Another (new) 130XE boots to red screen

[woj]

I am not questioning the quality of the mod, it works great. Only I looked at the schematics and it is not just quite as simple it seems, and perhaps there is more to be gained very easily, when I check I will report of course.

 

As for 74F08, it might have done something good in any case, it booted twice very well with SIDE3 this morning being totally cold. Another kind of RAM is also almost at my door step, I am hoping that with that all issues will be gone. The video issue was not there this morning after 40 minutes of running.

[Beeblebrox]

No worries. Didn't think you were.  Just providing the background to the mod should it help. 

 

74F08 nearly always improves stability over the inferior 74LS08 IC. I bought a load in bulk and just swap em over whenever I am in an A8 socketing, etc.

 

What brand and also speed or RAM chip are you using again?

 

Also I actually noticed that for ram chip sockets, contrary to what I normally use, (eg precision sockets for most ICs), I actually find double wipe sockets seem better. Decent quality ones of course. Weirdly what I found is using precision sockets for ram sometimes sees the Ram chips pop up a little when seated which can contribute to a slightly flawed connection. I've not seen this with other 16pin rounded precision sockets and logic ICs for example. It may well be that I've not pinched the ram legs in a little in those installs and there is extra "spring". 

 

Anyways thought I'd mention it.

 

I reckon 12ms to 15ms is the ideal for the RAM. As for brand Mostek, Oki and Samsung brands are good. 

 

Here is a list of compatible chips:

 

 

 

Edited October 11, 2022 by Beeblebrox

[woj]

It has the factory MT chips, 150ns, reportedly all OK according to SysCheck, but I have my doubts. The ones on the way are KM4164B, unfortunately rather fast, 100ns. And yes, I know there are divided opinions about the kind of sockets to use, for RAMs I have all double swipe ones. 

[Beeblebrox]

MT - Arrgghh defo get rid of em. 😁 Cool - I've used the Samsung Ram chips a good few times and great re double wipes. 100ns should still be ok I guess.

 

Double wipe sockets are great to install but if you ever have to remove them for any reason they can be a pain compared with round precision ones. (Thankfully removing ram sockets isn't something you need to do if all is well). Fingers crossed all good.

Edited October 11, 2022 by Beeblebrox

[woj]

Yes, they are a pain to remove, had to do it a couple of times already, but nothing tragic. 

 

Updates:

 

1. 74F08 indeed seems to have helped with the Side3 issue, apart from this one failed boot (or was it, perhaps I was inpatient, it takes a while anyhow) it keeps booting fine.

 

2. The video issue did not reappear, so far, fingers still crossed.

 

3. Lifting C55 does not change anything in the quality of the s-video signal.

 

4. SysCheck is very unstable as a memory expansion / rom switcher. I hooked it up to see if Crownland works (without memory expansion it crashes, it should, BTW, check this first), it does, but then, for example, River Raid Winter would not start, after disconnecting SC it does. I had similar problems with running Blinky on the 130XE+SysCheck as a rom switcher only. So I'd say good for tests, not so much for regular use.

 

5. You would not believe what I spent non-trivial amount of time doing today because of my Ataris. Since I was soooo annoyed that the cracked XEX version of Black Lamp that I use for testing the video issue requires pressing Start to get past the loading screen (and I work with the naked board), and otherwise the non-cracked ATR version does not load with Side3, I decided to look into the game's boot loader in the ATR to see if I can make a clean XEX version out of it. At first it seemed rather simple and I got it to work with the Atari800 emulator rather quickly, but it failed to load on a fully "equipped" (U1MB/Side3) Altirra. Quick investigation showed that the game conflicts with the Side3 loader code in the $700-$AFF region, so a little bit of working around this was needed. But I succeeded and also tested on real hardware. I would actually post the result here, but I am not sure if the forum rules would allow this, BL not being my game and all...

 

6. Socketed 4050 on the 800XL waiting for UAV. 130XE will need it too, but I am so afraid to even uncase that one...

 

It's finally becoming a lot of fun  

Edited October 11, 2022 by woj

[woj]

Nope, 74F08 by itself did not yet solve the Side3 booting problem, today cold it wouldn't boot, persistently, until I first booted it without the cart. Now waiting for the new RAM to arrive...

[Beeblebrox]

I wonder if you could try working out if this is more likley down to a bad solder connection somewhere. So for example once the machine has warmed up and solder point(s) or socket's contact have "expanded", any slightly dodgy connection is more likely to be stable I am guessing.

 

So how about you try leaving the machine off for at least 30mins then power on without the cart. Leave the machine on for 20mins, power off, plug in the cart and immedately power up. Does this result in a successful boot?

 

Incidentally on the SIDE3 is it set to SDX or loader mode on the cart's little switch? Try the above booting with SDX mode enabled and then later loader mode.

