647 replies to this topic

[Dropcheck]

Shouldn't some Pico PSU (like https://www.amazon.d...y/dp/B004QDP05C) suffice ?

 

That assumes some kind of external AC to DC power supply. 

 

The point here is to have an internal AC to DC power conversion.  No power bricks external to the case.  Just the power cord from wall to case.  At least that is the direction I am heading. 

[Dropcheck]

I just ran a 1088XEL external ATX PSU test, and I am glad to say it seems to work as advertised. So what am I talking about? As part of the XEL project, it was decided that it should be able to derive it's power via a standard PC type power supply so that we could also have a source of 12 volts available for things such as floppy drives, zip drives, or even a CD-ROM. So there is a built-in option to do just that via a 3-pin header on the XEL labeled 'PSU'.

 

 

An ATX PSU has two wires of interest on the main power connector. The purple wire (Standby Power) has an always active 5VDC @ 2 Amps available, and the green wire (/ON) is the main power activation signal, which when taken to ground turns on the high current 5 and 12 volt power supplies of the ATX PSU. The XEL has a latching power circuit that allows a momentary switch to toggle the XEL motherboard's 5VDC logic supply on and off, and at the same time provides the /ON activation signal through the PSU header (pin-2). Normally when the XEL is being supplied 5VDC power from a wall wart power adapter, this /ON signal is only used locally to turn on the power mosfet switch (Q1), which in turn directs the 5VDC power to the main bus. But in the event that we wish to use an ATX PSU instead, it can be wired up to the XEL via the PSU header (J1), so that it both provides power to the XEL (purple wire) and also controls the ATX PSU's Main Power ON/OFF action via the /ON signal (green wire) being toggled by the XEL's momentary power switch.

 

 

Here's a short video showing this in action...

 

- Michael

 

EDIT: I just wanted to point out that on the V1.0 XEL PCB (first run prototype) the labeling of the PSU header is incorrect. So what is shown in the first picture in this post is the correct labeling of that header.

 

  By the off chance do you know the 3 pin Molex connector part number?  If I have that I can find the TH female version on Molex's site.  

[Matej]

This is great news! Real ATX case with ATX PSU will be amazing!

Edited by Matej, Fri May 19, 2017 8:11 AM.

[Dropcheck]

 

That assumes some kind of external AC to DC power supply. 

 

The point here is to have an internal AC to DC power conversion.  No power bricks external to the case.  Just the power cord from wall to case.  At least that is the direction I am heading. 

 

I just wanted to add to my comment.  I don't want to imply that the current case power supply implementation is lacking.  For those who don't need the +12V extra power signal it is perfectly fine. 

 

It's just another option for power supply for most.

Edited by Dropcheck, Fri May 19, 2017 8:45 AM.

[JoSch]

The PicoPSU has a 12V line, so it should work.  My aim with this PSU would be more power efficient.

I searched Amazon for normal ATX PSUs. the lowest power rating was 250W. I would guess that's a lot power, which is not needed.

My guess would be that 90W are more than enough for the computer and your XF551 board.

Or am I wrong?

[Dropcheck]

The PicoPSU has a 12V line, so it should work.  My aim with this PSU would be more power efficient.

I searched Amazon for normal ATX PSUs. the lowest power rating was 250W. I would guess that's a lot power, which is not needed.

My guess would be that 90W are more than enough for the computer and your XF551 board.

Or am I wrong?

 

  I hear you.  ATX PSUs in many cases is overkill, especially over 250W.  I just don't like the idea of another power brick hanging outside the case. 

 

  As long as the PicoPSU has the standard 4pin Molex power connector for CD/5 1/4" and can do upto 1amp on the +12V rail it probably will work. 

[MrFish]

I'd love to throw one into an Antec Mini Skeleton 90 case.

 

 

That way I can see the motherboard all the time.

 

They're hard to find these days though. Only place I could find on a quick search was here: Server Parts Warehouse, and they aren't cheap.

 

The nice thing is, it would match my main system, which is in an Antec Skeleton (full-sized big brother).

 

 

[mytekcontrols]

 

  By the off chance do you know the 3 pin Molex connector part number?  If I have that I can find the TH female version on Molex's site.  

