Problems with build

[Hans Van Slooten]

Hi Everyone,

I just recently completed building my PAL-1 and while the instructions are clear and straightforward, I must've made a mistake somewhere because when I plug it in, I don't get any display.

I checked all my solder joints and the orientation of ICs/caps/diodes/etc and don't see any problems (that's not to say I didn't make a mistake, I just didn't see any).

I hooked up a logic analyzer to the 6502 and 6532 and there's a clock signal, there's activity on the bus (it's repetitive, like a loop), but it appears that there is nothing on the PA0-PA7 pins of the RIOT. I have a second RIOT from the expansion and I tried that with the same results, so I don't think it's that chip. 

I dumped the ROM and compared it to known ROM images and it matches, so the ROM is fine.

I'm no expert with debugging hardware, so if someone could point me in a general area to troubleshoot to track down my mistake, I would appreciate it.

Thanks!

Hans

[Neil Andretti]

Good morning Hans,

could you provide us with some photos of the board? Maybe it's something obvious.

Nils

[Magnus Olsson]

One thing that’s very easy to miss is the setting of the two jumpers for TTY vs keyboard and on-board vs external memory. If any of them is set wrong, you won’t get any display. 

[GN L]

Hi Hans,

I also think some clear photos of the board (front and back) will help.

And, I want to try my luck, is the I/O selector jumper OPEN?

> --
> You received this message because you are subscribed to the Google Groups "PAL 6502 computer" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to pal6502+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/1e106912-3295-4d11-9425-189803d70bdcn%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Best,
Liu

[Hans Van Slooten]

Sure thing. 

--
You received this message because you are subscribed to a topic in the Google Groups "PAL 6502 computer" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pal6502/KPfvKGdgOcQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pal6502+u...@googlegroups.com.

[Hans Van Slooten]

Yeah, I tried every combination of jumpers. I even tried to connect via serial, with no luck.

Thanks!

Hans

--
You received this message because you are subscribed to the Google Groups "PAL 6502 computer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pal6502+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/a0454039-a0db-439f-b0ea-1074198722f5n%40googlegroups.com.

[Hans Van Slooten]

I can also send logic analyzer output if it helps. I haven’t dug into the boot process to see what values should be on the bus yet, so I’ve only confirmed so far that there’s actually stuff going on.

[GN Liu]

From my eyes, the front side seems normal.

The back side looks dirty ;)  I think you can check all the solder joints again to make sure there are no shorts. From your photo I can see some possible shorted pins.

[Hans Van Slooten]

Thanks for your input everyone. As an update it still isn’t working, but I tested the ICs in my TL866II+ (everything except the RIOT, 6502, and 556) and they all checked out. The RIOT should be OK since I’ve tested two of them. I also cleaned up the flux on the board with some better cleaner and leaned off the “fuzzies” that looked like solder flags. I also did a ton of continuity checking to look for solder bridges, with no luck. I still have to do some more logic testing to see what the heck is going on. Here’s the cleaner board (I used a flash so it would be clearer).

Thanks!

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/6f62b0ed-afdc-4f73-a659-882efe7cf33en%40googlegroups.com.

[GN L]

All solder joints look beautiful after cleaning.

I see you added keycaps on the switches, did you install the original switches from the kit or use some other switches with keycaps? If the switch is replaced, you need to check that the upper two pins are connected and the lower two pins are also connected. Like this:

1 2
o-----------o

o-----------o
3 4

Thanks,
Liu

> On May 19, 2022, at 11:03 AM, Hans Van Slooten <vans...@gmail.com> wrote:
>
> Thanks for your input everyone. As an update it still isn’t working, but I tested the ICs in my TL866II+ (everything except the RIOT, 6502, and 556) and they all checked out. The RIOT should be OK since I’ve tested two of them. I also cleaned up the flux on the board with some better cleaner and leaned off the “fuzzies” that looked like solder flags. I also did a ton of continuity checking to look for solder bridges, with no luck. I still have to do some more logic testing to see what the heck is going on. Here’s the cleaner board (I used a flash so it would be clearer).
>
> Thanks!
> Hans

