Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Showcase builds, discuss cases, embedding MiSTer into existing computer cases.
LeeW
Posts: 28
Joined: Tue Jul 25, 2023 9:13 pm
Has thanked: 9 times
Been thanked: 26 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by LeeW »

I've verified almost all the other components now, hdmi, scart, vga, analogue audio out, ethernet, fan headers, switches, everything seems to work. As people have said the USB part of the board seems to be isolated from the rest. The HDMI and Ethernet also seem to be isolated pass-through connections.

EF3tXjc.jpg
vb3LGHF.jpg

Don't have anything on hand to test the optical digital audio, but the port is lit up so I'm guessing it works. The PSX SNAC adapter didn't work despite toggling to 3.3V mode. After researching this I think this is due to the multisystem not being compatible which is a bit annoying tbh as this was the main controller I wanted to use with SNAC.

Also assembled the case. I was considering getting it 3D printed from elsewhere but the prices elsewhere seemed more expensive than the official store, which makes sense considering bulk vs one-off.

You do not have the required permissions to view the files attached to this post.
DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

Nice! Congrats!

Can you try to run the memory tester and report back the maximum speed you are able to get out of the SDRAM? I'm curious how it compares to the MiSTer SDRAM expansion boards.

DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

emiliom wrote: Sat Aug 26, 2023 6:45 pm
DeKay wrote: Sat Aug 26, 2023 3:34 pm

What are the voltage for FLAG- and USB_EN on that chip?

  • If -FLAG (an output from U4) is around 0V, U4 detects a problem like over current (i.e. a short somewhere)

  • If USB_EN (an input to U4 from the FE2.1) is around 5V, then FE2.1 isn't happy for some reason

FLAG-: 3.27 - 3.30V
USB_EN: 4.33 - 4.56V

DeKay wrote: Sat Aug 26, 2023 3:34 pm

If, with everything powered off, you get some very low number of ohms between VBUS and Ground, you have a short circuit there somewhere. U4 is sensing that and shutting VBus down. Find that short and you're golden. If you are getting something well into the kiloOhms or greater, that is various parasitics and you're probably OK.

It had been looking suspiciously like a short. I was hoping there was some reason that I didn't understand that the VCC USB pins were showing continuity with GND, but obviously not. Resistance between each VCC pin and GND for USB1-7 is <1 Ohm. Voltages when powered on oscillate and range from 0.01-0.05V and 0.02-0.11V. So a very, very tiny amount of current is getting through, but most is leaking away.

JP11 (or USB0?) seems ok. VCC is 4.85V. So it looks like the short is somewhere after JP11. Would that be a reasonable assumption?

First of all, be careful with your terminology. VCC isn't a named net on the schematic. I am assuming you are referring to the 5V power supply that powers most of the chips on the board usually when you say VCC. BUT VCC (5V) and VBUS (5V when everything is working) are completely separate. VBUS comes from U4 from the 5V input to its pin 1.

It almost seems like you have two problems.

  • "Resistance between each VCC pin and GND for USB1-7 is <1 Ohm". This is a short circut between VBUS and Ground. This is bad. Very bad.

  • "USB_EN: 4.33 - 4.56V". The FE2.1 chip is not happy for whatever reason so it isn't letting U4 put power on VBUS in the first place. Might be related to the first problem. Might not.

I would start by searching the schematic for every instance of VBUS and find the corresponding point on the board, looking for the short to (probably) ground. Note that the shells of the USB connectors are grounded so perhaps the problem is creeping in there? Use an eye loupe or magnifying glass with lots of bright light. There is a problem there. You need to find it.

rezendes
Posts: 29
Joined: Thu Jun 29, 2023 12:57 am
Has thanked: 11 times
Been thanked: 9 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by rezendes »

LeeW wrote: Sat Aug 26, 2023 8:53 pm

I've verified almost all the other components now, hdmi, scart, vga, analogue audio out, ethernet, fan headers, switches, everything seems to work. As people have said the USB part of the board seems to be isolated from the rest. The HDMI and Ethernet also seem to be isolated pass-through connections.

Don't have anything on hand to test the optical digital audio, but the port is lit up so I'm guessing it works. The PSX SNAC adapter didn't work despite toggling to 3.3V mode. After researching this I think this is due to the multisystem not being compatible which is a bit annoying tbh as this was the main controller I wanted to use with SNAC.

Also assembled the case. I was considering getting it 3D printed from elsewhere but the prices elsewhere seemed more expensive than the official store, which makes sense considering bulk vs one-off.

I’ve heard to only really use the SNAC adapters for light guns that require it because the difference in lag is not perceptible when using a decent usb adapter. There’s the usb retropad and the newer version of it for example, also there’s a list out there someone made comparing the lag with many different adapters.

