1088xld problems

[chevymad]

Feel I shouldn't be filling up Mytek's build thread with my XLD problems. Some background posts are over there. Maybe a mod can move them to this thread?

 

Anyway, my XLD is getting improper keyboard inputs. Some characters will type as if the control button is pushed. Also, I have no SIO and Star Raiders won't boot(black screen). Today I reflowed all solder connections on the main board. I also removed every socketed chip and tried them in my 1088xel, and vice versa. My XLD now has every working chip from my XEL in it. This is the 3d set of Atari chips i've tried in it too. 

 

Pressing Alt F1 and Alt F2 print varying amounts of gibberish. Sometimes it's close to what it's supposed to say, other times not at all. 

 

Also, found I can go to the stock XL Memory, Keyboard, and sound tests.  Memory test shows 2 green ROM blocks, and 40 green RAM blocks. Same as my XEL. The keyboard test is interesting however. Keyboard almost works correctly on the test and only rarely will I see the Control key pressed as well. Maybe 1 time per alphabet or even less. While in Basic it messes up much more then that. 

 

Basic actually works. I can type in simple programs like hello world and they'll work fine. Frustrating to type in because of the keyboard input, but once typed correctly they work.

 

I've tested continuity between Pokey L and TKII chips, and TKII and keyboard input. All appears OK. TKII is in XEGS mode as well. 

 

Resistor networks appear to be installed in the correct orientation.

 

Plugging in my Sdrive Nuxx, the xld will try to access the disk and then boot error over and over.

 

I think that about covers what i've learned and tried so far. I do have a scope I can test with this weekend, I just don't know what to test! All help is appreciated. I'm sure there's a bad component or short somewhere, Just haven't found it yet.

[chevymad]

[chevymad]

Also as you see, i've got it down as simple as possible for the moment. Only UAV, no S2 board installed, and left the right Pokey out as well. 

[StickJock]

If the TK-II is communicating properly with the Pokey, and the Pokey is working properly, then perhaps the data lines from the Pokey are the problem? Check D7 on the Pokey. If this line is intermittent, it can appear as if control is pressed or not pressed. 

[+mytek]

I saw a recent post that said SDrive didn't work. So let's set aside the keyboard issues for a second and look at the SIO side of things. What other SIO related devices have been tried?

 

Seems to me since both the keyboard and SIO share a common chip, if SIO is also a problem then some other system connection to Pokey might be at fault for both issues. D7 has been suggested, and perhaps that would be a good place to start looking. Also earlier in the previous thread IRQ was also mentioned. So bottom line anything that disrupts a reliable connection of Pokey to the system bus could and would create issues with all of its individual parts (paddles, keyboard, sound, SIO). Might be time to run tests of each one those.

 

[+DrVenkman]

Since you have access to a ‘scope, it might be useful to simply do some basic checks of all the key signals on POKEY. Specifically, check Phi2 on pin 7 - make sure it’s not getting corrupted or otherwise distorted. The signal is buffered so you should see a nice well-defined square-ish wave with a rock solid 1.79Mhz. Check for activity on the keyboard data lines as you type (might need help with this - to hold the probes and watch the scope while someone tries to type).

 

From the photos of the solder-side of your board, I see what looks like some flakes of solder around - you might spray the board down with iso in a spray bottom and then compressed air to clean out any stray flakes. You could try the same on top of the board, especially around the base of sockets where it’s hard to see.  

[chevymad]

Yes, board is dirtier then I like after the reflow job. Running out of cleaner and being forced to move at the moment so my compressor is packed into storage. The sdrive is my only sio device.  I like the suggestion to come at this from another angle besides the keyboard. Sound seems to work fine, don't have paddles. Not sure how I would test paddles/joysticks without being able to run software, since cartridge and sio doesnt work. 

 

ISO might be a problem.. not sure how far I'll have to go to find a store not sold out. 

Edited April 26, 2020 by chevymad

[StickJock]

You could write a short basic program to display the stick, trigger & paddle values.  It may take you a bit of effort since you get control characters randomly, but it should be doable.

 

[+DrVenkman]

A can of compressed air should be pretty easy to find still, I'd think. Hope, maybe. Just to make sure there's no debris shorting a line under a socket or around your I/O riser.

[+mytek]

1 hour ago, chevymad said:

Sound seems to work fine...

