(Topic ID: 129030)

MMR Under the Microscope


By Crash

4 years ago



Topic Stats

  • 121 posts
  • 43 Pinsiders participating
  • Latest reply 2 years ago by Coyote
  • Topic is favorited by 23 Pinsiders

You

Linked Games

Topic Gallery

There have been 21 images uploaded to this topic. (View topic image gallery).

sounds_(resized).png
AFM_(resized).png
castlesheet_(resized).png
ram_(resized).png
damseldots_(resized).png
MMRdots_(resized).png
MMR_(resized).png
colors_(resized).png
corefiles_(resized).png
identical_(resized).png
dots_(resized).png
tree_(resized).png
joke_(resized).png
image_1.jpg
image.jpg
1566.jpg

There are 121 posts in this topic. You are on page 1 of 3.
19
#1 4 years ago

I finally got the chance to get a first-hand look at Pintucky's MMR #118 today. For those interested in the nerdy specifics of the electronics...

The CPU/Beaglebone Black "pinball controller" PIN-PCB-CONTRLR REV 1.

12v input is regulated to 3.3v via U8. May have isolated ground test point. Uses 3v button cell to maintain settings. Uses NXP 74LVC57AD ICs. RJ-45 CAT5 cable for playfield board interfacing. Unused J8 14-pin connector labeled "DMD." Beaglebone connects to LCD panel via micro HDMI. No micro SD card present (OS loads from onboard flash). Unused J8 6-pin socket labeled "ICP." External 19v input (unknown).

1563.jpg

1565.jpg

The LCD panel and controller.

M.RT2261.5B LCD panel controller. Accepts DVI and VGA inputs. An HDMI to DVI adapter plug is used. Power is supplied for flourescent backlight inverter (I expected this to be an LED backlit panel). Separate remote board installed in backbox. Receives 12v from custom harness (not DC barrel plug). Briefly displays "no signal" while CPU board boots.

1564.jpg

The playfield interface board MM-PCB-PLAYFLD.

Utilizes an ECS-100AX 64MHz crystal clock and onboard 3.3v regulator. 1000uf filter caps used. Switches connect to this board. Most likely does not use a switch matrix. Diodes not present on switches. Uses isolated ground? Expandable as an unstuffed "SOL4" solenoid driver board interface placeholder is present.

1567.jpg1568.jpg

The "solenoid driver" REV 0.2 board. Receives 78v unloaded "50v" input from line box board. This powers everything. 3 4A fuses protect drivers 1-4, 5-6, and 7-8. Coil damper diodes are onboard. Uses isolated ground? Driver MOSFETs have diodes installed upstream (why?). Boards are game specific as not all coil power headers are stuffed and serial number is printed on a sticker. Connectors hot glued to headers from factory. "94v-0" is printed near the version number. Does this suggest maximum input voltage tolerance? Driver MOSFETs appear to be secured to the board. Fuse rating warning sticker present (lol).

1566.jpg

#2 4 years ago

If you want to have some fun, remove the coil stop of any of the two flippers and then try to remove the coil.
It cannot be done without removing other items behind it.

When a transistor fails, You'll need to have a drill machine with you to get it loose.

#3 4 years ago

great photos and info guys.

#4 4 years ago

Is there only one RJ-45 interfacing with the playfield from the pinball controller?

#5 4 years ago

FYI, 94V-0 is a Fire rating for the PCB material and also related to testing. As I recall, it is a requirement for UL22 on PCB that carry more than 48V (or was it 24V?). So is the Fuse Rating warning sticker!

#6 4 years ago
Quoted from Revo76:

If you want to have some fun, remove the coil stop of any of the two flippers and then try to remove the coil.
It cannot be done without removing other items behind it.

You can loosen the front retainer and slide the coil forward, then rotate it and remove it. It is a PITA though.

Quoted from Revo76:

When a transistor fails, You'll need to have a drill machine with you to get it loose.

Yes, annoying, they pop riveted all the transistors.

#7 4 years ago
Quoted from stangbat:

Yes, annoying, they pop riveted all the transistors.

What was the rationale there?

#8 4 years ago
Quoted from gcp:

What was the rationale there?

