Thinking of getting a new Atari 800XL board made....

[+scorpio_ny]

That is very cool about using the W65C02 CPU! I will keep an eye on the information on this. So I see the theme of the board is not just a revise but a board for people who want to tinker with some ideas and concepts. When I talkinking about adding headers, I was thinking more along the line of adding holes to add jumpers along the PBI bus and some since most of the signals that people are looking for would be there and they would not need to solder to pads to pick up the signals. Also, the labeling of the signals pinouts on the PCB is extremely helpful. What I meant to add for Pokeymax upgrade was to add some by the Pokey area for easier pickup. Here is what I was thinking of from the XE remake:

[reifsnyderb]

9 minutes ago, scorpio_ny said:

That is very cool about using the W65C02 CPU! I will keep an eye on the information on this. So I see the theme of the board is not just a revise but a board for people who want to tinker with some ideas and concepts. When I talkinking about adding headers, I was thinking more along the line of adding holes to add jumpers along the PBI bus and some since most of the signals that people are looking for would be there and they would not need to solder to pads to pick up the signals. Also, the labeling of the signals pinouts on the PCB is extremely helpful. What I meant to add for Pokeymax upgrade was to add some by the Pokey area for easier pickup. Here is what I was thinking of from the XE remake:

My thought on adding headers is that this is a 2 layer board (to keep the cost down) and that it would take a considerable amount of time (probably days) to track down and add all of the headers, jumpers, etc.  There would be lots of re-routing just to squeeze things in, too.  With the breadboard area, almost any header can be added.  I remember that 2 years ago, I spent a couple weeks tracking down all of the possible upgrades and adding pins for them...and this was on a 4 layer board.

 

My original intent was to just add some missing features to the normal Rev. D board and get it made.  But after there was mention made of all of the soldering involved, I realized it would be very advantageous to reduce the component count.  So, this is pretty much what an 800XL would have been if it was made in the '90's....or something like that.  (Minus the Freddy chip, of course.)  Without considering the Atari LSI chips, I wouldn't be surprised if the cost of the components is now 1/2 what it would be on a normal 800XL board. 

 

I think it would be possible to use the breadboard area to add the extra logic to gain access to the "hidden" 64k on that SRAM chip, too.  Other tinkering or experimenting can be done as well.

 

Other changes can come later, too.  Right now, I need to find out if the board works as intended.  (I am pretty sure it will.   🙂   )  So, I already have 5 ordered.

 

 

 

 

 

[+David_P]

Power to the PBI?

[reifsnyderb]

20 minutes ago, David_P said:

Power to the PBI?

+5VDC on pins 47 and 48   🙂

 

Other changes:

/halt is on PBI pin 33.

PBI RAM capable as per the 1400XL & 1450XLD (Brings this board even closer to the Sweet 16 spec.)

Open collector refresh on PBI so as to avoid shorting out the ANTIC chip with PBI upgrades that use refresh to shut down the MMU

Improved DB-9 joystick ports

Copper ground planes almost completely covering both sides of the board

Video is on separate ground plane

Audio is on separate ground plane

Chroma is connected to the monitor port

Video is as per the 800's circuitry

Audio is @mytek 's audio circuit

OS and BASIC are on a single chip

No RF, RF holes left in place for optional SAVO installation.

Channel selector switch is used to select bank for OS/BASIC

Ferrite beads and unnecessary 1nF capacitors are removed

New power supply circuit with 2 rails; one for digital and one for audio/video

Experimental overvoltage protection in the power supply

Breadboard section that could fit up to 4 chips for experimentation, tinkering, mods, etc.

DRAMs have been replaced with a single SRAM chip

DRAM memory control chips are now unnecessary and, therefore, removed.

Many resistors replaced with resistor packs.

Greatly reduced component count.

 

Edited October 28, 2023 by reifsnyderb

[+Larry]

"OS and BASIC are on a single chip"

 

Eprom or flash rom?  Just glancing at the layout, probably conventional eprom.  But maybe your idea for an optional adapter board could be used later as an add-on?  Kudos for the component reduction!, and fingers crossed that there aren't too many issues for you to fix.

[reifsnyderb]

52 minutes ago, Larry said:

"OS and BASIC are on a single chip"

 

Eprom or flash rom?  Just glancing at the layout, probably conventional eprom.  But maybe your idea for an optional adapter board could be used later as an add-on?  Kudos for the component reduction!, and fingers crossed that there aren't too many issues for you to fix.

It's a flash ROM.  Specifically a readily available SST39SF010. 

[+David_P]

4 hours ago, reifsnyderb said:

It's a flash ROM.  Specifically a readily available SST39SF010. 

With 128K available, is that four 16K OSes and eight 8k cartridges?  Or can it be configured to offer a 16K cart as well? (Or does that get too complex?)