 

Also worth mentioning I have a stock 65XE I did a hell of a lot of socketing work on last year which works fine but absolutely refuses to boot to the SIDE loader with Side3 plugged in, where the same SIDE3 cart works in all my other XL/XE's with no problem, (same power supply). Said 65XE is my test rig as fully socketed and the machine I use to test other A8's ICs when problem solving them.

 

 

[woj]

Booting into SDX instead of the loader worked 2/2 when cold. This whole experimentation to find a cold solder point - even if I have enough patience and be able to do it in an organized way, it will tell me that I have one, but then what? As a last step I do plan to reflow the CPU socket (I think I have not done that one yet, I just soldered it in). But first, I want to eliminate other things. So pending the arrival of the RAM, I removed the Basic, removed the OS rom and connected SysCheck II as an rom switcher (my other OS rom from 130XE is shot it seems), and swapped the MMU with the other one. Will test this tomorrow when cold and through the day. If it is 100% good, then I will start bringing these chips back in one by one. (I actually do suspect this OS rom too, it works and checks out with SysCheck, but gets really warm, as warm as the CPU).  

[woj]

New RAM arrived (not without small drama, the package was quite damaged by someone on the way, down to bent chip legs, but everything sorted out now with the seller and 1-star rating for our postal services), works great, so far.

 

The experimentation towards narrowing down SIDE3 booting issues slowly diverge towards the OS rom, but not yet fully conclusive. Good I own an EPROM burner, so little costs to fix that.

[Beeblebrox]

Great news re RAM. (I've always been lucky when receiving ICs in the post - bad luck).

 

One other consideration I vaguely recall when it comes to SIDE3 is that you need to make sure you have a decent brand and spec of SDcard. Have you tried booting the cart without the SDcard in place a good few times? Notice any difference?

Edited October 15, 2022 by Beeblebrox

[woj]

It was the same without the SD card, so that's not it (also, I consider the card solid, it's an old one when they were still made good, it's a big format SanDisk from the Raspberry PI 1 times).

 

But, I think I am concluding this. One learns everyday, Google told me I do not need to have the exact size blank EPROM chip to burn the OS, as long as I burn at the right offset or multiple copies of the image onto a larger chip, so without waiting for the ordered 27C128-s I opened my drawer full of 27C512 (and also found one 27C256), got the EPROM UV eraser and the GS USB burner out and started playing around. The findings:

 

1. The computer obviously works with both the XL (rev 2 I think) and XE (rev 3) OS burned onto my spare eproms. These eproms BTW are probably too fast actually, they are all 80ns.

2. My PC read out the contents of my original XE rom from the 130XE without an issue. This original rom chip though does not work in the XL, I mean the memory SelfTest fails some memory locations, while SysCheck tests fine. This chip runs cold in the computer. Important in the scope of what's below: the chip read out on the PC at a high speed setting (+2) in GS USB programmer software.

3. My PC had huge problems correctly reading out the original XL rom from the 800XL with any speed setting but the slowest. This rom does work in the 800XL, but it runs hot. This together with the fact that the PC could not read it out at a normal or fast speed tells me this is the element that creates timing issues in the 800XL and failure to boot SIDE3 cleanly.

 

That's the theory at least, I keep testing.

 

For completeness now I need to find out on which kind of chip I can burn the Basic (which is also warmer than it should be), the 27C I have will not work for obvious reasons.

 

EDIT: Right, the research on the Basic eprom tells me that I either need to find the legacy 68764/66 chip, which my burner does not even support, or build an adapter board to use a 27 eprom.

 

EDIT2: Correction, the chips I actually used to burn the OS-es are 100ns, all others I have are either 80 or 90, not sure from the symbols and I now actually do not remember what they were supposed to be. Regardless, the ones that I picked for this work.

Edited October 15, 2022 by woj

[+kheller2]

Interesting.  The XE ROM should work in the XL, if it’s the newer one that tests (incorrectly) the expanded RAM, that should work too as an OS.  

[woj]

It's not the OS version that does not work, the same exact OS self burned onto a blank chip works fine. It's the chip itself that causes some issue, even though it works perfect outside of the Atari 800XL environment. (I swear it worked fine in the 130XE, but now I lost my confidence and I am too lazy to take the 130XE apart, take U1MB out and all just to check that).

[woj]

Silence here, because both computers just work 😮 After replacing the OS rom on the 800XL with one freshly burned on a new 27C128 chip, new RAM, and Basic chip from the XE (which stays cold) it all works fine, no Side3 booting issues.

 

While waiting for UAV-s to arrive (they now left the US), a necessary side project, times two:

 

 

[woj]

One of the joysticks is ready, they are actually quite nice, at first I had doubts, the travel is very little and the spring is very stiff. But hey, with it I made it as far as 17 lamps in ROF, that was a first  I am a bit reluctant a bit to rip apart the second one, when still with a USB connector I can use it on my PC with the emulator too, I should have bought 3, not 2.

