• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Design review: Analogous Board for SE and SE/30

obsolete

Well-known member
I've been working on converting one of my SEs from CRT to LCD. It all started with picking up one of @retro apple mods' excellent kits on Etsy, but that's only part of the equation. The other parts are how to power the system without the original analog board, and how to convert the monochrome TTL video to something a modern(ish) LCD will accept. I know that there are multiple ways to accomplish both of those, and I went through a few possible solutions until I hit on one I am happy with. I want a board that will mount in the place of the original analog board and slickly integrate with the rest of the system, so here's my design, which I'm calling the analogous board.

It uses guruthree's mac-se-video-converter firmware for RP2040, which I have tested on a breadboard and works excellently. I am using this panel and controller board with a 1 foot VGA cable. The barrel jack is intended for use with a 1 foot extension cable to power the LCD from 5V or 12V, selected by 0 ohm jumpers (corresponding jumpers exist on the controller board). My board has mounting holes for the SE power supply and fan in their original locations, though I still need to convert these to plated through holes. This is just a first draft, but I wanted to get it out here so that folks can see it and provide feedback.

The part I'm least sure about is my brightness pot implementation. I found a 1k ohm pot that will fit in place of the original, and it works great to provide brightness control by varying the peak voltage on the VGA signal from 0.5V (quite dim) to 1.0V (quite bright, and violating the VGA spec, but the panel seems okay with it). This works well on a breadboard but I'm worried that the long traces required to implement this on the PCB will act like an antenna and result in noise on the VGA signal. Does anyone have any thoughts? I'd like to hear everyone's comments on the rest of the design too. In addition to the screenshots below, the files are out here.

1728325221509.png
1728324310078.png
 
Last edited:

ymk

Well-known member
I would not run the video signals to/from the pot together.

You can avoid running them to the pot at all by using a transistor near the VGA connector to limit the output voltage. The pot would then control the base current.

If you must run video through the pot, put a strip of ground between those traces and have a ground plane on the bottom side.
 

aeberbach

Well-known member
Very cool - but you can get a rotary encoder in the same place as the pot, and control the output brightness a little more nicely that way. I do think you might get some issues with those long VGA lines, perhaps less so by shielding with copper tape over the traces but altering the color values within the Pico will remove that need.
 

obsolete

Well-known member
I would not run the video signals to/from the pot together.

You can avoid running them to the pot at all by using a transistor near the VGA connector to limit the output voltage. The pot would then control the base current.

If you must run video through the pot, put a strip of ground between those traces and have a ground plane on the bottom side.
Thanks. Yeah, I just threw the traces out there to the pot to illustrate the problem. I could certainly separate them and make sure they're surrounded by ground. This is a 2 layer board, so there is already solid ground plane underneath them. The length of each trace is ~179mm, so any noise they would pick up would be pretty high frequency, up in the UHF range. I still don't like the idea of running the video signal through a far away pot. It just looks like a dumb thing to do, but maybe it would work fine.

Unfortunately I haven't worked with BJTs since college, so I don't really understand what you're describing. If you're willing to spell it out for me a bit more, please do.

You got me thinking though, since this is monochrome video, each pixel is either on or off...I could just use a level shifter and set up the pot to control its output vref. Simpler still, I could also just run the 3.3V output from the Pico into the gate of an NFET and use the pot to control the voltage at the drain. Either of those solutions seem like they ought to work.
 

obsolete

Well-known member
Very cool - but you can get a rotary encoder in the same place as the pot, and control the output brightness a little more nicely that way. I do think you might get some issues with those long VGA lines, perhaps less so by shielding with copper tape over the traces but altering the color values within the Pico will remove that need.
Thanks. Yeah, I am not doing any of the resistor ladder DAC stuff with the Pico since I didn't see the need for monochrome video, but if I did, I could wire a rotary encoder to some GPIOs and do brightness control in software, which would be quite elegant...but I am trying to avoid having to do the software work right now :)
 

obsolete

Well-known member
@ymk Thanks for that. After spending a few frustrating late nights banging my head against the wall with ngspice in KiCAD, I switched to a Windows machine with LTspice and quickly came up with something that might not suck:
1728503477336.png

It's linear-ish through most of the adjustment range, and its current is well within what a RP2040 GPIO should be able to handle. Now to integrate it into the design.
 

ymk

Well-known member
@ymk Thanks for that. After spending a few frustrating late nights banging my head against the wall with ngspice in KiCAD, I switched to a Windows machine with LTspice and quickly came up with something that might not suck:

Looks good. (y)

If you do end up moving to software control, it looks like the Pico has some analog inputs available for the pot.
 

obsolete

Well-known member
Thanks. What surprised me is that I found very different values for R2 and R4 to work well in simulation depending on whether the transistor is a MMBT3904 (SOT23) or 2N3904 (TO-92), despite the fact that these parts are supposed to be interchangeable with only a packaging difference. Not sure whether this difference is real or a quirk in one of the models. I bought both parts and will play around with them in the breadboard before deciding anything.
 

ymk

Well-known member
The gain specs are fairly loose: 100-300 @ 10mA. A large difference in hFE between the models could explain what you're seeing.
 

CC_333

Well-known member
This is interesting!

Regarding the brightness pot, why not use it to control the brightness of the LCD's backlighting? Is there a reason not to?

c
 
Top