 

https://www.microchip.com/en-us/product/sst39sf010a

 

[reifsnyderb]

13 minutes ago, David_P said:

With 128K available, is that four 16K OSes and eight 8k cartridges?  Or can it be configured to offer a 16K cart as well? (Or does that get too complex?)

 

https://www.microchip.com/en-us/product/sst39sf010a

 

I just kept it simple.  There are 2 OS's and 2 BASIC's possible.  Each OS is tied to a BASIC.  I am going to configure mine as follows:

 

Bank 1:  Atari OS R.3 w/bank testing fix and Atari BASIC C

Bank 2:  OS 6.1 and Altirra BASIC 1.58

 

The channel select switch is re-purposed as a bank select switch.

 

I kept everything as simple as possible to keep the costs and number of components low.  Originally, it was just one OS and BASIC but since the channel select switch was now available, I figured it could be re-used.  Of course, the bank select switch could be replaced with a jumper and only one OS/BASIC combination made available.

 

While there is a lot of wasted space on the ROM, the cost and availability is the major factor.  The same goes for using an AS6C1008 for the SRAM.  With the SRAM, 64k is wasted but you'll be hard pressed to find a cheaper solution.

Edited October 28, 2023 by reifsnyderb

[reifsnyderb]

One thing nice about 2 layer boards:  They are pretty fast to manufacture!  I only ordered them Friday morning and I just checked the website.  The boards are done and awaiting pickup by DHL.   🙂

 

 

Edited October 29, 2023 by reifsnyderb

[Mathy]

Hello Brian

 

On 10/28/2023 at 6:51 PM, reifsnyderb said:

...since the channel select switch was now available...

 

Only on NTSC XL's IIRC.  Most PAL XL's come with neither switch nor the hole for it.

 

Sincerely

 

Mathy

 

 

[reifsnyderb]

1 hour ago, Mathy said:

Hello Brian

 

 

Only on NTSC XL's IIRC.  Most PAL XL's come with neither switch nor the hole for it.

 

Sincerely

 

Mathy

 

 

Good point.  With PAL cases, the switch pads can be jumpered to ground.  So, there will only be one OS/BASIC available.  Either way, even Atari OS R3 w/ Atari BASIC C is a decent upgrade.  Newer OS's can be loaded, of course.

 

Edit to add:  With the open collector refresh, on this board, other OS's could also be safely provided via the PBI.

Edited October 29, 2023 by reifsnyderb

[reifsnyderb]

I've been working on the board since Wednesday.

 

First problem:  I forgot to order a 74HCT14 for the clock.  A 74LS14 wouldn't work, probably due to voltage levels.  So, I modified an adapter, used a SMT 74HCT14 that I had from another project, and got the clock working.

 

Second problem:  Board wouldn't boot.  After much troubleshooting, I discovered the Reset was held low.  It turned out I missed cutting a trace on the modified adapter in #1 and it was grounding the Reset line.  After fixing this, it booted...to Black and White and to Self Test.  Sound works, too.

 

(If I would just have had the 74HCT14 DIP to begin with, problems 1 and 2 wouldn't have happened.)

 

Third problem:  I didn't have the onboard BASIC at the right offset on the flash ROM chip.  This was an easy fix.  Now I get a Ready prompt....in Black and White.  Testing also reveals sound works.

 

Fourth problem:  I forgot to put the 1k hold-up resistor on the color circuit.  After installation, still in black and white.  Testing the GTIA chip, in another board, shows the GTIA chip is good.

 

This is where I've been since Friday night.  It's only black and white.  Various troubleshooting and comparing between a working 800XL and this board reveals nothing.  Using the scope in dual trace.  I eventually ran the color circuit through a separate breadboard and it still won't work.  Using the 800 color circuit is an absolute failure.  Just as something to try, and since I don't have a 100uH inductor, I wired 4 22uH inductors in series and tried something like a 130XE color circuit.  Still nothing.

 

At some point, I'll order a 74HCT14 DIP and a few 100uH inductors and try the 130XE circuit soldered into the board.  If that doesn't work, I may cut my losses and scrap this project.  Too much time and money down the drain....with no color and absolutely no explanation as to why.

Edited November 6, 2023 by reifsnyderb

[woj]

If this was my project, this is usually the point where I would start questioning the basics - Is the cable good? Is the chroma pin soldered to the board? Aren't you sending a PAL signal to an NTSC screen or vice versa, etc. etc. etc.

[reifsnyderb]

5 hours ago, woj said:

If this was my project, this is usually the point where I would start questioning the basics - Is the cable good? Is the chroma pin soldered to the board? Aren't you sending a PAL signal to an NTSC screen or vice versa, etc. etc. etc.

I've been all through the basics.  Cable is fine....I use it all the time.  All 40 pins are soldered to the board.  W2 jumper is installed.  Signals are on all GTIA pins.  Color pin 21's output is comparable to that on a working board, etc., etc., etc.   I've even re-created the entire color circuit on a breadboard.  The adjustment POT circuit is outing the same voltage as that on a working board.  Adjusting the POT changes the voltage.  Etc.

