PCM Sound & Stability Fixes Multicarts - 120 in 1, 138 in 1, 161 in 1 and others.

wiseman9986

Krauser's Shoe Shiner
Joined
Jan 18, 2003
Posts
228
sorry but is there a list of everything i need to do all of these mods to my 138 in 1 i just dont want to forget to order something
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Take a look at the photos earlier in this thread.

I would order a pack of 50 x 100nF SMD caps, can't remember if they are 0805 or 0806 - should be mentioned earlier in thread. Get 2 x AMS1117 3.3v SMD regulators, 3 x small electrolytic caps - something like 220uF, 100uF or even 47uF will do, and 2 x 47pF ceramic disc type caps (small, something like the 50v ones). You might consider 2 x 100nF ceramic disc caps for the M1 and S1 as well. Without all the components mentioned it's still possible to get glitches. It improves with each component added.
 

wiseman9986

Krauser's Shoe Shiner
Joined
Jan 18, 2003
Posts
228
great thank you for all your work i cant wait to try this i have had problems with this thing for years and pretty much gave up on it
 
Last edited:

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
No problem! Just be aware that each component out of the list makes it a bit better which each step - no one component completely fixes the sound issues but the 47pF caps make the biggest difference when added to the 2 x PCM bus connections. I thought I had applied the fixes to all of my carts but when playing one of my 138 in 1's at weekend I noticed very rare sound glitches coming back. After I took the lid off I realized it was the only one of my carts that I had 'part fixed'. I was missing one of the 47pF caps on the PCB bus, a few 100nF caps underneath the CPLDs, and using a single regulator to power both boards. After I added those missing caps and a 2nd regulator it has been flawless all week and I've used it a lot each evening.
 

phoenixdownita

Cheng's Errand Boy
Joined
Jun 28, 2015
Posts
114
I modified my 161-in-1 with the suggested 47pF capacitors and results improved but not 100%.
My setting: AES unibios 3.2 with 3A power supply, MagicKey converter (onboard unibios disabled or the cart won't even boot).

Towards the end of Level 3 on Pulstar I still experienced a couple of seconds of sudden volume changes but that was it, much more stable than before.
I mounted my capacitors on the back of the PCB rather than on the front.
I ordered the 3.3 regulator, would that help with reducing even further the volume mishaps or ... not really?


Unrelated question: Is there any fix to the PCB to get the games on the 161-in-1 to play to a sane difficulty level, even when in AES modes and Easy difficulty some games are way over the top, it seems I must be in MVS mode and force the soft dip to get any semblance of difficulty settings. Pulsatr being a case in point I had to activate UniBIOS invincibility to get thru level 3 and 4 at easy level, there's way too much on screen no way that it is the easy setting. Even with invincibility the boss of level 4 managed to kill me (I guess there's an attach that simply kills no matter what), also I noticed that it resets if I activate all cheats (max power etc...), in any case on level easy with invincibility on I still can;t kill the level 4 boss and that after pounding him for a good 3 minutes non stop ... something wrong obviously.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
The position you mount the caps can have an effect! You might consider trying different values as well, maybe 22pF or 68pF to see what effect it has on your setup.

On the early MVS carts SNK used 150pF caps on those 2 PCM connections, but 150pF isn't good enough when used with some MVS, and depending on the game itself. I get no sound glitches on my MVS (MV-1FZ) using 2 x 120 in 1, 1 x 161 in 1, and 2 x 138 in 1. Pulstar on the 161 in 1 and 120 in 1 are fine for me =\

It's worth adding the regulator (and remove 1 diode) on the 161 in 1, maybe that will sort it for you. Check no caps are missing on your 161 as well. Double check you have both caps on the right PCM lines.