Heat sinking.

#9 4 years ago
Quoted from Revo76:

What was the rationale there?

I will take "less cost for a thousand, Alex".

#10 4 years ago

Lowest cost would be to stand them up like Stern, Bally , Williams....

#11 4 years ago
Quoted from gcp:

What was the rationale there?

Quick and easy.

#12 4 years ago

Williams did the same thing with their Fliptronics MOSFETs, except they used plastic rivets.

#13 4 years ago
Quoted from MrBally:

Lowest cost would be to stand them up like Stern, Bally , Williams....

Fair enough, how about lowest cost with better-ish reliability. Given the under playfield location I imagine cracked solder joints due to vibrations of a pendulum/lever like TIP package was a legit concern.

#14 4 years ago

Does anyone know how/where the OS is stored? There is one chip I could not identify.

#16 4 years ago
Quoted from stangbat:

You can loosen the front retainer and slide the coil forward, then rotate it and remove it. It is a PITA though.

Yes, annoying, they pop riveted all the transistors.

I would drill them all out and replace with a nut-and-bolt. That would make future repairs easier.

Remember, it's not a matter of *if* a driver transistor fails, but *when* a transistor fails.

Why did they put driver transistors under the playfield? Especially since that's where all the vibration and dirt originate. Were they trying to imitate Gottlieb System 80 or what?

#17 4 years ago

It probably is then. I was not aware there was onboard storage. Does this script modify the contents of the filesystem or use its own autorun.sh?

Edit: Nevermind, it just runs dd right from the shell upon boot. Sweet.

#18 4 years ago
Quoted from Magic_Mike:

Is there only one RJ-45 interfacing with the playfield from the pinball controller?

Yes, it's the only one I saw.

#19 4 years ago
Quoted from Crash:

Yes, its the only one I saw.

Thanks. That board is smaller than I thought!

#20 4 years ago

I always thought pinball manufacturers went to backbox stored boards to eliminate shock/vibration from cab areas. Why did PPS go back to pf mounted?

#21 4 years ago
Quoted from underlord:

Why did PPS go back to pf mounted?

Saves a bunch of wiring and assembly time.

Questions about this came up after prototype pics were posted long ago. Rick posted a video of the assembly having the crap shaken out of it for I can't remember how many hours on end with no issues. Maybe someone else can remember where to find the video. I don't think it is going to be an issue.

#22 4 years ago

That was the one for the LED boards.

#23 4 years ago

nerdy yes, but informative. thanks.

#24 4 years ago

What exact failure mode is this test trying to replicate? I see an attempt to create/perform an accelerated validation test, however, just not sure how this test results in useful information. Kind of like the Lexus commercial with all the champagne glasses on the hood. It makes for an impressive looking commercial but doesn't really mean anything.

Honestly, the best approach is to study past generations of failed boards to truly understand root cause. I have to believe others have studied this in great detail. Once root cause is understood, more meaningful accelerated validation tests can be performed. 6 Sigma is a beautiful thing.

#25 4 years ago

Guessing large difference in quality from 1993 through boards to 2015 SMB technology.

#26 4 years ago

All sorts of valid info would come from 250 hours of intense vibration. True, it may not exactly replicate real life vibration from normal playfield action but it would certainly make some problem spots appear. I am sure the intended purpose was to see if components would start falling off the boards or develop fractured solder joints. One would hope that following the test, the put the boards under a microscope to look for damage.

Yes, Tab transistors sitting up in the air without support would be a major no-no under the playfield. But then, we all have also seen cold solders to connector pins too. The larger/heavier a component, the more likely it is to develop a cold solder in the distant future. Riveting the larger components down was probably the choice for speedy assembly. I have even seen it done with plastic push rivets but that wouldn't hold up in a heat sinking situation.

#27 4 years ago
Quoted from CactusJack:

True, it may not exactly replicate real life vibration from normal playfield action but it would certainly make some problem spots appear. I am sure the intended purpose was to see if components would start falling off the boards or develop fractured solder joints. One would hope that following the test, the put the boards under a microscope to look for damage.

My point exactly. If parts fell off or joints fractured due to unrealistic system motions, it could drive wrong behavior. If they did their homework, I am sure the test is valid. Otherwise....it is just another cool YouTube video.

