The ESP32 SoftCard for the Apple II

Loving the way all this is going. Such a great card and am looking forward to the version 3 firmware when its all good and ready :) 

I do have some dumb questions though.

1. I have a Rev 4 motherboard in my apple ii and it doesn't do lowercase, which has never bothered me. I quite like the simpler earlier versions of the apple ii. But on a whim I decided to hotspot the esp32 card off my iphone and the current password has lowercase in it. I dont want to modify the computer, I want to keep it in all its Rev4 glory. But it got me thinking, is there a way to implement a key combination that would allow lowercase when the esp32 has control of the system. Just a thought. Lowercase would also be handy in the upcoming ibm pc emulation.

2. Was there some documentation for the esp32 commands? I thought there was but now think I was mistaken and its all currently provided onscreen by the "HELP" command only. Online only does make sense while the system is still being progressed.

Cheers

Dave

Maybe adjusting the joystick "dead zone" to a little larger would be good? Or adjustable somehow. Right now if I use the JOYSTICK screen to make sure it's centered, I can get it pretty solid about 128 / 128, but then if I move the joystick around (left/right and up/down) when it re-centers it may have drifted by a couple in some direction. And if I press one or both buttons, the center will drift down and to the right, which I assume is just some issue with the joystick using more power to close those connections which changes the resistance of the stick. Weird. It's still inside the center square, but just barely, touching the lower right corner.

It's a minor issue though... I don't mind unplugging the joystick if I wanted to just use the mouse and not worrying about the joystick causing things to drift.

I still can't figure out the audio issues... I tried several different 44.1K and 48K streams. The 44.1K streams all seem to work pretty well, even with higher Kbps (from 128K to 192Kbps or above, although at 192Kbps I might get a drop every now and then). But with 48K it seems like I consistently have dropouts.

I think some of those example URLs I shared before are geotargeted so you would hit them in Europe whereas I'm hitting them from a US server. That may or may not make a difference. I can setup my router to VPN all traffic to another location (or just for certain clients), so maybe I could try that out and force the ESP32 to pretend it's in Europe somewhere just to see if it changes anything.  In the case of that "Indiana Dunes" stream, I never had any issues, but they may geofence it to US addresses... that's too bad, it's a fun station. :)

I love this board.  One of the best creations of 2023 IMO.  Thank you for keeping us updated and for all your work on this board.  Pretty amazing.

Question:  Would it be possible to add VGA video to the next PCB iteration?  Is there room in the ESP32 module for that?

Here is version 3.0 of the ESP32 SoftCard firmware, which adds an IBM PC/XT compatible emulator to the mix, capable of running DOS and Windows 3.0

This is very cool. I haven't tried out a lot but here's a few things I noticed:

• - I found this on version 2.02 as well but only saw it today... when I "connect <myssid>" specifically, rather than picking from a list of SSIDs, it doesn't mask the password after typing it like it does when picking from the list or when it comes up from a "listen" or "ftpserver" command that also presents the list.
• - The 48K audio streams that were giving me trouble are still having dropouts
• - Probably expected, but I happened to be listening to an audio stream when I was copying over the new firmware, and the audio stream really started to stutter and dropout. That made me think that a high bitrate stream is just too much for the ESP32 to keep up with? I haven't done the math to see what the actual bitrate would be for some of these streams that have 48K... I think they're 16 bits per sample and just 2 channels so that would work out to about 1536Kbps (192KBps) and since I'm currently seeing an average FTP copy (using the DOS hard disk image 1 to test) of 180-200 KBps, that 192KBps audio stream may just drop too much to keep the buffer filled? I've even seen some streamers use 24 bits per sample but I didn't even bother testing those...  With a 44.1KHz stream, I calculate that around 176 KBps so that would tend to play nicer with what I'm seeing for transfer rates.
• - I stopped the audio stream and then noticed there is static coming from the speaker while that file transfer is in progress. It's not loud or anything, just interesting. Once the transfer stopped, the static went away. At least it's an audible cue when the file is done copying if I have that window hidden. :)
• - I tried out the GEOS image. It boots okay and shows the initial Geoworks screen, but then the video goes into a rapid roll. I tried again using NTSC50 and PAL instead of NTSC (my LCD monitor can handle all 3 formats) and also tried the different video settings when starting the PC emulator but it did the same thing each time. Maybe it's just something with my LCD though and a real CRT doesn't do that.
• - DOS runs fine but when I run Windows 3.1, as soon as it switches to graphics mode, it does the rolling screen. Same thing with NTSC50, PAL, or any of the options for color, mono or Hercules.
• - In fact, when I run the PC emulator with any video mode (NTSC or PAL) and use the Hercules option, the screen just rolls right away even in the text modes of the emulator.