I can't help with difficulty - I tend to use soft dips first on 161.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Hang on a sec!!!! MagicKey is your problem! Noise filtered out of cart, passes into magic key, and new noise introduced! You might want to add 47pF caps to SDRMPX and SDPMPX lines coming out of the MagicKey.
 

phoenixdownita

Cheng's Errand Boy
Joined
Jun 28, 2015
Posts
114
Hang on a sec!!!! MagicKey is your problem! Noise filtered out of cart, passes into magic key, and new noise introduced! You might want to add 47pF caps to SDRMPX and SDPMPX lines coming out of the MagicKey.

You mean in addition to whatever I already put in the 161-in-1?
Like double filtering?

I will give it a go as I find a little time after all adding capacitors between ground and the signal lines should not be too problematic.
The position I mounted the caps on the 161-in-1 is on the back of the PCB on the SAME vias you pointed out, so that part should not have any effect at all.

[Am I the only one on AES + MagicKey?]
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Even though on same vias, depends where the cap legs run - trace (or in this case, cap leg + nearby trace) makes a difference. Say on the other side of the board one of the cap legs is running very close to a clock or something that would give a problem, whereas same vias on other side might not.

That's why the caps are needed, noise from nearby traces. They tried to filter it on the 161 with those wavy tracks.

Edit: to be fair it's much more likely the adaptor introducing its own routing parasitic capacitance.
 
Last edited:

aha2940

AH, A, COLUMBIAN!,
Joined
Dec 15, 2013
Posts
2,528
Hang on a sec!!!! MagicKey is your problem! Noise filtered out of cart, passes into magic key, and new noise introduced! You might want to add 47pF caps to SDRMPX and SDPMPX lines coming out of the MagicKey.

Thanks for your help GadgetUK, but which lines would be these in plain English? :D

Regards.
 

phoenixdownita

Cheng's Errand Boy
Joined
Jun 28, 2015
Posts
114
But in our case we should "fix" the AES part right?

Afaik serpentine traces are used to delay a signal not filter it (albeit they sometimes may act as little inductors, but not capacitors).
The added 47pF capacitors really create RC filters aka low pass filter which may help smoothing out high frequency "glitch" at the same time said filters make raise time and drop time longer, I guess we're suffering from high frequency spurious noise hence why they work well in practice.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Yes, you are absolutely correct! I did wonder if a cap and inductor might be used to filter the noise, but the problem seems to be that it's related to the bus itself and how it is used, rather than a specific constant interfering frequency. This is why the 150pF works on certain games on older carts, but as soon as you install a game that never shipped on an older cart to an older cart, 150pF doesn't always work and where PCM is used a lot 47pF seems to work better than 150pF. It's dependent on the game as well, certainly with the 1FZ. 47pF might not be the ideal size, but it has removed the problem on 8 carts for my 1FZ.

EDIT: The other interesting point which ties in with the wavy tracks and general track layout - On one of my PROG-16 boards (Puzzle Bobble boot to Fight Fever conversion) I ended up having to sever the traces at the connection end and at the input to the PCM chip and route 2 seperate shorter wires (ie. eliminating the traces on the board completely), as well as the 47pF caps in order to eliminate the problem completely. I've not seen the MagicKey board itself so I've no idea if they thought to add any caps to SDRMPX and SDPMPX lines but I assume they probably didn't, which likely means with some games and some AES that sound glitches might occur when using the MagicKey as well. I have thought about buying a MagicKey for a while, maybe later in the year I might get one and see how badly manufactured they are.
 
