New Ultimate 1 MB with Extreme Birdies

[+Larry]

I just received a new U1MB from Lotharek, and put it into the same 800XL that had a V2-Candle unit installed. It is the same computer, same everything, except for the new cables and PCB (and new one does not require the Dropcheck +5V MMU adapter). Even the rom is the same -- swapped it with the old one to avoid re-flashing, but it is the identical type SST 39SF040 that came with the new unit. The pcb's and cables are mounted in the identical position. When I powered it on, I was blown away by the severity of birdies.

 

I know from past experience that I can affect the birdies by switching Antic's, but again, it is the same Antic used for both. Timing? Why would the timing be different? Was it altered for the updated version? That would be surprising. Can anyone shed any light on this. Clearly not happy with these results, and in fact, I've already put the old one back in the XL.

 

-Larry

 

Edit: I also have another V2 from Lotharek that I installed in an XE. It has a few birdies in the XE, but nothing like this.

[+bob1200xl]

It looks like the data fetch for the last line of character graphics is pulling $FF only for a <space> character. Can you try moving the OS under the ROM, into RAM?

 

In some instances, newer versions run faster that the older units. This is not always good. As you say, bad ANTIC.

 

Bob

[Marius]

I have no solutions for you, but I want to let you know that I have had this issue a lot with my Atari's.

It's sometimes a matter of being unlucky, and as you experienced swapping the upgrade into another atari will help.

 

Good luck anyway.

[+Larry]

I've done a lot of swapping, and the results say that it is the combination of the IDE+2 with the new Ultimate that results in the birdies. Without the IDE+2, the screen is perfectly clear. So I don't think that there is anything wrong with the Ultimate, itself. It is odd, but the Candle version has just a very occasional birdie present. However IIRC, I had to swap the Antic in this one when I installed it a couple of years ago for the same reason. It's like they need to be "paired." (?)

 

I swapped cables, roms, and tried 4 different IDE+2's. None of these changes made any difference. I'm going to keep the Candle version in this XL and when I install the new one in another machine, I'll do the Antic-swaps, if required.

 

Has anyone tried any of the Bob Puff Stable XL fixes with the Ultimate? I was wondering if they are compatible and/or might be helpful.

 

-Larry

[Marius]

U1MB and stuff connected to PBI is in more cases asking for trouble.

 

Although I absolutely love the Ultimate 1MB, already a long time ago I decided that it is definitely not an ULTIMATE memory expansion. This is one of these reasons.

 

[+Larry]

Hi Marius-

 

Do you use the IDE+2, and if so, what memory expansion would you use with it? I remember that you are not too fond of internal memory expansions, Using the SIDE2 in PBI mode would be workable.
Maybe you use the Freezer with the IDE+2? That would be a good solution, although I love the multiple, flashable OS of the Ultimate. I use XLOS, Fast Math, Omnimon, MyBios OS. (And Konrad's 65816 OS, but with a different computer.) Wish the Freezer had more to choose from than only the non-flashable "Old Runner."

 

-Larry

[MEtalGuy66]

Yep. It's PHI2-bus timing skew related to bus capacitance/propogation delay, etc... same old story. ATARIs suck in this regard. And probably alwayse will. Since they scrapped the plans for the 14xxXL(x) in favor of "cost reduced" machines which became the common "XL/XE architechture" (and with it, buffered expansion bus,) we have had to live with this.. and even if every single "expansion" that anyone makes has the famous "hias, et. al. PHI2 write timing compensation" circuit included, that does not keep that expansion from adding to the system-wide bus capacitance and therefore, PHI-2 to BUS timing skew.. This is an unfortunate reality of a SUCKY DESIGN that is the 8-bit atari's bus timing scheme. You can probably thank the price wars with commodore, the home entertainment electronics market hiccups of 1983-84, etc.. for this.. But, placing blame hardly fixes the problem.

 

Truth is.. Those who are aware of this have been pointing out the fact that it's a BAD IDEA to try to cram all these expansions in the same machine for years.. Nothing has changed.. And until they commit the entire system including these "expansions" to modern programmable logic, where none of this timing is an issue, your alwayse gonna have this problem on real hardware. You can "fine tune" a given "set" of Atari gear to work reliably by changing the timing characteristics of PHI2 one way or the other, cleaning up certain signals, etc. etc.. But as soon as you plug or unplug something from the data bus, you completely change it and all bets are off.. This includes PBI expansion devices, most anything that connects to the motherboard via a ribbon cable or pin header to various chip sockets, and yes.. even CARTRIDGES.. In stock configuration, without all the added "expansions", most Ataris have enough of a "happy medium" of PHI-2-to-bus timing that adding/removing a single cart will not change it enough to cause undesired operation. However, once you hang a bunch of "expansions" off the bus, that is no longer the case.

 

The number of times I've explained this on this forum is definitely in the double digits now. And that's not counting the times others have explained it.

 

I wish there was an eomoticon for "broken record". heh.