So I'm assuming you hopped into self test and ran the sound check from there. If all 4 voices appear to be working correctly that's a good sign that Data, Address, R/W, CLK (B02), and /CS0 & CS1 are all connected and reliable. Not sure off the top of my head if /IRQ is involved with the sound test. Maybe someone in the know can chime in on that.

 

No need to take it farther, since that rules out a good majority of the system connections to Pokey. Just need to verify if /IRQ is involved with the sound test.

 

Let's see if we can get the keyboard working without CTRL, SHIFT, or BREAK so you can at least verify that standard keys work and make it easier to write some BASIC for testing. Remove Pokey from the socket and slightly bend out pin 16 (KR2) and put it back in the socket. Temporarily tie pin 16 to +5V (pin 17 on Pokey). Now power back up and if all goes well normal keys should work.

 

 

[StickJock]

If mytek's suggestion above results in perfect characters from the keyboard, then you know that it is KR2 related.  If, on the other hand, you still get the Control corrupted characters, then it indicates that the issue is on the downstream side of the Pokey (like D7).

 

[chevymad]

Well, this is going to be a long slog. Local stores were all sold out of ISO, no electronic cleaner, nor canned air. So drove 40miles to Walmart, also sold out of ISO but bought a can of electronic cleaner and some air.

 

 I've discovered a couple issues, tried to correct and made it worse. First I tested PHi2 and get a steady 1.77volts.. no square wave no frequency.
Then I started testing continuity on PIA and discovered one of the joystick traces to rn7 was broken. While looking over the PIA socket I discovered a blob of solder under the socket. Had to remove the socket to remove the blob.. broke 2nd trace. Patched both traces under the board with wire, then tested continuity on all pins before retry. And now I just get a bright colored screen.  

 

Not sure what i've broken at this point. Bet I have too much solder under some other connector as well.  Going to take a break for awhile.

[+mytek]

I'm sorry to hear that you broke some traces. I now wish that you would have tried what I suggested before worrying about cleaning the board, since that would have provided valuable feedback as to what's going on. Unfortunately that won't be possible until you get back to the functionality you had prior to the attempt to repair traces.

 

The 1088XLD is not as robust as the XEL. It uses thinner traces and is made from a thinner board material. The thinner traces were to allow a higher density routing in certain areas, and the thinner board was to keep costs down, since the XLD is a rather large board. You'd be surprised how heavy those XEL boards get when you pack up 10 or more to be shipped, and of course that pushes up the shipping cost.

 

4 minutes ago, chevymad said:

First I tested PHi2 and get a steady 1.77volts.

Was this before or after the bright colored screen?

 

Since it would be impossible to have Phi2 be at a steady state, but still display a picture, even one with garbled characters.

 

[AtariGeezer]

52 minutes ago, chevymad said:

While looking over the PIA socket I discovered a blob of solder under the socket.

That was the exact reason I suggested checking for shorts between adjacent pins on each socket.  Had 2 of those on my XEL build, had to do some upside-down solder extraction to fix those...

[chevymad]

Before.. Left pokey pin 7, pia pin 25,antic 29, gtia pin 30.. all come up with steady 1.77volts.  I did get square waves on some of the lines between tkII and pokey though. If it's not a possibility, then it could be that my oscilloscope won't read that high of frequency. It is an automotive based handheld scope, mhz could be beyond it's range. The manual for it is long gone. 

 

First broken trace was at one of rn7's through holes. Rn7 had a bent pin I'd attempted to fix and should have left alone. 2nd trace I must have scratched with the soldering iron under the board. Trace runs between 2 socket pins and was broken right there.

Edited April 27, 2020 by chevymad

[chevymad]

Scope only works to Khz.. so there's why i'm seeing flat voltage. It's old enough I had a hard time finding an online manual. played with the 1088 a bit more, still don't have it working again yet. 

[+orpheuswaking]

@chevymad you asked for help from the designer of the board and have been asked several time to try troubleshooting steps, yet then gone off to clean up the board and try other things first. Honestly you are doing more damage at this point and I would suggest once you get the board back into a place you can test things that you actually follow what is being asked of you. I say this respectfully but at this rate you are going to completely destroy that board and have no chance for success. 

 

To be frank, the beta group and Macrorie had a fairly long discussion of whether these boards should be sold as kits or only full build for this exact reason. 