FPGA64
Top Contributor
Posts: 887
Joined: Mon Mar 01, 2021 3:10 pm
Has thanked: 48 times
Been thanked: 372 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by FPGA64 »

rezendes wrote: Sun Aug 27, 2023 12:57 am
LeeW wrote: Sat Aug 26, 2023 8:53 pm

I've verified almost all the other components now, hdmi, scart, vga, analogue audio out, ethernet, fan headers, switches, everything seems to work. As people have said the USB part of the board seems to be isolated from the rest. The HDMI and Ethernet also seem to be isolated pass-through connections.

Don't have anything on hand to test the optical digital audio, but the port is lit up so I'm guessing it works. The PSX SNAC adapter didn't work despite toggling to 3.3V mode. After researching this I think this is due to the multisystem not being compatible which is a bit annoying tbh as this was the main controller I wanted to use with SNAC.

Also assembled the case. I was considering getting it 3D printed from elsewhere but the prices elsewhere seemed more expensive than the official store, which makes sense considering bulk vs one-off.

I’ve heard to only really use the SNAC adapters for light guns that require it because the difference in lag is not perceptible when using a decent usb adapter. There’s the usb retropad and the newer version of it for exampl, also there’s a list out there someone made comparing the lag with many different adapters.

https://rpubs.com/misteraddons/inputlatency is the list

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

DeKay wrote: Sun Aug 27, 2023 12:11 am

Note that the shells of the USB connectors are grounded so perhaps the problem is creeping in there? Use an eye loupe or magnifying glass with lots of bright light. There is a problem there. You need to find it.

That had occurred to me. I wondered if I put too much solder in and one or more pins had bridged where I couldn’t see them. I couldn’t see any obvious shorts with a magnifying glass and then I even checked with a microscope (cheap but does the job) and it really doesn’t look like there are any shorts where solder joints are visible on either side of the board. I started checking continuity between the middle 2 pins of the USB connectors, and then between each of those pins and pin 1 and on one I got a tone. I’m not sure if that could be the issue as a short between pin 1 and 2 wouldn’t cause a short to GND would it? Regardless, it’s still a short so I tried to remove the connector, but without success so far. I’ve used wick and a de-solder pump to remove as much as I can, and it barely looks like there’s any visible solder holding it in place, but it still appears to be rock solid.
I suspect I may end up having to remove the connectors one by one, if for no other reason than to eliminate them, but it seems it’s going to be very difficult.
I’ve already looked on YouTube for advice on de-soldering connectors but I’ve not found anything different to what I’m already doing. Would you (or anyone) have any tips on removing the Ethernet connector?

DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

emiliom wrote: Sun Aug 27, 2023 10:11 am

That had occurred to me. I wondered if I put too much solder in and one or more pins had bridged where I couldn’t see them. I couldn’t see any obvious shorts with a magnifying glass and then I even checked with a microscope (cheap but does the job) and it really doesn’t look like there are any shorts where solder joints are visible on either side of the board. I started checking continuity between the middle 2 pins of the USB connectors, and then between each of those pins and pin 1 and on one I got a tone. I’m not sure if that could be the issue as a short between pin 1 and 2 wouldn’t cause a short to GND would it?

By "I got a tone", I assume you mean a few ohms of resistance on a continuity check. I would not expect a short between pin 1 and pin 2 to create a short to ground, because that would imply pin 2 is shorted to ground.

emiliom wrote: Sun Aug 27, 2023 10:11 am

Regardless, it’s still a short so I tried to remove the connector, but without success so far. I’ve used wick and a de-solder pump to remove as much as I can, and it barely looks like there’s any visible solder holding it in place, but it still appears to be rock solid.
I suspect I may end up having to remove the connectors one by one, if for no other reason than to eliminate them, but it seems it’s going to be very difficult.
I’ve already looked on YouTube for advice on de-soldering connectors but I’ve not found anything different to what I’m already doing. Would you (or anyone) have any tips on removing the Ethernet connector?

I got a desoldering gun because of my frustration with wick. But I think you mentioned you had a hot air gun, so I'd try that.

If upon careful inspection the connector doesn't appear to have a short on it, then it seems unlikely that the connector is the problem. Are you certain you are looking in the right spot? You didn't mention which connector had the problem, but if we are talking about the electrical net USB7_D_N for example, that shows up in three places right next to VBUS: P5, J23, and J1.

LeeW
Posts: 28
Joined: Tue Jul 25, 2023 9:13 pm
Has thanked: 9 times
Been thanked: 26 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by LeeW »

The reason the SNAC not working is kind of annoying is because I wanted the option of using my G-Con45 for Point Blank.

DeKay wrote: Sat Aug 26, 2023 11:46 pm

Nice! Congrats!

Can you try to run the memory tester and report back the maximum speed you are able to get out of the SDRAM? I'm curious how it compares to the MiSTer SDRAM expansion boards.

