1088XEL Atari ITX Motherboard DIY Builders Thread

[Dropcheck]

16 hours ago, Bikerbob said:

Just to give everyone who contributed closure.. The U1mb board is sturdier than I thought.

 

Works just fine.

 

Now I just have to get a joy2pic to program the pic chips (that all my other programmers wont.. LOL) So.. the build is on hold until I can get some chips programmed.

 

James

Yep, definitely need to update the U1MB Bios. 

[+mytek]

On 8/15/2019 at 4:02 AM, Dropcheck said:

The PIC chips can be programmed with INDUSTRY standard PIC programmers as long as they support that particular PIC.   Some do, some don't.   The PIC chips used are chosen for their specific purpose, not for whether they are supported by a third party's PIC programmer.  Since there is not room on the mainboard to include InCircuit programming capability, your programmer has to be able to do out of circuit programming.   If you want to use Microchip's InCircuit programmer you'll have to provide a configurable hardware device to accommodate the different programming pinouts of the different PICs.

Thank you for covering all the key points on this .

 

When I first started development on the TK-II aspect, I was planning on using a PIC16F88 which I believe is programmable by the TL866II programmer, but as the code quickly grew I passed the available memory that chip had to offer and was forced to use its bigger cousin the PIC16F1847 instead. Part of the problem was my choice of code development suites. I used something called FlowCode which is a very high level code development tool that makes writing code extremely simple, but comes at the cost of efficiency (it writes very bloated code). As for the choice of the PIC12F1571 used for the VGATE chip, the application had a very high speed requirement that this particular chip just squeaked by to meet, and that was even factoring in that I wrote the code virtually at the machine level. Then along came MIDI, and it's brother the PIC12F1572 was chosen since it contained a UART which was needed, while still keeping the package size to a minimum. I have since gone to the 12F1572 for the VGATE chip in the 1088XLD as well, just to keep things simpler in the BOM, and it's actually sometimes cheaper than the 12F1571.

 

The funny part is that all of the chips I selected have been in production for at least six or more years, and due to their popularity are produced in numbers that allow the chips to be bought for less than $2 a piece (often times closer to a buck and a half). The PICkit2 which is the older Microchip programmer (PICkit3 being the current one), didn't support the 12F1571 or 12F1572 out of the gate, although it did support the 16F1847. However unlike the popular TL866II programmer, 3rd parties stepped up and modified the original device file, adding in the 12F157x series, as well as quite a few other newer chips. I also have  the newer TL866II Plus model of that programmer and was greatly disappointed to not only see that the parent company hasn't updated their PIC device files for like forever, but that no 3rd party has come along to do so either.

 

So this brings us to why the JOY2PIC was developed. And to tell you the truth even though I have both a PICkit2 and a PICkit3 I prefer to use the JOY2PIC method of flashing the chips used in my products. It's just plain easier. This means I only use the PICkit programmers for firmware development.

 

 

[wildstar87]

I'm almost ready to start my board up.  I'm only using a single pokey for now, which pokey slot should I put it in?

[+DrVenkman]

4 hours ago, wildstar87 said:

I'm almost ready to start my board up.  I'm only using a single pokey for now, which pokey slot should I put it in?

Left. That is the “main” POKEY that handles SIO, keyboard and mono sound. 

[+mytek]

The LEFT Pokey is actually marked as U5 on the board silkscreen (I probably should have marked the Pokey's as LEFT and RIGHT ).

 

[wildstar87]

Thanks, on another note, I'm toying with the idea of using a 600XL keyboard as the keyboard for this.  Is there a relatively simple way to interface the keyboard with the 1088XEL, or would I basically have to create a PS/2 interface for the keyboard?

[Kyle22]

6 hours ago, mytek said:

The LEFT Pokey is actually marked as U5 on the board silkscreen (I probably should have marked the Pokey's as LEFT and RIGHT ).

 

I wonder why the sinistral side became the default for mono, ignoring the dextral side completely...

 

[+mytek]

4 hours ago, Kyle22 said:

I wonder why the sinistral side became the default for mono, ignoring the dextral side completely...