> <IMG_5778.jpg>
>

[Ronny Ribeiro]

Hi Hans,

Some usual things to test: voltages and currents (to see if any of the components is shorted; you can also remove socketed components to rationalize this); swapped transistors (NPN by PNP and vice versa; I remember that there were different types in the circuit); if you have an oscilloscope, check for clock signals (pin 38 of the 6502 must have a signal).
I see from your picture it seems you messed with the rom (did you?). Be sure that it is working and properly burnt. If you are not sure, get the original (http://pal.aibs.ws/assets/PAL1_ROM_v1.0.bin) and burn it on an erased, blank tested EPROM chip (this last step is important, otherwise you may end up with garbage on the chip).
Good luck!

Ronny

[Magnus Olsson]

This is a bit of a tangent, especially since it didn't help in this case, but is there any practical reason for cleaning off the flux residue, except making it easier to see bad solder joints? I hear conflicting opinions on this: some say that cleaning off the flux residue is absolutely essential to ensure that the board works, while others don't seem to care. Apparently, at least some sorts of flux are corrosive and can damage the connections over time, and it is also claimed that it can attract moisture.

It obviously depends on which type of flux you're using, but I'm talking about the resin included in ordinary lead-free solder (the one I'm using leaves a clear, glass-like residue, not the brownish one that I remember from the bad old days of lead-based solder).

[GN L]

Thanks Ronny’s constructive instructions. But I just want to say the download link is for the ROM expansion, the right download link for the 27C64 on PAL-1 is http://pal.aibs.ws/assets/kim.bin

> --
> You received this message because you are subscribed to the Google Groups "PAL 6502 computer" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to pal6502+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/b7db8f47-81c1-43a7-b20b-f3b8d1586f40n%40googlegroups.com.

> For more options, visit https://groups.google.com/d/optout.

Best,
Liu

[Ronny Ribeiro]

Oh, I'm sorry for that mistake! I simply forgot about the other ROM image XD

[Hans Van Slooten]

Per the suggestions, I did the following:

1) Tested that my switches were correct as described by Liu. They were, but to be extra sure, I desoldered mine and soldered the stock buttons. No difference.

2) Check the resistance of all resistors. They were correct. 

3) Checked for even more solder bridges, nothing found.

4) Tried swapping in a 65C02 chip to see if my processor was bad. No difference (not even sure that model would work in this board, but I'm getting desperate).

5) Checked the clock signal, that is correct. 

6) I can even get one of the displays to show an "8" by shorting pin 7 on U1 to GND. So, I'm sure the LEDs are working.

7) I did take the sticker off the EPROM before, but that was so I could get the chip version for reading. I wanted to get an image of the chip to compare to other ROM image to make sure it wasn't corrupted. It was fine.

Looking at the different signals on the board, there is clearly activity, it just, for whatever reason, isn't the right activity. I have to trace how exactly a digit is sent to the display so that I can try to trace where in the process it is getting thrown off.

Thanks,

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/51f19e02-674c-4ccc-96a3-860758e2b354n%40googlegroups.com.

[GN L]

The old R65C02P can work on the board except for the SST function and some activities need using interrupts, you can light up the display and do basic operation like key in a program, terminal operation even play the lunar lander with R65C02P. But the newly produced WDC W65C02S cannot work with this board.

Thanks,
Liu

> On May 20, 2022, at 2:12 PM, Hans Van Slooten <vans...@gmail.com> wrote:
>
> Per the suggestions, I did the following:
>
> 1) Tested that my switches were correct as described by Liu. They were, but to be extra sure, I desoldered mine and soldered the stock buttons. No difference.
> 2) Check the resistance of all resistors. They were correct.
> 3) Checked for even more solder bridges, nothing found.
> 4) Tried swapping in a 65C02 chip to see if my processor was bad. No difference (not even sure that model would work in this board, but I'm getting desperate).
> 5) Checked the clock signal, that is correct.

