1088xld problems

[_The Doctor__]

At this point I'm looking for a chopper pilot, we'll pick Stephen and Michael up... then fly to your house and beat this 1088 into submission!

Edited May 17, 2020 by _The Doctor__

[chevymad]

Rn's read correct. I did find one more incorrect single resistor. Not sure how after being checked at least 3x.. but it wasn't even close. Of course it was in the PAL section R13 and made no difference anyway. 

[+mytek]

Once again I've run out of ideas, but if I think of anything I'll let you know .

 

BTW, a 1N4148 substitutes perfectly for a 1N914. Although based on what you were seeing, it was pretty certain that neither the transistor or the diode were at fault. The key thing to keep in mind is that the only reason the system worked with just those 2 items removed, has more to do with the 74F138 (U25) coming online when it shouldn't and providing the GND for D14 and Q6, thus bringing either data line D6 or D7 LOW (with the later also requiring that the Swap Latch is set).

 

And it does appear from your checks that you do have a problem with your Swap Latch being set when it shouldn't be, or its not getting cleared after boot. Although that in itself would not do a thing without the cooperation of the 74F138 decoder to read it, which should only happen when the U1MB accesses it via a $D1Cx read, unless there is a problem with the decoder or the decoder addressing, or the main enable input (pin 4) is going LOW when it shouldn't as has been mentioned.

 

I know lots of redundant info, but this really has to be the area of contention. As to what specifically is the issue, that's pretty much up in the air based on the tests you've run thus far.

 

[chevymad]

Replaced the rest of the cheap sockets, except U6 and sram. No change of course, didn't expect one. 

 

Now i've been doing some experiments. 

 

If I pull just pin 10 of U25, keyboard functions but cart does not. 

Put that back and pull pin 2 of U22. Keyboard fails, cart fails but tries then locks at first.

Pull pin 3 of U22 same as pulling pin 2.

Pull both 2 and 3 of U22 same.

Pull leg of D14 up. Keyboard fails, cart fails but tries.

Pull both pin 10 of U25 and leg of D14. Keyboard works, cart works, SIO works.

Pull pin 9 of U26, nothing works. Pull pin 8, nothing works-light turns green. (by nothing I mean keyboard,cart,sio)

 

 

Diode and all other pins reinstalled. Pull pin 3 of U24 then both keyboard, sio and cart work. 

What is interesting however, is that if I pull pins 1 and 5 of U24 and put pin 3 back in place the keyboard, sio and cart problems return. I've tried the HCT00 from my XEL as well. 

Edited May 18, 2020 by chevymad

[chevymad]

Ok, going over function of nand gate, checking voltages on pins 1-6 with pin 3 lifted.

pin 1 5v

pin 2 1.7v

pin 3 3.7v output from chip with pin lifted

pin 4 1.25v

pin 5 4.34v

pin 6 1.7v

 

Pin 3 re-inserted

Pin 1 5v

pin 2 0v

pin 3 5v

pin 4 5v

pin 5 4.3v

pin 6 0v 

[chevymad]

Way back earlier Mytek mentioned either hooking up a cf adapter or tying pin 9 to 14 of u24. I've tried both and the led stays yellow. 

 

Is pin 6 of u13 ever going to be low? Joy1 and Joy2 are the indicator lights correct? So only 1 is ever on at a time. It cycles both off, right on, both off, left on. 

[chevymad]

Well, actually played a game of Star Raiders on it. The Atari part of the computer seems to work fine with pin 3 pulled. Another oddity however. If I put my PokeyOne in the right pokey slot, the stereo selection of the U1mb is grayed out. With nothing in the slot, stereo is enabled. This works the same with or without pin 3 of U22. 