 

 

You'd have to ask Chuck Steinman, since he established the standard BITD.

 

[Kyle22]

No, that is the standard of EVERYTHING. If you're mono, then plug into the left side jack.

 

More info: https://www.gearslutz.com/board/newbie-audio-engineering-production-question-zone/813925-there-reason-behind-why-left-output-stereo-keyboard-l-mono.html

 

 

Edited August 31, 2019 by Kyle22
added info for clarity

[+mytek]

5 hours ago, Kyle22 said:

No, that is the standard of EVERYTHING. If you're mono, then plug into the left side jack.

 

More info: https://www.gearslutz.com/board/newbie-audio-engineering-production-question-zone/813925-there-reason-behind-why-left-output-stereo-keyboard-l-mono.html

 

 

I thought you were referring to why the original Pokey was assigned as the left channel, and the added Pokey as the right channel, instead of the other way around. That's why I suggested that the original guy who implemented this idea would be the one to ask.

 

Edit: In my mind it could have gone either way, despite what is considered standard in the audio world, because people are always free to create their own standards .

 

[_The Doctor__]

most items work from left to right in the world of standards and electronics country of origin... so everything kind of follows along...

if the fervor and frenzied pace of electronics had spread as quickly and greatly elsewhere, let's say a country that wrote from right to left or top to bottom right to left we all might be talking about why things get slotted at the top first or the right first... sometimes it's just this simple.

[ebiguy]

Hi,

 

I built a MIDI XEL II from @mytek but I had no luck making it work successfully in my 1088XEL.

I think I don't connect the board to the right headers in the 1088XEL.

SIO part is easy to connect as there is an exact match with the SIO-AUX.

But I am puzzled with the AUDIO part.

I did not use the L-OUT and R-OUT from AUDIO-AUX header but L-IN and R-IN from the MPBI - maybe I should have.

I did it this way because I am not in electronics and I found it quite logical to connect the MIDI-XEL II L-OUT and R-OUT into the XEL L-IN and R-IN as the MIDI board is generating sound as an input for the XEL.

Something strange also when both audio wires are connected this way, there are some vertical lines appearing on screen.

 

Could you tell me what is the right way to connect the MIDI XEL II board ?

 

And if L-OUT and R-OUT should be used, can you tell me what is the use of L-IN and R-IN (explanation at feature level not electronics level, please) ?

 

Thanks

[+MacRorie]

1 minute ago, ebiguy said:

Hi,

 

I built a MIDI XEL II from @mytek but I had no luck making it work successfully in my 1088XEL.

I think I don't connect the board to the right headers in the 1088XEL.

SIO part is easy to connect as there is an exact match with the SIO-AUX.

But I am puzzled with the AUDIO part.

I did not use the L-OUT and R-OUT from AUDIO-AUX header but L-IN and R-IN from the MPBI - maybe I should have.

I did it this way because I am not in electronics and I found it quite logical to connect the MIDI-XEL II L-OUT and R-OUT into the XEL L-IN and R-IN as the MIDI board is generating sound as an input for the XEL.

Nope, that is the correct connections.

1 minute ago, ebiguy said:

Something strange also when both audio wires are connected this way, there are some vertical lines appearing on screen.

 

Could you tell me what is the right way to connect the MIDI XEL II board ?

 

Can you send a picture of your board? 

[ebiguy]

29 minutes ago, MacRorie said:

Nope, that is the correct connections.

 

Can you send a picture of your board? 

Here they are.

The first one gives an overview and the other ones focusses on each side of the wires

 

 

 

 

 

 

Edited September 1, 2019 by ebiguy
Good pictures this time !

[+MacRorie]

31 minutes ago, ebiguy said:

Here they are.

The first one gives an overview and the other ones focusses on each side of the wires

 

 

 

 

 

 

Looks like your have your wires swapped.  /RST is green (coming off of the MIDI bd), yes?  You have that connected to R-IN on the MPBI

Your Ground is connected to +5

R-IN is connected to /RST and L-IN is connected to /HLT

 

[ebiguy]

35 minutes ago, MacRorie said:

 

Looks like your have your wires swapped.  /RST is green (coming off of the MIDI bd), yes?  You have that connected to R-IN on the MPBI

Your Ground is connected to +5

R-IN is connected to /RST and L-IN is connected to /HLT

 

 

Such a BIG BIG mistake !!!!!