Last edited:

MtothaJ

Host for Orochi
10 Year Member
Joined
Apr 15, 2013
Posts
756
First of all many thanks to GadgetUK for sharing these fixes - the Queen should give you a knighthood for your efforts:)
I have a 161-in-1 and proceeded today to doing the 2x 47pf cap fix - found it quite difficult to solder onto the traces but got there eventually. On the subject of the 161 cart, one thing I am wondering about is whether there are two caps missing from the cart - talking specifically about these two places with the solder pads in place:

missing caps.jpg

Also, concerning power consumption of the 161 vs a standard cart I have a volt meter hooked up to my MV-1FZ board and as was mentioned many times the 161 cart draws more power than a standard cart. Here are the readings:

161 (it actually cycles between 5.00 and 5.01)

161.jpg

NAM-1975 (5.05)

nam.jpg
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Yes, well spotted - looks like 2 caps missing there - not sure if they are supposed to be bypass spots or what. If its working fine without them don't worry about it, if you want to tempt fate - try putting a 100nF cap into each of those 2 positions and see if anything changes. Chances are you won't notice a difference. When I get time I will follow the traces to see where they connect but it's probably VCC and GND. With regards to voltage drop, yes there is a small difference - especially when using a 2 amp PSU like I did. I see the same small voltage drop as you, from say 5.05 ish with a stock cart to 5.01v on a multicart. Some of the larger chip count carts or carts using encryption also draw more current from what I've seen, eg. KoF 2003.

EDIT: And thanks for the thanks lol =)
 
Last edited:

MtothaJ

Host for Orochi
10 Year Member
Joined
Apr 15, 2013
Posts
756
Yes, well spotted - looks like 2 caps missing there - not sure if they are supposed to be bypass spots or what. If its working fine without them don't worry about it, if you want to tempt fate - try putting a 100nF cap into each of those 2 positions and see if anything changes. Chances are you won't notice a difference. When I get time I will follow the traces to see where they connect but it's probably VCC and GND. With regards to voltage drop, yes there is a small difference - especially when using a 2 amp PSU like I did. I see the same small voltage drop as you, from say 5.05 ish with a stock cart to 5.01v on a multicart. Some of the larger chip count carts or carts using encryption also draw more current from what I've seen, eg. KoF 2003.

EDIT: And thanks for the thanks lol =)

100nf caps added :)

100nf.jpg

And these are the 47pf caps in place:

47pf.jpg

The main thing is that after the mods the cart still boots ;)
I will have to do some testing since this evening I was mainly preoccupied with doing the mods, but - maybe it a placebo effect thing - its seems that games such as Pulstar do sound a LOT better.

PS. would this be the correct voltage regulator package - https://elty.pl/pl/p/Stabilizator-napiecia-AMS1117-3,3V-SOT223-5-sztuk/972 ? Is there any other mods you would suggest for the 161? Would be keen to do a proper job on this :cool:
 
Last edited:

donluca

Ninja Combat Warrior
Joined
Aug 19, 2015
Posts
529
wow, I completely missed those 2 caps too. Adding them in right now and will report back as well.

By the way, judging from the pads they don't seem like the usual 100nF SMD caps. The pads are bigger, I'm wondering what the heck they removed from them. I have those pads with solder on them as well, I have a feeling they removed them for some reason.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
The regulator you've linked to there looks OK to me! When you fit the regulator don't forget to remove the top diode (one furthest from the cart connector edge). Not sure if you've seen the video I did on the 161 in regulator? Once you remove the top diode and fit the regulator you need to check that the cathode (where the 2 diodes join each other) of the bottom diode you leave in is still connected to the solder bridge nearby. The way it works is 4.3v is derived from the first diodes voltage drop and fed to the CPLD responsible for the M1 ROM decoding. If you miss that solder bridge, or remove both diodes (as I did to start with) you will get a Z80 error reported on boot as the Z80 cannot read the M1 ROM. If you get a Z80 error double check that the first diode is in place correctly and its cathode is bridged to the little trace which runs off to a via and they to the CPLD as shown in my video:- https://www.youtube.com/watch?v=5a-67DIteMk
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
On a separate note, one of the other reasons to put a voltage regulator on there is related to the tolerance of your 5v level. With diodes and a nice clean 5v level you will expect ~3.6v to be fed to the VCC pins of the 3.3v logic, if your PSU is running at say 5.2v the VCC to the 3.3v logic will be 3.8v. That's why its a good idea to use a 3.3v regulator to ensure that the chips will last longer on the cart.