The regular DOS (and I just barely tried the CPM) work great so it seems like it's the graphic modes that cause the rolling screen, but if it's just something that happens with my LCD, I wouldn't be too surprised... I haven't seen the rolling screen with any of the other game console or Mac emulators though so it was a surprise. :)

CVT wrote:

 

The 48K audio stream fix in v3.00 only applies when you don't see "Slow stream, dropouts are possible". ...

I thought about instrumenting up some network capture at my firewall level to get a better picture of what's going on, or at least looking at my access points to see what it's reporting for the RSSI to confirm the connection quality. I've just been too lazy to do that so far. :)

Looking at my access point, it's showing the same -53 dBm RSSI level that I saw when doing a CONNECT and looking at the SSID list. that's a stronger signal than some other items on that same acces point like some wifi cameras, a couple of Rokus, etc. A Roku probably has a similar tiny antenna, and those are further away from the access point so I'm not surprised by that. But those Rokus are able to stream multi megabit per second (up to 4K video). :) But then again, those Rokus (they're 7th gen) have ARM A53 processors in there which are probably much more powerful than the ESP32.

If I get a chance I'll try to setup a port mirror and capture the ESP32 traffic while streaming and see if anything stands out.

I did a quick look at the traffic while streaming. I put ntopng on my OPNsense and watched the traffic from "icecast.walmradio.com"...

Couple of interesting things... first, I realized (and I should have checked this earlier) that "walmradio.com" is indeed located in Europe (Moldova). My latency to their streaming server is 212ms.

Second thing I saw is that while streaming (and dropping quite a bit) it was receiving at a rate of about 172 KBps.

I'm going to do another test in a bit with the same stream but using VLC since I know that works, and see what the bitrate is for it that way, so I can get a baseline of what it should be when it works right.

So, something I learned (which I should have known) is to check the source of the streams. Because the latency can be a killer. Especially if it the buffer is small, the latency is going to kill the stream.

I'm happy to pick different streams. I don't know why but I just assumed that "walmradio.com" site had hosts in the US... the idea it was only hosted in Moldova didn't even cross my mind. 

EDIT:

Oh, I posted before seeing your last post. Seems like you saw 272 KBps so that must be the actual rate it needs. Yeah, those are old timey, and probably not very well compressed.

3.01 seems to be working okay for me, with those things mentioned being patched.

The other day I picked up an old CRT (an old Sylvania 13" TV/DVD combo). It works great and the 4:3 picture looks so much "better" than my 16:9 LCD screen did, for the ultimate retro experience.

Unfortunately in my case it doesn't work well with the 50 Hz (or lower) signals. It shows the image but rolls. There are no external controls on this at all (it probably dates to 2006-2008... even has a DTV tuner) and even the service modes don't have any controls. When I have time I may pop the case off and see what kind of vsync controls exist. It would probably need to be some compromise between working well at 60 Hz for the regular Apple II NTSC, and working for some of those emulated graphic modes.

