1088XEL Alternative Mother-Board Project

[+mytek]

Looking on this project from time to time, I am asked how much it would cost completed in a case, with all the bell and whistles -till now- for the potential builder?

 

It won't be cheap But realistically it depends upon so many different variables when you are talking about a complete turn key system, that an exact number isn't possible. Also quantity buys would definitely influence the cost.

 

- Michael

[danwinslow]

Well, not being hardware handy, I just want to put it out there that I would definitely buy several completed kits, should someone produce them.

[+mytek]

First the bad news... Stephen your board will be coming back to you with my U1MB, since I blew yours up by accident (2nd U1MB kill for me). Looks like I took out the Xilinx chip, but there might have been some collateral damage as well. Everything else appears to be ok on your board. So what happened??? When I was running a no-color polarity test PAL vs NTSC (I'll be discussing that in a minute), I went to connect a lead to the +5v connection on the MPBI with the board still live, slipped, and shorted across to I believe D0 for an instant. Stupid me Anyway the best I can do is give you my Lotharek board for your now dead Candle board. Like I said this is the 2nd time I've killed a U1MB, and I'm hoping that the guy I gave that one to, who was able to fix it, will volunteer to fix this one as well (I'll pay for the parts). I'm sooooooooo sorry .

Now let's get to what I was experimenting with when this happened.

So if you guys recall, I was working on a better solution to the 'no-color' noise issue, and had opted to tie pin 5 (a formerly floating input on the analog chroma switch) to +5V, which appeared to be working fine (reference link). Well it was, if you were running as an NTSC system. But when I switched over to using Stephen's board which was setup for PAL and tried the same MOD, lo and behold the colored stripes appeared in the over scan area (WTF?!). So with a bit more investigation (and one fried U1MB ) I landed on the solution. Apparently the 'no-color' level output of the GTIA switches polarity for PAL vs NTSC, with NTSC = +5V and PAL = GND. So what I'll have to do is implement a 'no-color' polarity selection header. If you look at the following updated schematic page you'll see what I am proposing (please note that I re-purposed what once was a PAL/NTSC clock config header (turned out to be unnecessary), and made it into a no-color polarity select header instead).


Updated 8/21/2017: corrected terminology used in describing clock signals

Change of topic... Audio Outputs.

While I was experimenting with driving the SC1224 series RGB monitors, I discovered that the audio loading on the JVC version was causing the output to sound distorted. So apparently the audio input of that monitor is a fairly low impedance, and was too much for my High 'Z' output stage to handle. So after some experimentation, I was able to correct this situation, as well as provide headphone capable output from the XEL. To make this happen, I changed from a 100K collector resistor down to a 100 390 ohm one instead, thus converting it to a Low 'Z' output stage. It won't do a very decent job of driving speakers, so you'll still need amplification, but it will pump out a lot of volume on a set of headphones without an amplifier. When used with an amp, the volume setting will need to be slightly higher than before to obtain the same sound output, but I think this is preferred, since it used to be almost too hot. Note: This same change can be made on the TK-II-Stereo Board by tacking on some 100 390 ohm resistors underneath between pins 5&6 of RN1 and RN2.

 

Note: Because of too much attenuation of the audio outputs discovered during later tests, I increased the value of the collector resistors from 100 ohm to 390 ohm. Still has more than sufficient drive to run headphones.


Updated 8/23/2017: changed value of R51 & R52

 

Note: These changes have now been incorporated into the V1.1 design, and the complete updated schematic can be downloaded from the site listed in my signature line. Also beefed up the power supply filtering, upping the final filter cap by a factor of 10. This results in less audio background noise.
Here is a fresh look at the revised pcb layout (updated 8/25/2017: added two additional mounting holes).



- Michael

Edited August 28, 2017 by mytekcontrols

[+orpheuswaking]

Anyone in the Chicago area on the weekend of September 9th will be able to come out and see at least one (maybe two) XEL in person @ VCF Midwest.

[+Stephen]

Anyone in the Chicago area on the weekend of September 9th will be able to come out and see at least one (maybe two) XEL in person @ VCF Midwest.

I'll have mine running as a PAL configured machine. Don't think I'll have time to get the custom top and back panels made. I'll have to rig up something to keep the CF card stable, as I definitely want to demo that. Also need to order some SanDisk cards - my Transcend 2GB cards are perfect for everything except the video player which of course is probably the biggest "wow factor" when demoing.

[+orpheuswaking]

Yeah I wont have the custom panels at this stage so I'll be running with the top off as well. So we will have a NTSC and a PAL unit running different demos and software.

[+mytek]

I'll have mine running as a PAL configured machine. Don't think I'll have time to get the custom top and back panels made. I'll have to rig up something to keep the CF card stable, as I definitely want to demo that. Also need to order some SanDisk cards - my Transcend 2GB cards are perfect for everything except the video player which of course is probably the biggest "wow factor" when demoing.