Edited September 7, 2015 by MEtalGuy66

[Marius]

Hi Marius-

 

Do you use the IDE+2, and if so, what memory expansion would you use with it? I remember that you are not too fond of internal memory expansions, Using the SIDE2 in PBI mode would be workable.

Maybe you use the Freezer with the IDE+2? That would be a good solution, although I love the multiple, flashable OS of the Ultimate. I use XLOS, Fast Math, Omnimon, MyBios OS. (And Konrad's 65816 OS, but with a different computer.) Wish the Freezer had more to choose from than only the non-flashable "Old Runner."

 

-Larry

 

Indeed I use The Turbo Freezer and the 576K XE memory upgrade (Ctirad).

 

I have one U1MB setup, in one of my 800XL's. That setups seems to work fine, since I only use SIDE2 on that machine (Although I had the issue with the birdies with my BlackBox connected)

 

I love the 512KB upgrade from Hias (1 gal, and 512KB SRAM ). That one is internal but works very well.

 

I try to stick as much as possible with original atari 8bit. I find the computers too cool to spoil my fun with the problems you did run into again. When one is not into high speed I/O even a simple SIO setup is great.

[+Larry]

Well, the problem has vanished (for now), and I did nothing, save moving the installation point inside the XL by about 2". When I say that the issue is gone, I mean using the IDE+2 both direct connect and using about a 4" ribbon cable. Not a trace of a birdie. Same original IDE+2.

 

I thought that Candle had said that this upgrade had the Phi2 issue fixed, so that that it used the proper Phi2. I know there is a lead to the Ultimate board which I believe feeds Phi2 to the board and not from the OS rom. But anyway, I'm not changing a thing -- it now is "not broke."

 

-Larry

[MEtalGuy66]

I thought that Candle had said that this upgrade had the Phi2 issue fixed, so that that it used the proper Phi2. I know there is a lead to the Ultimate board which I believe feeds Phi2 to the board and not from the OS rom. But anyway, I'm not changing a thing -- it now is "not broke."

 

You don't read very well..

 

Any upgrade that "has the PHI2 issue fixed" is only "fixing" PHI2 timing where writes to devices located onboard that upgrade are concerned. It doesn't fix PHI-2-to-bus timing skew system wide. In fact ALL upgrades that attach to the data bus INCREASE bus capacitance and therefor change the relationship of PHI2 relative to bus timing. It doesn't do any good to have "the PHI2 issue fixed" on board your upgrade if the whole rest of the system now has unreliable bus operation.

 

That crap you were seeing on your screen most certainly was not the result of any erroneous reads/writes to devices on board the IDE plus 2.0, or the Ultimate 1meg. If anything it was reads/writes from motherboard base ram. Specifically, the area in which the custom chips maintain the display data.. This is VERY typical of PHI2 timing issues.

 

I don't know what you moved or how it effected the system/bus capacitance/inductance/timing wise... Issues such as high resistance/bad connections can obviously also effect system stability/integrity in various ways.. But whatever you did has made the situation better. Lucky you.. I hope it stays that way and works well in all operational instances.

Edited September 7, 2015 by MEtalGuy66

[+Larry]

Powered on all night, and still good.

 

Thinking about this some more, I think the original problem had to result from some type of contact issue that was altered when I moved the component. I wish I knew what it was -- Harting connectors in their sockets? Cables/pin connectors? Rom/socket? I'm not used to seeing birdies in my systems. I've certainly seen other types of screen disturbance issues, The only other spate of birdies I've seen was with an accelerator board. But if there is a "next time" with an Ultimate, contact issues is the first place I'll look.

 

-Larry

[flashjazzcat]

Yep: poor contact will cause birdies too, so we shouldn't jump to conclusions regarding bus timing and device compatibility, although these are equally legitimate concerns.

[Marius]

My biggest concern is that when you have the birdies you actually can see by eye that you have somewhere an issue, but there are also troubles like this one, that you can not see. That you have a steady screen doesn't mean the atari is running stable. I'm rather neurotic when it comes to stability. I use my Atari for some serious stuff too (of course that's my own choice) and I really have to trust the computer.

[flashjazzcat]

Indeed, which is why you need to run RWCRC and any other HDD stability tests before you can trust the reliability of the machine to any extent. And then the machine may still exhibit issues with cartridges (as I'm experiencing right now on one machine). And on the subject of IDE Plus: I have one IDE Plus which works well with a heavily upgraded 1200XL, and one which exhibits birdies. So there's variance all over the place, and it's a real lottery. MG is absolutely right about the state of the Atari bus, but equally Phi2 sometimes gets the blame for marginal connections and other installation issues. The other time I experienced birdies was during a rushed VBXE install on a 600XL, and it was down to shorts and broken traces on the Antic socket.

[danwinslow]

What *I* want to know, is why they are called 'birdies'.

[Marius]

What *I* want to know, is why they are called 'birdies'.

