After a few hours of use, the field found A500 broke again! (+let's talk about RAM tests)
![Adrian's Digital Basement ][](http://yt3.ggpht.com/ytc/AIdro_k7CrQRe71qN6c0Ua7A1P_I8KkxjqLa70u0zYeCvdvnKw=s48-c-k-c0x00ffffff-no-rj)
ฝัง
- เผยแพร่เมื่อ 28 พ.ย. 2023
- Well that didn't take long! The Field Found A500 had barely been back together and working when it broke again, this time while I was using it!
So let's figure out what broke so it can be fixed, and let's talk about some of the ways RAM chips fail and how to detect those errors.
Field Found A500 series:
Part 1: • This Amiga 500 was lef...
Part 2: • The Field Found Amiga ...
Part 3: • From Death's doorstep ... (where I we work on an fix the external RAM card)
-- Video Links
Adrian's Digital Basement Merch store:
my-store-c82bd2-2.creator-spr...
Support the channel on Patreon:
/ adriansdigitalbasement
Adrian's Digital Basement (Main Channel)
/ @adriansdigitalbasement
-- Tools
Deoxit D5:
amzn.to/2VvOKy1
store.caig.com/s.nl/it.A/id.16...
O-Ring Pick Set: (I use these to lift chips off boards)
amzn.to/3a9x54J
Elenco Electronics LP-560 Logic Probe:
amzn.to/2VrT5lW
Hakko FR301 Desoldering Iron:
amzn.to/2ye6xC0
Rigol DS1054Z Four Channel Oscilloscope:
www.rigolna.com/products/digi...
Head Worn Magnifying Goggles / Dual Lens Flip-In Head Magnifier:
amzn.to/3adRbuy
TL866II Plus Chip Tester and EPROM programmer: (The MiniPro)
amzn.to/2wG4tlP
www.aliexpress.com/item/33000...
TS100 Soldering Iron:
amzn.to/2K36dJ5
www.ebay.com/itm/TS100-65W-MI...
EEVBlog 121GW Multimeter:
www.eevblog.com/product/121gw/
DSLogic Basic Logic Analyzer:
amzn.to/2RDSDQw
www.ebay.com/itm/USB-Logic-DS...
Magnetic Screw Holder:
amzn.to/3b8LOhG
www.harborfreight.com/4-inch-...
Universal ZIP sockets: (clones, used on my ZIF-64 test machine)
www.ebay.com/itm/14-16-18-20-...
RetroTink 2X Upconverter: (to hook up something like a C64 to HDMI)
www.retrotink.com/
Plato (Clone) Side Cutters: (order five)
www.ebay.com/itm/1-2-5-10PCS-...
Heat Sinks:
www.aliexpress.com/item/32537...
Little squeezy bottles: (available elsewhere too)
amzn.to/3b8LOOI
--- Links
My GitHub repository:
github.com/misterblack1?tab=r...
Commodore Computer Club / Vancouver, WA - Portland, OR - PDX Commodore Users Group
www.commodorecomputerclub.com/
--- Instructional videos
My video on damage-free chip removal:
• How to remove chips wi...
--- Music
Intro music and other tracks by:
Nathan Divino
@itsnathandivino - วิทยาศาสตร์และเทคโนโลยี
*_Thanks Adrian for showing another fight with this rusty Amiga._*
That stash of bad parts is STILL surprisingly SMALL considering that thing is a RUST BUCKET! I knew you'd fix it! YOU Da MAN!!
Totally agree! The MARCH test is probably the best one, because it scans for minute chip defects.
I've fixed many a computer that exhibited rare problems, all due to ram chips failing MARCH tests (PC Check 7 and 8 are great tools when I was using them).
Shadow of the Beast does actually use the RAM expansion. It preloads some sections of the game in the extra 512kb so it doesn't have to load them further hence decreasing number of loading times during the game. That's probably why it crashed first, because it was trying to access the extra ram.
Lemmings, which is by the same folks, does the same thing.
Shadow of the Beast was programmed by Paul Howarth, Lemmings by David Jones, Russell Kay, and Mike Dailly. The former was based in the midlands somewhere (i forget), the latter in Scotland up in Edinburgh. No overlap, just a common publisher.