 

Originally I just went with a break-away SIP Header (3-pin). But I think due to the power polarity thing and not wanting to fry anything, I think it would be better to go with something more like this for the header...

 

Digi-Key P/N: WM2701-ND

 

And this for the mate...

 

Digi-Key P/N: WM2001-ND

 

 

 

- Michael

[GlowingGhoul]

Maybe I missed it. Is their a header to attach a real Xl or XE keyboard?

[mytekcontrols]

Maybe I missed it. Is their a header to attach a real Xl or XE keyboard?

 

Yes this was discussed quite a while back (I know... this is a long thread  ). Short answer, no there isn't. As to why... this project has never been about replacing the motherboard in a stock A8, and has always been about creating something for a completely different form factor where using a stock A8 keyboard never made much sense.

 

- Michael

[mytekcontrols]

The PicoPSU has a 12V line, so it should work.  My aim with this PSU would be more power efficient.

I searched Amazon for normal ATX PSUs. the lowest power rating was 250W. I would guess that's a lot power, which is not needed.

My guess would be that 90W are more than enough for the computer and your XF551 board.

Or am I wrong?

 

As Dropcheck pointed out already those supplies require an external power brick which ends up being a little clunky, especially if you want to take your system to a club meeting or some other event. For the smaller console style cases having to use a 5 VDC power adapter is acceptable because it'll easily fit in your pocket. But when you want to have real floppy and hard disk drives, the bigger cases are the way to go, and they usually come with a micro-ATX PSU. And although they have massively more power than needed (250W), it usually isn't an issue since the space for that has already been allocated.

 

Here's a good example of a case suitable for mounting and powering the XEL motherboard, Dropcheck's XF551 board, a floppy drive and even a 'real' hard drive...

Rosewill Mini ITX Tower/ Desktop Computer Case with 250-Watt PSU ($61 at Amazon)

 

 

And for interfacing from the ATX power supply to the XEL's PSU header something like this would probably do the trick...

Cytron ATX Power Supply Breakout Board (Right Angle)
Product Code : RB-Cyt-101

 

EDIT: Good article on using this breakout board here

 

- Michael

Edited by mytekcontrols, Fri May 19, 2017 1:37 PM.

[Dropcheck]

 

Originally I just went with a break-away SIP Header (3-pin). But I think due to the power polarity thing and not wanting to fry anything, I think it would be better to go with something more like this for the header...

 

Digi-Key P/N: WM2701-ND

 

And this for the mate...

 

Digi-Key P/N: WM2001-ND

 

 

 

- Michael

 

 

Were you thinking about doing the same type of keying on the SIO AUX connector?

[mytekcontrols]

 

 

Were you thinking about doing the same type of keying on the SIO AUX connector?

 

I'll have to investigate to see if I have enough room to do that, and get back to you. It's kind of tight quarters in that area, so it'll be a bit of a challenge, but I'll give it a try and see how things go. 

 

- Michael

[mytekcontrols]

SIO AUX connector J5 has now been changed to a polarized version same as was done for J1...

 

 

 

And it will mate with Digi-Key P/N: WM2003-ND

I also converted the 2-pin 5 VDC Accessory Power connectors to polarized versions as well. So basically anything that has 5 VDC power on it can only be inserted one way, and in all cases pin 1 = +5 VDC. Resistor networks RN1 and RN2 did have to be squished together to allow for the larger header to fit. But since RN1 is directly soldered to the board, there is no socket and its associated clearance requirements to contend with.

 

- Michael

[nikolaisen]

Would it be possible to extend the cartridge port to a 3.5" drive bay using a ribbon cable? This would be great for using a larger PC case. Sorry If I missed this topic already. 

 

I see this question has been answered already. Sorry for the trouble.

Edited by nikolaisen, Mon May 22, 2017 3:50 PM.

[mytekcontrols]

Would it be possible to extend the cartridge port to a 3.5" drive bay using a ribbon cable? This would be great for using a larger PC case. Sorry If I missed this topic already. 

 

I see this question has been answered already. Sorry for the trouble.