In the meantime I may just need to keep an eye out for other, older, CRTs. The place I got this one from (RePC in the Seattle area) had a couple other older models, TV/VCR combos, but I wouldn't hold out much more hope for those either. It might just be better to find out which brands for sure support the multi frequencies and an lock on solidly, or have external vsync knobs, and bide my time looking for one of them.

The other emulated things (NES, Sega, TG) all look great. Better than the LCD, I think. Just a crisper, more vibrant looking picture.

I think I saw that when I was looking for a service manual, but the one I ended up getting is a Sylvania CD130SL8 so it's a little different. Plus it didn't come with the remote (I think if I went back there and dug in the pile of remote controls I might find it, or a similar one). I was just using my universal remote to switch inputs.

I'll have to look back at the monitors you were using and see if I can find something like that, or I was looking in ebay for monitors that list themselves as both NTSC and PAL compatible. I figured if they support both of those, that's a good start and may work at lower rates as well. But that's a gamble... you never know. :)  At least I can put this Sylvania on my other Apple II if I find one that's better to use on the system I have the ESP32 card in.

I was just thinking back to all of the old TV's I had 20-30 years ago and how those would probably all be awesome at this. A 27" CRT with all kinds of controls, inputs, etc. Oh well. They're in a landfill somewhere by now.

I remember seeing those videos from Adrian with that replacement TV board. It's a pretty impressive mod for sure. I'm afraid if I tried it, I'd end up with his first version where it looked kind of janky with the controls just hot-glued to the front of the set. :)

The LCD I had been using is a 19" Jensen TV (made for RV installations) and I was glad to see it supported NTSC50 and PAL modes just fine. The Mac emulator displayed okay except the top/bottom were missing and this TV has no controls for vertical size unfortunately. But I could see enough to try things out a bit. It just couldn't handle those other resolutions you mentioned where it needs 45Hz... the screen would just roll, no vsync at all.

I wondered about this before, and I'm thinking about it again now, but would the Retrotink be able to work on those signals? I've heard about them but I haven't looked into the specs on it to guess if it could handle those frequencies or not. If it could, I'm guessing someone could create a profile to handle this output and display it? From what I understand, the Retrotink has profiles (some come with it, but you can customize them as well?) to help properly display the signal from various devices. The 2x pro isn't cheap (I see it listed for $139.99 right now) but if it worked, it might be something easier to setup and then use an LCD, plus have a lot of other use for retro nerds. I think the nice thing would be the (potential) ability to lock on to different signals just by changing the profile to match, rather than fiddling with controls on a CRT.

On the other hand, like I mentioned, the CRT picture looks so cool compared to the "softness" or whatever that the LCD shows. I think I'd probably stick with the CRT if I were just playing things that work on it without modification. I need to do the RCA jack modification so I can pipe the internal speaker to the CRT and control the volume easier.

This Sylvania set is funny... it has the analog and digital tuners, a composite video input, and just a single audio input. I get that it's a small 13" TV but it was funny it didn't even have stereo inputs. It probably only has a single speaker in it, I haven't really checked. :) In some ways it seems modern like the ATSC tuner, but in other ways it's like a relic from the 80's (which makes it great for retro systems I guess). I'm definitely keeping it and I might just toggle back and forth for now between it and the LCD, depending on what I'm doing. :)

I was checking ebay yesterday for those Sony Trinitron PVM models and yeah, they're ridiculously expensive at the moment. $400 for a 14" monitor? Nah. The 8" models are cheaper and I see a few for about $100 USD right now (that includes shipping) but I don't know if I'd want to go all the way down to an 8" screen. 

Just for fun over the weekend, I hauled my Apple II setup to different rooms of the house, trying different TV's that have composite inputs.

My results are that an older Panasonic Viera 42" (plasma screen) didn't work at all ... not even regular Apple II composite video. Oh well.

I also tried a Mitsubishi 82" DLP (rear projection) and it displayed the regular video just fine (NTSC 60).