MARCH test is one way to do it, but it is slow. A fast RAM test is to use a pseudorandom number generator seeded with a randomly chosen but known value. Seed the PRNG, write the output, reset it with the same seed, read back and compare. It may not catch the error on every pass but it should catch it eventually as long as the seed keep changing.
The "rug pattern" a lot of Williams/Midway arcade boards do at power on is the result of a PRNG test on screen RAM. They typically do three passes and consider it good.
I love seeing the NES controller on top of the Amiga! :)
Adrian, can you please write me some excuse for my boss? It's midnight here and I have to get up in 5h 😢
Haha what’s the expression? You can sleep when you’re dead? (Or you have a day off) That’s what I always told myself when I got way too little sleep. :-)
Tell your boss that you don't talk about fight club.
@@DavePoo2rule one of geeky fight club is don't talk about geeky fight club
@@lindoran it's also rule 2
Bad RAM chip. Also 'no clock'.
entranced at adrians hand puppet action during intro
ROM socket is definitely a candidate for a ZIF.
You just reminded me that the original BBC Micro has a cutout for a ROM socket originally intended for a speech module. Third parties sold a ZIF socket to fit there instead so we could have loads of (sailing on the high seas) ROM's to change at will.
It would get in the way if you wanted to play with one of the accelerator/ram/IDE boards that use the CPU socket.
+1 for a ZIF A500 😁
@@bweebar not if you use the compact ones, though they are much pricier.
@@bweebar Get one of those DIP-to-IDC ribbon connectors and put a ZIF ROM board elsewhere in the case?
Hey Adrian. It's been a few years but I think if you just select one of the "empty" autoconfig memory test entries, it will try and detect the layout of the expansion installed.
That’s why we always did burn ins back in the 80s with hardware
In 89 or maybe 90 we got an MSX2. I loved that thing and I still do. But a pal got an A500, and soon enough he showed me SOTB. I had never seen anything like it and other reasonably priced systems wouldn't either for a few years to come.
This year I've been fidling with my A1200 which I bought a few years ago. It's a bit pumped with an 8MB 30MHz 68030. And it still amazes me what it can do, and what basically the original A1000 could do when it arrived. The architecture is so strong for what it's made for. By the way I also have a very early chicken lips A500 (the one with mechanical keyboard, it's awesome) which is the one retro-machine that's in my actual office and not in my retro-room. From time to time I just power it up with either SOTB or Turrican II, just to hear the music again and float away a bit to much simpler, yet way more exciting times.
By the way this must have been the longest ramble (pun intended) about ram chips I didn't just skip through😂
I predict another video where all the RAM chips are removed, socketed and replaced. 🙂
Adrian, thank you so much for this follow up - Whilst it was the shortest of the series, you presented some invaluable information into how this stuff actually works under the hood, it was a lot of smoke and mirrors to be before I discovered your channels! Thank you for your awesome brand of Edutainment Sir!
The trash to treasure that keeps giving great repair adventures! 💪
Thank you Adrian for not giving up on this battered old man!
Adrian, I think you have a pretty solid justification here to turn the art project Amiga 500 board into a ZIF 500… with ZIF sockets for all major chips for quick and painless testing.
Thanks for posting, I was hoping for a Thanksgiving Adrian video!
Great video! You made me think of issues surrounding memory testing that I never considered before. Love your channel.
Concerning 4164 vs 41256, it is not as simple as having 4 mirrored areas since it is a row / column addressing ! On a 512KB A501 wrongly equipped with 4164, I would expect a 256KB upper/lower mirroring and a 512 bytes mirroring every 1KB. Moreover, OCS/ECS Agnus does row addressing with the address LSB and columns addressing with the address MSB, which is quite unusual. However, Alice AGA does it the right way (otherwise the double CAS fetch mode would not work).
Well said!!
Wow! This one just keeps going after all this, but even if it doesn't I'm confident your fix it again.
"Ruff n tumble" is pretty nice, too :)
Found another one: "Wolfchild", has even Parallax scrolling, aside from Dual scrolling :)
The phenomenon Adrian speaks of where addresses get mirrored to multiple locations in RAM is sometimes used deliberately, and which is referred to as 'partial address decoding'. I believe this can be observed in the C64 when examining the memory range $D000 - $DFFF (VIC, SID, color RAM, CIAs).
the sinclair machines, zx80 and zx81 have mirrored addresses, due to very simple (one line) address decoding! and actually make use of it for display generation
Shadow of the Beast was one of my all time favourite games, along with Gods, on the A500. The music was fantastic.