#28 4 years ago

You should see the shit that safety-related Automotive Electronic devices are put through. Besides vibration testing under both temperature extremes, thermal shock testing is performed as well as tests simulating unprotected drivetrain elements hitting a nasty pothole with a fully loaded vehicle, voltage transients, load dumps etc.

1 week later
#29 4 years ago

Has anyone tried imaging the Baglebone's flash memory using the software ecurtz linked to?

#30 4 years ago

Having the small driver boards makes it easier to get the board out to repair them rather than the huge board with dozens of connectors in the backbox. It also saves a bunch of wiring, unneeded complexity and manufacturing/labor cost.

As for the driver transistors, I don't like the idea of riveting them down or holding them down. If you look at the SPIKE boards Stern left the transistors standing. To me this will be the better way in the long run. There is little or no vibration under a pinball playfield except from the speakers and that vibration is present in the backbox too. There is shock under a playfield (shock is different than vibration). Shock comes from the coil plungers hitting their stops and in this case is not enough to affect something so light as a transistor that is supported in place by three heavy soldered leads which are basically holding it flush on the board. Something like a ceramic resistor that is not glued down would be a problem, but not a properly soldered-in transistor. Personally I would be more worried about heat expansion and contraction of the transistors and their leads with them riveted to the board. That will cause more stress on the solder joints over the years as they heat and cool and fight against the rivet and solder joints than the shock (or vibration) ever would. Also they will stay cooler with them mounted upright. The board is not a good heat sink and I don't see any metal there...

Anyway, I am splitting hairs here. Both the MMr system and the new SPIKE system are advances in pinball technology the will make them cheaper and easier to manufacture, more reliable, and easier to maintain and service.

2 months later
#31 4 years ago

This is an interesting thread. I appreciate people people taking the time to give facts without bashing others. I'm a rather new member here. I have owned machines for about 15 years now, but consider myself a newbie compared to most of you. I, for one, am excited about MMR and anything else that can come about. One thing I don't understand is why there is animosity towards those trying to make money in pinball? I realize that someone owning an original MM would be upset about losing value, but supply and demand is what drives markets. We need more machines in the hands of more people to bring prices down. The more people - the better. If pinball numbers continue to go down all the good machines will only be owned by the very wealthy.

As for whether the new electronics are better or worse - time will tell; however, I can tell you that it is refreshing to open the bock box and not see 6 different boards with 50 fuses, 20 connectors, and thousands of board components that are 20-30 years old. My biggest concern is that technology will now change fast enough that the "new technology" today will not be supported in the future. With the old stuff, it was a little easier (I think) to repair an audio board vs having to source a complete board today because it is not repairable and not supported. We currently have this problem in aviation. For years the technology did not change. This was nice since avionics could be repaired as needed for decades. The downside is that innovation and options were few. We now have touchscreens and wonderful tools in the cockpit, but 10-15 years after the product is installed for the tune of $100k it becomes unsupported because it is 3 generations old and the factory decides to move on.

Looking forward is it probable that these new smaller boards will be able to be troubleshot and repaired with their smaller surface mount technology and if not will technology enable someone to simply copy the design and build a new board ready to install?

1 week later
#32 4 years ago

John_I post is insightfull . I must point out that You do have serious playfield vibration and shock between the trolls slamming up and down and the shaker motor shaking . Playfield solenoid board is right in the center of this violent action

#33 4 years ago

Mine plays still like new after over 1,000 plays not one issue, no sling damage, no switch problems, no ball swirls, no lights burnt out play field is like a mirror finish still!! This game kicks ass!

image.jpg

image_1.jpg

#34 4 years ago

Ball swirls and nonworking LEDs are a problem?

#35 4 years ago

I just saw one of these for the first time today and played a quick game on it. I could tell no difference between an original and MMR. Congrats to PPS and Chicago Gaming. I am excited to get mine.

