memory expansion won't....

[eslapion]

Also, as far as the discussion is concerned here, the information disclosed by dimebag had been very sparse indeed. Not much to infer from it other than 'it doesn't work'. No mention of what programs exactly were used, how the errorneous (screen) output looked like, etc.

A standard test scenario as given by my program would have cleared things up a lot for me.

Same thing here.

There were a couple of boards with manufacturing errors in that specific batch. I suspect that's the issue.

In that case you're zeroing in on the error anyway. If the culprit has been identified, please be so kind to inform the people here, so they get a closure upon this.

I can't say anything for sure here.

I did have one specific board in this batch with a short circuit between 2 lines and this was fixed.

I did test dimabag's RAM expansion by loading/executing programs with it before shipping.

I can only suggest to dimebag he runs your software (or that he ships back the expander so I do it).

[dimebag]

Hi Mike & François

i have done the test from Mike

the first check loop have a weird behiour....... full of errors, stop anywhere, continu with a cont at weird loop count....

i have done a video.....



Cheers

Laurent

[Mike]

For this test (and documentation) to be of any significance, it would be necessary you'd show:

- this program was typed in into a freshly powered on VIC-20, and
- written without any type-in errors.

Any kind of type-in errors or something else done before you entered the program is likely going to not work as intended.


The test program I gave eslapion which he then has relayed to you specifically writes a stream of RND() numbers into the two areas 8192..32767 and 40960..49151, and then checks whether that exact sequence can be read back by restarting the RND() generator. Here's the program:

Code: Select all

10 X=RND(-2):CLR
20 FORT=8192TO32767:POKET,INT(256*RND(1)):NEXT
30 FORT=40960TO49151:POKET,INT(256*RND(1)):NEXT
40 X=RND(-2):CLR
50 FORT=8192TO32767:IFPEEK(T)<>INT(256*RND(1))THENPRINT"ERROR AT:";T:STOP
60 NEXT:FORT=40960TO49151:IFPEEK(T)<>INT(256*RND(1))THENPRINT"ERROR AT:";T:STOP
70 NEXT

On the start-up prompt, first protect BASIC memory by typing in

Code: Select all

POKE55,0:POKE56,32:CLR

Then, type in the program EXACTLY AS SHOWN, and start it with RUN. When the program ends with just a "READY." prompt, all is O.K. thus far. Otherwise, you will be prompted with a *single* error address (not a list!), which would give a first hint something's wrong with the RAM.

[eslapion]

The test program I gave eslapion which he then has relayed to you specifically writes a stream of RND() numbers into the two areas 8192..32767 and 40960..49151, and then checks whether that exact sequence can be read back by restarting the RND() generator. Here's the program:

If you carefully look at the beginning of the video you see 'erreur' messages for addresses 1050 to 1060 which is a definite proof Mike's program is used improperly by dimebag.

The generated RND() numbers must be generated exactly the right way to match read and writes.

On the start-up prompt, first protect BASIC memory...

This can also be done by disabling the expander at power-up then enabling it once the BASIC V2 message is on the screen but he doesn't seem to understand the procedure or even the necessity to do so.

[dimebag]

Hi

i'll do another try this evening

Cheers

Laurent

[dimebag]

Hi

This is the video of the results of the lastest test code.



Cheers

Laurent

[Mike]

Dimebag,

the program either completes without any error message, and just a "READY." prompt - or it prints out a *single* address, NOT A LIST!

Once again:

Type in the program AS SHOWN, with no alterations whatsoever! And show the program listing in the video before you execute the program, so eslapion and I actually can verify that you typed in the program correctly.

[dimebag]

Hello

i have done the test, thes are the picture of the listing and of the results





Cheers

Laurent

[Mike]

Following observations:

1. Why the ?SYNTAX error in line 30? The statement in the upper picture looks fine. FYI, it is generally not possible to CONTinue a program after an error, or when a line has been changed after an END or STOP command or after pressing the STOP key. Even just going with the cursor to a program line and pressing [RETURN] "over" it disables the CONT command.

2. Line 60 is *missing* the NEXT command at the start. Note this omission would once again have invalidated this attempt, if it weren't for the fact that the test was stopped dead in its tracks right when reading back the first byte at 8192!

What you didn't include in the photos was the start up banner with the BYTES FREE value. For now I assume, that you actually activated all RAM in BLK1..3 and BLK5 and got 28159 BYTES FREE on start up.


For the result:

As you see, the byte value in address 8192 did not "survive" the subsequent write actions to 8193..32767 and 40960..49151. This implies cross-talk *is* happening here. I'll follow up with another test program later this day.

[dimebag]

Hi

This is the latest try. i have exactly typed the program like it does.

i have an error at the first test location: adresse 8192

These are two pictures:





Cheers

LAurent

[eslapion]

@dimebag
Please disable BLK2 and BLK 3 and run the test.

This should give you 11775 bytes free at power-up.

Change line 20 and 50 to:

Code: Select all

20 FORT=8192TO16383:POKET,INT(256*RND(1)):NEXT
50 FORT=8192TO16383:IFPEEK(T)<>INT(256*RND(1))THENPRINT"ERROR AT:";T:STOP

[Mike]

Change line 50 to:

Code: Select all

50 FORT=8192TO16383:IFPEEK(T)<>INT(256*RND(1))THENPRINT"ERROR AT:";T:STOP

... this also requires a correspondent change in line 20 so that one reads:

Code: Select all

20 FORT=8192TO16383:POKET,INT(256*RND(1)):NEXT

Anyhow, I've prepared a new test program, but first have asked another Denial fellow to beta-test it.

[eslapion]

@Mike
Yeah, didn't have the time to post the edit!! Was about to.

... done!

[Mike]

No worries!

[dimebag]

Hello

i have done the test with the right setting, 11775 bytes free, BLK 2 & 3 disabled

i have the same error at 8192.



same behaviour with my other motherboard



Cheers

Laurent