 

Today I also opened both computers for a silly reason, the U1MB Sally adapter I have in the 130XE and that I also made for the XL does not fit under the keyboard in the XL. So I hooked up the 4 wires directly to the board to have the 800XL U1MB ready, tested and it works fine. While having the U1MB out of the XE I could test if the XE rom works in the XE. And it does!!! But as said earlier, when run in the XL the memory Self Test fails. And it does work when read on a PC using an EPROM burner. So I am absolutely lost if the chip is shot or not. One thing that could explain this would be that the speed rating is too high / too fast for the XL cannot handle it, while XE can do so. But I do not even know what the chip's speed rating is or should be and why it would matter if it is faster than recommended...

 

Oh, and I desoldered the RF-modulator out of the XL, I need the space and passages for cables for when the UAV arrives (and they are not out of the US as I thought, they just flew from L.A. to N.Y., so I won't get them within next week, no way I guess).

[xrbrevin]

8 hours ago, woj said:

from L.A. to N.Y.

just like the great american cross country road race! 😁

[woj]

5 hours ago, xrbrevin said:

just like the great american cross country road race! 😁

Precisely! Seriously, the tracking history has 21 items (so far) on its list, all stops so far in the US.

[Beeblebrox]

I've had this when ordering from - for example - Atarimax in the States. Interestingly the tracking takes a while to update in the first 3-4 days after placing an order. It passes through many points in the US but as soon as it leaves the country within days I've received it here in the UK. I've heard scare stories of waiting 2-3 weeks for orders from the States but generally 10 days or less is the norm I've experienced from my Atarimax orders for example. 

Edited October 22, 2022 by Beeblebrox

[woj]

Well, I am in Sweden, which means i should expect good 1w+ of more waiting after it lands here, mostly due to inefficient customs procedures (though they got considerably better recently). It all depends if they decide to communicate with me over text messages or posted letters.

[Beeblebrox]

Sure. We'll good luck.

[woj]

The 800XL has the UAV mounted. The picture is indeed better in many ways, but (a) I had to turn the UAV pot all the way to one end to get all the patterns on the UAV control picture visible and get rid of flickering on some of the vertical line patterns (and it felt a bit like it could use a bit more turn), (b) some vertical lines (half pixels) are thinner than the others in a few paces (that I guess is due to the up-scaler in the TV, but I did not have it with the original 800XL circuit + chroma wire), (c) the "floating sea" noise when the brightness / backlight are set at a certain configuration is still there, but a bit more controllable (disappears quicker when fiddling with the TV settings). This is despite the fact that made the ground connection the most "aggressive" I could.

 

The 130XE is half way done, the most sensitive part of soldering the 4050 socket in is done and working, so now it's about wiring up the UAV. A very unpleasant surprise was to discover that the green connector on one of the UAVs was soldered in front side back, quality control seems to be lacking at TBA 😕 I managed to repair it quick and clean, tested on the 800XL, it works fine, still... On the 130XE that has the simple s-video mod done I do not expect any difference with the UAV to be honest (now that I know the floating sea noise effect will not disappear), apart from the black hair line between the shades of gray disappearing on the control picture, and getting back the composite output that vanished with the RF modulator gone. So it's still worth for my OCD  

 

Oh, and UAV or not, the s-video picture on my living room TV still sucks sooooo bad, but at least now I have a decent composite for this one.

 

[Beeblebrox]

Check this out which includes a uav output review. 

 

 

 

[woj]

Right, so I have bit of a pickle with the 800XL. I mounted the other UAV into the 130XE and there the picture and the UAV pot calibration is spot on - in the exact middle position of the pot the lines are solid thickness, patterns OK, everything is spot on. On the XL I need to turn the pot all the way to the left to get to almost the same state as on the XE (on the XE if I turn the pot slightly to the right I get the same picture as on the XL with the pot at max left). I swapped around the UAV-s, it's not the fault of the UAV, it's the computer (in the process I managed to snap one of the UAV legs off, fixed now, but a second casualty today, earlier I snapped one of the legs of the 130XE MMU when I removed the U1MB for testing the UAV).

 

The first picture is the XE with pot in the middle, the second one is the XL with the pot at max left. It is hardly visible on the pictures, but the x and l on XL are slightly thinner than the rest of the picture and a bit gray.

 

Any clues? The video cable is the same, the power supply is the same, the same TV...

 

EDIT: Thanks,yes, I have seen Jon's video already yesterday  

 

Edited October 29, 2022 by woj

[woj]

Huh, I rewatched Jon's video now, he had the same problem with the UAV on that XL machine, even though that one was NTSC... Interesting, very interesting...

[woj]

I had a moment today when I thought I figured out the UAV issue on the 800XL, but no, not yet, but I am onto something there... In the meantime, I professionally repaired what I broke recently, the MMU has two brand new legs