5 months later
15
#36 3 years ago

Software Analysis

Ok, this was quite a lot of fun. Apparently PPS has had full SD card restore images floating around via a publicly accessible link for a while, and people have been sharing it around the board...

https://drive.google.com/file/d/0B6aZTLdJ97CmOExya29LbThHWjg/view?usp=sharing

I grabbed it and did an analysis of the filesystem, something I've been itching to look at for several months but I didn't have access to since I don't own a machine. Basically mmr.zip contains an "install disk" image file that can be plopped on a micro SD card with Win32 Disk Imager or similar program. The install script runs and unpacks a second img file, which is then installed to the Beaglebone's eMMC onboard memory.

joke_(resized).png

I loaded mmc_image.img in FTK Imager and it basically all boils down to this.

tree_(resized).png

It's a Debian distro with AppArmor for security just like JJP's Ubuntu-based platform. Nothing is actually encrypted here, so I'm assuming it's for network control access by looking at the install image's etc directory. The game program is called emumm and runs out of home. There are 3 ELF executables: a main version, a version compiled from .NET, and a third version I'm not sure on. Can anyone identify the "REL" language/library? The appdata directory contains DMD dot masks, the MM 1.0 game ROM, a single TTF font, sound and music files, RAM snapshots, and color DMD images.

On a side note I found something interesting in gpio.sh, supposedly used to map inputs and outputs to and from the "pinball controller" board. Is there something for Stern SAM in the works...?

corefiles_(resized).png

I found some bitmaps that determine how the dot mask appears. We have 6x6, 9x9, and 10x10 pixel dots in each of the 4 shades of red.

dots_(resized).png

To those that may be skeptical of the use of emulation, here's your scoop. The original unmodified/unhacked MM CPU ROM, identical to the one from IPDB.

identical_(resized).png

Next is a Courier New TTF font for the menus. Nothing exciting. Time for some meat and potatoes: the samples directory, which contains all the sounds assets: music, sound effects, voices, and even DCS menu sounds. The vol_25perc directory means the sounds are leveled at 25% volume. There has been some debate lately on the true quality and reproduction of these sound recordings, so let's clear the air a bit here.

The original DCS files are highly compressed, a mix of 12kHz and 31.25kHz sample rates, and have a maximum bitrate of a couple of hundred kilobits per second. The files found in MMR are, contrary to belief, not 44.1kHz full "CD quality." Like many of the original DCS recordings, they are all 31.25kHz quality but these files are uncompressed 16-bit PCM at 500kbps, not a huge step up but the difference in sound reproduction to the naked ear is very noticeable. Even the 22kHz WAV samples from the original website sound better than the DCS bites. I am not sure if the "fatality" callout from Mortal Kombat II is actually used, I'm assuming so considering it was also remastered. Because the sample rate was not increased to full CD quality I'm assuming the masters were originally stored on digital media, not analog studio tapes. This means they cannot be upsampled. They are also named in the same order as the WPC95 ROM, a nice touch.

sounds_(resized).png

When Rick claimed an improvement over using uncompressed sound bites I think that statement was taken a little out of context since most knowledgeable people would associate uncompressed with "CD quality," but the remasters are still a big improvement over what we had for the last 15 years. Listen for yourself:

https://soundcloud.com/titan91-1/mmr-pinball-demo

After that there are 3 NVRAM files, probably captured from actual NVRAM chips running on the real WPC95 hardware. There's a "clean" image which is probably following a factory reset and 2 more MM 1.0 images. I'm not sure what the difference is but here's a status report at the beginning of one. Look familiar?

ram_(resized).png

I'm sure if someone put enough work into it they can write a save state system for PinMAME that allows these images to be used as a working snapshot which will allow us to see the exact system state when the game was turned off to image the NVRAM...

The last thing sitting in an unsuspecting src directory is... color DMD frames. The high resolution stuff! As in the very images used in CGC's color DMD announcement video. Everything is stored in a single file, rgb16.bin. This file contains a series of 256x64 16-bit RGB bitmap frames. The first few are used during attract mode surrounded by strips of palette colors, which aren't necessary for rendering but are most likely used for artist's reference. There really isn't anything in here that hasn't been seen already save for a few variants of the damsel rescue scene. I also made a mask from the 6x6 dots.

colors_(resized).png

MMR_(resized).png

MMRdots_(resized).png

damseldots_(resized).png

castlesheet_(resized).png

Hmm, wonder what this is?