[+mytek]

Quote

Before.. Left pokey pin 7, pia pin 25,antic 29, gtia pin 30.. all come up with steady 1.77volts.  I did get square waves on some of the lines between tkII and pokey though

Since your scope isn't any where close to fast enough, you are seeing it average the transition between high and low states on the faster clocking lines, hence the steady 1.77 volt reading. The key scan counter on Pokey is clocking in the Khz range, with K5 being the slowest, K0 being the fastest, getting into the range that your scope can handle.

 

Good news is that Pokey is still alive.

 

 

[chevymad]

2 hours ago, orpheuswaking said:

@chevymad you asked for help from the designer of the board and have been asked several time to try troubleshooting steps, yet then gone off to clean up the board and try other things first. Honestly you are doing more damage at this point and I would suggest once you get the board back into a place you can test things that you actually follow what is being asked of you. I say this respectfully but at this rate you are going to completely destroy that board and have no chance for success. 

 

To be frank, the beta group and Macrorie had a fairly long discussion of whether these boards should be sold as kits or only full build for this exact reason. 

I'm sorry, I know Mytek,s the designer and would have done exactly as he suggested, but finding a big blob of solder under a socket meant I needed to get that out of there. At one point he did indeed suggest going over all of the solder joints, not just the one's I previously had. My previous work of checking shorted pins had concentrated on the pokey and TKII. Since that wasn't going anywhere, I looked at the pia. If the board had booted up after removing the blob, I'd be trying Mytek's suggestion.

[_The Doctor__]

have you swapped ultimate 1 meg?

double checked jumper config on main board and sub boards?

verified stability of clock (use scope)

check continuity with all chip out that you can from all pins in common from socket to socket furthest away from each other. If opens exist, work your way in until opens disappear. then look between the last two sockets after the open resolved for breaks and cuts.

Edited April 27, 2020 by _The Doctor__

[+orpheuswaking]

5 hours ago, chevymad said:

I'm sorry, I know Mytek,s the designer and would have done exactly as he suggested, but finding a big blob of solder under a socket meant I needed to get that out of there. At one point he did indeed suggest going over all of the solder joints, not just the one's I previously had. My previous work of checking shorted pins had concentrated on the pokey and TKII. Since that wasn't going anywhere, I looked at the pia. If the board had booted up after removing the blob, I'd be trying Mytek's suggestion.

are the pictures up there ☝️ before or after you reflowed the board?

[+Stephen]

Not to pile on, but the soldering on the board is not-passable.  When I did this for my full time job, I would fail my own work from below.  However, note all the joints are smooth, shiny, consistent and the solder has cleany flowed all the way to the top.  There should be no buldges (too much solder), no indents (too little solder), no flux residue left after cleaning.

 

 

 

[chevymad]

You know. I'm doing this for fun. I'm not a pro, I'm not selling it. I'm not complaining about it not working. I haven't said there's anything wrong with Mytek's work, I know the stuff will work when done right, and eventually I'll get it there. If this was only sold as a full build, I'd never have one. Once this is working, it probably won't see tremendous use. The reason I bought it, is because it's a kit of something I like. It's the hot rod of the Atari world and I get to build it myself.

 

The whole reason for the post was for tips on what to check. So apparently it's my soldering, I'll go from there. Thanks for your help.

[+Stephen]

Look - nobody is trying to put down your efforts.  If you start a thread asking what may be broken and someone gives you advice on what they think is broken, don't get pissy about it.  Everyone that has posted here has done so nicely and is trying to help.

[chevymad]

One person came across as elitist and just rubbed me wrong. They probably didn't even mean it like that. Stephen thanks for the constructive criticism on the soldering. It definitely doesn't flow as easily as on the XEL. I'm going to try a few things and see if I can pick that up. Temp,solder, flux etc. The "electronic cleaner" I used also seems to dull/gray the solder. Just for curiousity, what temp is everyone using for the iron? This was bumped up a bit from what I used on the XEL.  Also have better magnifying glasses coming. I think my vision is even crappier then it was when I built the XEL. Even with stronger readers. Until then I won't be doing any more troubleshooting. Stepping back for a few days. I've been avoiding saying anything else here, to not make the hole deeper. The old saying If you can't say anything nice.