And now that I look at my pictures, I see that the pin meanings were written on the main board...

Thank you for pointing out my mistake.

I am going to try with the right connections but now I fear that either the MIDI XEL II board or the S2 board has been damaged with +5V instead of GND !!!

I let you know.

 

EDIT: I understand now why I did this mistake: I used the XEL logical diagram to know where the connections should go and the pin 1 appears on the right on this diagram and on the left on the board. I should have looked at the mask on the board itself.

Edited September 1, 2019 by ebiguy
Added the reason for my mistake

[ebiguy]

With the right connections, it works like a charm.

I am surprised (but happy) that nothing was damaged.

[+mytek]

My apologies ebiguy for the transposition (left to right orientation) of the signals as they're depicted on the schematic vs. reality. But in my defense at least I clearly labeled them on the silk screen of the board as you discovered.

 

Just a warning to others, serious damage can be done to the U1MB if +5v or GND is ever accidentally connected to M1 or VB on the MPBI header. I know this from personal experience . The other stuff is pretty robust, including the Atari VLSI chips, but it's always good to triple check connections before powering up the system after an upgrade.

 

BTW, I like the new board you developed that flashjazzcat has a sample of. Do you plan to release that design?

 

[Mr Robot]

46 minutes ago, mytek said:

BTW, I like the new board you developed that flashjazzcat has a sample of. Do you plan to release that design?

Was going to say this. 

I'm hoping for a release of this board also!

 

[+Stephen]

3 hours ago, mytek said:

My apologies ebiguy for the transposition (left to right orientation) of the signals as they're depicted on the schematic vs. reality. But in my defense at least I clearly labeled them on the silk screen of the board as you discovered.

 

Just a warning to others, serious damage can be done to the U1MB if +5v or GND is ever accidentally connected to M1 or VB on the MPBI header. I know this from personal experience . The other stuff is pretty robust, including the Atari VLSI chips, but it's always good to triple check connections before powering up the system after an upgrade.

 

BTW, I like the new board you developed that flashjazzcat has a sample of. Do you plan to release that design?

 

Well, I sadly learned that plugging in ANTIC backwards gets it so hot it will blister a finger and it obviously dies in the process

[+mytek]

5 minutes ago, Stephen said:

Well, I sadly learned that plugging in ANTIC backwards gets it so hot it will blister a finger and it obviously dies in the process

I don't understand, since I can drive my car backwards all day without it melting down (assuming I keep it below 15MPH ).

 

[ebiguy]

16 hours ago, mytek said:

BTW, I like the new board you developed that flashjazzcat has a sample of. Do you plan to release that design?

 

 

I am happy to ask as I was unsure anyone would be interested in Rapidus as everyone thinks it is too expensive (hence the other thread you started about overclocking).

 

I send everything to @flashjazzcat so he can publish it on his website but maybe the best is if you can add it in your XEL page.

Would be great !

BTW, feel free to fix any minor issues that you could see.

As I am not in electronics, I did not expand GND on one side and +5V on the other side of the PCB as you did for the MIDI XEL board, because I don't know the benefits or the drawbacks.

But without any modification, it seems to work for Jon and me.

 

I would be pleased to see the board on your XEL-CF3 web page.

Here are the KiCad files and the Gerber files to build the board without dealing with KiCad.

 

And now for the explanations (maybe obvious for expert but there are the others like me):

- You can use the original XEL-CF3 BOM files as the board uses the same components (only layout has been changed and headers have been added).

- The 4 components 74F00, 74F32, 74F138 and 74F245 are not numbered the same way as in the BOM. So do not refer to U1, U2, U3 and U4 but look at the component name (74F00, 74F32... which is written on the board).
- There is a misfit between PN2222 form factor and the holes on the board (I did not find what I wanted in KiCad). The component listed in the BOM has the 3 legs aligned, but on the board the middle leg is behind. Using pliers, you can adapt the legs or use another PN2222 you have with the right form factor.