When I tried NTS50 and PAL, it kinda worked: It would lock in okay on the signal but the screen size didn't adjust and the top of the image was about halfway down the screen. I didn't look to see if the Mitsubishi had any kind of screen adjustments for position, size, etc. since this was just for fun, to see.

Further, when I tried something like GEOS in the PC emulator, it was similar... it locked in to the signal just fine but did the same thing where the image was halfway down or so and cutting off the other half. But still, this was way better than on the LCD TV (Jensen) I'd been using where that particular mode doesn't lock at all, or the CRT where even the 50-Hz modes don't work.

Mainly I was curious to try it out for fun and build up a little more data on what kinds of TV's might work. Both the Panasonic and Mitsubishi are probably mid-2000's or so (can't remember when I bought them) so I figured it would be a long shot. I was surprised more that the DLP was locking on okay, and only needed some (possible) screen position/size to get it to render correctly.

Oh, hmm... I forgot I could adjust the screen placement with the hot keys. Whoops.

The 82" TV that showed the image, just in the wrong place, I was actually doing exactly that, using the Y input since that TV didn't actually have a composite input so I gave that a try. That might be why it actually showed better than I expected. I think my 42" TV has component inputs as well as the composite so that might be interesting to try.

In the meantime, when I'm not trying any of the games on there, I still enjoy just listening to some streaming station while I'm working on something else. It gives off all of those good "retro vibes" :)

I was going to try out the new firmware this morning but when I turned my Apple IIe on, it's doing the same thing it did originally where it shows it found the ESP32 in slot 1, finds the mouse in slot 4, and then just hangs there. It doesn't show the text about making sure the monitor is connected or whatever.

I haven't turned it on for a few days, maybe a week. The Apple IIe itself seems fine. I'm running the Apple IIe diagnostics right now and besides a high pitched whine from the CRT it's all great.

When that happened before, I didn't do anything and it just started working on its own after a while, so I'll do some other things on there today and try again, or maybe try unplugging the joystick and mouse (both tested okay) in case something is going on there.

Tried the latest firmware this week, excellent work!!!

I must say as a complement that may seem weird to some people, it's fantastic that your CLI actually seems to handle spaces in directory names themselves. I usually do not use spaces in file/dir names, deliberately, but I had a space in the name of my PC images directory name. I already had set the monitor atop the IIe so I tried it anyway, and it works great. 

CVT wrote:

Try running the Apple II RAM Test Utility: http://www.ivanhogan.com/kfest/MemUtil/

I did some extended testing and it's interesting...  If I kept powering off and on, eventually it would go ahead and boot to the ESP32 prompt, maybe 1 time out of 5. But then about half the time it did that, the keyboard wouldn't work. When the keyboard did work, I could do some things, listen to streams, I launched GEOS in the PC emulator, etc. but eventually it would just freeze up.

So I did something that I didn't do last time because it started working reliably and I didn't bother. But this time around, I popped the cover off and removed all of the extra cards in there so all I had was the 80-column card, the ESP32 (in slot 1), and a disk II controller in slot 6 (with the FloppyEmu).

With just that, it works every time, not a problem.

Then I started plugging cards back in one at a time to see what caused the issue...

First I put the super serial card back in (slot 2)...works.

Second I put in my 2nd Disk II controller (with 2 disk II drives attached) into slot 5, and again, that works fine.

Last, I put my Mouse Interface card in slot 4, and... that was it, it doesn't work reliably when that's in.

So, I wonder if it's something that happens when the mouse card is there, because I would always get the message that it detected the mouse in slot 4, and then nothing else after that. By removing the mouse controller, does it just skip whatever routine might have been hanging it? I'm just guessing.

I tried again with the mouse controller installed and just pulled out the super serial card instead, thinking maybe all those cards was reducing the voltage enough that the ESP32 is on the margin of getting enough, but it just wasn't happy no matter what, as long as that mouse card is installed. I also just tried it with the super serial and the second Disk II card removed, and yeah... it's only when the mouse interface is in there that it locks up right after it shows the "MOUSE CARD FOUND IN SLOT 4" message. I did try it with the mouse unplugged but that didn't help either, it still froze up in the same place.