 

Just to add to this... the CART and ECI ports are non-buffered and so the absolute minimum extension should be used in order to see good results. I remember building an A8 into a monitor many years ago (1984?) and having problems with certain cartridges due to the DIY extension CART/Cable that I had made to put the cart input where it could still be used. I learned the hard way that there is no clear cut answer that I am aware of that will let you know how long is safe for reliable operation. I even tried using independent ribbon cables for each side of the CART, and even went so far as to double the number of conductors to provide a grounded one in-between each active signal. Nothing fixed the problem until I figured out how to reposition the board in such a way that I could reduce the cable length down to 6-8", after which things started behaving much better.

 

I never did try twisted pair ribbon cable, but the only way that would have probably helped is if I still used the grounded conductor between the signals. Bottom line is no cable or extension is the best way to go, but that becomes a logistics problem as to how to access the CART from within the case. On some of my proposed XEL incasements (is that even a real word?) I plan on creating a custom top panel to access the port directly on the board.

 

- Michael

 

EDIT: There is also the possibility of using a right-angle edge card connector which might open up some possibilities

Edited by mytekcontrols, Mon May 22, 2017 4:21 PM.

[Van]

Would it be possible to extend the cartridge port to a 3.5" drive bay using a ribbon cable? This would be great for using a larger PC case. Sorry If I missed this topic already. 

 

I see this question has been answered already. Sorry for the trouble.

It should be, and I intend to do something similar. The main restriction would be the length, but under ~12-16" should be OK.

Yogi

[Van]

 

Just to add to this... the CART and ECI ports are non-buffered and so the absolute minimum extension should be used in order to see good results. I remember building an A8 into a monitor many years ago (1984?) and having problems with certain cartridges due to the DIY extension CART/Cable that I had made to put the cart input where it could still be used. I learned the hard way that there is no clear cut answer that I am aware of that will let you know how long is safe for reliable operation. I even tried using independent ribbon cables for each side of the CART, and even went so far as to double the number of conductors to provide a grounded one in-between each active signal. Nothing fixed the problem until I figured out how to reposition the board in such a way that I could reduce the cable length down to 6-8", after which things started behaving much better.

 

I never did try twisted pair ribbon cable, but the only way that would have probably helped is if I still used the grounded conductor between the signals. Bottom line is no cable or extension is the best way to go, but that becomes a logistics problem as to how to access the CART from within the case. On some of my proposed XEL incasements (is that even a real word?) I plan on creating a custom top panel to access the port directly on the board.

 

- Michael

 

EDIT: There is also the possibility of using a right-angle edge card connector which might open up some possibilities

I can imagine in the case of a Monitor, you would be fighting some nasty RFI. Along while back I got some round PATA cables, they have the ribbon cable folded and wrapped in a foil shield inside the 3/4" sheath.

The TRS-80 CoCo had a 4 Cart expansion box (actually, there were a few third-party boxes also)  that connected via ribbon cable, I think it was  ~16", but some wrapped it in a GND foil to help with noise. The TRS model interface also buffered all the cart signals, but some of the other designs didn't bother. 

So yes it is tricky

Yogi

Edited by Van, Mon May 22, 2017 5:00 PM.

[mytekcontrols]

Project Update

 

It's been a while so I thought I'd bring everyone up to date on how things are going. The first two boards being built by the BETA Test Team were failing to boot-up. After a lot of inspecting there didn't appear to be anything wrong with the PCB assembly. However there was one difference between my working XEL V1.0 board and the ones that the BETA Tester's had built, and that was the fact that they were using the original Candle U1MB's and mine had a Lotharek version. At first whatever differences between these two U1MB versions we could find didn't seem to be an issue. But thanks for AA member Van's due diligence the problem was discovered, and that came down to Candle's board needing A13 on the OSROM header and the Lotharek board picked this up solely from the MMU header. In the V1.1 production design A13 was intended to appear on both headers, but this wasn't so on the V1.0 prototype boards which I had borked by putting the wrong signal to it. Anyway suffice it to say with a quick trace cut and a jumper to put A13 on the OSROM header, everything started to work   

 

Another unrelated problem that was discovered had to do with the 556 dual timer Reset/Power Control circuit. We had already modded this somewhat from what was originally done, but apparently my mods were only good for certain 'brands' of 556 chips (not good  ). Essentially what we needed to see, was for the /RST line to be held low on initial power-up, so that the system would have a clean start-up when power had stabilized. So some alternative solutions were tried...

