ST: G-Stream G2020

[trap15]

I should add this to the ST since I noticed it the other day:

When an enemy drops a powerup, the powerup type is initialized to a random (probably time-based) powerup. However, when it spawns it always starts at red, then will change to the next color in the actual order. So you could get Red, Red, Blue, Green; or Red, Blue, Green; or Red, Green, Red, Blue.

A bit silly

[rtw]

knarzing

Fantastic onomatopoetikon

Even though they did not change the revision date, can you tell me if your labels are the same as mine ? My program ROM seems to be labeled 02 ? Of course a quick dump of the 27c160 would be best

http://gameordie.com/rtw/g-stream_pcb.jpg

@BPzeBanshee, thanks for the info on how you viewed the sprites!

[BPzeBanshee]

So there was no difference in the date listed in the test menu? Damn.

I'm no expert, but according to the documentation in MAME's gstream.c:

Code: Select all

+------------------------------------------------+
|        GS_SND_01   RAM4  16MHz     RAM1 S2 BT1 |
|  AD-65 GS_SND_02 GS_PRG_01 E1-32XT RAM1        |
| VOL    GS_SND_03                               |
|J AD-65 GS_SND_04                      GS_GR_07 |
|A    1MHz     GS_PRG_02 XC95288   RAM2 GS_GR_08 |
|M             GS_GR_01  RAM2           GS_GR_09 |
|M      RAM4   GS_GR_02   XC95288 27MHz GS_GR_10 |
|A      RAM4   GS_GR_03           54MHz GS_GR_11 |
|              GS_GR_04       S3        GS_GR_12 |
|    GAL       GS_GR_05       RAM3 RAM3 GS_GR_13 |
|    GAL       GS_GR_06       RAM3 RAM3 GS_GR_13 |
+------------------------------------------------+

- it would seem that there are two program chips. Going by this layout your GS_PRG_01 is the one marked with a black circle, which I find a bit odd considering how well marked the rest are.

[rtw]

it would seem that there are two program chips. Going by this layout your GS_PRG_01 is the one marked with a black circle, which I find a bit odd considering how well marked the rest are.

What I find odd is that PRG1 is loaded as 8 bit but PRG2 is loaded as a 16 bit. And I just noticed that the Xilinx chips are not FPGAs they are CPLDs i.e. the code is stored inside them.

[BPzeBanshee]

Interesting. I had no idea that this was the case. Yet another strange oddity about the board's design.

I guess the best way to figure out if it's immediately any different from what's in MAME is to fire up the thing and see what the title screen looks like. So far these two pictures are our only absolute evidence of differences in the boards (left is MAME, right is from iyashi.blog.so-net.ne.jp:


EDIT: Replaced second image with the other one I found for posterity. I've got this one downloaded in case the blog ever removes it.

[THE]

@BPzeBanshee
I only see one pic.

If could contact the try tower guy. Is it known if they still have a copy?

The 8-bit P1 ROM is most likely related to the (horrible) sound.

My eprom burner setup is currently in storage. Can take a while...

[BPzeBanshee]

The other one takes a bit to load, and is loading for me here. Another's on Google Images if you type in the game name. In short: the G-Stream ship is on the title screen, in large.

[THE]

Doesn't work for me.

But here is the link for future reference:

http://file.trytower.blog.shinobi.jp/P1010420.JPG

My copy's title screen doesn't show this cool big ship either.

[EmperorIng]

So I would assume that scoring well in this game has less to do with keeping up a consistent medal chain, a la Raizing, and more to do with building up the chain to just the right point to let out a cancel bomb and rack up a bunch of 1000 medals, no?

[system11]

What size are those chips some people have said overheat? Akasa sell northbridge/southbridge heatsink kits which have 2 sizes of heatsink and thermal pads to fit them with. Might be worth a look - example:

http://www.overclockers.co.uk/pimg/HS-0 ... 19_600.jpg

[BPzeBanshee]

So I would assume that scoring well in this game has less to do with keeping up a consistent medal chain, a la Raizing, and more to do with building up the chain to just the right point to let out a cancel bomb and rack up a bunch of 1000 medals, no?

Seems that way. Medal chaining is almost good for parts where it's sanely possible ie. with few ships. Most of Stage 1 can be chained with a bit of careful management along with Stage 2, but it's 3-A/3-B where the emphasis seems to go to cancel-bombing, or getting +5000 hidden chain bonuses.

EDIT:
So I was on IRC with a few folks (trap15, system11 et al) last night and I managed to figure out a special segment of Stage 1.

The trick so far seems to be to make sure you kill EVERYTHING in the segment when the popcorns come spinning in from the left at the beginning, even using bombs if you have to (sometimes it's a must, bonus enemies like to appear here which may be linked to the first popcorns that appear as the stage starts).

