NEOSD MVS Support thread !

[Razoola]

These games have a special banking system. Have you flashed a new beta firmware? If so I would suggest it is related to that.

 

[GadgetUK]

Thanks!! Despite that concerning you, it's kind of positive news in that I am not alone with CRC issues. It could be related to the firmware update, or another update might be required, but the NGDEV team are looking at this already, so I wouldn't worry too much at this stage.

 

[neodev]

Sure!
As a matter of fact, I've been running CRC checks for EVERY game, and I usually repeat them again every time I swap a game. They have always resulted in OK for all ROM regions and banks, except for the games/bootlegs/homebrew not included in UniBIOS 3.2 CRC database.

However, I've just performed the CRC checks you asked, and I'm afraid this time it's been different... I'm kinda freaked out...

Firstly, I flashed Metal Slug 5, then ran CRC check: ROM bank 0 was NG, strikingly as I've run the very same CRC check several times a few days ago, and all ROMs were OK at first try. Repeated the check immediately: ROM bank 0 was NG, and with a different value from previously. Again: 0 NG, completely different value.

Next, I went back to game select menu and flashed KOF2003. In a first CRC check, ROM banks 2 & 4 were NG, the rest OK. Repeated the check immediately: ROM banks 2, 3 & 4 NG, and 2 & 3 with different values from previously. Again: 2, 3 & 4 NG, completely different values.

Time for switching off the MVS and re-inserting the NeoSD, looking for a perfect and smooth insertion. Whitout re-flashing, I performed 3 consecutive CRC checks for KOF2003: Now all ROMs were OK and passed the check.

Finally, I re-flashed Metal Slug 5 to run new CRC checks. ROM bank 0 was again NG, with the same value than in the first test. Again, and at last all ROMs were OK. Once again, and bank 0 was again NG, with a different value from previous ones.

What surprises me most is that last week I performed such CRC checks for both these games several times, always obtaining OK for all ROMs.


Any insight into what can be happening here? Why CRC checks were OK before and now they measure variable values for some ROM banks? The only condition that has been altered is the new beta firmware and my recent testing with handmade .neo files from homebrew and new hacks.

Let me explain. The new beta firmware contains several changes to try to solve the crc issues gadgetuk was experiencing. It's still under testing, and more changes are being done, as it didn't solve the issues for him, so we'll try to find a different approach.

 

[GadgetUK]

These games have a special banking system. Have you flashed a new beta firmware? If so I would suggest it is related to that.

You can get the same issues with other games not using that banking system though - for example, after 10 CRC checks of Burning Fight I see the problem, or Ghost Pilots. On lower meg count carts you have to repeat CRC check many times, but often it will glitch. I am not too worried, it might even be solved with some grounded copper tape if it cannot be fixed in firmware. Chances are it's just a minor timing glitch that needs ironing out.

 

[Razoola]

Yes, it needs to be remember here that neodev has been trying to fix the crc32 issue that GadgetUk has with beta firmware. One connot rule out that GadgegUK has a faulty PROG board given he has identified slight damage on the PCB. I think it was mentioned a replacement NeoSD was going to be sent. Did that happen already and there is still a fault with GadgetUK?

 

[GadgetUK]

Yes, it needs to be remember here that neodev has been trying to fix the crc32 issue that GadgetUk has with beta firmware. One connot rule out that GadgegUK has a faulty PROG board given he has identified slight damage on the PCB. I think it was mentioned a replacement NeoSD was going to be sent. Did that happen already and there is still a fault with GadgetUK?

That slight damage is related to PCM! But that doesn't mean my PROG board doesn't have a fault, but I think unlikely. I think the CRC issue might have come into play with the lines fix. No point in speculating I think, just best to let NGDEV do more testing.

EDIT: Yes, a new cart is on the way - but I am fully expecting that to behave the same way if I am honest, unless perhaps the code on it is different to the code I am testing at the moment.

As part of testing it might be useful for me to 'rewind' one of these carts back to the original firmware so I can re-test CRC issues (accepting that the lines and KoF2003 graphical issue would return). I think that would be a useful test.

 

[Razoola]

That slight damage is related to PCM! But that doesn't mean my PROG board doesn't have a fault, but I think unlikely. I think the CRC issue might have come into play with the lines fix. No point in speculating I think, just best to let NGDEV do more testing.