 

Photo by Van

 

 

Example of analog approach to a sustained reset on power-up (R46 and C51 form an RC time delay circuit, with DCHG from the Power-Control timer re-initializing it when powered OFF)...

 

 

And this did work quite well, and we also discovered at the same time that there was another possible good side effect. By lengthening the reset even more, thus allowing the PS2 keyboard to fully boot before the XEL booted, it was possible to have things like holding down the START, or OPTION, or HELP keys be recognized and acted upon when the XEL did boot-up. So just like on a real A8, Basic could be disabled, a cassette could be booted from, and the U1MB Setup Menu could be directly entered upon start-up. So we gave that a try and sure enough it worked! Then after thinking about it a bit more, it struck me that the built-in TK-II keyboard chip could handle this all by itself, thus eliminating the reset timer all together. So a revised TK-II firmware build was done and tested yesterday incorporating a sustained reset upon first time power-up, and I'm happy to report this test was successful. Now I just need to tweak around with how long reset really needs to be sustained to allow time for all the various PS2 keyboards to boot-up. At any rate, this TK-II sustained reset feature should also work on a normal A8 installation, and in most cases not require any hardware modifications (however the XEGS will be an exception).

 

There has been a lot of work behind the scenes by six very dedicated and busy individuals as they one by one get their XEL boards up and running. And a lot of good stuff has come out of this synergy that just wouldn't have been possible on my own, and I am extremely thankful to all of their efforts and ideas in this process.

 

By early next week I suspect that all six BETA boards will be up and running, and then the real testing will begin. Meanwhile I've been busy implementing changes into the schematics and PCB layout to keep up to date with the problems that have been discovered and fixed, as well as a incorporating a few additional features into the design (nothing crazy). My plan is to have a 2nd small run of boards made strictly for test purposes in another 2-3 weeks. This will hopefully be the final production design. If all goes well when testing that version (V1.1), the design files and gerbers will be released. So possibly we are looking at late June or early July for that to happen.

 

- Michael

[mytekcontrols]

I just wanted to add besides the six people building up XEL's for testing, there are also 3 other individuals in the BETA Test Group that have contributed at least one or more of the following: components, insights, and recommendations that have also helped keep this project on track. One those 3 prefers to be on the sidelines and is a bit shy so he'll remain nameless, but the other two are well known on AA, that being bob1200XL and flashjazzcat. Bob and I go back a long time, and he has always been nothing but supportive and kind, and I remember even when I was a very young man (boy was that a long time ago ) with hardly any money, he graciously gave me his time, computers, parts, dinner (well his wife Sharon did that ) and a free oscilloscope which I desperately needed at the time. Jon I just got to know over the last year or so, and as most of you already do know he is an incredible resource and wealth of knowledge on all things Atari, and is always willing to lend a hand. Simon, Brent, Rory, Marlin, Stephen and Yogi I also got to know by my time spent on AA, and they too impressed me with their abilities, perseverance, and willingness to jump in and git it done. Those last six are the actual builders of the XEL prototypes, and they put a lot of faith, money, time, and personal risk into the XEL project.

 

 

 

- Michael

[Firedawg]

Thank you 1088XEL Team!!!  Looking forward to this project and all the cool uses that this community of Atari dreamers can come up with.  Michael thanks for sharing some of your history.  These 8-bit machines are kinda like time machines in a way aren't they;--) 
 
Mike

[Stephen]

Thanks for the kind words!  I got my 1st Atari back in 1982.  Never imagined I'd have my name on two device silkscreens (even if only as a tester)!

[Van]

Thank You Michael, for your incredible work and vision. I feel very lucky to have the opportunity to come to know you through these posts and to be a small part of this great project. It's truly been great fun, even when there was a hiccup or two. It's very satisfying when the computer burps and the Ready prompt pops up on screen, knowing that I started with a box of parts.

  For everyone reading, this is one of the best projects I have done with our beloved Atari in years. The schematics are clear and straight forward. The assembly is not complicated as it's through hole. Parts are readily available, even NOS Atari VLSI chips are still sold. The most popular mods are already incorporated in the base design and  some of the more optional mods are easily added with dedicated headers. You'll end up with a first class Atari with a modern form and feel and the pride of true DIY.

Yogi