FYI:

I'm in the process of making a device that won't require the use of any ROMs at all.. so you won't need a chip programmer either. It will also support running 3rd party ROMs too, including the original game ROM so you can leave it in the machine if you so desire. It's still not ready to be manufactured but the prototype works like a charm!

At the moment it will support the Williams 3-7 games, but it should be compatible in the future with other games as well. More details as they surface

Interesting. I just tested my system 6 board and get 8.36ms on the 02-blanking test, and almost exactly that with the GREEN ROM.. What game/board are you troubleshooting?

For what it's worth, the blanking test does nothing but loop until it's time to toggle the input pin (pin 5) of IC7. then it uses an EORA to invert the correct bit in the PIA1_A register.

If your clock speed checks out then it is peculiar that the bit in the PIA register would show signs of toggle so much faster than the CPU even asks.. I can't say I've ever seen that before. Please let me know what you find

Nice to see a variation on the chip used, and that you have made a board that splits things by 4K chunks

My system 6 is giving me 3.2hz on the LED's test.

The sys346 02-blanking ROM in the latest beta download (pincoder_roms_2019.10.29.1803_beta) can be modified at the locations shown in blue.. Currently the value is 01D0 (msb to lsb) and the value is simply a 16 bit "sleep" counter number used in the X register for creating a simple delay. At this stage of the test ROMS I don't want to rely on the external interrupt for calculating a proper delay as it requires the proper functioning of additional circuitry (the IRQ circuitry), which gets tested in later ROMs anyway.

So I will focus on getting a suitable counter value for system3-4. However, I don't have any 3-4 boards to test with.

On a system 6 board counting down to zero from 01D0 (hexadecimal) yeilds a toggle period of 8.3ms on IC7 pin 5. This is what the system 6 Williams ROMs are also refreshing at.

@beaver, if you have time would you mind playing around with the counter value in the beta image to see where it best matches the 8.3ms on the system 3 board? I will then test that number on my system 6 and see if it's still within reason. I may have to create a different binary for systems 3-4.

Thanks
02-blanking (resized).png

I should have thought of this sooner, but thanks to some digging by beaver I remembered that the 02-blanking test in the current release version (2019.04.07.1941) does not have a delay built in for the blanking signal refresh. Apparently this causes system 3-4 boards to show incorrect results when running this ROM.

The latest beta download (2019.10.29.1803_beta) does have the correct delay loop, and therefore correct results with the 02-blanking on a system 3-4 board.

Sorry for the confusion folks!

Merry Christmas everyone. Hope you all had a good one!

Check out the latest release version of pincoder ROMs: http://pincoder.ca/ccn/roms/pincoder_roms_2019.12.26.2048.zip

Changes since the previous release:

* Slowed pulse rate (PIA1 Pin 4 - PA2) on 02-blanking from 2500Hz down to 120Hz to match williams ROMs refresh rate.

* Eliminated ghosting from neighboring segments on displays.

* Added 11b-sounds test to allow sound card testing on the bench.

* Modified 10-solenoids to also fire a solenoid using the DIP_01 switch together with the COMMAND ENTER button on the MPU board. Some games do not have a HIGH SCORE RESET button.

* Temporarily removed sys7/2K images.

* Rewrote the documentation for "01b-bus". The ROM itself remains unchanged.

Enjoy!

Excellent research! This ROM has been driving me a little nuts as I don't have any system 3 or 4 machines to test with.

According to beaver the latest ROMs (2019.12.26.2048) work in his system 3, but I've also had someone tell me that the same version didn't work in theirs.

I would very much like a copy of your sim files. Thank you. I'd like to come up with a ROM that will work for all of them as it would be disheartening to have to split the sys346 set out into their own binaries.

So yes please send me anything you think I will need to do some testing. I'll post an updated version as soon as I come up with something. It sounds like it'll be just a matter of changing the length of low side of the cycle but I'll need something that works with or without the voltage divider and still resembles the output of the Williams ROMs.

Thanks again for the good work and the detailed update! I'm looking forward to running some spice tests!

Oh and you're right, the blanking test isn't supposed to depend on RAM. I'll look into that. In the mean time please let me know how the 2019.12.26.2048 version works for you. I'm glad these ROMs helped you

Awesome! Nice Going! and thanks for letting me know they helped you! Did you also run through the rest of the tests? You might find switch problems, etc that could improve the game play (if it's unnoticably lacking).