Ok, I did not know that issue was tied back to PCM and not PROG lines. But the issue with kof2003 attract mode and the girls hair... From the video you had it is clear incorrect sprite data going into VRAM rather than the NeoSD passing incorrect sprite data to be displayed. That is quite different than the hortz lines that the first firmware change fixed. If I remember at that time you were the only person with crc32 issues or were there others also?

 

[neosd]

Hello GadgetUK,

You were the only one reporting crc issues, we have been trying a lot of different things to solve your issue, even there are high chances your neosd is faulty, thats why we have sent you another neosd in advance without return us yours, by UPS yesterday.

We have sent you a BETA firmware, we have sent this same BETA firmware to Sceptre_JLRB to have a contrast.

As neodev has stated, any results of this BETA firmware has to be taken in consideration as what it is, BETA.

So we are glad that you guys are reporting issues with this BETA firmware. But please, don´t panic finding a crc issue on the BETA firmware. This is normal, we are trying different things and some things that were working, may get broken.

Thanks

That slight damage is related to PCM! But that doesn't mean my PROG board doesn't have a fault, but I think unlikely. I think the CRC issue might have come into play with the lines fix. No point in speculating I think, just best to let NGDEV do more testing.

 

[GadgetUK]

Ok, I did not know that issue was tied back to PCM and not PROG lines. But the issue with kof2003 attract mode and the girls hair... From the video you had it is clear incorrect sprite data going into VRAM rather than the NeoSD passing incorrect sprite data to be displayed. That is quite different than the hortz lines that the first firmware change fixed. If I remember at that time you were the only person with crc32 issues or were there others also?

I think part of the problem is I haven't been posting updates on here because I thought I was the only one experiencing problems. The last update I applied fixed the girls hair problem - that for me is resolved but CRC issues remained. I worked out today that the slight inperfection in the trace on my PROG board was related to PCM and there's no way that is causing the CRC issue. Then as Sceptre_JLRB confirmed he's getting the CRC issue too leads to one of three possibilities:-
1) A recent firmware update (which isn't finalised yet - it is being tested by myself, Sceptre_JLRB and a few others) - might be the cause.
2) Maybe on some MVS this problem occurs and perhaps not on others.
3) The problem has been there all along, but its not easy to replicate unless you test the CRC over and over and over (like 10 or more times).

 

[GadgetUK]

Hello GadgetUK,

You were the only one reporting crc issues, we have been trying a lot of different things to solve your issue, even there are high chances your neosd is faulty, thats why we have sent you another neosd in advance without return us yours, by UPS yesterday.

We have sent you a BETA firmware, we have sent this same BETA firmware to Sceptre_JLRB to have a contrast.

As neodev has stated, any results of this BETA firmware has to be taken in consideration as what it is, BETA.

So we are glad that you guys are reporting issues with this BETA firmware. But please, don´t panic finding a crc issue on the BETA firmware. This is normal, we are trying different things and some things that were working, may get broken.

Thanks

Thanks for everything you've done so far - don't worry, I am not panicking yet =D Might be worth rolling my current firmware back to the very original release firmware - just to see if the CRC issue disappears. Also worth anyone else with a Neo SD to test their CRCs (in particular KoF2003 and MS5 - it's easier to get an error because of the volume of prog code checked, on a single bank P ROM game you have to repeat the test many many times to see a glitch in CRC, and in some it always comes back OK).

 

[neosd]

GadgetUK,

You were the ONLY reporting CRC issues, we thought something was wrong in your setup or with your NEOSD. But we decided to investigate this instead just replace your neosd and forget about it, cause we wanted to know whats happening.


We have done deep changes just to know if we had a software issue, sending you firmware updates for that, trying so many things that we knew may broke working things, thats why we havent sent this update to everyone else.

We have sent you a BETA firmware update, ONLY you and Sceptre_JLRB got this BETA for a reason :

We needed someone else to get contrast on the tests, Sceptre_JLRB getting CRC issues now when he didnt have those in the previous firmware is NORMAL, we were expecting that based on the changes we have done.

This is NOT related with your issue.

I hope you understand that.