I ran the mem test and it settled to 146MHz for a while and then dropped to 145MHz. To be honest though, I thought it was kind of a waste of time carrying on the test as the memory frequency is reported as 143MHz by the manufacturer and the maximum frequency used by cores is 126MHz. It's definitely worth running the test to prove that it's working, I just don't see the point in doing it for an extended period of time.

emiliom wrote: Sun Aug 27, 2023 10:11 am

Would you (or anyone) have any tips on removing the Ethernet connector?

Obviously I'm not the most experienced haha, but I think the braid is good to remove most of the excess and then a desoldering pump with a heat resistant tip would help as it can be directly placed on the soldering iron tip briefly, so the solder doesn't have as much chance to cool down. Looking online I found you could get a heated desoldering pump which looks even better. The next step up from that is a heat gun or desoldering gun as DeKay mentioned. But these are expensive. For me I don't think it's worth it with how infrequently it will be used.

I can't really offer too much help with your connection problem as I also have my own USB problem, waiting on the replacement parts for U4 to arrive. The only things I can think is that the USB controller has very thin pins and seems extremely easy to bridge neighbouring pins when soldering. Other than that, U4 and the LEDs are a bit difficult to tell the polarity of, may be worth double checking they are the right orientation (although I don't think the LEDs would affect the circuit?).

Edit: Also the other thought is that continuity detection may be misleading if it can find a connection the through a short much further away, the "wrong" way through the circuit.

DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

LeeW wrote: Sun Aug 27, 2023 4:57 pm
DeKay wrote: Sat Aug 26, 2023 11:46 pm

Can you try to run the memory tester and report back the maximum speed you are able to get out of the SDRAM? I'm curious how it compares to the MiSTer SDRAM expansion boards.

I ran the mem test and it settled to 146MHz for a while and then dropped to 145MHz. To be honest though, I thought it was kind of a waste of time carrying on the test as the memory frequency is reported as 143MHz by the manufacturer and the maximum frequency used by cores is 126MHz. It's definitely worth running the test to prove that it's working, I just don't see the point in doing it for an extended period of time.

Thanks for this. This was indeed a worthwhile test and not a waste of time. Earlier versions of the "official" 128M SDRAM modules would not operate at frequencies this high because of the board layout, not the memory chip speed rating. That you are getting 145 MHz indicates the layout for this design is solid. I think people are seeing issues with the WIP Saturn core with slower SDRAM expansion modules.

LeeW wrote: Sun Aug 27, 2023 4:57 pm

The only things I can think is that the USB controller has very thin pins and seems extremely easy to bridge neighbouring pins when soldering. Other than that, U4 and the LEDs are a bit difficult to tell the polarity of, may be worth double checking they are the right orientation (although I don't think the LEDs would affect the circuit?).

Backward LED's won't create a short. They just won't ever light up (they pass current only one way and are "reverse biased" if in backwards).

Goodness knows what problems U4 in the wrong orientation might create, so double checking that is definitely a good idea.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

DeKay wrote: Sun Aug 27, 2023 2:41 pm
emiliom wrote: Sun Aug 27, 2023 10:11 am

That had occurred to me. I wondered if I put too much solder in and one or more pins had bridged where I couldn’t see them. I couldn’t see any obvious shorts with a magnifying glass and then I even checked with a microscope (cheap but does the job) and it really doesn’t look like there are any shorts where solder joints are visible on either side of the board. I started checking continuity between the middle 2 pins of the USB connectors, and then between each of those pins and pin 1 and on one I got a tone. I’m not sure if that could be the issue as a short between pin 1 and 2 wouldn’t cause a short to GND would it?

By "I got a tone", I assume you mean a few ohms of resistance on a continuity check. I would not expect a short between pin 1 and pin 2 to create a short to ground, because that would imply pin 2 is shorted to ground.

No I mean the multimeter beeps when put on continuity check setting. But yes, if I switch to resistance check I get <1 Ohm. As I said previously I really can't see any obvious shorts / bridges on either side of the board by visual inspection with magnifying tools, so I'm thinking the most likely thing is that one (or more) of the USB port power pins (1 & 5 on the dual connectors and just pin 1 on USB 7) is shorted to the shell/s, and then pin 2 of the above connector is either also shorted to it's shell, or to pin 1. So I think I've got at least 2 shorts.

DeKay wrote: Sun Aug 27, 2023 2:41 pm
emiliom wrote: Sun Aug 27, 2023 10:11 am

Regardless, it’s still a short so I tried to remove the connector, but without success so far. I’ve used wick and a de-solder pump to remove as much as I can, and it barely looks like there’s any visible solder holding it in place, but it still appears to be rock solid.
I suspect I may end up having to remove the connectors one by one, if for no other reason than to eliminate them, but it seems it’s going to be very difficult.
I’ve already looked on YouTube for advice on de-soldering connectors but I’ve not found anything different to what I’m already doing. Would you (or anyone) have any tips on removing the Ethernet connector?