[GN L]

Are you checked the resistor network under the processor is in the right direction? The white dot needs on the right (the print on the resistor network needs face to the IC socket), it’s seems wrong direction from your photo (very hard to identify because of the photo angle)

> On May 20, 2022, at 2:12 PM, Hans Van Slooten <vans...@gmail.com> wrote:
>
> Per the suggestions, I did the following:
>
> 1) Tested that my switches were correct as described by Liu. They were, but to be extra sure, I desoldered mine and soldered the stock buttons. No difference.
> 2) Check the resistance of all resistors. They were correct.
> 3) Checked for even more solder bridges, nothing found.
> 4) Tried swapping in a 65C02 chip to see if my processor was bad. No difference (not even sure that model would work in this board, but I'm getting desperate).
> 5) Checked the clock signal, that is correct.

> To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/CANf3X5X-73gSNbf%2BMmfX%2BHCM3scm%2BL5ATAmMRzy4W2cTJiABrg%40mail.gmail.com.

[Hans Van Slooten]

Yep, that's fine. 

[Magnus Olsson]

The LED segments are connected in a matrix with segments being selected by the outputs of U13 (pins 2, 4, 6, 8, 10 and 12) except for segment D which is connected to pin 5 on U2, 

while the digits are selected by pins 5-11 of U1. Did you check that you get signals on those pins?

[Magnus Olsson]

Correction: pin 8 on U1 is GND so the digits are selected by pins 5-7 and 9-11.

[Hans Van Slooten]

So, minor update. I hooked up a logic analyzer to the 6502 address pins so that I could see what the processor was doing. It looks like it's getting stuck in a loop in this code:

I've attached the logic analyzer output in CSV format. Not sure why it would loop there yet.

Here's what it looks like visually after reset:

Thanks,

Hans

[Hendrik-Jan Megens]

Hi Hans, 

great work on finding where your PAL-1 gets stuck. Which logic analyzer did you use for this?

I'm not sure if I can be much help - what strikes me is that there is a reference to two memory addresses of the RIOT there. You mentioned you tested both of them, but did you swap them out?

I know this is very frustrating for you, but I'm sure you will solve it! If it means anything I'm following this thread closely and learning from every step. So thanks for posting!!!

Hendrik-Jan

[Magnus Olsson]

OK, that code is listening to the RS232 input. You should only end up there if the keyboard/TTY jumper is closed.

If you follow the program flow from a reset, you can see that the state of this jumper is checked by the BIT instruction at 1C31. If the jumper is open, you skip this code and go to START instead, which interacts with the keyboard and display. 

So if the jumper is open and you end up where you are, you have a hardware problem either with the jumper or the circuits connected to it. 

If the jumper is closed and you are connected to an RS-232 terminal, then there seems to be something wrong with your RS-232 interface.

This is my reading of the code, anyway. I may have missed something.

On Friday, May 20, 2022 at 5:39:10 PM UTC+2 vans...@gmail.com wrote:

[Jim McClanahan]

You're spot on. That code is for initialization of the RS-232 timing constant. So that would seem to indicate 1) the jumper to enable serial communication is shorted, 2) the RIOT is bad, or 3) U1 is bad. If U1 is bad, nothing "stand alone" is going to work--the serial detection jumper, the keyboard, and the multiplexed LED driver all go through that. Since you seem to be dumping to the serial routine, I'd hook a terminal up and see if you can get to the monitor prompt. If so, that would narrow it down to the LED & Keyboard / Serial Port selection and probably mean U1 is bad if there are not problems with solder bridges or cold solder joints.

Thanks,

Jim W4JBM

[Hans Van Slooten]

PA0 is held high even before reset, and the jumper isn't shorted. When
I short the jumper, PA0 is held low. I believe that is the port that
it looks at for determining IO source.

I wasn't able to open a serial connection with the jumper in either
position, although introducing the USB-to-Serial and null modem cable
into the mix also increases the chances I made a mistake.