 

Both Stephen and Simon will be demoing an early prototype of the XEL-I3 board. The production version is going to be based on a new design, where the interface electronics will stay close to the MPBI port (direct plug-in PCB), and have the CF card connector attached via a ribbon cable. This allows for much longer cable lengths (I've tested with a 16" long cable without a problem), thus taking advantage of the buffered aspect of the chips being used on the interface connection to the CF card.

 

The new board will be called: XEL-CF-][ (more info HERE)

 

 

J1 = MPBI connection (will use a female header with extra long legs to double as both a plug and a socket, giving it pass-thru qualities)

J2 = Activity LED connection

J3= IDE44 connection (connects to a CF card via an inexpensive carrier board)

 

- Michael

[Dropcheck]

Hi Michael,

 

Are you still shooting for a mid to late Sept 1088XEL release?

[+mytek]

Hi Michael,

 

Are you still shooting for a mid to late Sept 1088XEL release?

It'll be more like late September, but yes this goal should be achieved. Currently I am just doing the sanity check on the board layout, double checking things like trace clearances, and redundant paths (easy to do on an all manual routed board such as this). But all the functionality has been verified to be 100%, and other than what I just mentioned, the design is locked in.

 

- Michael

[Lavalamp]

This is the coolest thing I have ever seen

[marty]

Is there a way to do a straw poll on this thread as to numbers of people who would be in for buying this lovely item at several tiers, ie. (1) board only, (2) fully populated board, (3) populated board in enclosure, etc.? It'd be interesting to see now that the design is nearing readiness.

[+orpheuswaking]

With the length of this thread it might be beneficial to start a separate poll thread... As for my answer I would go for a Board and a full set of parts as a Kit (I have a second Realan case so I want at least one, but I might actually go for two full sets) I'll need to order two sets of Realan case tops and backs along with the cart tunnel as well

[+mytek]

Test order placed with EasyEDA.

 

 

I went with green solder mask this time around because it's considerably cheaper in these smaller quantities.

 

I'm going to be stingy with these, although a few will be going to people that are involved in manufacturing accessories for the 1088XEL, and also a couple will go to the guy that contributed 50 SIO jacks to the project (you know who you are).

 

I tell you it's always nerve wracking to push the order button on something of this complexity , but it's done!!!

 

Here's a simulation of what the board will look like in green (produced by EasyEDA's on-line gerber viewer).

 

 

Pretty frickin cool looking

 

Also have some XEL-CF-][ boards coming as well.

 

 

- Michael

[+MrFish]

Test order placed with EasyEDA.

 

 

 

Actually, I like it. It looks like it will keep things from looking "too modern".

[+mytek]

 

Actually, I like it. It looks like it will keep things from looking "too modern".

 

Good point. Looking at some other boards I've had done in the past with green solder mask, I noticed the traces pop out at you much more so than any other mask color. I always liked that, makes it more a work of art than just a piece of electronics .

 

- Michael

[+mytek]

Rapidus compatibility was dropped from design considerations...

 

Due to space constraints and the desire to embrace affordable reliable technology that enhances the 'existing' software experience, Rapidus support is difficult at best. When considering that this upgrade is close to the same cost of the entire 1088XEL motherboard (assuming user supplied Atari VSLI chips), it just wasn't high on my priority list of things I needed to support. Not saying a Rapidus can not be used, but you will give up easy MPBI port access in the process. Of the two prospects, I would choose the MPBI over the Rapidus any day. Reason being, is that it will bring the end user so much more. We already have two MPBI projects in the works, with working prototypes for each, and the promise of more to come.

 

I know there will be people that don't agree with my decision, but at the end of the day it was my decision to make. The 1088XEL project was born out of my desire to have a very specific machine in a small footprint, that had maximum compatibility and operation as a stock A8, while also supporting the U1MB in a direct plug-in format. And I always wanted the HDD aspects of the SIDE2, but in a form that could be easily integrated into the 'cased' package, hence the MPBI expansion capability utilized by the new XEL-CF-][ compact flash interface board.

 

So supporting the new MPBI standard that was born in the 1088XEL project, took center stage in my decision to encroach on the space originally designated as free and clear for the Rapidus.

 

- Michael

[marty]

Any chance you'd be willing to create modern mini Atari 2600 and Amiga boards next?

 

[+mytek]

Any chance you'd be willing to create modern mini Atari 2600 and Amiga boards next?

 

 

Hmm... let me think about that

 

No I'm pretty sure this will be my first and last computer motherboard design. Although it's been a fun ride, it's also been a tough one as well. I don't think I have it in me to do something like this again. And I really do feel completely satisfied to have done just the one .

 

- Michael

[+S.D.W.]

I'm going to be stingy with these

 

I kind of want to troll this post... hmmm... actually maybe I just kinda did.

[+DrVenkman]

If anyone does a production poll, I suggest several options:

 

1. Bare board only

2. Board plus really hard-to-get parts like SIO jack and cartridge connector (and any others not easy to source commercially from current electronics houses)

3. Complete board + full BOM (minus case/PSU)

 

Any others? Personally, I'd be interested in #1 or #2. I'm not above salvaging some stuff from a junky XL/XE to build up a system, though I'd prefer not to, funds allowing.