I use to use the dollar store hot glue as well, the difference is quite insane.
Careful with the gorilla hot glue if you get it on your fingers, don't pull on it, it'll take your skin off with it, use IPA.
Ask me how I know... hehehehe
Nice video. I do wonder could there not be a voltage problem with the machine that is killing those ram chips as well ?
The system runs off the PSU's 5 volts so it cannot really generate higher voltage to kill chips. Maybe if the supply rails were broken some chips could get a lower voltage than required or have a bad decoupling cap. Then the chip would not work with the board but would test fine in other systems.
A great video! My first Amiga was the A500. Today I have two Amiga 1200. I haven't booted them up in awhile. I sure hope they still work.
The field found Amiga strikes again! 😂
Try the up/down arrows in the Logica #2 option . See if you can change the prefs that way.
Interesting how the extra ram card gives the biggest troubles for this machine :)
While watching, took better look what's in my Amiga. My extra ram card is Alfa-Data RA5-512K+ and the ram chips are socketed from the factory. The pcb has 9333 marking on it, I believe it means 33rd week of 1993, and the chips have Mitsubishi logo in them, and numbers M5M44256BP and 338GEOG -7.
There's 4 chips and 8 sockets. Probably because other model of this card is for Amiga 500 Plus with 1Mb extra ram. Model number for that is RA5-1MB+, but it doesn't have the real time clock or switch.
I have the original cardboard package for this ram card, instruction manual and a little brochure about Alfa-Data's other Amiga products :)
My A500 motherboard ram chips are Siemens HYB514256B-60 made in W-Germany, time stamp 9045.
the temperature grade of the hot glue dictates a lot of things like that initial stickiness, it's kind of surprising. when i bought a fancy industrial glue gun i went down a bit of a rabbit hole. the common type is good for crafts and stuff, but less good at things that take a lot of heat away fast & cause shrinkage at the interface. also tends to be easier to control
Awesome work, as usual! I love these field/barn/wherever find adventures. Also love that you keep them looking largely as they were, just cleaned up. Just one thing though, did you ever replace the 1000uF caps you used in Part 1 with the stock 3300uF values again? If not I guess it seems to be fine with less capacitance :)
if you look at newer 500. the 3300 is actually replaced with 1000uf
When i thought i had seen everything Amiga had to offer, when i first saw shadow of the beast my jaw drapped in amazement ...
Great T-shirt 🤎
a useful thing is you can use 4164s and 41256s in place of 4116s if you lift up pin 1 and 8, so they dont go in a socket, connect pin 8 to 9, connect pin 1 to pin 16,
Hi adrian, I just sent an email to the publisher in order to ask for the manual. It's based in Pordenone, Italy but I don't know if they're still in the business. I'll keep you update.
Psygnosis was the software company that seemed to get the absolute most out of the Amiga hardware. They were known to make the best looking and sounding games. Play? Dunno about that, the ones i played were pretty good. But pretty graphics and awesome sound do not a good game make. They just help.
I remember when Shadow of the Beast came out and it was gobsmacking. I didn't think it had bad gameplay at all. Maybe as a programmer I was in awe at what they achieved back then. I will just add the haunting music was also excellent.
Another totally amazing Amiga game was Another World (Out of This World in USA) by Delphine. I've played it using emulators and totally fail compared to when I was a young man. I blame the screen and keyboard lag involved. It is probably just because I'm an old fart with bad reactions now, but I like to pretend it requires real hardware or an FPGA device like MiSTer. ;)
I remember having shadow of the beast back in the day and I did not play it far at all. Started plenty of games and listened to the tunes and everything but I just never actually played the game far. Sure if was difficult but it must have not been very good either. Others won over it I guess.
Yep! Loved my amiga as a kid..school full of pirate kids and x-copy ! Imagine that now! 😅
🚔
The overwrite explanation you gave for 4164 used in system expecting 41265 reminds me of the same situation for the fake 2TB thumbdrives that are really 16GB with controller programmed to report as 2TB.