Hendrik-Jan: For logic analyzer, I just use Salaea Logic 1.2 with a
cheap 16-channel logic analyzer from eBay. I think you can get them
for like $50 US (I got mine a few years ago, but they haven't really
changed.) I think Salaea is changing their software with version 2.0
to stop these kinds of knock offs. I would happily buy one from them,
but at $1400, it's a bit much for a hobby. I'll be using version 1.2
for a while, I think.

> --
> You received this message because you are subscribed to a topic in the Google Groups "PAL 6502 computer" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/pal6502/KPfvKGdgOcQ/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to pal6502+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/084adc51-811b-4450-b15b-6e0388c40a38n%40googlegroups.com.

[Jim McClanahan]

I'm going to have to think on that a bit. U1 should be HI on pin 4 until the ROM sets PB4, 3, 2, and 1 to 0011. That pulls it low and then the RIOT would check PA0.

That kind of makes U1 suspicious. But if the serial link really doesn't work, that would point to the RIOT.

I'd probably also do some continuity checks. A bad via might cause something like this.

The good news is the CPU, RAM and ROM all seem to be working.

Thanks,

Jim W4JBM

[Hans Van Slooten]

I’ve tested all the logic chips in my TL866 II+, but that doesn’t necessarily mean they are good. I also tested using my second RIOT from the daughterboard, with no luck, so I’m still leaning towards my dodgy soldering skills. I’m going to do a complete once over with continuity checks just to confirm I didn't miss anything.

Thanks!

Hans

[Hans Van Slooten]

Well, I continuity checked everything, scoured the board for any possible bridges and I’m pretty sure those aren’t the issue. 

Since I was able to test the logic/sram/rom ICs and they came back good, and the processor seems to be doing its thing properly, I have to assume the RIOT is the issue. I should say RIOTs, because I have two and it doesn’t work with either of them (which is the one thing keeping me from saying I’m sure that’s the issue). At the very least, I’m going to get some 74LS145s to see if replacing those helps, but I’m also going to order some 6532s as well.

Thanks,

Hans

[GN L]

Very hard check!

But that’s weird, I checked every recycled vintage chip (6502 and 6532) when I received them from the seller and only kept the good ones. I got one report before that a 6532 went wrong after a few days of use, because these chips are old, the reliability may be at risk. But it doesn’t work from the first time and two 6532s are all failed, the possibility is very low.

On the other hand, from your testing process and phenomena, it is most likely that there is a problem with the 6532. I checked your order information, the purchase is within 90 days, so I will send you two double checked 6532s for the replacement hopefully that will help you to figure out what’s wrong.

Sorry for the bad experience.

Best,
Liu

> On May 22, 2022, at 2:51 PM, Hans Van Slooten <vans...@gmail.com> wrote:
>
> Well, I continuity checked everything, scoured the board for any possible bridges and I’m pretty sure those aren’t the issue.

> <IMG_5802.jpg>

[GN L]

And the TL866 II+ can also check the 74LS145.

Btw, the 74LS145 with the kit are brand new chips.

Best,
Liu

[Hendrik-Jan Megens]

Hans, are you by any chance living in The Netherlands? (wild guess re your name). 

I have spare parts I might be able to send you quickly. I can send you one of my RIOTs to try it out, if you DM me your address. Catch: I'm away for work all week so that would have to be before 16:00 this Sunday (or it will be sent next wekend but might still be faster than Liu sending you spares, esp. with current logistics challenges in China).

I was wondering - can you be sure the RAM is not compromised? Looking at the startup sequence you seem to be hanging just after a return from subroutine which means use of stackpointer and therefore RAM. If the RAM is faulty all sorts of weird things can happen I guess? I don't know by heart I have to look it up but I think the RAM can be checked with the TL866II+ 

Btw, for the most awesome chiptester, this is very sweet: 

https://8bit-museum.de/sonstiges/hardware-projekte/hardware-projekte-chip-tester-english/

Not being too experienced that was a difficult build (as was, btw, my PAL-1, I was even less experienced then!). But it is insanely useful for testing chips. Unfortunately it doesn't test larger chips like the 6502 etc. 

