first test was simply copying C2 rom into C1 rom, that didn't load, so I restored C1 and rezipped it, but that didn't work either. The original zip file worked of course.
 
yes, just unzip to a folder, mame can run from the unzipped folder as long as it's named the same as the zip (so "pulstar.zip" becomes a folder called "pulstar").

All files MUST exist be the correct size, however MAME will prevent you from loading from the menu if the CRCs don't match so you must launch the game directly from the command line.

you might also need to launch with the debugger flag? I don't remember
 
There was nothing wrong with my EPROM. The rom file was bad! I tried 3 different rom sets from MAME and they all showed exactly the same issue I saw on my conversion cart under MAME. Then I tried a FBA rom set, it looked just fine. So looks like I'll be burning all of them again. :)

actually, looking at the md5 sums, only C1-C6 and V1 are different. So I only have to burn those again.
 
Last edited:
MAME2003_Reference_Set_MAME0.78 and MAME v0 37b5 roms non-merged also tried an unknown set. FBN v0 2 97 44 roms non-merged worked in latest MAME release.

Just so I don't desolder a 3rd time, not sure if the pcb can take anymore abuse. Is my V rom splitting done correctly? The mame rom has v1 and v2. I split them into V1 -> v1 v3, and V2 -> v2, v4, to burn.
 
Last edited:
wow that's ancient. we're on MAME 231 and you were using ROMs from MAME 78 and 37

I think 37b5 is from 2000, so it's over 20 years old at this point.
 
twistedsymphony said:
wow that's ancient. we're on MAME 231 and you were using ROMs from MAME 78 and 37

I think 37b5 is from 2000, so it's over 20 years old at this point.
Click to expand...
Yea most of the MAME stuff I have are old because that's the ROM set that works for most of the games I used to play. Now I got another issue, the adapter seems to not be working anymore. It programs both bank 1 and 2 at the same time for some reason. Not sure if it's the adapter's issue or maybe the programmer's issue. At this point, I may have a bad adapter, a bad burner, or a batch of damaged EPROMs. Leaning and hoping towards the first. So this conversion is on hold again. :(
 
rcboosted said:
Now I got another issue, the adapter seems to not be working anymore. It programs both bank 1 and 2 at the same time for some reason. Not sure if it's the adapter's issue or maybe the programmer's issue.
Click to expand...

Didn't you replace the jumpers with a rotary switch? I bet it's that.
 
rcboosted said:
Yea most of the MAME stuff I have are old because that's the ROM set that works for most of the games I used to play
Click to expand...
that's not how that works.

the newest version of MAME should have the exact same game ROM data as an old version, but sometimes there were bad dumps and they get fixed, and as time goes on other versions of the dumps are found and added.

So maybe version 78 only has 1 dump from a Chinese version with a bad ROM dump, but version 228 will have dumps from several regions with the bad dumps fixed.

somtimes they get ROM naming wrong too, so even if the data is the same the names might be incorrect when compared to the real hardware, and that is also fixed in newer versions.

because of that last point, if you're playing, you should be using a ROMset that matches the version of MAME you're running. I believe the only reason those old ROM sets are still kicking around is because of some Raspberry Pi builds that only work on ancient versions of MAME. But if you're doing conversions or using MAME to troubleshoot things on PC you 100% want to be running the latest MAME version and the latest ROM set because that's going to be the most complete and the most error free.
 
nem said:
Didn't you replace the jumpers with a rotary switch? I bet it's that.
Click to expand...