[+mytek]

16 hours ago, chevymad said:

1. If I pull just pin 10 of U25, keyboard functions but cart does not. 
2. Put that back and pull pin 2 of U22. Keyboard fails, cart fails but tries then locks at first.
3. Pull pin 3 of U22 same as pulling pin 2.
4. Pull both 2 and 3 of U22 same.
5. Pull leg of D14 up. Keyboard fails, cart fails but tries.
6. Pull both pin 10 of U25 and leg of D14. Keyboard works, cart works, SIO works.
7. Pull pin 9 of U26, nothing works. Pull pin 8, nothing works-light turns green. (by nothing I mean keyboard,cart,sio)
8. Diode and all other pins reinstalled. Pull pin 3 of U24 then both keyboard, sio and cart work. 
9. What is interesting however, is that if I pull pins 1 and 5 of U24 and put pin 3 back in place the keyboard, sio and cart problems return. I've tried the HCT00 from my XEL as well.

I decided to map out the pulled pin tests that chevymad did, as can be seen above. And I have made observations listed below.

 

Step 1 should be a complete disconnect of any kind of energizing signal for D14 or Q6 which is derived from U25 pin 10, and thus be the same as if those components are not present. However the Step 1 note says: "keyboard functions but cart does not".

 

Steps 2, 3, and 4 should not affect anything other than a connected IDE/CF device. However the notes disagree: "Keyboard fails, cart fails but tries then locks at first". Although I think that's really the U25 issue revealing itself once again.

 

Step 5: not much I can say, other than there is an issue with U25.

 

Step 6 supports what I've been saying all along, and that is that U25 is coming online when it shouldn't, and thus energizing the diode (D14) and if the Swap Latch is SET (which it apparently is) also transistor (Q6), which has the end result of screwing with the state of data line D6 and D7 respectively for any other OS reads whenever U25 has brought pin 10 low when it shouldn't.

 

Step 7 should only affect the Bi-Color LED and nothing else. However the note for step 7 says: "nothing works-light turns green. (by nothing I mean keyboard,cart,sio)". I underlined the last part of that note which suggests the OP sees otherwise. Most likely a U25 issue, since U25, D14 and Q6 are still in circuit.

 

Step 8 would totally break the flip-flop latch formed by two gates of U24, and leave one half of that circuit with a floating input, as well as the LED logic that's downstream of it. This would result in unpredictable behaviour.

 

Step 9 would simply leave the latch SET input floating, resulting in unpredictable behaviour.

 

15 hours ago, chevymad said:

Way back earlier Mytek mentioned either hooking up a cf adapter or tying pin 9 to 14 of u24. I've tried both and the led stays yellow.

Since /DASP on the IDE connector is floating without a IDE to CF adapter present, the color of the Bi-Color LED will be unpredictable if the adapter is not present. However if the LED remains yellow with the adapter plugged into the IDE port, then it points to a problem with the Swap Latch or it's trigger which has already been established many times already.

 

15 hours ago, chevymad said:

Is pin 6 of u13 ever going to be low? Joy1 and Joy2 are the indicator lights correct? So only 1 is ever on at a time. It cycles both off, right on, both off, left on.

The TK-II chip momentarily (almost faster than the eye can see) brings both JOY1 and JOY2 signals high in response to ALT+N in order to set the Swap Latch via the inversion by NAND gate U13.

 

Bottom Line...

 

U25 is at the heart of the problem as far as I can see. Yes there are other problems on the board (e.g., Swap Latch), but the crux of the problem is U25 misbehaving. If anyone cares to dispute that or present there own interpretation, please do so.

 

Edit: when I say that U25 is at the center of the problem, that doesn't necessarily mean the the 74F138 chip is at fault, which I think has been previously verified as good by swapping for a known working chip already. It could be a floating (disconnected) input to U25, or any of the 3 enable pins (4, 5, 6) being permanently engaged by a short to ground in the case of a /EN pin (4 or 5) or a short to +5v on a EN pin (6), or something to do with the signals coming into those pins. So this is what the tests should concentrate on first. We don't care about any of the other issues with the Swap Latch circuit, or the Bi-Color LED, or D14 and Q6 (which appear to be ok) for the moment, since those are all non-issues to what is really the cause of the problem.