Let me know if sending you a spare is useful. Whatever happens, I *almost* feel jealous because by the time you have the PAL-1 working you'll be an absolute expert on it's workings!!!

Cheers,

Hendrik-Jan

[Hans Van Slooten]

Hendrik-Jan,

I actually live in the U.S. My great-grandfather moved to the U.S. and
my parents gave me a relatively Dutch first name, but that's about the
extent of my "Dutchness". :)

I've ordered a pack of RIOTs from eBay but, yeah, it'll probably be a
month before they arrive. I have a few logic chips coming from Mouser
that I'll be able to test this week, though. I have plenty of
projects, so this one might just have to sit on the shelf for a bit. I
would like to get that chip tester eventually. It looks pretty useful
for retrocomputing.

I've actually built a 5150 clone board, C64 clone board, PiDP-8, and
PiDP-11 before with surprisingly few issues. I was overdue for a build
issue. I've probably learned more from this little board than I did
from any of those other soldering projects.

Thanks for the offer, though.

Hans

> --

> You received this message because you are subscribed to the Google Groups "PAL 6502 computer" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to pal6502+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/abb0bc61-f23f-4b70-b80a-375c755316e9n%40googlegroups.com.

[Hans Otten]

Hope the problem can be found!

Two bad 6532's is possible but unlikely., as they were tested before. If the first one had a short circuit on an output  due to a solder bridge/PCB error/other logic IC error it may have died The second one then shared the ill fate 
So make sure all pins are not short circuited! before you insert another one.  

So check all pins of the RRIOT for short circuits to ground/power/other outputs of other IC's or not connected.  

[Hans Van Slooten]

That’s a good suggestion, thanks!

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/c167d364-58e5-4ad9-989c-841c8e03808cn%40googlegroups.com.

[Michael Doornbos]

Interesting, mine went to get her with zero hitches whatsoever. I do have two spare 6532s but I don’t actually need for anything at the moment. 

--
Michael Doornbos
Imapenguin.com

Evadot.com
michaeldoornbos.com

myzenith750.com <—- yes I’m building an airplane in my garage (hangar)
828-450-7418

HAM: N4LNX 

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/c167d364-58e5-4ad9-989c-841c8e03808cn%40googlegroups.com.

[John Kennedy]

I don't know where you are based, but if you are in the US and want someone to compare your build with a known working one, I am happy to help.

I can put them side-by-side and test with my 'scope.

[Hans Van Slooten]

I’m in the Twin Cities in Minnesota. Thanks for the offer, I’ll let you know if my current troubleshooting efforts fail.

Thanks,

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/51ee58dc-381a-4c5d-b369-276d5abedcb4n%40googlegroups.com.

[Hans Van Slooten]

As an update (nothing particularly interesting), I tried some different 74LS145s, with no luck (unsurprisingly). I did sone more end-to-end tests PA0 of the RIOT, and it definitely looks like it’s being held high. Here’s the logic diagram with lines annotated. The logic changes are when I press RS and then release it. 

I’m still waiting on the RIOT replacement to test, but I thought I’d share this with you all if it helps troubleshoot (I’m still not 100% convinced it’s the RIOT since two bad one’s just seem unlikely unless I hit them with static electricity or something). I also checked that no pins were bridged to VCC.

Thanks,

Hans

You received this message because you are subscribed to a topic in the Google Groups "PAL 6502 computer" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pal6502/KPfvKGdgOcQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pal6502+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/006b8b19-4ac9-4ced-acdb-c0a3c476f6fb%40Spark.

[Hans Van Slooten]

Here’s a full capture of U1 from start through reset (pressed three times) It seems odd to me that Q0-Q3 are all low and the rest are all high regardless of the inputs, but none of those are connected by default (JP1 is open as are all the switches connected to Q0-Q4) so maybe that’s expected (I admit to not being a hardware expert). I’ll note that I know that the LS145 outputs high for “off”.

 

Thanks,

Hans

[Will Senn]

Hi Hans,

I got nothing concrete by way of help, at this point, but I did want to say that you inspired me to figure out how to use the el-cheapo logic analyzer I have laying around. I got it working and recognized with Pulseview, so I can do tests on my working PAL-1.