If you pull it off, the stage will scroll to the right, and two large green ships shooting propellers which can then in turn be killed for medals will appear. Kill those green ships quickly enough and you'll lock yourself in to a nice score-friendly series of segments before returning to a modified version of Segment 3.5 where a green ship will be around the very edge of the top of the screen. Killing them will spawn the large 'rush' planes like in the normal segment, followed by the boss.

Here's a video of the above in action made by mtune: http://www.youtube.com/watch?feature=pl ... bBnvoGnDbw

[rtw]

The trick so far seems to be to make sure you kill EVERYTHING in the segment when the popcorns come spinning in from the left at the beginning, even using bombs if you have to (sometimes it's a must, bonus enemies like to appear here which may be linked to the first popcorns that appear as the stage starts).

Managed to pull this off today, thank you for the information

The Xilinx XC95288 HQ208 CPLDs measure 25mmx25mm. They are HQFP (Heat Sink Quad Flat Pack) so they have a heat sink in them already. According to Xilinx docs the HQ208 packages have the heat sink facing down i.e. into the PCB.

Noticed one more thing, after you die or the level ends there is a set of gold tiles but one of them changes colour ? I have seen yellow, red, green to name a few. Anyone know what they signify ?

[trap15]

They are HQFP (Heat Sink Quad Flat Pack) so they have a heat sink in them already. According to Xilinx docs the HQ208 packages have the heat sink facing down i.e. into the PCB.

That would explain why these boards get all warped then

Noticed one more thing, after you die or the level ends there is a set of gold tiles but one of them changes colour ? I have seen yellow, red, green to name a few. Anyone know what they signify ?

I don't think they actually mean anything, they're probably just there for decoration. All of the screen transitions are like that.

[system11]

The Xilinx XC95288 HQ208 CPLDs measure 25mmx25mm. They are HQFP (Heat Sink Quad Flat Pack) so they have a heat sink in them already. According to Xilinx docs the HQ208 packages have the heat sink facing down i.e. into the PCB.

Hm, do you think it would be a good or bad thing to attach an external heat sink? The only other real option is to build it a frame with cooling fans on.

[rtw]

The Xilinx XC95288 HQ208 CPLDs measure 25mmx25mm. They are HQFP (Heat Sink Quad Flat Pack) so they have a heat sink in them already. According to Xilinx docs the HQ208 packages have the heat sink facing down i.e. into the PCB.

Hm, do you think it would be a good or bad thing to attach an external heat sink? The only other real option is to build it a frame with cooling fans on.

Disclaimer, if you add a heatsink and it ruins your board, don't blame me

I spoke to a hardware designer and he said that the thermal resistance might prevent a passive heatsink working effectivly i.e. prevent heat from flowing into the PCB. He said the only option was to try it!

[system11]

Another option would be to put the heatsink under the PCB. If it's on legs that might work. It's looking like fans though..

[rtw]

Another option would be to put the heatsink under the PCB. If it's on legs that might work. It's looking like fans though..

In order to do this properly we would need to power on the PCB for 1 hour and measure the temperature on the chip itself and underneath

Then we would need to mount a heatsink and perform the measurement again. When I have something I believe should work I will post here.

[trap15]

Fantastic, I'll add it to the ST when you've figured it out Strategy for making your board not self-destruct

[EmperorIng]

Has anyone writing this strategy played the board?

I only ask because it seems a given that MAME emulation is a bit off, perhaps given the hardware. I anticipate the 360 release to do some real comparisons (and have a soundtrack that doesn't sound like it's filtered through a DAC sampler on a Sega Genesis).

The amount of secrets and variables this game has really boggles the mind... And it seems as if sometimes enemy spawning is randomly determined on the game-screen.

And when I found out there was variable stage order? Say good night, Gracie.

[rtw]

I spent some time researching CPLD cooling and according to various CPLD manufacturers the best heat sink is a pin grid array. The pin grid array allows omni-directional airflow to maximize heat dissipation.

Mouser part number: 532-374324B35G or 532-374424B35G should fit the task.

532-374324B35G 27x27x10 mm pin grid array with self adhesive for plastic surfaces, thermal resistance: 30.6
532-374424B35G 27x27x18 mm pin grid array with self adhesive for plastic surfaces, thermal resistance: 20.3

Less thermal resistance is better.

I will run a few temperature tests when I get mine to determine if it's worth the effort.

[BPzeBanshee]

Has anyone writing this strategy played the board?

Trap hasn't, I haven't, system11 will soon?, rtw THE and Dave K have. As I understand it's not exactly a common board to find.