I got a desoldering gun because of my frustration with wick. But I think you mentioned you had a hot air gun, so I'd try that.

I was worried about using the hot air gun. As I said in a previous post I initially want to try and reflow the SMD components, but in the end I didn't. The week I spent doing some soldering practice, the through hole soldering was ok (a bit blobby in places), the SMD soldering with the iron went really well, but the reflow test was a non starter. I was just using a few little practice boards and I couldn't even get the solder paste to melt. And it was lead based so it should have been easier. I had the air gun up to 300C at one point and it still wouldn't melt. I put a thermocouple on the board and where I was heating it reached over 220C so that should easily have been enough to melt the solder, but no success. I just ended up damaging part of the board which is why I just used the iron to solder everything on the MMS board. So yeah, I was a bit wary of trying to desolder using the hot air gun. I did try it though, and I taped up everything around with heat resistant tape to try and protect it, but I still can't melt the solder. All that happened is the protective tape started to melt, and that's supposed to be resistant to 200C so I stopped out of fear of damaging the board. I think I'll have to get a desoldering gun nothing else seems to be working at the moment.

DeKay wrote: Sun Aug 27, 2023 2:41 pm

If upon careful inspection the connector doesn't appear to have a short on it, then it seems unlikely that the connector is the problem. Are you certain you are looking in the right spot? You didn't mention which connector had the problem, but if we are talking about the electrical net USB7_D_N for example, that shows up in three places right next to VBUS: P5, J23, and J1.

I'm not certain I'm looking in the right spot, but this is what I've narrowed it down to based on my throughs as explained above. JP4, JP9, JP10 and J23 all have shorts on pin 1 (and 5 for the dual connectors), but JP9 is the one that also has a short on pin 2.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

LeeW wrote: Sun Aug 27, 2023 4:57 pm
emiliom wrote: Sun Aug 27, 2023 10:11 am

Would you (or anyone) have any tips on removing the Ethernet connector?

Obviously I'm not the most experienced haha, but I think the braid is good to remove most of the excess and then a desoldering pump with a heat resistant tip would help as it can be directly placed on the soldering iron tip briefly, so the solder doesn't have as much chance to cool down. Looking online I found you could get a heated desoldering pump which looks even better. The next step up from that is a heat gun or desoldering gun as DeKay mentioned. But these are expensive. For me I don't think it's worth it with how infrequently it will be used.

Funnily enough a plastic connector would have been easier. At least I could have just melted it and then removed the pins! I think for me it'll have to be a heated desoldering pump or gun as I don't think I can get the connector off another way without damaging the board. As I said to DeKay I've not had much success with my hot air gun.

LeeW wrote: Sun Aug 27, 2023 4:57 pm

I can't really offer too much help with your connection problem as I also have my own USB problem, waiting on the replacement parts for U4 to arrive. The only things I can think is that the USB controller has very thin pins and seems extremely easy to bridge neighbouring pins when soldering. Other than that, U4 and the LEDs are a bit difficult to tell the polarity of, may be worth double checking they are the right orientation (although I don't think the LEDs would affect the circuit?).

Hopefully your problem will just be the wrong part and swapping out U4 will fix it. I'm sure my pins aren't bridged as I checked them with a microscope. My LEDs are definitely the right way round. I had to check them with the magnifying glass but you can just about make out which end is which, plus I put 2V through all of them before fitting to make sure they lit up and were the right way round. U4 was a little bit suspect but I think I have it the right way round. I don't see a 'proper' dot on pin 1 (although the datasheet shows one), but from looking at the datasheet there's only one 6 pin package with a 6 letter code and pin 1 is bottom left when the letters are the right way up, so that's how I've installed it. That dark thing under the A might be a dot, and the one under the R, which would also fit with the description on the datasheet.

U4.jpg
You do not have the required permissions to view the files attached to this post.
LeeW
Posts: 28
Joined: Tue Jul 25, 2023 9:13 pm
Has thanked: 9 times
Been thanked: 26 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by LeeW »

For U4, in respect to the printed board and the board connections, the bottom left pin (next to white dot) is pin 1 (5V in). Bottom middle is pin 2 (ground). Bottom right pin is pin 3 (USB_EN). And the pin numbers continue counter clockwise with 4 is the top right (-FLAG), 5 is top middle (ILIM), 6 is top left (VBUS). So the datasheet for the part does seems to match up and looks like it's correctly installed in your picture.

(EDIT: In respect to the reverse side of the board) For JP11 and JP4 the rightmost of the 4 pins is ground. JP9, JP10, J23 are facing the opposite direction so the leftmost of the 4 pins is ground.
The leftmost pin of JP11 and JP4 is VCC. JP9, JP10, J23 are again the opposite direction so the rightmost pins are VCC. And J1 is just a header for J23, so the pins are basically the same. Same for J22 and JP4, J20 and JP9, J21 and JP10.