 

[_The Doctor__]

It's pretty clear u25 is the eye of the storm...  what would a failing, flaky u26 do to it?.... it's got itself intertwined there

[+mytek]

3 hours ago, _The Doctor__ said:

It's pretty clear u25 is the eye of the storm...  what would a failing, flaky u26 do to it?.... it's got itself intertwined there

Yes it could, and there were some preliminary checks that suggested things were ok. Of course if $D1xx remains high, except for when it's suppose to go low, and that migrates though the 74F32 (U26) to the /EN pin on the 74F138 (U25), then it wouldn't even matter what the input (A3) on U26 was doing, since there could be no enable without $D1xx going low. I don't know if it was done, but lifting each input pin (1 and 2) on the associated gate in U26 and then starting out by jumpering both to GND should also give a GND on it's output (pin 3) and on the /EN pin (4) of U25. Next does bringing either input of the U26 (1 or 2) high while the other is grounded make the output also go high? And at the same time does the /EN pin on U25 also go high simultaneously. Both U26 inputs should be verified in this manner.

 

Some of this comes down to the only way to truly know is if I had the board sitting in front of me. Not trying to say anything bad about chevymad, but I was a service tech for quite a few years, and the number one rule was don't believe anything the customer tells you without proving it for yourself. Or in other words do your own troubleshooting with no preconceptions. I can't tell you how many times this way of approaching things saved me countless wasted hours. And bottom line this could just as easily be due to a manufacturer defect in the PCB itself, which has been seen in past board runs.

 

However just to be clear, I'm not volunteering to work on chevymad's board . And this really takes us back about a year ago when the initial decision was to not offer bare untested boards for sale. MacRorie caught a lot of flak for that announcement, and ultimately conceded to selling DIY boards to the public. I can't speak for MacRorie, but watching this scenario unfold over the last several weeks convinces me that if I were the one selling the 1088XLD I would flatly have to say no thanks to selling the unstuffed/untested boards. There's good reason why people like Lotharek don't, and its not always about money, but more often to do with the best use of a person's time.

 

[_The Doctor__]

All true, however, it is much of the fun to have your own kit computer, stereo, amplifier, weather station like you would build back in the day...

 

-snip- if what you suggest hasn't been done... maybe it will be---

lifting each input pin (1 and 2) on the associated gate in U26 and then starting out by jumpering both to GND should also give a GND on it's output (pin 3) and on the /EN pin (4) of U25. Next does bringing either input of the U26 (1 or 2) high while the other is grounded make the output also go high? And at the same time does the /EN pin on U25 also go high simultaneously. Both U26 inputs should be verified in this manner.

 

Worst comes to worst he could bread board the section you highlighted and sub it for the section on the PCB, if it's all good... that leaves him with bad trace, bridge, open, or broken.... If it's all the same... defect component... or incredible universe improbable calculation on a waitresses check pad... to have mistake replicated on bread board also...

Edited May 18, 2020 by _The Doctor__

[+mytek]

37 minutes ago, _The Doctor__ said:

All true, however, it is much of the fun to have your own kit computer, stereo, amplifier, weather station like you would build back in the day...

 

-snip- if what you suggest hasn't been done... maybe it will be---

lifting each input pin (1 and 2) on the associated gate in U26 and then starting out by jumpering both to GND should also give a GND on it's output (pin 3) and on the /EN pin (4) of U25. Next does bringing either input of the U26 (1 or 2) high while the other is grounded make the output also go high? And at the same time does the /EN pin on U25 also go high simultaneously. Both U26 inputs should be verified in this manner.

 

Worst comes to worst he could bread board the section you highlighted and sub it for the section on the PCB, if it's all good... that leaves him with bad trace, bridge, open, or broken.... If it's all the same... defect component... or incredible universe improbable calculation on a waitresses check pad... to have mistake replicated on bread board also...

It would be much easier to substitute a modified XEL-CF3 board for the same section on the 1088XLD. You do give up a tiny bit in the process (additional address buffering), but for the most part it's identical in the IDE and swap circuits. However even that would feel like a kludge to me.

 

[chevymad]

You can see I was trying to isolate d14 & q6 from all input/outputs. 

I doubt it's soldering or PCB since I've tried 2 different pcb's of different colors even. 

Alt + N does make the 2 joystick led's quickly flash. My dvom is just too slow to see the voltage change. I will have a virtual scope later this week. 2 channel 100mhz Osos. Just have a tool fetish. 