[pajero_pn]

Can I ask for a list. Devices tested and compatible with XEL

VBXE yes
SIDE2 yes
KMK +IDE ? http://atariki.krap.pl/index.php/IDE_Plus
Rapidus maybe in future ?
Sio2SD yes
Turbo Freezer ? http://www.horus.com/~hias/freezer/turbo-freezer-2011/

[+mytek]

Can I ask for a list. Devices tested and compatible with XEL

 

VBXE yes

SIDE2 yes

KMK +IDE ? http://atariki.krap.pl/index.php/IDE_Plus

Rapidus maybe in future ?

Sio2SD yes

Turbo Freezer ? http://www.horus.com/~hias/freezer/turbo-freezer-2011/

 

VBXE has not been tested, but it is assumed that the XE version of it will work (NO).

 

SIDE2 has been tested and works (YES).

 

KMK +IDE has not been tested (NO).

 

Rapidus should also work, but has not been tested, although it will partially cover the MPBI port and require a short 34 pin male/female cable to attach MPBI devices to the bus (NO).

 

SIO2SD has been tested and works (YES).

 

Turbo Freezer has not been tested (NO).

 

 

Note: Since the XEL-CF creates a fast parallel IDE44 interface, there is very little reason to use an SIO2SD or any other similar CF/HDD solution. Especially considering that I just bought 20 bare PCB's for close to $1 a piece, and the components to stuff it probably are in the area of $5. Since it is a Through-Hole board, it is very easy for the DIY person to build, and due to the limited amount of parts required to do so a relatively quick job. With Jon's XEL-CF specific BIOS flashed into the U1MB, it is the same as accessing a CF card on a SIDE2, and just as fast. Theoretically any IDE device could be accessed, assuming a supporting BIOS for such were in use. The IDE44 header is directly compatible hardware-wise with 2.5" laptop hard drives, and a 2 mm to 0.1" header adapter with a suitable PSU could be employed to access larger devices.

 

- Michael

[pajero_pn]

Thank you for your answer.

We follow the topic in Poland

 

http://www.atari.org.pl/forum/viewtopic.php?pid=233963#p233963

[Van]

Note: Since the XEL-CF creates a fast parallel IDE44 interface, there is very little reason to use an SIO2SD or any other similar CF/HDD solution. Especially considering that I just bought 20 bare PCB's for close to $1 a piece, and the components to stuff it probably are in the area of $5. Since it is a Through-Hole board, it is very easy for the DIY person to build, and due to the limited amount of parts required to do so a relatively quick job. With Jon's XEL-CF specific BIOS flashed into the U1MB, it is the same as accessing a CF card on a SIDE2, and just as fast. Theoretically any IDE device could be accessed, assuming a supporting BIOS for such were in use. The IDE44 header is directly compatible hardware-wise with 2.5" laptop hard drives, and a 2 mm to 0.1" header adapter with a suitable PSU could be employed to access larger devices.

 

- Michael

Let me ask you Michael, or anyone, about the CF HDD speed? I'm debating the best handling of a large DRO music file, either direct from CF or pre-loaded into ext RAM. Banking the whole file is the fastest access at runtime, but the down side is the load time but that would be before play back starts. So for a 50-100k file, SIO based media seems prohibitive for pre-loading the RAM banks. Transferring from CF to ext RAM seems like it would cut the load time but if the CF access time is fast enough would rather buffer small blocks during runtime direct from the CF.

 

Background:

A DRO file is a log of register writes to an OPL chip. Each entry in the log are two bytes composing either the register number and value or command and value. The commands are delays, either 1mS or 256mS x Value+1. So during play back the main loop could be handling back to back register writes or register writes inter-spread with delays. Over all song timing depends on (register writes + delays commands)+loop/system overhead.

Yogi

[gozar]

Note: Since the XEL-CF creates a fast parallel IDE44 interface, there is very little reason to use an SIO2SD or any other similar CF/HDD solution. Especially considering that I just bought 20 bare PCB's for close to $1 a piece, and the components to stuff it probably are in the area of $5. Since it is a Through-Hole board, it is very easy for the DIY person to build, and due to the limited amount of parts required to do so a relatively quick job. With Jon's XEL-CF specific BIOS flashed into the U1MB, it is the same as accessing a CF card on a SIDE2, and just as fast. Theoretically any IDE device could be accessed, assuming a supporting BIOS for such were in use. The IDE44 header is directly compatible hardware-wise with 2.5" laptop hard drives, and a 2 mm to 0.1" header adapter with a suitable PSU could be employed to access larger devices.

 

- Michael

 

I know this was brought up earlier in the thread, but now that the XEL-CF is pretty much finalized, my question is, will it work on U1M machines such as my 800XL? I dislike the SIDE2 taking up a cartridge slot...