Flipping your serendipity on its head - the bad chip being one of the ones you'd already socketed, another way to look at it could be that it was your heat soaking when removing said chip that finally did for it! Of course, chances are it's days were probably numbered anyway, but interesting to think you may have accelerated its demise.
Thanks
Very nice
hahaha Perfect for this channel Lets face it new sockets The way it keeps flaking out at the slightest thing "im just starting the video"
We have a fixed field found Amiga again... for now 🙂
As to your rant, yes, march test is definitely desirable, and in case of an expanded A500 like this one shouldn't be difficult to implement either. It can be tricky to implement on machines with only 'one bank' of ram, ie, every ram chip in the machine handles one or more bits for the entire address space, like a C64.
The issue is you want some ram to keep track of the current state of the test, and errors you find, and the ram chips in which those are stored should not be the ones you are testing. Not only because it would prevent testing the entire address range for those chips, but also because ram errors, esp things like addressing errors, have a good chance to corrupt the state and error data.
Which reminds me of one of my pet pieves, the way the C64 and C128 diagnostics cartridges do their ram test. They run the actual ram test code from ram, which means a completely defective ram chip will just cause the test to crash. That is why you should always use the dead test cartridge first. It will at least test the bit of ram from which the diag cartridge runs the first part of its ram test. This won't catch addressing errors, but at least ensures the code can be copied there without ending up corrupted from the start (it may still corrupt itself if there are addressing errors)
68k has enough registers to keep the state of the test, me thinks.
They have to run the ram diagnostic from ram. Otherwise, couldn't check the ram under the cartridge ROM.
Are the programs smart enough to check the ram before transporting itself to ram? Don't know.
@@williammentink That is correct for the C64 but certainly not for the C128 (and I can know because I wrote a ram test for the C128 which can test all ram and doesn't have to copy itself to ram, through it does use a little bit of ram in the 'alternative' bank to keep track of things, and tests that ram before doing anything else, and can do that without using any ram).
Are the official diagnostics smart enough to test the bit of ram they copy the ram test to?
The ram test in the 'official' diagnostics consists of 2 parts, and for the first part, the ram to which the test gets copied is not tested. That is why you first need to use the dead test cartridge.
For the second part of the ram test, that runs from a bit of ram that has been tested by the first part.
Another annoyance is that despite doing a 'low' and 'high' test, the ram test in the c64 diagnostics does not test all ram, and nor does the one for the C128 (and it is even worse for that one, it fails to test a large chunk of the 2nd bank of ram, and instead tests a large chunk of the 1st bank twice).
@@akkudakkupl should have indeed. I suppose the tests which end in just a screen color do exactly that.
@@c128stuff Is there a plug in card for the user port that can take over the system to test system components and memory?
Memory test in the Atari XL's is almost useless. Plug in cards are useful there.
I'm curious what those MiniDV tapes on the bench are all about. :)
It's hard to believe, but OKI still exists today. I own an Okimate 20. Not sure it it still works.
You could try magic eraser on those rust stains. It does wonders with those.
If you suspect bad sockets every time it hiccups, why not change them out? 😊
Gorilla glue is wonderful, but those poor gorillas.
I was wondering if a wiring harness could be made so that the entire expansion board could be plugged into the Retro Chip Tester Pro so that it could test all the chips at once. That might make it quicker to identify which chips are bad (assuming the Retro Chip Tester Pro would have an good way to indicate which chip is bad).
The piggy back repairs used to see that with the cheap repair companies especially the magazine offering in the 80's
Every time I've had a problem with an A500, it's always the trap door memory. The first thing I do when testing a dead system is take that card out!
You ahould proably use dual wipe sockets since you reuse crippled chips they are better for flat legged IC's.
Deffo STB plays on a original 500 as I had it working on mine!
Btw, did you notice that your two channels are not linked together in the Channels tab? Would maybe increase traffic to this one (and maybe stop TH-cam from thinking you are impersonating yourself ;))
its the same problems with my A500 Rev6a. three ram 512kb ram boards later it still is not working but booting without. the problem eventually finding out after spending loads of money was the soldering on the 44 pin connector to the board is broken. I do not have good soldering skills so left it basic original machine, now got a A600 recently just hope it goes well with this one and spend less time fixing and more time playing :)
Oki is the Amiga's MT
If there is space, you can you yet another spare socket solder the memory chip with short legs on that socket and stick the socket with soldered RAM to the socket that's soldered on the board.