AFM_(resized).png

While they're not easy to work with (I had to use Tile Molester to manually rip these), everything is now stored as full color frames, not layered mini-sprites like the original game. That method was primarily due to the limited storage space and computing power of 1995-era hardware. Man, we've come a long way. Now it's all starting to look like the bane of pinball, the dreaded 16-bit era consoles!

If you have the image file, poke through it and explore stuff. If there's anything specific you want to ask, post it.

#37 3 years ago
Quoted from Crash:

Software Analysis
Ok, this was quite a lot of fun. Apparently PPS has had full SD card restore images floating around via a publicly accessible link for a while, and people have been sharing it around the board...
https://drive.google.com/file/d/0B9ZImdZr3-d4cHNIRjdjTUdLYnc/view

That's not PPS's mmr image. It's mine, that I made available so people could update their MMR's without the hassle of going through CGC.

Edit: removed at the request of PPS.

#38 3 years ago

I'll buy a drink for the first person to swap an AFM ROM in that package, then install it into an MMR and see if it fires up. Might have to make the checksum match the MM ROM checksum if the mmemu program looks at that before starting... also unplug the playfield coil harnesses first just in case

#39 3 years ago

Nice work Crash. Thanks. The fatality callout from MK2 is used. I heard it the other day while playing my MMR.

#40 3 years ago
Quoted from Pablito:

Nice work Crash. Thanks. The fatality callout from MK2 is used. I heard it the other day while playing my MMR.

Is "Toasty" in there ?

#41 3 years ago

Haven't heard toasty yet.

-1
#42 3 years ago

So I've been waiting for some smart dude to attempt this. My interest is that I have young kids and parents in my office and id like to put mmr in my waiting room. But Williams "family mode" doesn't take out some stuff some parents would object to (yes, yes, YES!). And it removes the dragon beheading which I am totally cool with. So,...

This is a noob question. But now that the file system is better understood, is it something like pinsound where you can find a sound file you don't like, and maybe copy a different file and rename it to the same name as the deleted file? So it would still have a callout but it would be a g-rated one all the time? (Or at least pg). Is it that simple? Probably not, but is there a pathway to that?

#43 3 years ago

I haven't tried it. You would have to write your own script file to replace the sounds, but like JJP there doesn't appear to be any authentication in place. Just give it the same file name.

#44 3 years ago

I was able to swap in the profanity rom. It works fine.

#45 3 years ago

Hehe! Just the opposite of what I wanted to do! Could this also be used elevate the volume of just the voice call outs relative to the rest of the sounds to make them clearer?

So can you elaborate? How did you identify individual sounds? The op said you'd need a script to do it. Probably a noob question, but why? Can I have your script?

#46 3 years ago

Yes, but you would need to modify each callout file. There are over 800 total sounds. I think there is already a way to do this in the menus but I'm not sure. I can't identify them without listening. You would need to write a Bash script to replace whatever files you want to swap. So you would generate a Beaglebone bootable SD card image, put your custom files in a folder, and write a script that copies and overwrites those files into the samples folder on the onboard storage. Then burn the card, boot in the game, wait a sufficient amount of time and restart the game with the card removed.

#47 3 years ago

Awesome. This is wonderful.

#48 3 years ago

Outstanding deconstruction work, Crash.

#49 3 years ago
Quoted from paynemic:

But Williams "family mode" doesn't take out some stuff some parents would object to (yes, yes, YES!).

So they want the puritan edition of the ROM? I feel sorry for those kids when they finally hit real life...

-1
#50 3 years ago
Quoted from metallik:

So they want the puritan edition of the ROM? I feel sorry for those kids when they finally hit real life...

Maybe they want 11 different censor levels like the spoooky pinball people want

Promoted items from the Pinside Marketplace
$ 50.00
€ 159.00
Cabinet - Toppers
FlipperLED
There are 121 posts in this topic. You are on page 1 of 3.

Hey there! Got a moment?

Great to see you're enjoying Pinside! Did you know Pinside is able to run thanks to donations from our visitors? Please donate to Pinside, support the site and get anext to your username to show for it! Donate to Pinside