Thanks

I think part of the problem is I haven't been posting updates on here because I thought I was the only one experiencing problems. The last update I applied fixed the girls hair problem - that for me is resolved but CRC issues remained. I worked out today that the slight inperfection in the trace on my PROG board was related to PCM and there's no way that is causing the CRC issue. Then as Sceptre_JLRB confirmed he's getting the CRC issue too leads to one of three possibilities:-
1) A recent firmware update (which isn't finalised yet - it is being tested by myself, Sceptre_JLRB and a few others) - might be the cause.
2) Maybe on some MVS this problem occurs and perhaps not on others.
3) The problem has been there all along, but its not easy to replicate unless you test the CRC over and over and over (like 10 or more times).

 

[Razoola]

I think part of the problem is I haven't been posting updates on here because I thought I was the only one experiencing problems. The last update I applied fixed the girls hair problem - that for me is resolved but CRC issues remained. I worked out today that the slight inperfection in the trace on my PROG board was related to PCM and there's no way that is causing the CRC issue. Then as Sceptre_JLRB confirmed he's getting the CRC issue too leads to one of three possibilities:-
1) A recent firmware update (which isn't finalised yet - it is being tested by myself, Sceptre_JLRB and a few others) - might be the cause.
2) Maybe on some MVS this problem occurs and perhaps not on others.
3) The problem has been there all along, but its not easy to replicate unless you test the CRC over and over and over (like 10 or more times).

Yes I though this is the case in you and others testing a beta firmwares. The kof2003 issue you need to be carefull of, in a way it was very lucky that corruption in the bank caused an easily visable game glitch. Much of the time that would not be the case so its better to say the kof2003 issue is not fixed until the crc32 test passes 100% of the time.

My opinion at the moment is the attemps to fix your crc32 issue via beta firmwares has broken what was previously working firmware in relation to banks for other beta testers of that firmware.

It may well turn out that there is indeed an issue to be fixed via firmware in relation to banks but there is just not enough to be sure ATM. I feel until you reveive the replacement NeoSD and it is shown to give the same results on your system firmware changes in your case are shots in the dark. It may have to also wait until the board in question is returned to the neosd guys. So it could be in the firmware, a fault on the NeoSD or maybe something related to the NeoGeo.

Personally I think the best cause is going to be the 2nd (related to the NeoSD board you currently use).

 

[GadgetUK]

GadgetUK,

You were the ONLY reporting CRC issues, we thought something was wrong in your setup or with your NEOSD. But we decided to investigate this instead just replace your neosd and forget about it, cause we wanted to know whats happening.


We have done deep changes just to know if we had a software issue, sending you firmware updates for that, trying so many things that we knew may broke working things, thats why we havent sent this update to everyone else.

We have sent you a BETA firmware update, ONLY you and Sceptre_JLRB got this BETA for a reason :

We needed someone else to get contrast on the tests, Sceptre_JLRB getting CRC issues now when he didnt have those in the previous firmware is NORMAL, we were expecting that based on the changes we have done.

This is NOT related with your issue.

I hope you understand that.

Thanks

OK, don't get stressed =D

It's just a case of the more feedback people provide the easier it will be to identify a fix.

 

[fille1976]

i just checked twice my crc of metal slug 5,all banks ok.
how many times did you test it gadget?
and just tested a third time,all banks ok.
this is on mv1b unibios 3.3
checked fourth time,all ok.

 

[neosd]

Hello,

Take in mind that maybe we have been hunting ghots for more than a week. The time we put on those things, is time we dont put into adding more features.

I know that is frustrating being the only one with crc issues and that you would like to not be only one with those just to be 100% sure there is an issue there. But right now, appart of spectre that got the same beta firmware you have, you are the only one reporting a CRC issue, this is a fact to atleast take in consideration.

Anyways, we have sent you another neosd just to be sure where the problem is.
We are going to solve that, no matters where it is, just allow us to work, to deliver a solution, we cant keep rushing things anymore.
This whole thing has stressed us so much and we should have taken much more time before sending you a beta firmware.

I hope you understand that we are just doing our best to address the issue and deliver you a solution, maybe our mistake here has been not sending a replacement from day one just cause we wanted to know where the issue was.

Thanks

OK, don't get stressed =D

It's just a case of the more feedback people provide the easier it will be to identify a fix.