The description of Shadow of the Beast is why we don't have Amigas today. Plays like garbage but looks amazing.
Sort of unrelated to the topic, but for potential seizure flashing in your videos, did TH-cam get rid of the feature that let creators pause the video and put up a warning on screen? I remember a few years back some videos would pause with a countdown and some text on screen displaying why it was pausing, typically indicating a warning of some sort, like "flashing images", but you could write whatever you wanted I believe. No need to have the warning in-edit, and much safer and attention grabbing for people
Wouldn't surprise me if TH-cam axed it years ago, as they seem to do with every useful feature on the site.
Fascinating, I never noticed such a feature. (And I do have a special interest in weird A-B testing and retired features, plus another one in accessibility!)
But in 2020 and 2021 I was sick enough that I only ever used TH-cam on my TV, so it might’ve been web-only (or phone-only) and would’ve passed me by.
Did you consider buying a Heat camera? I think it would save you some time troubleshooting.
Every time I see a Video of yours I want a hobby cellar so bad. 😁
Heatcamera does not help THAT often. sure you can find a shorted cap or so. but not that often
hell yeah!!!!!
New Channel: Adrian's Field Found Repair channel. Lol
do you have a tl866 or something like it? they can directly confirm your hypothesis about the address line with the component tests
i did not read all the comments. if u have spare sockets u can solder chip with short legs into one. than put a socket with a soldered chip in a socket. Maybe it is an expensive solution. Also takes more space but i have done that a several times. All the best i love everything u do !!!
Amiga legend John (Chucky) Hertell wrote diagrom, look him up.
I have already left a comment :)
I want to eventually add an Amiga and a Commodore 128 to my collection
I recommend Lionheart, a later game of the amiga library. It has even better graphics and fun gameplay.
Lesson learned: Play Shadow of the Beast III, it's a much better game! :)
Is the voltage on that memory card correct?
What about creating an adaptor board that can use DDR3 RAM modules?
Is there a schematic for that ram board? I'm pretty sure by now someone's done a new sketch and has a design you can get a new one to repopulate from one of the big PCB makers like JLB or PCBWAY
Better yet there are much smaller modern replacement boards from Amiga vendors like Sordan, AmigaKit, etc. They're much easier to get out of the bottom trap door. :) I'm guessing Adrian wants to stick with that original board since it came in the machine.
May I suggest that you just go ahead and swap out those OKI ram chips with better ones just to save yourself a ton of trouble later on ?
and replace the sockets
Or just get a modern memory expansion for $20-$30. Instead of messing about with an old rusty memory board. Oh wait, then he couldn't make a bunch of videos about it failing.
Yes, I want a video for every failure. Keep them coming. What else can he try to crash the thing?
Shadow of the beast a good game, that's the first time I've heard that one...
My first A500 developed a fault where it would crash if the machine were bumped. THrough experimentation, I discovered that I could make it crash by just lightly pressing down on the middle of the motherboard. I couldn't figure out how to fix it, so...
I bought a used A500 bundle. That machine worked great until it too developed a crashing problem. This time I discovered that it would crash if the power cord was moved. Investigating further, I discovered a loose solder joing on the power socket. That was an easy fix.
are the Commodore 64 CPU chip pins short too? I have an issue with my C64C CPU doing funny things
I’ve seen the clock fail on these boards on other videos due to an issue with a dirty trimmer capacitor?
The tests in this video didn't even find the clock. I noticed in video 2? from the main channel that the clock was showing a different random time each second, so I think it's not working properly.
You can use all the diagnostic tools you like and some are better than others but the only real test - when it's passed all those diags - is to run beefy programs and games that stress all the architecture and then use the computer on a daily basis.
I wonder if the reason the machine was dumped in a field in the first place was because some of that expansion RAM failed and the owner didn't know how or didn't want to put in the time to diagnose and fix the issue.
It's hard to tell, but the fact that the chips were handled might have been part of the cause. Also, I'd probably use dual wipe sockets.
If you think Shadow of the Beast looks good you should see Agony. It's a much later game but still runs on the A500.
Agony was beautiful, but I found the gameplay a bit clunky. My favorite game of the era was maybe Killing Game Show, or Awesome.