I don't know if you have the patience for it, but if you tell me what to hook up where (as in, hook ch1 to pin XX of the 6502 or whatever) and what to press when (as in press RS three times after everything's hooked up) along with any settings you set (as in rising edge, falling edge, whatever), and what to look for, I might be able to tell you how mine acts - be it the same or be it different. My analyzer has 8 channels.

Keep in mind that I'm a quick study, but I know next to nothing, so you need to be explicit...

Later,

Will

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/CANf3X5WAFooF%2BgCa-xkaynM5D_JuN21njZJGZcGEWbdQvCdf4g%40mail.gmail.com.

[Hans Van Slooten]

So, this morning I decided just to see what PB0-7 and PA0-7 on the RIOT are. Theoretically, the RIOT sets all of those to inputs as, according to the datasheet:

"During system initialization a logic "0" on the RES_bar input will cause a zeroing of all four I/O registers. This in turn will cause all I/O buses to act as inputs thus protecting external components from possible damage and erroneous data while the system is being configured under software control. The Data Bus Buffers are put into an OFF-STATE during Reset."

So, it seems everything should be zeroed out, then in the code, DDRA is set to #$00 (which shouldn't change anything. PA should remain inputs) and DDRB is set to #$3F, so PB0-5 should be HIGH (outputs) and PB6-7 should be LOW (inputs). The output registers should still be 0, so that shouldn't change any data lines. Then, Output Reg B is set to #$07 which should set PB7-0 to LLLLLHHH.

With all of that in mind, I tested both of my current RIOTs. If they are both bad (already unlikely), the likelihood that they are bad in the _same way_ is exceedingly unlikely. So, I put the logic probe on the PA and PB lines and tested both:

EXACTLY THE SAME (The first edge drop is me lifting my finger off the RESET). PA is always $7F after reset and PB is $40 then transitions to $47. If I assume that input lines behave differently and PB6-7 can be ignored, then you could squint and say that PB out is $07, which is what you would expect. And then, if you are ignoring PB6-7, you could ignore PA0-7 since they are all inputs. The datasheet says:

"When in the input mode, the peripheral output buffers are in the "1" state and pull-up device acts as less than one TTL load to the peripheral data lines". 

Does that mean that for input mode you should assume that those pins will be high? My hardware knowledge is reaching its limits here.

Unfortunately, all of this is making me think that the RIOT isn't the cause. They both seem to be responding the same (at least on their data ports) which isn't what you'd expect for bad chips.

If you made it this far, thanks for listening to my ramblings. I think I need to track this further back than the RIOT (or at least to the inputs rather than the outputs of the RIOT). This would be a lot easier with more than 16 lines at a time for testing. :)

Hans

[Hans Otten]

The RIOT PA PB etc are after reset set by the KIM-1 monitor, and look what they should be.

It looks indeed the RIOTs are both fine. 

I suggest you read the monitor listing and compare with what you are measuring.

[Hans Van Slooten]

Thanks, Hans. Yeah, I did that a bit before (not sure if you saw that email) and the consensus was that it was dropping into the serial code when it shouldn't be. Based on the logic analyzer output, I believe it was getting stuck in this code:

If you have an insight into why the bit test on SAD would fail (when the jumper is open and tested to not be bridged) I would appreciate it. This is baffling to me (admittedly not a difficult task).

Thanks,

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/80906ffc-c8ca-4e53-8561-82e20bc352f7n%40googlegroups.com.

[Hans Van Slooten]

Ok, interesting update: I was looking at Hans’s simulator and noticed that it jumped at the exact instruction where mine was getting confused (I assume this is because it doesn’t have a “real” serial port). This gave me the idea to try it myself. I also happened to have some AT28C64B’s around, so I thought I’d try using an EEPROM as it has a faster development turn-around time. Well, I swapped out those BIT tests with JMPs, et voila:

(That’s “0802 DF” since the camera has trouble picking up the LEDs.)

