Julien

Twitter-Acc for my hobbies. Including NTSC/PAL encoding/decoding, reverse engineering 1980s discrete logic chip MMUs, and cocktails near beaches (pre-COVID).

[Julien @booldown]

And we've got color! Fully compliant PAL, modulated in the FPGA.
Resolution is still crappy, because I don't have a frame buffer yet, and the available ROM (on-FPGA block RAM, actually) is limited.

[Julien @booldown]

@drtune It was like that on Wednesday already. I gave up and turned the heat on, in July. Even with a sweater it was just too cold, at least in my part of the city. Let’s see how I will fare today. 😬

[Julien @booldown]

It was of course the boring thing: A broken 4116. Two actually, but of course 0x02 is not a very thorough test pattern, so only once I replaced the chip that caused the lowest bit to get stuck to 1, with the startup code now proceeding, did I realize some other errors in bit 4.

[Julien @booldown]

Used my new logic analyzer to debug why this ZX81 16k memory expansion does not work. It bypasses onboard RAM, so won't start and can't write a memory check program (I could replace the ROM but don't have an EPROM handy). Just tracing bus activity when the regular ROM starts...

[Julien @booldown]

@AlanAtTek Well deserved!

[Julien @booldown]

@isdixon The blutack being indeed most crucial with that pack. %)

[Julien @booldown]

Maybe. This 4116 requires refresh within 2ms. Quick back of the envelope calculation says that's much shorter than it takes that loop to count up to 0xa00. If e.g. a mux does not work, that row might be decayed. But could still just be a broken chip, e.g. borked sense/refresh amp

[Julien @booldown]

Used my new logic analyzer to debug why this ZX81 16k memory expansion does not work. It bypasses onboard RAM, so won't start and can't write a memory check program (I could replace the ROM but don't have an EPROM handy). Just tracing bus activity when the regular ROM starts...

[Julien @booldown]

And here, right after, is a 2nd decrement of same memory. The byte read on bus is still 02 from the 1st DEC, so this time no bit was stuck! Maybe it takes time to decay? The location is also interesting, it's exactly 0xa00. Could the pack's custom refresh circuit have a problem?

[Julien @booldown]

@0x56 @Leeborg_ Ordered some! Can never have enough. The color map does not matter for me, my current clip/cable situation is so insufficient (I have certain types of clips only in red or black for example), the colors long ceased to have any meaning anyway. 😅

[Julien @booldown]

We'll see if I regret it. For now, I think I can do with 160 Msamples (no compression though, so sparse signals will need good triggers for segmentation), and it effectively gives me 4 more channels+1 trigger channel, good UI, and I can still use it with logic analyzer software.

[Julien @booldown]

I ordered the MSO upgrade (RTB-B1) for my scope, instead of the Saleae. Not an easy decision. In the end, the only advantage to the Saleae was that it has virtually unlimited buffer size because it streams via USB, while my scope has 20 Msamples/160 MSamples segmented) max.

[Nelsoncoindigger @b1tcminerpljy]

Doc writing tips; Instead of confessing to a minor technical fudge with tedious heritage, consider instead drawing a veil of mystery, danger and intrigue...

[Julien @booldown]

@drtune I know, right? There's something about "solid pop" that really makes me appreciate their talent. Singers, writers, music producers... Even if it's not someone's taste, there's no denying how competent it sometimes is.

[Julien @booldown]

@0x56 Back to the two keyboard "cables", they are literally just two extruding parts of the very brittle plastic keyboard membranes that get plugged into sockets on the board, not separate cables.
But I guess I could have tested by putting probes on two ends of the visible surface...

[Julien @booldown]

Fixed the keyboard ribbon "cable" of a Sinclair ZX81. They are horribly thin, brittle, and break a lot.
I stuck clear packing type on kitchen tin foil and cut out lanes. Took way too long, but works... though if you have conductive ink, better use that I guess.😬

[Julien @booldown]

@0x56 But I also designed a video mod for it from scratch (I'll explain why it's necessary and with more pics in another tweet), and there I did some bode plots. Here before and after I compensated with some caps. Without, the bump in the frequency magnitude caused some bright halo.

[Julien @booldown]

@TubeTimeUS I guess those are fine, since selecting the right character seems to work fine. I'm placing a (not super confident) bet on the 4th bit of a text mode specific shift register not latching correctly (maybe that's part of the PCB damage).