 

Because when you look at the screen they move like birds in the sky. It's not a static screen as what you see in a picture.

[Marius]

Indeed, which is why you need to run RWCRC and any other HDD stability tests before you can trust the reliability of the machine to any extent. And then the machine may still exhibit issues with cartridges (as I'm experiencing right now on one machine). And on the subject of IDE Plus: I have one IDE Plus which works well with a heavily upgraded 1200XL, and one which exhibits birdies. So there's variance all over the place, and it's a real lottery. MG is absolutely right about the state of the Atari bus, but equally Phi2 sometimes gets the blame for marginal connections and other installation issues. The other time I experienced birdies was during a rushed VBXE install on a 600XL, and it was down to shorts and broken traces on the Antic socket.

 

By the way... I remember when MyIDE ][ was developed I had a CPLD-firmware that caused those birdies on some of my XE's .... I told Tucker, and he came with a new firmware and the birdies were gone. So obviously in THAT case it was not just a hardware-thing, but a timing thing that could be cured with different logic-code.

 

I'm not sure HOW he did it, but it worked.

[MEtalGuy66]

Yep: poor contact will cause birdies too, so we shouldn't jump to conclusions regarding bus timing and device compatibility, although these are equally legitimate concerns.

My point is that it is all relative and by nature, every Atari experiences this type of issue to one degree or another. The more expansions you cram on it, the more prone it becomes to experiencing one of these issues to the degree that it causes operational problems. Noones jumping to any conclusions.. What I said most certainly is the case with EVERY Atari.. and moreso with "expanded" ataris. Like I said. Im glad whatever steve did managed to put that particular conglomeration of Atari+expansions back into the margin that operates trouble free.. At least for the moment..

[flashjazzcat]

Well the thread rapidly reached the stage at which the validity of using Ultimate 1MB with any PBI device was called into question: that's all I was referring to. Everything you've said is indisputable.

[Marius]

Well the thread rapidly reached the stage at which the validity of using Ultimate 1MB with any PBI device was called into question: that's all I was referring to. Everything you've said is indisputable.

 

I'm rather responsible for that, and it's my personal experience unfortunately. Perhaps indeed a wrong conclusion, but I have built U1MB in a few of my huge atari collection, and except for one, they all -suddenly- had issues with peripherals connected to the PBI. I can't repeat too many times that I really love the Ultimate 1MB upgrade, and that -especially with your PBI driver- it makes a fabulous combo with SIDE2, but for me it is no longer the 'ultimate' upgrade. Too many stability issues. Probably not due to U1MB, but due to what MetalGuy66 describes.

 

Perhaps it is a conclusion too soon, but let me state it different then.

 

In my atari's, on my location here on the world U1MB is fabulous, as long as no other equipment is connected to the PBI.

Perhaps it has something to do with the fact that the south-west window of my atari room is headed towards the Bermuda Triangle

[danwinslow]

So. The obvious conclusion is we need to re-implement the entire architecture on FPGA or something. Get on it. I envision an Atari 1400XL case, a little teensy tiny board in the middle, with ribbon cables going to edge connectors. It would be beautiful.

 

Plus, I am disappoint with the birdies explanation. I originally thought it might be some kind of new game for the U1MB. Has anyone tried to implement the ability to shoot these things?

Edited September 8, 2015 by danwinslow

[flashjazzcat]

I can't repeat too many times that I really love the Ultimate 1MB upgrade, and that -especially with your PBI driver- it makes a fabulous combo with SIDE2, but for me it is no longer the 'ultimate' upgrade. Too many stability issues. Probably not due to U1MB, but due to what MetalGuy66 describes.

I wonder if the Ultimate's long ribbon cables compound the bus capacitance issues? Certainly you wouldn't see so much wire in other internal RAM upgrades.

 

Anyway: nothing stays "Ultimate" for long.

 

Plus, I am disappoint with the birdies explanation. I originally thought it might be some kind of new game for the U1MB. Has anyone tried to implement the ability to shoot these things?

Heh: finally the inspiration I've needed for an Easter Egg for the Ultimate BIOS: Birdie Hunt!

[Frankie]

I thought extreme birdies was a new VBXE game or something.

[+Larry]

Actually, with the light gun, you can try to shoot the birdies as they "fly" across the screen. But at 60 Hz, it is pretty challenging... Maybe easier at 50?

 

-Larry

[Kyle22]

Actually, with the light gun, you can try to shoot the birdies as they "fly" across the screen. But at 60 Hz, it is pretty challenging... Maybe easier at 50?

 

-Larry

OK, ANTIC knows the position of the electron beam at any given time. A light gun "hit" on a birdie should provide information as to the location in screen RAM where the glitch occurred.

 

So, IF a person has a super-hero combination of eyesight and steady grip, and perfect timing, they might, once in a million years, "hit" the birdie. That's only the beginning. A intermittent glitch at a particular location, (most likely repeating locations) would still be impossible to trace using that method.

Edited September 9, 2015 by Kyle22