I hope that helps, or makes you think of something that it could be? It's a real Apple branded "MOUSE INTERFACE" 670-0030 and it has been working reliably since I got it. The mouse is an official Apple II mouse that was one of those "new, old stock" floating around in eBay, fresh out of the factory plastic.

I thought I'd mention, I tried the mouse interface card in other slots (3 and 7) and same result. It sees the card in those slots and then freezes.

CVT wrote:

Try running the Apple II RAM Test Utility: http://www.ivanhogan.com/kfest/MemUtil/

I just let this run for 100 iterations testing all RAM and it tested fine. I've used other utilities too when I first got this system (but not this one) to make sure the RAM was all okay and they all gave it a passing grade too.

I can try pulling the 80-col card and see if it matters... I don't know if the ESP32 card works with just the basic memory? I assume it would be okay? That would help eliminate any issues with extended memory at least even though it's all testing okay.

If I really need to, I have this other Apple IIe I can test it in, but right now it needs a few keyswitches replaced (I have them, I just need to solder them in) and I haven't really done a lot of testing on it. It's non-enhanced but I don't think that would matter.

One other thing to mention is that the Apple IIe I'm using now has the ROMxe installed, but it sounded like at least one other person is using that as well since he romified the boot disk for this, and I assume it's working okay for him. :) Maybe I should try booting the ROMxe with the unenhanced ROM, or try some other things just to see if I get any different results. I have the original ROMs I can put back in if needed but I don't really think that would be the problem.

EDIT: Yeah, after letting it run 100 passes of the memory test and everything was okay, I tried the ESP32 again and it at least brought up the command prompt but then it froze up, no keyboard output. That happened before about every 1 out of 10 times...  On the ROMxe I noticed that one of the included options is that same memory test utility, so I actually have run it before, I just didn't remember. And if I boot the unenhanced ROM it didn't boot, just gave me the ] prompt. Maybe I was using the Prodos disk though.

I also just pulled the 80 col card and got the same error... it's booting more reliably to the ESP32 command prompt but then I can't type anything and I assume it's frozen/locked up?

The only thing that's guaranteed to make it work is by removing the mouse interface card. Works every time. I guess it's possible I got a bad mouse controller but it works in regular Apple II apps just fine, never had any issues with it.

CVT wrote:

Is your mouse always plugged-in to the Apple II Mouse Interface card?

That's correct. I tried it yesterday by just unplugging the mouse (and joystick, just in case) and leaving the card installed, but that didn't change anything.

EDIT: Just to see what happened, I just swapped the ESP32 and mouse cards around, so the ESP32 is in slot 4 and mouse is in slot 1, but it did the same thing... freezes up right after it said it found the mouse in slot 1. If I remove the mouse card, it sees the ESP32 in slot 4 just fine and starts up normally. So it really does seem to only be an issue when the mouse card is installed... very weird. :)

I'm happy to provide any other info you can think of that might help. I wish I had another mouse interface card to test, but I don't. :( 

CVT wrote:

No, the mouse should be plugged-in to the interface card. When it's not there are issues.

 

So, at this point perhaps the best test would be to try both the ESP32 SoftCard and the Apple II Mouse interface card in your other Apple II, even though it's missing some keys.

LOL - I was afraid you might ask me to do that. :)

I moved the ESP32 and mouse interfaces, and one of the Disk II controllers, to my other Apple IIe. It boots fine *without* the mouse controller, and if I plug the mouse interface card in, it does the same thing that it did on the other Apple... it says "MOUSE CARD FOUND IN SLOT 4" and then just hangs there. So... same problem but on a different Apple. Different memory, different 80-column card. The ESP32, mouse card, and disk II card are all the same but everything else is different.