@@nicholas_scottI was obsessed with Killing Game Show back in the day. I even wrote up a full review of it for the local Amiga Users Group in Kansas City back then. Great times.
We had servers at work where if one ram is bad out of 20-30 slots, it would not boot the server properly.
CPU, RAM, Mobo, any cards can be problematic.
How does one extra address line quadruple the capacity?
The address lines are used to select both the row and column (RAS and CAS) so if the RAM were addressed linearly, it would be like adding two more address lines. (Like an EPROM)
@@adriansdigitalbasement2 ah, I didn't realize it was addressed as a matrix.
Let me know when it's going back in the field ;-)
I saw the order you took them out and it was D10 the one who was causing the issue.
First thing that confuses me is that the power LED is red. I seem to remember that green was the Power / Guru LED and red was for the internal floppy? Am I remembering wrong?
Earlier models had red power led. Depends on the keyboard I think.
🎶Noone knows what it's like to be the bad RAM....🎶😁
can you fix my Commodore 1094S monitor?
Level with us - are you breaking it so you can make another video? LOL!
LOL! Haha definitely not, as much as I like this computer, I don't want to fix it again, at least not for a while. :-)
That field found amiga case is a prime candidate to be painted a funky colour, or even sprayed black. those ram modules are going bad because they are absorbing moisture, water gets inside them under the legs, before powering them up they need to be dried out completely with a hairdryer for a few minutes
Then call it The Cursed Amiga.
If you were to write an algorithmic pseudorandom non-repeating pattern to memory for an entire address range, then read back memory verifying the value read back matches the algorithm for that address, you could catch addressing errors such as you describe.
There is at least one method which would work for RAM testing (although for other uses such as for games, you'd want to be able to randomise the seed number). You simulate a 16-bit shift register in RAM. You start with a non-zero 16-bit number. The logic is that it takes the lower byte and does a shift (not a rotate) left through carry, saves that byte, then takes the upper byte and does a rotate (not a shift) left through carry and saves that byte. The next steps are to load the upper byte and XOR the 13th bit with the 11th bit then XOR the result of that with the 7th bit from the lower byte. The resultant bit then replaces bit 0 in the lower byte. Repeat for the next pattern of bits.
This should produce a pseudo-random pattern, the only restriction on which is that it can never be 0 either at the start or at any point during the run, since a 0 would stop the pattern from ever producing anything but 0 all of the time.
This can be extended if you want to test RAM which is wider than 16 bits. The key is that you should XOR together bits which are prime numbers, such as 2, 3, 5, 7, 11, 13, 17, 19, 23, 29 and 31. Use a search for "prime numbers 1 to 100" or something similar like I did if you need more. XORing more bits will produce different sequences. Three bits is generally sufficient.
You can write the output each time through the sequence to the RAM. After the simulated shift register has been "clocked" 65535 times, it will have produced every possible bit sequence except for zero.
Be warned that larger shift registers will take exponentially longer times to get through all combinations.
@@melkiorwiseman5234 In my ill-spent youth I studied the mathematics of shift-registers for generating pseudorandom sequences. I had in mind something like what you described.
Adrian I have a request for a new video: Legally obtain the latest version of the Logica Diagnostic, program it into two (even and odd) chips, and run the code on an NTSC Amiga 1000 with a Kickstart Daughter Board.
soon that list of patrons is going to have to start at T-minus 5 minutes in order to fit them all in.
So a 41256 with a bad A8 could still work just fine as a 4146 replacement in a c64.
Interesting, let me dig through my stash and see if i can verify this ...
Just for comprehension: couldn't you test the RAMs with the Retro Chip Tester? Wouldn't that be faster?
I think the legs might be too short to work in the Retro Chip Tester's ZIF socket.
I think testing in place is partly to show you can test without desoldering (I know some were socketed already but ignore that) or needing extra equipment like a chip tester.
That’s not always accurate with (N)MOS chips, because it’s CMOS levels.
Too bad a March test is only usable one month out of twelve. So make sure to schedule your troubleshooting... ^-^
This video is playing fine, but the preview shows with the broken-link icon.
Shadow of the Beast... wonderful? 😅
@Adrian's Digital Basement ][ I've got a challenge for you if you're up to it. Build an entirely breadboard computer and get it to play Zork (the text-based Infocom game).