Of course, it’s still not working. Specifically, none of the keyboard commands work except for ST and RS which are directly wired to processor pins.

So, that’s fun.

Oh, the EEPROM works, for what it's worth.

Hans

[John Kennedy]

I can't believe a pair of Hans can't fix this. I thought many Hans make light work?

[Neil Andretti]

Let's suppose the 6502 and the 6532's are both working. And the IC's are also tested (U1 is sus).

What's left? The RAM? (I have pleanty of that, I can send you for checking)

Could it be a bad transistor?

What other parts could possibly fail?

[Hans Otten]

Can i still post here?

[Hans Otten]

All messages I enterred were deleted immediately, sorry for testing,

Check all inputs wit a scope, It could be mismatch between TTL IC and NMOS RIOT thresholds.

On Wednesday, June 8, 2022 at 8:19:05 AM UTC+2 master...@googlemail.com wrote:

[Hans Van Slooten]

I saw all your posts, thanks for the detailed troubleshooting suggestions. I’ll try that the next time I’m debugging (might be tomorrow)

Thanks again,

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/75f0c363-ab7e-4719-80f2-5730cc84a5ban%40googlegroups.com.

[Hans Otten]

Patching the ROM is an excellent idea! Indeed, I patch the ROM to replace the bit banging serial port (In fact I replaced it with a 6850 ACIA emulation!) 

 

Since the LEDs now light, means the SCAND routine succeeds in outputting to a working 74LS145. All transistors are fine, LEDs are fine.
What is not working is reading the keyboard or serial, which means the inputs of the RIOT are not seeing inputs.

And the same goes for the TTYin detection at PA7?

So it could be a mismatch between the 74xx IC's and the RIOTs at the analog level.  

Like I told before, Bob Corsham had this problem in his KIM clone with 74LS145 and replacing with 74HCT145 fixed it)

Pairing NMOS with TTL can be tricky,  are the thresholds of 0 and 1  reached?  

 

It would be interesting to see measurements of the ANALOG voltages (not LOGIC!) with a scope of:

- power supply voltage at the IC's itself (especially U11 and U1 and U7

- outputs O0 -O3 of the 145 (pulses)

- PA0 when TTY jumper installed (puls)

- PA0 - PA6 when pressing a key on the keypad (note that the quick scanning means it will be pulses)  

- PA7 when serial input is arriving (that is pin 11 of the 74LS00 U11) 

[Hans Van Slooten]

Hi everyone, I haven’t had time to test the voltages yet, but I did receive a batch of 6532s I ordered off eBay and IT’S ALIVE! 

I still want to dig into why these 6532s didnt work. Since BOTH of them didn’t work, I wonder if it’s voltage-levels issue and not a “bad” chip issue. These are R6532AP’s rather than the UM6532’s. Liu has sent replacements that I will test as well when they arrive (besides a great kit, he gives great customer service, so none of this should be read as a knock on him).

Woohoo! Can really get going now. Thanks for listening to my ramblings through this. I will follow up when I’ve tested the other chips.

Hans

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/40b26b04-1b8d-41e0-87d8-4d37e37ee138n%40googlegroups.com.

[Hans Van Slooten]

Oh, I wouldn't be surprised if they were rebadged. Most chips from eBay are pretty dodgy, but you take what you can get when you are dealing with discontinued chips. :)

Hans

On Wed, Jun 8, 2022 at 8:01 PM Ronny Ribeiro <ronny...@gmail.com> wrote:

That's fantastic news, Hans!

I'm very happy for you but I'm also sad that the adventure is over XD
These Rockwell 6532's you've bought seem rebadged... Anyway, don't do the acetone test to discover they are good old UM6532's!
Best Regards From Brazil,

Ronny

PS: It seems my posts are being deleted, like Hans Otten's...

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/6f65f7f7-654c-481e-91f4-0451bfc7f7cdn%40googlegroups.com.

[Hans Otten]

Congratulations! Good to see another PAL-1 alive!

So after all it was the 6532. Perhaps both functional 6532s, but not in this PAL-1.