 

[GadgetUK]

i just checked twice my crc of metal slug 5,all banks ok.
how many times did you test it gadget?
and just tested a third time,all banks ok.
this is on mv1b unibios 3.3

Just do it several times to be sure. It takes ages but its worth ruling it out. In my case I can find that it passed OK for 2 or 3 tests and then NG on one bank.

 

[GadgetUK]

Hello,

Take in mind that maybe we have been hunting ghots for more than a week. The time we put on those things, is time we dont put into adding more features.

I know that is frustrating being the only one with crc issues and that you would like to not be only one with those just to be 100% sure there is an issue there. But right now, appart of spectre that got the same beta firmware you have, you are the only one reporting a CRC issue, this is a fact to atleast take in consideration.

Anyways, we have sent you another neosd just to be sure where the problem is.
We are going to solve that, no matters where it is, just allow us to work, to deliver a solution, we cant keep rushing things anymore.
This whole thing has stressed us so much and we should have taken much more time before sending you a beta firmware.

I hope you understand that we are just doing our best.

Thanks

I totally accept what you are saying and highly appreciate the time taken to investigate the problem.

 

[fille1976]

did it a fifth and sixth time,all ok here.
will try later kof 2003 and again ms5

 

[GadgetUK]

Thanks! I've got a 2nd MV-1FZ setup here now, just waiting on a 27C1024 so I can put a copy of my unibios on it (temporarily) to allow me to test checksums.

 

[neosd]

I totally accept what you are saying and highly appreciate the time taken to investigate the problem.

Hello,

We really apreciate the effort and time you have put here, trying to help us. You decided to help us trying things instead of just ask for a replacement, thunbs up for that, you have our infinite gratitude.

Both of us have put a lot of hours here, we should be near to the conclusion now with the replacement NEOSD.
We will know more next week, for sure.

Thanks

 

[neosd]

Thanks! I've got a 2nd MV-1FZ setup here now, just waiting on a 27C1024 so I can put a copy of my unibios on it (temporarily) to allow me to test checksums.

Hello,

Take in mind that with the current BETA firmware, you are probablly going to see the same CRC fails.
Lets wait until the new NEOSD arrives, so we can have a picture of whats happening.
Thanks

 

[GadgetUK]

Hello,

Take in mind that with the current BETA firmware, you are probablly going to see the same CRC fails.
Lets wait until the new NEOSD arrives, so we can have a picture of whats happening.
Thanks

Thanks, I suspected you were going to say that lol. So the chances are the other MV-1FZ will prove nothing =/ It will be useful for testing never the less. I've also got 2 other MV-1FZ boards which I've ordered some parts for - so in theory I can test on 4 x MV-1FZ boards =D

 

[neosd]

Hello,

You cant roll back to the previous firmware aswel, cause BETA firmware wont allow you to do that. We can send you a firmware to come back to the previous version, but for now, lets stay the board on the BETA firmware so we dont mix things

The other MV-1FZ will prove a lot, with the old firmware, we will try it, but lets not open so many fronts at the same time. Also i hope those boards you have purchased for parts are working

Thanks

Thanks, I suspected you were going to say that lol. So the chances are the other MV-1FZ will prove nothing =/ It will be useful for testing never the less. I've also got 2 other MV-1FZ boards which I've ordered some parts for - so in theory I can test on 4 x MV-1FZ boards =D

 

[GadgetUK]

I already had 2 boards that were working but needed RAM chips - never got around to replacing the RAM on them, but they do work. The other MV-1FZ is fully working and tested. So 2 definitely work 100% now, the other two just need the video SRAMs replacing.

 

[Sceptre_JLRB]

Ufff, what a relief... I was thinking my NeoSD or, even worse, my MVS itself was going awry. ^_^U

No problem at all with this testing, it'll help tracking the problem. All I can confirm is I'm having those CRC issues ONLY with this beta firmware update (I am a truly CRC-check maniac, and never before had this issue in many MANY CRC checks carried out). The horizontal-lines glitch is gone (as it did with the previous line-fix firmware), but there is no girl-hair KOF2003 glitch. So the latter one could be an issue exclusive from GadgetUK setting.

Glad to be contributing to your testing and problem solving, guys. Now, would it be possible to return to the previous firmware, with no CRC issues (and probably SVC Plus2 & Super Plus working properly)?

Thanks!