Here's an image with the VCC pins highlighted red and ground highlighted brown, the unmarked pins are +-data:

Jthnso6.jpg

The red USB "VCC" pins are connected to U4 pin 6 "VBUS" so you should get continuity between all of these (excluding pin 1 of JP11 which is not connected to anything).

Confusingly the pin layout of the FE2.1 on the multisystem pdf is 90 degrees rotated pin1 is bottom left while on the board it is top left. You should get continuity from the data pins on the FE2.1 and the associated USB components.

You do not have the required permissions to view the files attached to this post.
DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

emiliom wrote: Mon Aug 28, 2023 12:04 pm

I was worried about using the hot air gun. As I said in a previous post I initially want to try and reflow the SMD components, but in the end I didn't. The week I spent doing some soldering practice, the through hole soldering was ok (a bit blobby in places), the SMD soldering with the iron went really well, but the reflow test was a non starter. I was just using a few little practice boards and I couldn't even get the solder paste to melt. And it was lead based so it should have been easier. I had the air gun up to 300C at one point and it still wouldn't melt. I put a thermocouple on the board and where I was heating it reached over 220C so that should easily have been enough to melt the solder, but no success. I just ended up damaging part of the board which is why I just used the iron to solder everything on the MMS board. So yeah, I was a bit wary of trying to desolder using the hot air gun. I did try it though, and I taped up everything around with heat resistant tape to try and protect it, but I still can't melt the solder. All that happened is the protective tape started to melt, and that's supposed to be resistant to 200C so I stopped out of fear of damaging the board. I think I'll have to get a desoldering gun nothing else seems to be working at the moment.

DeKay wrote: Sun Aug 27, 2023 2:41 pm

If upon careful inspection the connector doesn't appear to have a short on it, then it seems unlikely that the connector is the problem. Are you certain you are looking in the right spot? You didn't mention which connector had the problem, but if we are talking about the electrical net USB7_D_N for example, that shows up in three places right next to VBUS: P5, J23, and J1.

I'm not certain I'm looking in the right spot, but this is what I've narrowed it down to based on my throughs as explained above. JP4, JP9, JP10 and J23 all have shorts on pin 1 (and 5 for the dual connectors), but JP9 is the one that also has a short on pin 2.

JP9 pin 2 is USB1_D_N and only goes to JP9, J20, and the FE2.1. There aren't a lot of places for a short to happen if you have carefully inspected FE2.1 because a short around J20 would be obvious. That seems to leave the potential problem at J9. So this is my guess as to what is going on (apologies if this is your line of thinking already).

USB.jpg

When you solder from the bottom, the solder wicks up through the via to top. There looks to be a lot of solder blobbed around the these ground pins on the outside of the USB connector. If there is just as much of a blob underneath the connector and it is bridging to the adjacent pin, that could create the problems you are seeing. I think the shorts are coming from underneath the connector on the top side of the board. It wouldn't take much extra solder to create a bridge, especially between the two top shield pins as the distance there is shortest. You have a short between VBUS and GND, so the highest chance of a short is Pin 9 (GND) and Pin 1 (VUSB) on one of the USB connectors (you wouldn't care about a short between Pin 4 and 10 on the same connector because both are GND). You probably have another bridge between Pin 1 and 2 on JP9 on the top layer underneath the connector as well. Maybe you are "lucky" and the connector on JP9 is responsible for both shorts.

Your hot air temp of "up to 300C" is (probably much) too low for removing a through hole part like this, built with substantial amounts of metal and numerous connections to a big fat ground plane that is wicking away the heat from the part like crazy. Remember that all the pins need to be heated across all the pins of the connector from top to bottom. I usually run my hot air around 350C IIRC. Do a lot more reading on this and watch some videos. I think what you'll find is that some people will blob also solder all over the pins to increase the thermal conductivity between all the pins while heating. Then they'll go in with wick and clean up each individual hole.

Good hunting.

Edit: looking again at the PCB picture, maybe the amount of solder blobbing isn't that excessive given the size of the hole and the outline on the silkscreen. But it is the only thing I can think of that adds up.

You do not have the required permissions to view the files attached to this post.
emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

DeKay wrote: Mon Aug 28, 2023 4:16 pm

JP9 pin 2 is USB1_D_N and only goes to JP9, J20, and the FE2.1. There aren't a lot of places for a short to happen if you have carefully inspected FE2.1 because a short around J20 would be obvious. That seems to leave the potential problem at J9. So this is my guess as to what is going on (apologies if this is your line of thinking already).

That's what I'd been thinking. There definitely isn't one around J20, and I've checked my FE2.1 under the microscope several times now and I don't see any bridges there either so JP9 is my prime suspect for the short to pin 2.