I tested an input replay in MAME today with G-Stream and found that despite not resetting the NVRAM at all the replay worked dandy, the coloured blocks in the transitions were in the same place, and the end score was exactly the same. I doubt much of anything important in this game is truly random: it's a matter of needing players to figure out how it all works. Persevere EmperorIng and help us find the secrets of this game!

I also did some further tests with the second player and I learnt a few things:
1. I once continued using it and found it did not reset the score! Could replicate that since though, not sure what happened but I was on Stage 3A when I died
2. The bomb icon IS used: it's spat out when you do a continue only!
3. There is a way to increase your bomb stock without continuing as I've seen when I played but I'm not sure exactly what - score? stage clears? medals?
4. The bug on the highscore screen where it asks for entry for 1P and 2P appears ONLY on the first run after the game is powered on - subsequent runs appeared to operate just fine
5. Couldn't really tell if the attack power was any greater or worse, but you'd think there'd be something to offset its increased overall speed (and it's not like Raiden DX where vertical speed is lowered)
6. The bomb when used collides with whatever's in front of it and explodes, I find. Great for panic bombing, maybe not so much for detonating a specific area if a popcorn decides to get in the way.

rtw, even though I don't own anything of arcade hardware yet I'm glad you're doing these tests: I'm currently in the mindset that a G-Stream board purchase for me is impossible not because of lack of money or a way to run it, but that it would meltdown and die almost immediately where I live in humid northern Australia. I'm very curious to see whether your modifications will make a difference and change my mind on that.

[system11]

My board arrived, it works fine despite a filter cap having a hole in the top (about to go and replace that now). I notice one of the GPUs has already been reflowed along one side, but thankfully my board is not warped. Even after 3 minutes in the open the chips were hot to the touch, this board definitely runs too hot, poor design. For now, I will aim a desk fan at my board bay.

[rtw]

My board arrived, it works fine despite a filter cap having a hole in the top (about to go and replace that now). I notice one of the GPUs has already been reflowed along one side, but thankfully my board is not warped. Even after 3 minutes in the open the chips were hot to the touch, this board definitely runs too hot, poor design. For now, I will aim a desk fan at my board bay.

Do you have a picture of the PCB, what is the date in the config and do you see the big ship on the title screen ? ... please

[system11]

1) http://www.system11.org/temp/forumposts/0504haul8.jpg

2) Forgot to check - will try to remember next time.

3) Same as the MAME set - although I notice it doesn't make the startup noises MAME does, which I assume is an emulation bug - does yours play them?

[BPzeBanshee]

mtume did another video showing off the purple sea segment along with the submarines in Stage 4: http://www.youtube.com/watch?v=Bq4MDHqkYmM

Note that it's using an older version of MAME so the music's even more garbage than it normally is (stage clear/gameover/stage 2's music playing in stage 5A). Considering that it's quite likely the sound effects playing on MAME launch may in itself be a bug too if it doesn't happen on the PCB.

[rtw]

1) http://www.system11.org/temp/forumposts/0504haul8.jpg

2) Forgot to check - will try to remember next time.

3) Same as the MAME set - although I notice it doesn't make the startup noises MAME does, which I assume is an emulation bug - does yours play them?

1. Thank you

2. Please do

3. No noise on the real PCB, I think this is an emulation artifact.

[BPzeBanshee]

Anyone got a clue how the bombs work? mtume and I have been trying to figure out how extra bombs are obtained, and it seems to be related to the amount of powerups you get, but not consistent.

EDIT: Okay, bomb icons. https://www.youtube.com/watch?v=4sTymE9ZLGo

Seems to be related to skimming the powerup as it's on the edge of the screen. I can't explain it. This game's got some fucking obtuse secrets.

[rtw]

Initial setup


Thermal image after initial power-on


Thermal image after power-on for 1 hour


Increase the hostility of the environment, use a plastic box to encapsulate the PCB


Thermal image after power-on for 1 hour inside the box


Mount the heat-sinks

Thermal image after power-on for 1 hour inside the box with heat-sinks mounted


The heat-sinks decrease the temperature by 10 degrees C inside a plastic box by dispersing heat away from the PCB

Recommended:
532-374424B35G 27x27x18 mm pin grid array with self adhesive for plastic surfaces, thermal resistance: 20.3

[system11]

Brilliant - thanks a lot for the testing!

[BPzeBanshee]

Reposted from the DeltaZeal/XIIZeal Shmups Chat thread: http://www.youtube.com/watch?v=qdkIj1dlScM

We now have a stage showing all of the (significant) alternate segments of Stage 1. Nothing real new, unless there's something in the Japanese text that I can't read, but it shows that among other things the transition screen is different/more animated and the music's definitely 5% faster.