Quote from Bob Applegate about his problems with 6532 in a Corsham KIM clone, in groups.io CorshamTech:

 "I have never found a 6532P that works in the KIM Clone, while the AP version works perfectly! The P parts won’t even let the KIM Clone come up.

According to the Rockwell and Commodore spec sheets the AP indicates 2 MHz clock in a plastic package, while the P indicates 1 MHz in a plastic package.

Given that the clock speed on a KIM Clone is 1 MHz it certainly seems either device will work, but only the AP does."

Liu, did you test the 6532 you use in the kits in a PAL-1? 

If yes, it is perhaps my theory of mix of some LS types brands and some 6532 voltage level thresholds. You did try different brands of 74LS145 and the other 74XX IC's?

Since the success rate of build of PAL- kits seems high, it is an uncommon problem.

[GN Liu]

Congratulations!!! Waiting for the double checked UM version to be test!

Hans, I test all the main chips when I received it, one by one. And all the 74 series with the later kit (I think about after the 30th kit I sold) are brand new and original, except the 74LS38, the 7438 is NOS (can't find brand new seller...), the brands are TI, Renesas and NS (7438 only). 

The test board I used is mixing LS brands, and changed over time, I can't remember the brands exactly... I think next time I purchase 65xx, I will document the testing board configuration here ;) I Have no idea what cause this happened.

[Allen Evenson]

I've been following this discussion with great interest. My PAL-1 is currently in LA enroute to my meager lab (and Ham shack) in Kent, WA.

I do not have a scope or logic analyzer at this time... just a multi-meter. Should I be on the lookout for test gear?

Hoping for a good build!

Cheers,

Allen E.

On Fri, Jun 10, 2022 at 5:31 PM Hendrik-Jan Megens <hjme...@gmail.com> wrote:

Congrats, Hans, on getting the board up and running!

As Ronny mentioned earlier, it is almost sad that the quest is over, I've been following this quite closely and learned a few things. Your meticulous analysis of where the board gets stuck also made me realize how valuable scoping out an ailing board can be, especially if you know how to translate that back to the ROM code. Since most of the time I only half-know what I'm doing 16 channels will not be enough for me so decide to make the plunge and bought myself a (relatively cheap) 32-channel logic analyzer. I think I'm going to dump the entire startup sequence of the PAL-1 somewhere, in time, when I figure out what I'm doing, which may be useful for others that are troubleshooting.

So, again, thanks for posting! And I'm looking forward to any further insights into this potential conflict of different generations of logic.

Cheers,

Hendrik-Jan

To view this discussion on the web visit https://groups.google.com/d/msgid/pal6502/cc0f626c-b3dc-4aad-b78d-40cf033c96abn%40googlegroups.com.

[Will Senn]

Hi Allen,

Since no one else chimed in, I thought I'd give you a newb perspective... I'm hardly an expert, but I really don't think you "need" a logic analyzer or a scope, but you probably do need a Volt Meter to check continuity, voltage levels and such. When you solder this baby up, there's quite a few connections you can flub up. I didn't have any solder issues, but I did plug 2 of the IC's in backwards - try not to do that :)

That said, I already had the scope from a while back when I was thinking I'd get into electronics and had some spare cash (a bonus). It has been in a box in storage for the last few years, gathering dust. A few weeks back, I started reading the Micro Cookbook v1 and v2 by Lancaster. In the books, he is always using a scope in his explanations and experiments, so when I got the PAL and started doing experiments, I tried the scope stuff. Seriously, after doing a few successful micro experiments, I'm hooked.

The logic analyzer is gravy, that, you should get for fun, cuz they're cheap (mine was less than 10 bucks, but now it's like 13) and they work great. Lancaster also talks about using a Voltmeter (analog) as a poor man's scope - which I'm trying along the way and it actually does give you interesting and useful information on what the micro is doing. Jury's out on whether it's all you need.

I'm a programmer by trade and a CIS prof by education, so this is brand new territory for me, but I am having serious fun with it.

Will