DeKay wrote: Mon Aug 28, 2023 4:16 pm

When you solder from the bottom, the solder wicks up through the via to top. There looks to be a lot of solder blobbed around the these ground pins on the outside of the USB connector. If there is just as much of a blob underneath the connector and it is bridging to the adjacent pin, that could create the problems you are seeing. I think the shorts are coming from underneath the connector on the top side of the board. It wouldn't take much extra solder to create a bridge, especially between the two top shield pins as the distance there is shortest. You have a short between VBUS and GND, so the highest chance of a short is Pin 9 (GND) and Pin 1 (VUSB) on one of the USB connectors (you wouldn't care about a short between Pin 4 and 10 on the same connector because both are GND). You probably have another bridge between Pin 1 and 2 on JP9 on the top layer underneath the connector as well. Maybe you are "lucky" and the connector on JP9 is responsible for both shorts.

It will be lucky if both shorts are under the same connector, but it wouldn't surprise me if I've shorted VBUS under more than one connector - they're all a bit blobby. But I'll start with JP9 and then I can eliminate them one at a time.

DeKay wrote: Mon Aug 28, 2023 4:16 pm

Your hot air temp of "up to 300C" is (probably much) too low for removing a through hole part like this, built with substantial amounts of metal and numerous connections to a big fat ground plane that is wicking away the heat from the part like crazy. Remember that all the pins need to be heated across all the pins of the connector from top to bottom. I usually run my hot air around 350C IIRC. Do a lot more reading on this and watch some videos. I think what you'll find is that some people will blob also solder all over the pins to increase the thermal conductivity between all the pins while heating. Then they'll go in with wick and clean up each individual hole.

Yes, I think this is going to be a case of watching some more videos and getting a lot of practice in. I had another go on a test board at 400C. This time the solder started melting but I still burnt the board. But at least I got the solder hot enough and I didn't burn the board as badly as the first couple of times. So some progress.

DeKay wrote: Mon Aug 28, 2023 4:16 pm

Good hunting.

Edit: looking again at the PCB picture, maybe the amount of solder blobbing isn't that excessive given the size of the hole and the outline on the silkscreen. But it is the only thing I can think of that adds up.

I would agree. In the absence of any obvious bridges on the surface and given that VBUS is clearly shorted it seems logical the short is under a connector that connects to VBUS.

DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

emiliom wrote: Mon Aug 28, 2023 6:18 pm

Yes, I think this is going to be a case of watching some more videos and getting a lot of practice in. I had another go on a test board at 400C. This time the solder started melting but I still burnt the board. But at least I got the solder hot enough and I didn't burn the board as badly as the first couple of times. So some progress.

Sounds good. 400C sounds on the hotter side so you might back that off some to start. One key to prevent burning is keep the hot air gun tip moving all the time. You don't want to blast one spot - you want to bring one area up to temperature at the same time. You'll get there with more research and practice.

jca
Top Contributor
Posts: 1911
Joined: Wed May 27, 2020 1:59 pm
Has thanked: 145 times
Been thanked: 454 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by jca »

Kapton tape is also very useful: you tape around the part you want to desolder in order to reduce the heat on the surrounding components while you move the tip.

DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

jca wrote: Mon Aug 28, 2023 9:53 pm

Kapton tape is also very useful: you tape around the part you want to desolder in order to reduce the heat on the surrounding components while you move the tip.

He said he used heat insulating tape in his first tries and I assumed he meant kapton tape. That stuff will melt if things get hot enough (at least the stuff I bought from Ali does :lol: )

Aluminum foil works really well too if in a spot where you can press it around something to keep it in place.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

jca wrote: Mon Aug 28, 2023 9:53 pm

Kapton tape is also very useful: you tape around the part you want to desolder in order to reduce the heat on the surrounding components while you move the tip.

Yes as DeKay said I did try that, but things got a bit melty. My technique needs improving. It wasn’t branded Kapon though. I see that’s rated up to 280C. The tape I got was rated up to 260C, so not far off.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

DeKay wrote: Mon Aug 28, 2023 11:23 pm

Aluminum foil works really well too if in a spot where you can press it around something to keep it in place.

I didn’t think of that - that’s actually a really good idea. I’ll try that on my next test.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

Hooray!!! Praise the Lord!!! After 2 days I finally got that bleedin Ethernet connector off! :D In the end I decided to try the heated desoldering pump. It was only £30 and it takes up no space on my desk as I can put it away in a drawer. I’m sure it wasn’t as good as a gun would be but it does a good job of cleaning out the central pins. The shell connectors…not quite as good because the legs are bent and holes are big. I had to resort to a bit of wick there. But after about 15 minutes of desoldering pump, wick, and soldering iron I finally managed to pull it out.

C17B308F-DF20-426D-8528-613F222FA0C9.jpeg

It did lose a couple of legs in the struggle but it wasn’t too badly damaged in the end.

