Hi, I’m new here, but I’ll been reading this forum for years. I’ve been looking for a good way to test the ram on my various Apple II’s, Clone, and 16K RAM cards. Since I didn’t find a readymade solution, I decided to create my own and I would like your thoughts on this project. It is based on the work of several others:
To keep it simple, I use a step-up voltage module and a reverse voltage module. It’s powered through the Nano (I covered the Nano LEDs as I find these really annoying). It seems to work well, and after testing almost a hundred ram chips, I found several bad on an unused 16K card I have. I totaled up the parts cost, and it’s around $18.
It looks like you've done a good job of it.
Are you planning on selling the PCB's?
Hi, yes, eventually. But I'm really looking to help support the vintage computer community (Apple II) as so many others have (and so many on this forum have).
Could it be easily adapted to test the 4K x 1 RAM chips found in the Apple I ?
(During the past 4 months my own Apple I build was blocked most of the time by running DRAM tests. So far I found 1 truly bad DRAM (which even spat into the bus so that the monitor program produced lots of superflous stars but almost worked) and 6 suspect / flakey ones out of more than 200 of various brands (Intel, Mostek, Motorola).
Seems that 40+ year old NOS DRAM really need some thorough testing !
Beware that some of the faults / bit disturbances take a lot of observation time to find (I use 1 week running Mike Willegals RAM test 24/7).
Back in the day when the radiation sensitiviy issue with DRAMs was discovered (around early 1980s ???) some daring engineers used DRAMs as an altimeter and it worked well enough to marvel at the perils of living at high altitude.
Anyway, if you make PCBs please consider to test 4k x 1 and 64k x 1 DRAMs, too. And some option to run tests for a long time without any user intervention.
Of course there are lots of other DRAMs. But for 1970/80s vintage microcomputers the 4k x 1, 16k x1, and 64k x 1 as used in the Apple I, II, IIc ... will cover a lot of ground.
Have you had a chance to work on any kits? I would frankly be happy to pay good money for multiple, pre-assembled testers, as testing 4116s is something that I need to do quite often and I have never used Arduino kit, so I would be wading in to an ocean with a rubber duck as a life vest. I should invest time into learning the Arduino kit, but I am three years behind in so many projects that the sheer convenience of being able to buy an off the shelf tester for 4116s is to me, quite justified.
Did you ever come by this project: https://8bit-museum.de/sonstiges/hardware-projekte/hardware-projekte-chip-tester-english/
This device can test multiple memory chips, also the 4116 and onther ICs.
I would love a device like this that tested more than just 16K chips. If it could do the 64KB chips in the IIe/iic (etc) or even the 256Kb chips on the IIgs I would buy one! that would be great.. But honestly just 16KB makes it too specialized. And most machines I have that use those chips have sockets so I can test rather easy anyway..
Hi, I hope to add the ability to test multiple types of RAM chips.
-Also, sorry for not responding sooner, we were away.
Put me down for one :-)
Have you thought of publishing it under an Open Source license? Then everyone can simply source their own from their nearest PCB manufacturer.
Here is one more version of arduino 4116 dram tester. Including: Part List, gerber files, arduino code, driver for clone arduino on video description. And you can power the arduino from 9v dc power supply, or from arduino 5v usb. I have not personal tested, but know peoples that did.
Arduino 4116 dram tester FINAL ~Muttley Black~
Group buy for PCBs ?
I've ordered 10 PCBs and am planning to build one out to test it.
I'll sell off the remaining PCBs if it works properly. They should be cheap enough to build, especially if you have a stach of components lying around. But even if you have to buy everything the total cost should be about $60 or so.
The most expensive component is the ZIF socket which is criminally high at $26 CDN.
Post here the results, i think i will order too
I just did some searching to see what was out there as to 4116 testers. Someone in Canada is selling on eBay multiple units of a 4116 RAM tester that looks very similar in circuitry and part count to your design. They are asking $13.45 per unit plus shipping of $6.53 .
See the tester offer here: https://www.ebay.com/itm/114707724765/
So if you intended to make these to re-sell, do note there is a competitor out there. Or is the person in Canada selling these now on eBay you?
I'm the guy selling those PCBs on eBay.
Muttley Black graciously put his project into the public domain, and I built one successfully.
It works great.
I'm now selling the remainder of the bare printed circuit boards at what I consider a reasonable price.
Several have been sold - I have maybe 5 left.
I bought an Arduino based DRAM tester.
I have verified that some DRAMs, surely broken, are recognized as working by the tester.
I believe that the problem is in the timing of the program, common to all these testers offered on EBAY.
Has anyone encountered the same problem as me ?
I haven't run into that particular issue. Which Arduino tester did you buy on eBay?
And I'd love to verify your situation with my own tester if you're willing to mail me a questionable chip.
the installed software is the standard one, available on the internet.
The chip is definitely defective despite 2 different testers, both based on Arduino, have given the OK result on the chip.I cannot ship the bad chip, because I want to understand the tester problem, too.My first impression is that Arduino is not sufficiently precise in timing and that the software algorithm must be completely rewritten since it executes Read/Write on every single cell and not as I imagine correct with Write on the entire DRAM and then Read as verification.
This is the project of one of the 2 testers I tried:
You may need to reassign the pins but have you tried with this sketch?
Hi Baldrick,the author of the site of the previous link, informs me that she has fixed the bug contained in the zip file you posted.The algorithm does not recognize the failed cells of a DRAM, only the dead chip.
I didn't think there was a bug in that code.
But if there's an improvement I'd love to see it.
I'll ask him if he'll post the fix.I received the tester and now it recognizes the faulty cell.
Muttley's tester works perfect.