I'm looking at the mouse interface card with my magnifying visor and everything looks okay. All the slot contacts are good, pins all look fine on the ICs (no corrosion). The only thing that looks odd is one of the jumper pads, X1, has been cut through and I'm not sure if that's normal or not... about to look that up. Okay, yeah, just looked it up and X1 is supposed to be open, X2 closed, so that's all normal. The socketed chips on it all look to be original with the (C) APPLE stickers on there. The mouse cable and socket all look good too... no tears/rips/etc and the connector is nice and clean.

And as mentioned, the mouse has been working fine... even when it was working okay previously with the ESP32, I used the mouse test on there to try it out and it was all good, and using the Mac or GEOS emulator, there weren't any issues moving the mouse around.

Well... this is interesting...

I mentioned I had my magnifying goggles on and everything looked okay. I got a flashlight on there as well and noticed a little piece of white "something" stuck between a couple pins of that non-socketed IC on the interface card (the AMI 8402 chip). 

It looked like just a piece of styrofoam peanut, honestly, and may have been lodged in there by the ebay seller during packaging, or who knows what.

At any rate, I got a pick and managed to pull it out... sure did feel like styrofoam but it was just a small little thing. I don't remember which pins it was inbetween, sorry, but probably around the middle of the chip on the "lower" side of it.

After removing that, I tried again on the other Apple II and it actually worked!

I just moved everything back over to the first Apple and it's working there as well.

I really doubt this thing was conductive, so I wouldn't think it was shorting a couple pins, but maybe just barely enough to make a difference in some cases? And I could see how it would be intermittent, just based on whether or not gravity had pulled it closer to the pins, or maybe it was able to build up a tiny capacitance that made enough of a difference in whatever pins it touched?

I'm still too nervous to say for sure that was the problem all along because it seems too easy, and also too weird, but it sure seemed to fix it. It went from not working at all to working 100% the past few times I've power cycled it, and I'll just keep doing that until I'm confident. I'm leaving the case open and the cabling just temporarily in place since I want to be certain it's fine before I put everything back together.

Well, if that's all it was, I really need to apologize for wasting your time when it was a flaky mouse card the whole time. Sorry about that. :/ Now I want to look up the specs on that chip and see if I can guess what pins were being kind-of bridged.

CVT wrote:

No apology is necessary! I am glad you found the issue and nothing turned out to be defective. I don't consider this wasted time, but if it is, then it's time well wasted. :)

I looked up the datasheet on that IC and that strange white piece of whatever was in the PA0-7 or PB0-7 pins, for the peripheral data lines. I guess it could have caused some corrupted data in/out which could cause strange results when it polls for mouse activity. I saw you've posted previously about that chip (it's the 6821... I was reading the wrong part of the chip id) so you probably know more than I do, regarding what that could do if a couple of those lines were shorting intermittently. I think it's more likely that that crud was in the PB0-7 pins but it could have been around PA7 - PB0.

If I had to guess, it must have been something that didn't really cause problems when apps were doing normal mouse polling since it was working fine in native Apple programs, but if it was causing some strange things in the bus that normal apps would ignore, I guess that could cause problems with other things like the ESP32. Since the issue always seemed to happen right after it said it deteced the mouse in slot 4, perhaps there's something it's doing to check if a mouse is actually plugged in, or if it merely enables those mouse features like you suggested, and that's when unexpected noise coming from the interface would make it trip up.