Swapping the CPU's may have been enough to point a fault at the socket as well. You could always go back to the old chip and see if there's any change. At any rate, swapping the socket wouldn't be a bad idea, especially if you happen to have a proper heat gun to pull the old one out. If not you can always pull it using the old fashioned way

Glad they helped you and I'm always looking for feedback so its good to hear good things I've always wondered if those names actually resembled the same thing to everyone else. It's too bad about your friend.. he could have made them millions had they followed his advice!

Depends on whether you are tying them high or low using the resistors. I like to tie them low, and when a switch is closed, it goes high. This allows for setting the bank to the natural binary value.. ie. slot 0 is with all switches open (0000), and slot 15 is with the 4 least significant switches closed(1111).

I would double check the 8t28 jumpers, and also remove and jumper the other 8t28.

LEDs come on by default on a board that has no code to boot. They only go out once the code initializes PIA1.

Address zero is located either on CPU1 itself, or IC13 depending on what CPU chip you are using, and whether it is jumpered properly for the CPU you are using. See documentation for 04-ram-IC13 for more info..

While running 01b-bus, if you are seeing activity on the address and/or data bus lines continually, I suspect the CPU is continually rebooting. This is usually caused by a faulty reset circuit. Verify the reset signal as per the instructions in 01b-bus.

At this point, I don't think the transceivers test will do you any good since you are removing them. The test is meant to verify the input and output lines of the transceivers: Data out should equal data in.

The images for the test ROMs can be downloaded for free:

http://pincoder.ca

The images for the games ROMs can be found here:

http://www.planetarypinball.com/mm5/Williams/tech/roms.html

As the physical ROMs are not produced any more you will have to search the net. Ebay typically always has some, but they're not reliable ads - many are scams.

Another option is to order an interface board that uses a newer and bigger chip (that is still in production) and will allow you to use a programmer to write multiple images to the chip. You can order one here:

https://dirtypcbs.com/store/pcbs/buy/107605/2716to27256-bank-switch-zip

You can also contact @thewiz. He had some boards made up and may have some for sale.

I am currently working on a board that doesnt require a programmer or any chips, but it is not yet in production.

Hope this helps.

It was designed by @barakandl. He may have the actual BOM for that. Keep in mind that this board was designed a while back and the number of ROMs has since increased so it isn't big enough to hold all of the latest test ROMs. You might want to order two or even three of the 27256 chips and use a ZIF socket instead of a regular socket.

This problem is obviously going to come up again before my board is ready, so I'll make a bigger version of @barakandl's board this week and make it available as a short term and least expensive option.

I'll keep you posted

Hi everyone!

I've got an intermediate adapter board (the "Pincoder Adapter") available now that comes preloaded with these test ROMs. It isn't the "Pincoder Controller" I've been working hard to bring to market, but it is an intermediate solution that still gives you all of the Pincoder ROMs preprogrammed into a single adapter that supports 2K and 4K ROMs simultaneously AND it is available today!

It's especially great for those of you without a chip programmer and chips because it eliminates the need for those items so you can start troubleshooting right away!

For those of you WITH a programmer there's plenty of space for third party ROM images like the Andre and Leon test ROMs.

You can find all about the Pincoder Adapter on this page:

https://pincoder.ca/index.php/2020/07/10/adapter-2020-06-23-0506

I've also created another pinside thread for questions and answers related to the board itself:

https://pinside.com/pinball/forum/topic/new-pincoder-adapter#post-5743793

Thanks for helping me support the preservation of your classic pinball machines!

Yes I've cross posted to a few Facebook groups. Thanks for taking the time to add your same comments here. As I also mentioned there I have not seen any socket stretching as such this far, but you do make a good point

If I have some spare sockets available I'll include them with the board. They might not be dual leaf, but they'll work.

Good News! Online purchase of the Pincoder Adapter is now available everywhere (click and scroll down to "HOW TO GET ONE" here):

https://pincoder.ca/index.php/2020/07/10/adapter-2020-06-23-0506/

(I'll also post this at the top of this thread)

Thanks again for your support!