That's the first thing I suspected as well, but continuity checks out and I was able to dump the entire cart, reassemble the bin files and verify the md5sum to the original just fine. But to completely eliminate the possibility, I swapped back to the dip switches, and it was still appearing to program 2 banks at a time, it could just be read issue. I messaged the creator of the board, and it might be the circuit for handling 322 chip or maybe my EPROMs are bad now. I accidentally programmed a 160 with the 322 switch flipped or vice versa. That for sure would damage the chip being programmed. But it may also damage the circuit or the burner, so subsequent burns are bad or damaging. He's checking the circuit to see what effects/consequences are. This is making me want to get the 610p, the adapter is too easy to make mistakes and break things as I burned maybe 120+ banks for Pulstar. But then there's a $300+ price tag. :)

twistedsymphony said:
that's not how that works.

the newest version of MAME should have the exact same game ROM data as an old version, but sometimes there were bad dumps and they get fixed, and as time goes on other versions of the dumps are found and added.

So maybe version 78 only has 1 dump from a Chinese version with a bad ROM dump, but version 228 will have dumps from several regions with the bad dumps fixed.

somtimes they get ROM naming wrong too, so even if the data is the same the names might be incorrect when compared to the real hardware, and that is also fixed in newer versions.

because of that last point, if you're playing, you should be using a ROMset that matches the version of MAME you're running. I believe the only reason those old ROM sets are still kicking around is because of some Raspberry Pi builds that only work on ancient versions of MAME. But if you're doing conversions or using MAME to troubleshoot things on PC you 100% want to be running the latest MAME version and the latest ROM set because that's going to be the most complete and the most error free.
Click to expand...

I understood how the ROM sets and they must match the MAME versions. (it did take me until recent years to figure that out) But NeoGeo has always been the "easiest" of the emulations to get working, so I just used 2003 pack due to familiarity and how it always worked. Pulstar was not one of the game I played before. :) I do want to get the latest ROM set for future conversions now that I've learned my lesson. :) So I'll get that going.
 
Does anyone have the patched p1 for Ironclad?

I know theres the dim Wii VC rom, the proto default spanish on euro one and a patched one.
 
Just some updates. After 3 bad 322s, 1 bad 160, a new adapter and a new burner, I have a (mostly) working Pulstar. At this point, I don't really know what happened to the double bank burning issue. I attempted to eliminate everything by getting new equipment one by one. In the end, all I know is:

1. original burner is not damaged (now I have 2 TL866 :) )
2. original adapter probably was not damaged (now it is damaged after desoldering headers with very tight fit)
3. original ZIF socket might be iffy coupled with #4 below
4. I used a new lead-free(SN100C) solder, this became a MAJOR pain in the butt as it would NOT clean off the legs to get good contact with burner
5. this adapter will always show OE pin 21 to be not detected. I used a new burner, new adapter, new ZIF, new 322, it showed OE pin not detected, but burns fine.

What I think might have happened is the lead-free solder might have left some legs with bad contacts to the burner, it might have been ok during initial check but during programming it's intermittent which caused the chip to get damaged or routed signals to 2 banks, and this caused it to program 2 banks at a time. What confused me the most was because multiple 322s and 160 were showing the same issue, so it appears it's a burner/adapter issue. Although reads are perfectly fine from masked rom as a test. Eventually with the new burner and adapter and ZIF socket, multiple cleaning of EPROM legs, I found some EPROMs were damaged while most survived.

Lessons learned are, clean the legs, double check (322 vs others) programming switch, double check pin contact before programming, double check burned eprom.

As for the game, I saw some slow down during game play with lots of sprites and some weird graphical issue during the slow down, not sure if this is normal as I've never played Pulstar before. I will play the MAME version to confirm. The graphical issue looks like a big chunk of the screen is skewed/shifted. Like someone sliced the screen at 45 degree, slide that chunk a bit so now you have some part of the screen that is shifted. It only happened during the slow down, so it could be just caused by the slow down? Background music is also strange as if it's not playing the correct track almost, but again, I've never played Pulstar before, so I don't know if that's normal, to be confirmed. It's possible that my V rom interleaving is wrong. I split V1 file into V1 V3 EPROM and V2 file into V2 V4 EPROM. Worst case I swap V2 and V3.
 
You must log in or register to reply here.
Back
Top