[Overange]

Well done for getting the main part working  

 

Its more than likely something very silly.

 

Have you tried either the UGV or UAV?

[reifsnyderb]

2 minutes ago, Overange said:

Well done for getting the main part working  

 

Its more than likely something very silly.

 

Have you tried either the UGV or UAV?

After days of working on this, I'd choose the term "obscenely stupid"....not "silly".  But I get the message and agree.

 

I don't have a UGV or UAV.  Also, I'd prefer that the board can be built without any upgrades.

[Overange]

5 minutes ago, reifsnyderb said:

 

 

I don't have a UGV or UAV.  Also, I'd prefer that the board can be built without any upgrades.

If anything just to see where the fault could be

[reifsnyderb]

5 minutes ago, Overange said:

If anything just to see where the fault could be

It would create another data point.  Yes.  But I don't have a UGV or UAV.

[+scorpio_ny]

I would like to share an experience I had recently with a manufactured PCB for keyboard. When I first assembled it, everything worked except a row of keys. I went a week testing everything and trying to track down my mistake in cabling or on my motherboard. In the end, it turned out the PCB had defect somewhere I could not see. I grabbed another board from the batch that I was sent and did the same tests and it worked fine. Do not rule out that there may be an unseen defect on the PCB. Maybe grab another one and try it out see if you still have issues the issue.

[reifsnyderb]

16 minutes ago, scorpio_ny said:

I would like to share an experience I had recently with a manufactured PCB for keyboard. When I first assembled it, everything worked except a row of keys. I went a week testing everything and trying to track down my mistake in cabling or on my motherboard. In the end, it turned out the PCB had defect somewhere I could not see. I grabbed another board from the batch that I was sent and did the same tests and it worked fine. Do not rule out that there may be an unseen defect on the PCB. Maybe grab another one and try it out see if you still have issues the issue.

Good point.  I had a defect in a 320k SRAM board, once, that took a couple of hours to find.  In that case, they drilled a hole through +5VDC power trace on the 3rd layer of the board. 

 

In this case, it would take at least 4 hours to assemble another board.  (Maybe even more.) 

 

I put everything from the GTIA color pin 21 out to the connection to the monitor jack on a breadboard with no success. 

Edited November 6, 2023 by reifsnyderb

[+scorpio_ny]

I just did some research on this. The symptoms you are having are indicative of either using PAL or defective clock crystal. Can you swap that out with a known good one?

[+kheller2]

And since you have the entire color circuit on a breadboard, does that breadboard work on a stock 800XL?  Trying to narrow down the culprit here to the board or the color circuit. 

[ClausB]

Have you traced the color subcarrier signal from the GTIA all the way through the circuit with a scope, to see where it's getting lost? Is there maybe something in the circuit that limits bandwidth below 3.6 MHz?

Edited November 6, 2023 by ClausB

[Overange]

3 hours ago, reifsnyderb said:

It would create another data point.  Yes.  But I don't have a UGV or UAV.

Could you post your schematic of the GTIA/Video Circuit and if possible the GTIA/Video Section as shown in the PCB editor.

 

Also, I guess you have no PAL colour burst parts installed and you have put in the NTSC link??

 

[reifsnyderb]

3 hours ago, scorpio_ny said:

I just did some research on this. The symptoms you are having are indicative of either using PAL or defective clock crystal. Can you swap that out with a known good one?

Thanks for the idea.  It's a brand new crystal but I'll swap it out with the one from the donor board (it worked before I took the LSI chips) and see what happens.

 

3 hours ago, kheller2 said:

And since you have the entire color circuit on a breadboard, does that breadboard work on a stock 800XL?  Trying to narrow down the culprit here to the board or the color circuit. 

 

I didn't try this.  However, I took the breadboard circuit apart and re-enabled the circuit on the board.  (More on this at the end.)

 

3 hours ago, ClausB said:

Have you traced the color subcarrier signal from the GTIA all the way through the circuit with a scope, to see where it's getting lost? Is there maybe something in the circuit that limits bandwidth below 3.6 MHz?

I traced the color signal through and didn't find anything.

 

1 hour ago, Overange said:

Could you post your schematic of the GTIA/Video Circuit and if possible the GTIA/Video Section as shown in the PCB editor.

 

Also, I guess you have no PAL colour burst parts installed and you have put in the NTSC link??

 

There is no PAL and the NTSC is jumper is installed.

 

The schematic is being experimented on, so there's nothing to post yet.

 

After re-enabling the circuit on the breadboard, the video synch and color circuit now line up.  I have no idea why they didn't before.  Still no color.  The amplitude of the unloaded color was way too high at around 4V.  It looks like a working board is around 500mV so I am trying to get this down.