So, the good news is my short on pin 2 is gone! The bad news is the short on VBUS remains. :( Ah well, I was never that hopeful that I’d get them both on the same connector.
I did have an idea though. I’m thinking before trying to remove another connector, I should just try and desolder pin 1 (and pin 5 on the dual ports) and then check for the VBUS short after each pin is desoldered. I might get lucky and desolder the bridge without having to remove the entire connector.

You do not have the required permissions to view the files attached to this post.
DeKay
Posts: 34
Joined: Sun Apr 02, 2023 10:11 pm
Has thanked: 16 times
Been thanked: 25 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by DeKay »

emiliom wrote: Tue Aug 29, 2023 12:47 am

So, the good news is my short on pin 2 is gone! The bad news is the short on VBUS remains. :( Ah well, I was never that hopeful that I’d get them both on the same connector.
I did have an idea though. I’m thinking before trying to remove another connector, I should just try and desolder pin 1 (and pin 5 on the dual ports) and then check for the VBUS short after each pin is desoldered. I might get lucky and desolder the bridge without having to remove the entire connector.

Nice! One thing I might suggest though if you haven't already is to check the connector itself for shorts now that it is off. It is highly unlikely but you might have gotten a bad batch.

You can see from the layout that the spacing between Pin 9 and Pin 1 is about the same as that between Pin 2 and Pin 1. Pin 9 will take a lot more solder though, so the chances of a bridge are even higher from that pin to Pin 1 than between Pin 2 and Pin 1. So I think you are headed in the right direction.

But... I might suggest a modification to your idea and desolder Pin 1 (VBUS) and Pin 9 (Shield / Ground) on each of the dual ports, starting with the "blobbiest" ones first. A short from Pin 5 to the shield pin is less likely given the greater distance.

User avatar
multisystem
Posts: 135
Joined: Mon Jul 05, 2021 11:34 am
Has thanked: 114 times
Been thanked: 172 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by multisystem »

rezendes wrote: Sat Aug 26, 2023 8:24 am

@multisystem I was wondering if there are any plans to release the PCB files for the Pi-MMS as well?

Yes, we will.

LeeW
Posts: 28
Joined: Tue Jul 25, 2023 9:13 pm
Has thanked: 9 times
Been thanked: 26 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by LeeW »

USB! All ports working and LEDs lighting up when active. Also was able to test the optical audio and it is working as well.

jn7YOMb.jpg

A couple of hiccups with mispicked parts but got there eventually. Have gone through and desoldered all the components that were mispicks. I've found the best way to desolder a component is to actually add MORE solder. Just make a few big blobs of solder which then heat up quickly and retain the heat a bit better and you can pull the part out quite easily. Then remove the excess solder with a wick or pump. For the 40 pin headers there's just too many pins and too large an area to attempt to remove it by heating all the pins, I think destroying the shroud is the best approach, then remove the pins one by one... That's 160 pins I've removed today. What a tedious mistake.

You do not have the required permissions to view the files attached to this post.
Missus
Posts: 168
Joined: Fri Jan 27, 2023 12:45 am
Has thanked: 38 times
Been thanked: 37 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by Missus »

multisystem wrote: Tue Aug 29, 2023 7:33 am
rezendes wrote: Sat Aug 26, 2023 8:24 am

@multisystem I was wondering if there are any plans to release the PCB files for the Pi-MMS as well?

Yes, we will.

And the Composite / S-video expansion? Pretty please? :)

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

Finally found the short this afternoon! Ah the relief... :lol:

DeKay wrote: Tue Aug 29, 2023 2:08 am
emiliom wrote: Tue Aug 29, 2023 12:47 am

So, the good news is my short on pin 2 is gone! The bad news is the short on VBUS remains. :( Ah well, I was never that hopeful that I’d get them both on the same connector.
I did have an idea though. I’m thinking before trying to remove another connector, I should just try and desolder pin 1 (and pin 5 on the dual ports) and then check for the VBUS short after each pin is desoldered. I might get lucky and desolder the bridge without having to remove the entire connector.

Nice! One thing I might suggest though if you haven't already is to check the connector itself for shorts now that it is off. It is highly unlikely but you might have gotten a bad batch.

No I hadn't checked it but subsequently I did and the connector was fine. In fact I cleaned it up and was able to put it back. 2 of the legs were half gone but there was still enough left to solder it properly so it's all good now and connection seems sturdy.

DeKay wrote: Tue Aug 29, 2023 2:08 am

But... I might suggest a modification to your idea and desolder Pin 1 (VBUS) and Pin 9 (Shield / Ground) on each of the dual ports, starting with the "blobbiest" ones first. A short from Pin 5 to the shield pin is less likely given the greater distance.