Thanks for being gracious enough to look into it for me. I'm glad you don't consider it wasted time. Same with me... even though that's probably all it was, it gave me a chance to clean up some of my internal cards and wiring, and it forced me to get that other Apple II going. Now I've pulled the keyboard out so I can get those replacement switches installed... something I've put off doing for the past month since I got the new ones in the mail. :)

Oh yeah, FYI, version 3.02 is working good. I think I mentioned that GEOS is showing up now without the vertical rolling I had before (on my LCD monitor). So it's the same as how it was on the Mac emulator. I can use the Ctrl keys to move the screen up/down. At some point if I keep doing that, it actually starts scrolling again so I can only go up or down a little bit before that happens, but there's probably nothing that could be done about that. I guess unless the code actually changed where it starts/stops rendering the screen based on those offsets, which is probably too complicated and doesn't really solve the problem of the resolution simply being too high for a "240" monitor to show. :)

I'm going to keep an eye out for a used Retrotink 5x ... I'm hoping someone will upgrade to a 4K and sell off their 5x for a decent price. Or just keep looking for a decent Trinitron PVM on ebay that won't cost an arm and a leg.

Oh... now that it's working better, I had a chance to try the game emulators again with 3.02 (I tried NES, Sega and TG16) and they all roll the image vertically. Worked fine in Doom, Wolf3d, etc...

I was freaking out, thinking I damaged something when fussing around, and then I remembered that when I was trying out GEOS, I was using the Ctrl keys to move the image up/down and I guess it "remembers" those settings even with a power cycle? :)

I hit Ctrl-R to reset and then it was fine again. Whew. I even pulled all the cards back out in case I did something weird when I put them all back in. LOL

I guess it's a general question then... those screen up/down/left/right settings are persistent? Would it be possible to have different settings for each screen mode so that Hercules, for example, would have one setting, and then other modes like the 320 x 240 or plain 320 x 200 would have their own? If not, it's not a big deal... it was one of those things where I changed it and then forgot about it, and didn't realize it was sticky between reboots.

If anyone hasn't bought one yet, you should.  It's really one of the most interesting cards for the Apple II that has come out in years and it just keeps getting better.

Great work as always! Never did I think I'd be happy to see interlaced 480i over 240p; the surprises never stop :D 

Nice work on the 480i support.  Hopefully most HDMI up converters can handle the new video modes, that will make things a lot more flexible.

CVT wrote:

Thanks! On a CRT monitor 480i is still pretty bad and the flashing can give you a throbbing headache! :)

 

You get it; I just learned how to enable 240p on an arcade board that for some reason defaults to 480i and is installed in cabinets with roughly 27" CRTs. The interlaced mode is definitely... "shakey" when you're sitting 18 inches from a tube that big :'D 

CVT wrote:

Interlaced 480i to the rescue.

Thanks!  This will be a big help and make the card even more enjoyable.

This is very welcome news. I tried every LCD/plasma in the house as I mentioned and didn't really come too close except for this 19" LCD TV I've been using (which locks on but doesn't have the resizing necessary). For fun I may hook up this CRT I got just to see how bad the flashing really is, but I believe you, it wouldn't be pleasant. :)

Now I'm chasing another problem which may or may not be related to the funky mouse interface card issues I had. The card has been starting up reliably ever since I dug out whatever crud was on the interface card's chip, but every now and then when I'm doing something with the ESP32 (or even nothing at all like it's just sitting there at the prompt, or when streaming audio), the keyboard will lock up and won't accept any input. The card seems to be running fine because if I'm streaming audio, it keeps on streaming and will update the screen with whatever comes in, but no keyboard input. Today it did that (just a few minutes ago) with a brand new symptom: one of the physical floppy drives did a head seek and just started spinning. I had a disk in there but didn't have it closed, so that was unexpected. Happened when I was streaming some online station, which kept playing, until I heard that drive seek and saw the disk activity. No keyboard input... I just shut it down.

Of course with all my tinkering, and knowing that the mouse interface card had the previous issue, I'm suspicious that it's still doing something weird so I'll probably just pull it out again for now. I could just get another interface card. I was hoping I'd resolved the issue but there may be more to it. I could even desolder that entire chip and give it a good scrub in case more crud got under it. Of course, just my luck, it's the one chip on there that's not socketed... but I do have a desoldering gun so it hopefully wouldn't take too long to do that just to make sure, recheck all the other traces.