- Beware that the MIDI header MUST be right angle (I mean horizontal) otherwise it will be impossible to use when Rapidus will be above !
 

Then Rapidus should sit on top of 4 sockets to go over the new XEL-CF3 board.

 

xel-cf3-1.6 KiCad.zip xel-cf3-1.6 gerber.zip

[+tf_hh]

20 hours ago, mytek said:

Just a warning to others, serious damage can be done to the U1MB if +5v or GND is ever accidentally connected to M1 or VB on the MPBI header. I know this from personal experience . The other stuff is pretty robust, including the Atari VLSI chips, but it's always good to triple check connections before powering up the system after an upgrade.

 

Just for the records... Connecting any I/O to ground of the Xilinx XC95...XL series is harmless. Connecting an I/O directly to +5 volts without current-limiting resistor will kill the XL series (3.3 volts models with 5 volt tolerance) immediately.

 

[+mytek]

2 hours ago, tf_hh said:

 

Just for the records... Connecting any I/O to ground of the Xilinx XC95...XL series is harmless. Connecting an I/O directly to +5 volts without current-limiting resistor will kill the XL series (3.3 volts models with 5 volt tolerance) immediately.

 

 

Yes accidentally shorting 5v to an I/O is exactly how I killed one of my U1MB boards. It was the third such incident, with a total of 3 U1MB boards that I killed during my 1088XEL development cycle. Of course I was doing a lot of unusual stuff during this process, constantly removing and re-inserting boards,connecting test probes and makeshift jumpers, and worse of all sometimes forgetting that the board was live when doing so. Luckily those days are mostly behind me, because now I'm down to only one working U1MB. Hardware development is expensive .

 

[+mytek]

6 hours ago, ebiguy said:

 

I am happy to ask as I was unsure anyone would be interested in Rapidus as everyone thinks it is too expensive (hence the other thread you started about overclocking).

 

I send everything to @flashjazzcat so he can publish it on his website but maybe the best is if you can add it in your XEL page.

Would be great !

BTW, feel free to fix any minor issues that you could see.

As I am not in electronics, I did not expand GND on one side and +5V on the other side of the PCB as you did for the MIDI XEL board, because I don't know the benefits or the drawbacks.

But without any modification, it seems to work for Jon and me.

 

I would be pleased to see the board on your XEL-CF3 web page.

Here are the KiCad files and the Gerber files to build the board without dealing with KiCad.

 

And now for the explanations (maybe obvious for expert but there are the others like me):

- You can use the original XEL-CF3 BOM files as the board uses the same components (only layout has been changed and headers have been added).

- The 4 components 74F00, 74F32, 74F138 and 74F245 are not numbered the same way as in the BOM. So do not refer to U1, U2, U3 and U4 but look at the component name (74F00, 74F32... which is written on the board).
- There is a misfit between PN2222 form factor and the holes on the board (I did not find what I wanted in KiCad). The component listed in the BOM has the 3 legs aligned, but on the board the middle leg is behind. Using pliers, you can adapt the legs or use another PN2222 you have with the right form factor.

- Beware that the MIDI header MUST be right angle (I mean horizontal) otherwise it will be impossible to use when Rapidus will be above !
 

Then Rapidus should sit on top of 4 sockets to go over the new XEL-CF3 board.

 

xel-cf3-1.6 KiCad.zip 49.22 kB · 5 downloads xel-cf3-1.6 gerber.zip 101.62 kB · 3 downloads

 

I will look this over when I get a chance, and decide how best to present this on the AtariBits website since you have generously said this would be okay to do. I do not use the same PCB tools as you, so I would not be looking at changing anything myself, but if you or someone else wanted to fine tune and/or fix any known issues before I posted this on my site that would be great. I will make sure that you are given full credit for the design when I post it, and provide an OSHpark purchase link if that is okay with you. Thanks for releasing this .