That's logical and would be the way to go except for one thing; even with the heated desoldering pump it's a lot easier to check the central pins before the shell pins as they come out clean in one hit. The shell pins still need wicking so I thought it was a better bet to go for pins 1 & 5 first, that way if I'm lucky I save a lot of mess with the shell pins.
In the end it worked out well. Initially I was suspicious of J23 because my SNAC connector was a bit blobby and even though I could see underneath it to an extent, because of the USB7 connection to the SNAC VBUS pin I just had a bad feeling about it. So I desoldered pin 1 on J23 and the SNAC VBUS pin, but still had the short. Then I moved onto pin 1 from JP10 but still had the short, then pin 5 and it was gone! I resoldered the 4 pins, this time with a more modest amount of solder, and all was good with the continuity check on all the USB GND and VBUS connection pins. I also checked the resistance and now instead of <1 Ohm I was getting around 5MOhms.
I did have one more issue. When I powered everything up and tried a USB keyboard, all the ports worked except USB4 on JP10, which was odd because I had previously checked continuity of the data lines to FE2.1. So I rechecked the USB 3 & 4 data lines and everything still seemed ok, although sometimes I was having to touch the pin 2 or 3 times to get a signal. I thought maybe when I desoldered pins 1 & 5 on JP10 that perhaps I had affected the solder on the adjacent pin, so I took the opportunity to desolder (and deblob) the 4 data pins, resolder them and then everything worked fine.
It was definitely worth investing in that heated desoldering pump. I wish I had more room on my desk as I would have gone for the gun - I'm sure that would work better on the shell pins as you can keep suctioning for a bit longer until all the solder is out. But lesson learned - be less blobby with my through hole soldering in future.

emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

MMS_Working.JPG

Everything seems to be working ok now - HDMI, VGA, USB, Ethernet, Audio, fan - all good. I can't check the SCART or SNAC connector at the moment but I checked all the pins for continuity and shorts and all seems well. Ran the memtest for a while and got 141MHz.

MMS_Case_half_done.JPG

The case is proving a slight pain. The front 2 buttons are too tight so when you push them they don't pop back out so it looks like I'm going to have to make the case holes a little bigger so that they work correctly.
My power switch is slightly askew so the case switch doesn't slide all the way to the end. It still works but I'd still like to fix it so I need to resolder the power switch to bring the left side forward a bit which should fix it.
Also the fittings supplied for the fan aren't great. It's 4 short 3mm screws. There's a good chance you will overtighten one or more given that the fan is plastic and it takes very little force to trash the thread. I got 3 in ok but the last one I can tell is not right, and I really only tightened it a little. It would have been better to have longer screws and 4 nuts, which is what is shown in the manual but looks to be no longer supplied. Anyway, I'll get some screws and nuts myself and sort that easily enough.
And the last thing is the SD card / reset button mount / LED light guide. That had to be lined up pretty much dead on on the top front panel doesn't fit on the right hand side. It might be better to leave the underside screw a tiny bit loose and use the top screw to fix it in place so that you can line it up exactly where it needs to be.

You do not have the required permissions to view the files attached to this post.
emiliom
Posts: 104
Joined: Mon Jul 03, 2023 5:43 pm
Has thanked: 62 times
Been thanked: 27 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by emiliom »

Done! :P

MMS_Case_Complete_Rear.JPG
MMS_Case_Complete_Top_1.JPG

Just for information, a 20cm ethernet cable is fine. I found them easier without bulky connectors because there seemed to be more of them about than 15cm cables. I think longer than a 15cm USB cable will be tricky to fit in the case unless it's very flexible.
Also I notice that with the case on, the green LED isn't very visible unless you look straight down. The red LEDs you can see from a mile away. The buttons and switches all work well.
For the fan screws I got 16mm M3s with nuts. There's just enough room between the fan and the left side of the board to fit them in.

You do not have the required permissions to view the files attached to this post.
rezendes
Posts: 29
Joined: Thu Jun 29, 2023 12:57 am
Has thanked: 11 times
Been thanked: 9 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by rezendes »

Nice! Waiting for my official case to come from the UK to the me in the states.

LeeW
Posts: 28
Joined: Tue Jul 25, 2023 9:13 pm
Has thanked: 9 times
Been thanked: 26 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by LeeW »

I've completed all boards now.

Bottom:

0hjemoD.jpg

Top, in case. The camera really picks up the flux residue huh, not this noticeable to the naked eye.

QBdEhjz.jpg

Front cover on:

Hq3dOhi.jpg

The fan is very loud and appears to be a 30mm fan so the smallest Noctua fan won't fit... Also if the screws are tightened the fan seems to be in contact with the enclosure while running. Will probably be running this fan-less.

You do not have the required permissions to view the files attached to this post.
dcubed
Posts: 51
Joined: Sat Sep 03, 2022 5:44 pm
Has thanked: 26 times
Been thanked: 24 times

Re: Open Source release of MiSTer Multisystem Console hardware all-in-one I/O design

Unread post by dcubed »

Great job with it!

I got my MiSTer Multisystem about a year ago and absolutely love my one! :D

Post Reply