There's not many components left that have been on both pcb's and that I don't have working units in the XEL. I may just figure out which ones those are and order them.

 

I want to thank Mytek, Simply Graham and everyone else who has contributed and helped here. I'm afraid to ask how much time you've wasted on my problem Mytek, it's very much appreciated however. As an automotive technician I too use your NUMBER 1 rule. It's saved me probably years worth of time by now. 

 

You can see from one of my posts to the next, that I had to look up how nand gates function. Got to thinking earlier today, it's been 32 years since I took a course in logic circuits at WSU. At that time I was going for an EE major(they didn't call it Computer Engineering til the next year). I haven't had a chance to use what I learned in all that time til now, and it's basically all gone.

 

I'll keep plugging away. Some of my posts are more like "thinking out loud" then anything. Thanks for commenting on which direction to head afterwards. But I hope no one feels obligated to help like it's a job or anything. This is supposed to be fun. (i'll keep telling myself that!)

 

So my pulling of pin 3/u24 is cutting off the power supply to Q6. But the real problem is U25 turning Q6 on when it's not supposed to. Why would pulling pin 10 of u25 not make the cart/sio work too then? Pin 10 and d14 have to be lifted before all 3 work, yet the diode should stop flow in that direction. At least that's where I was thinking when I went to pin3/u24. The diode must backflow just enough to activate q6? These questions are more for my understanding here, since q6 and d14 have been replaced they should obviously function as intended. 

 

I'll be trying your additional tests after awhile Mytek. Thanks again.

chevymad/Brandon

 

[+mytek]

20 minutes ago, chevymad said:

I'll keep plugging away. Some of my posts are more like "thinking out loud" then anything. Thanks for commenting on which direction to head afterwards. But I hope no one feels obligated to help like it's a job or anything. This is supposed to be fun. (i'll keep telling myself that!)

As long as you are having fun and not cursing at me for inventing the thing that is giving you so much trouble, I'm perfectly fine.

 

20 minutes ago, chevymad said:

So my pulling of pin 3/u24 is cutting off the power supply to Q6. But the real problem is U25 turning Q6 on when it's not supposed to. Why would pulling pin 10 of u25 not make the cart/sio work too then?

Yes that is odd. Perhaps without a connection some sort of strange back feed is occurring between the transistor and the diode, since the transistor is being switched on because of the stuck Swap Latch. Normally when connected, and assuming that the decoder is working properly and connected, you would have 5v present that would inhibit that. 

 

20 minutes ago, chevymad said:

Pin 10 and d14 have to be lifted before all 3 work, yet the diode should stop flow in that direction. At least that's where I was thinking when I went to pin3/u24. The diode must backflow just enough to activate q6?

And yes perhaps the back flow is in the opposite direction then what I just described .

 

20 minutes ago, chevymad said:

These questions are more for my understanding here, since q6 and d14 have been replaced they should obviously function as intended. 

 

I'll be trying your additional tests after awhile Mytek. Thanks again.

You're welcome ? .

 

[chevymad]

Brain's fried.. I either ordered the wrong scope or forgot the specs of the one I did. So the correction to my previous post, I have a 25mhz 2channel  Owon VDS1022i scope coming. Spent much too much time comparison shopping between too many scopes.

[Simply_Graham]

Lots have been happening whilst I've been slaving away on a new garden shed.

As Mytek stated, issue with [D1xx] line driving Low (enabling U25 via U26) when it should not; this is pin 14 of U11 that feeds into U26 of pins 2, 4 & 5

 

Please re-check if Pin 2 (or 4, 5) ever goes low with U26 removed (With your new scope whilst computer is on)

D1xx only accessed by PBI BIOS as far as I'm aware (confirmation would be good here)  so make sure PBI disabled in U1MB setup.

If it ever goes Low (possible 74LS138/ U11 fault) or if its permanently Low with power disconnected pull U11 and check if pin 14 on U11 is short to ground or maybe with adjacent pin 13 (Pokey select thinking issues with SIO, Keyboard etc.)  With a multimeter, leaving out U26 as well as this should isolate that track.

