It begins! (Wren Prototype build log)

[Dr Jefyll]

allisonlastname wrote:

the 62256s I have don't play nicely with the 6502 for whatever reason

Was there any followup on this matter? In this post you mentioned fixing an error in a RAM test jig you created. And the 62256s then tested successfully in the jig, but still failed when installed in your project. One possible explanation is, the inputs on the test jig recognize TTL voltage levels, whereas the inputs on a modern WDC CPU do not. It's plausible that the 43256 you got from your dad significantly *exceeds* the spec (as many TTL-rated chips do), but the 62256s... not so much.

allisonlastname wrote:

exploded

I assume you're just being colorful with your speech... although chips truly can and do sometimes explode! Check out the DC-to-DC buck controller IC in the Keyence industrial vision controller I repaired earlier this year! A surge on the incoming 24VDC left a crater in the IC and also vaporized one of the tiny PCB traces underneath.

-- Jeff

[barnacle]

allisonlastname wrote:

And now the RAM has somehow exploded.

That's right up there with the quote 'At this point, the princess does not die'...

As Jeff suggests, chips generally explode only if they exceed their substrate's current rating, usually because of an over-voltage situation. Are you still working on breadboards? If so, that's a likely explanation.

Neil

[Paganini]

I really like these "skinny" SRAMs:

https://www.mouser.com/ProductDetail/Re ... cyCode=USD

They're being discontinued by the manufacturer, but Mouser still has quite a few in stock.

For my first project I used one of these:

https://www.jameco.com/z/HM62256LP-70-H ... 82472.html

I think I got it because it's the same part Ben Eater used, and I didn't know what I was doing yet. This seems to be yet another case where the datasheet specifies TTL voltage levels, but where the actual parts are over-specced sufficiently that they don't have any trouble working with the 65C02.

Like others, early on I encountered problems using a "skinny" Alliance 15ns SRAM. Switching it out with one of the IDT/Renesas ones did away with the glitches.

[allisonlastname]

I think I might be haunted.

First off, I have no idea how the 43256 (figuratively) exploded. I was running the board on 5v from a USB-UART adaptor, so there shouldn't have been anything more than 5v on the board. It could have been ESD, but I didn't go anywhere near it. Who knows.

Secondly, I put a 62256 in the board to have another crack at debugging it - and this time it works. I don't know why, or how, but this time it works. Is this what they felt like designing the BBC micro?

[drogon]

allisonlastname wrote:

I think I might be haunted.

First off, I have no idea how the 43256 (figuratively) exploded. I was running the board on 5v from a USB-UART adaptor, so there shouldn't have been anything more than 5v on the board. It could have been ESD, but I didn't go anywhere near it. Who knows.

Secondly, I put a 62256 in the board to have another crack at debugging it - and this time it works. I don't know why, or how, but this time it works. Is this what they felt like designing the BBC micro?

Who knows - do watch the TV drama based on real facts-ish on youtube though

https://www.youtube.com/watch?v=XXBxV6-zamM

but they already had at least 2 systems under their belts - the Acorn system 1 and the Atom.

And to let you know you're not the only one with frustrations - my current project is on breadboards and I've so-far had 2 issues that caused me much head scratching - both wires in the wrong places... but it's now running very well at 2Mhz, so don't give up!

Cheers,

-Gordon

[allisonlastname]

Except now it doesn't. It worked for a few minutes but now all reads are returning nonsense. I didn't touch the board at all in that time.

[barnacle]

I've seen that symptom (or similar) where a control input is floating: sometimes the level is such that it operates, at other times, charge accumulates/dissipates at the gate until the voltage reaches a value that stops things.

This is a default CMOS logic fault; I'd be looking for unconnected pins, possibly in the control logic or enable lines. Or indeed the power; while memory chips usually need reasonable amounts of power, most CMOS chips will self-power through an input held high while the chip is unpowered.

Neil

[allisonlastname]

I've tested all the control pins with a multimeter, and they're definitely connected to the stuff that produces those control signals. Unless there's something disconnected internally, it's not that.

[BigEd]

Sometimes if there's an unreliable connection, putting your probe onto the thing makes it work. That does make it more difficult to locate the problem.

I think the thing to do is to keep firmly in mind that there is a problem - ruling everyone out logically can't solve it. Check all your assumptions.

[and3rson]

drogon wrote:

And to let you know you're not the only one with frustrations - my current project is on breadboards and I've so-far had 2 issues that caused me much head scratching - both wires in the wrong places...

I can relate. I even angrily destroyed one of my older Deck65 prototypes which I was SURE had all issues fixed, but still refused to work as intended... I promised myself to never come back to this hobby again...

And yet here I am, after 5 more revisions and some crunching, it finally worked. Those dreaded open-drain lines... Good thing that JLCPCB sends 5 copies of the PCB.

[barnacle]

Tsk. We all should know to remove the expensive components from the board before breaking it angrily.

Neil

[BillO]

Just a reminder..

https://getyarn.io/yarn-clip/1a552bd8-9 ... 9b7595be54

[allisonlastname]

So.

I acquired some W24512AK RAMs following a previous recommendation in this thread, and installed one in Wren. But it doesn't seem to work. I'll test another one on a breadboard with an arduino, because it's entirely possible that they're all dead on arrival, seeing as I got them on aliexpress.

My leading hypothesis is that the write cycle isn't working, as on a read cycle, it's outputting all zeroes. Namely, 0.4v on each data pin, which the datasheet lists as the maximum output low voltage.

Thankfully, everything else is working. With any luck the RAM issue will just be something stupid I've missed and I finally be able to get on with programming.