Having said that if your 5v line is running at 5.20v I would tempted to swap out your PSU as that's a little high in my mind. Up to say 5.15v is reasonably safe, and anything under 4.95v might cause possible problems.
 
Last edited:

donluca

Ninja Combat Warrior
Joined
Aug 19, 2015
Posts
529
I've been using a PSU which outputs 5.3V under load and so far I had zero problems (been playing around for a week).

I've even left it on for 5-6 hours straight and didn't have any issue.

EDIT: I stand corrected. It's 5.3V with no cartridge inserted and 5.2V with the 161-in-1 in.
 
Last edited:

MtothaJ

Host for Orochi
10 Year Member
Joined
Apr 15, 2013
Posts
756
Thanks, I will report back once I get and add in the regulator. If there is anything else that can be done to the 161 I would be keen to give it a try. I had a look at the electrolytic capacitors and these seem fine IMHO in terms of quality etc. so don't think there is much value in recapping unless they screwed up the values of the caps.
Regarding voltage related issues I am using an arcade PSU (so probably like 15 amps available on the 5V line). While its good to be covered for amps (after all MV-1FZ technical manual states you need 7amps (!)) IMO a key advantage of an arcade PSU is that you can adjust the 5V line, which comes in handy especially if one is using a long or crappy quality jamma harness (high AWG wire). These result in quite a bit of resistance and in turn voltage drop which may mean that a standard 5V PSU may not be man enough to supply 5V by the time it reaches the PCB. I found that anything below 4.8 / 4.9 generally results in pretty unreliable operation - screen blacks out, random resets etc.

PS. just to say that with regard to the two missing caps, my solder pads didn't have any solder on them.
 
Last edited:

donluca

Ninja Combat Warrior
Joined
Aug 19, 2015
Posts
529
I've tried some games and I didn't notice any difference adding the two caps honestly.

I'll investigate a bit more in the following days to see if something comes up.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
I've been using a PSU which outputs 5.3V under load and so far I had zero problems (been playing around for a week).

I've even left it on for 5-6 hours straight and didn't have any issue.

EDIT: I stand corrected. It's 5.3V with no cartridge inserted and 5.2V with the 161-in-1 in.


The maximum safe voltage on some of the chips - in particular the Z80 is 5.25v. At 5.30v you could kill the Z80 or other chips with same max working voltage.
 

GadgetUK

Ace Ghost Pilot
Joined
Sep 27, 2013
Posts
1,323
Thanks, I will report back once I get and add in the regulator. If there is anything else that can be done to the 161 I would be keen to give it a try. I had a look at the electrolytic capacitors and these seem fine IMHO in terms of quality etc. so don't think there is much value in recapping unless they screwed up the values of the caps.
Regarding voltage related issues I am using an arcade PSU (so probably like 15 amps available on the 5V line). While its good to be covered for amps (after all MV-1FZ technical manual states you need 7amps (!)) IMO a key advantage of an arcade PSU is that you can adjust the 5V line, which comes in handy especially if one is using a long or crappy quality jamma harness (high AWG wire). These result in quite a bit of resistance and in turn voltage drop which may mean that a standard 5V PSU may not be man enough to supply 5V by the time it reaches the PCB. I found that anything below 4.8 / 4.9 generally results in pretty unreliable operation - screen blacks out, random resets etc.

PS. just to say that with regard to the two missing caps, my solder pads didn't have any solder on them.

The 7 amps in the MV-1FZ manual is a typo - It pulls around 1.2 amps without a cart, and something like 1.3 with a cart. There's no way it would ever draw close to 7 amps, unless there was an electric heater connected to it =D
 

MtothaJ

Host for Orochi
10 Year Member
Joined
Apr 15, 2013
Posts
756
One more question about the 3.3V regulator - would one rated for 0.8amps be fine or do I categorically need to go for the 1 amp version?
 
Top