Hi everyone,

I've published a beta version of the latest pincoder ROMs:

https://pincoder.ca/ccn/roms/pincoder_roms_2020.12.04.1236_beta.zip

In it, the 09-lamps test has been given an added "delay" feature that may help in troubleshooting strange behavior with bulbs in the lamp matrix. As it clearly states in the 09-lamps.txt file, USE WITH CAUTION. This is because the test gives you the power to hold constant power to a column of bulbs for longer than intended in the original design of the Williams hardware and software.

Everything you need to know about the delay feature is documented in the 09-lamps.txt file.

Because this is a beta release, future orders of the Pincoder Adapter will be shipped with the current Release version installed (2020.06.23.0506) unless you specifically ask for this beta version by messaging me (@pincoder) before you order. For those who already have the adapter and would like an updated EEPROM chip sent to them you can also message me and I will ship you one for $10 CAD plus shipping.

If you have a chip programmer that can write an ST39SF040 chip you will find the updated adapter image (pincoder_adapter.bin) and the updated DIP settings file (pincoder_adapter.txt) in the beta download. Since this is only a change in content, and not a change in the number or order of tests included on the chip, the DIP settings on the adapter remain the same as the Release version already installed on the adapter.

As always, your feedback is welcome. I am not certain how useful this feature is as I have no suspect hardware to test it against. However, it would be nice to know how this test works out in the field

Also, created a poll at the top of this thread if you wouldn't mind ticking a few boxes Thanks!

Awesome! Thank you for your donation, and taking some time to give your feedback I appreciate it!

First off, I'm glad to hear the adapter is found to be useful! Thanks for the feedback!

It sounds exactly like you have issues with IC15, 16, 17, 18, and possibly IC11 (though I'm betting its somewhere in those first four). With 06-switches running, 2J2 and 2J3 disconnected you should get 00 in the match display and nothing anywhere else. Take a look at the numbers on the displays and map them to the switch matrix chart in your manual. See if any one entire row or column is shown. Chances are you'll see more than one row or column that will turn up. These rows and columns refer to the pins on 2J2 and 2J3.

The last section in 06-switches.txt ("ABOUT THE MATRIX SWITCHES (and PIA2)" will walk you through tracing signals so you can verify which ones actually need to be replaced, but in my experience IC15 and IC16 fail before IC16 and IC17 ever do. For the price, it's worth replacing all four anyway. Don't forget to install sockets first, it makes life easier.

Sometimes, when IC15 and IC16 fail they cause damage to the pins on IC11. If you replace IC15 through IC18 and the test still fails, you should replace IC11.

As for the lockup issue, that's a bit odd at this stage in the tests.. It may indicate a bad IC11 (or some other chip on the address/data bus), but the tests should continue to run even with a totally dead IC11. If you can properly remove IC11 without damaging it, you could try running the test with IC11 removed. In this case you should get "64" in the match display and numbers from 01 to 12 in all of the player displays - and it should still run forever.

Do any of the other tests crash if left to run for extended periods? You could try 01a-leds and see if the blink-rate ever "stutters".

Can I quote you on that? That's an excellent statement, thank you.

The IC numbers I'm referring to are on the driver board. IC17 on the driver board is a 7406.

You don't need a working driver board for these tests.. that's what the tests are designed for.. However, some people have had driver boards that lock up the MPU board. When that happens you can still run MPU board tests with the driver board disconnected, and then reconnect as you go down the list of tests.

Yes, please post some pics of the numbers. At first glance they look seemingly random, but are in fact sequences of two digit numbers starting from the left side of P1 working to the right.. to the right side of P4. if you look at them you will see they are actually sorted from lowest to highest. Each two digit number corresponds to the switch number in the switch matrix. Numbers higher than 64 refer to switches outside the matrix. Those numbers are assigned by me and relate to other switches, like the DIP switches on the MPU board for example.

When the Pincoder Controller comes out, you won't have to run back and forth to your PC to look at the documentation, so hang in there

As for system hangs, disconnect the driver board and run the displays test again to see if it hangs. If so, perhaps you also have a power problem. Check rectifiers and caps etc with respect to the power supply. It's worth mentionting here too that cold solder joints can also be a culprit here, as can the 40 pin connector so if you haven't already, check for both.

If it doesn't hang with the driver board removed then you likely have issues with either the 40-pin connector, and/or the circuits related to the PIAs on the driver board.