 

Assuming something is found or not it's another part or track that's eliminated, but as Mytek mentioned there appears more than one issue.

Has pulling all U22 to U26 been tried?  Only issue would be D14 and Q6 state (as base floating) with all chips removed.

 

I'll always agree with you on good test equipment! 

[+mytek]

Was this been tried yet?

 

This would be the easiest way to completely take the 74F138 decoder out of the mix, while still leaving all of the other components including the decoder still in circuit. If this allows everything to work other than the IDE, it would really narrow down the search for the problem.

 

[chevymad]

8 hours ago, mytek said:

Was this been tried yet?

 

This would be the easiest way to completely take the 74F138 decoder out of the mix, while still leaving all of the other components including the decoder still in circuit. If this allows everything to work other than the IDE, it would really narrow down the search for the problem.

 

When I do this, tying pin 4 of U25 to pin 14 of U26 causes it to boot to a rolling bright yellow screen full of noise. disconnecting the jumper and pressing reset will bring back the basic ready prompt. 

[+mytek]

1 hour ago, chevymad said:

When I do this, tying pin 4 of U25 to pin 14 of U26 causes it to boot to a rolling bright yellow screen full of noise. disconnecting the jumper and pressing reset will bring back the basic ready prompt. 

My initial reply is that is not logical. Bringing U25 pin 4 high should have completely taken the decoder out of the loop, basically the same as if you removed the chip. The only exception is that removal of the chip allows everything connected to its outputs to float. But with it responding like that, there's really nothing else I can recommend at this point, besides blow some magic dust over it and see what happens. I know that's not the response you would like to hear, but with U25 defying it's own operational parameters, and the fact that you said you substituted another decoder earlier in the troubleshooting stages with no difference in results, means there really isn't an answer for what you are seeing.

 

If you look at the logic diagram below, you'll see that pin 4 = /E1, and if that input is high (1st line in chart), then it doesn't matter what any of the other inputs are doing, because all of the outputs will remain high irregardless.

 

 

So in other words, nothing will be selected (all decoded states are active low). And since you don't have a CF card in play, there are only two outputs that can even influence anything (pin 10 & 11 of U25). One of those is for resetting the Swap Latch, and the other is for reading the signature and status bits (D6 & D7). Both of those require that there respective connection to the decoder goes low in order for those operations to be carried out.

 

Oh there is one more thing you can do just to rule out the decoder as a problem... Bring pin 4 of U25 high once again and then take voltage readings on all the outputs. They should all be in a steady state of high (+5V). If any are not, something fishy is going on.

 

[chevymad]

Ok. Testing voltage on all outputs shows all high at 4.3v. Except pin 10 @ 3.7v. Like it has more load on it.

 

Disregard video, boot issue was poor connection on stacked socket. 

 

Edited May 20, 2020 by chevymad

[_The Doctor__]

voltages look very low to me...

[Kyle22]

What is your power supply voltage [under load]? Please measure supply voltage on other chips.

 

[chevymad]

Ok.. I have a stacked socket under U26 to get it high enough to pull pin3 since the sockets on the board touch. Evidently that connection was preventing boot. Messed with that and it will boot to the ready prompt. However, board still has the same issues. Also it did not change the voltage readings on U25 outputs. Input voltage to U25 is 4.94v @ pin 16. 

[chevymad]

To me this is saying there's a load on pin 10 of U25 that the chip can't supply. It's causing a voltage drop across the entire output and especially pin 10. I think I need to start isolating that circuit while watching for the voltage to come back up. 

 

Anyway, tomorrow maybe. Today was my 50th birthday and i'm ready for bed. Thanks again everyone. 

[Kyle22]

1 hour ago, chevymad said:

To me this is saying there's a load on pin 10 of U25 that the chip can't supply. It's causing a voltage drop across the entire output and especially pin 10. I think I need to start isolating that circuit while watching for the voltage to come back up. 

 

Anyway, tomorrow maybe. Today was my 50th birthday and i'm ready for bed. Thanks again everyone. 

Happy Birthday.