Project Logs

Collapse

• Summing up

  KJoh • 09/21/2026 at 00:08 • 0 comments

  That 486 was a ton of fun to work on, came with a half-dozen repairs and mini-projects, and the Turbo display was the topper to giving this computer a new lease on life.

  As I mentioned, I did eventually sell this computer.  In the bottom of the case, I bundled up the printouts of the schematics and arduino code in case anyone wondered what the heck this computer was about.

  All of the work done could be easily reversed in case anyone objected to the "non-standard" configuration of the machine.

  After retrobrighting, and the addition of a new case badge:

  I hope you enjoyed this write-up.  Thanks for reading, and happy retro-computing!

• From breadboard to protoboard

  KJoh • 09/20/2026 at 23:55 • 0 comments

  The way I chose to implement this was to install a 2-digit display where it was meant to go, and have the arduino mounted in the bottom of the case, connected together with a cable with sufficient conductors for power and signal between the arduino and the display.  Below are some photos of the protoboard version:

  At this point, I had transferred the code to an Arduino Pro mini.  Because reasons.

  Here's some video of me testing the circuit.  You can see the external USB to serial adapter I used to upload the code, necessary because the Pro mini did not have an on board interface of its own like most other Arduinos.

  https://youtu.be/g-BjOMZr67c?si=kI5Xqb4GFN_D65Uo

  Unfortunately, I can't find any photos of the install into the case.  To sum up, it was a small bracket to mount the arduino in the bottom, a run of cable, and lots of hot glue.  What hackaday project doesn't call for hot glue?

  If I find any install photos, I will upload.  Unfortunately I had passed the machine on to a new owner a couple of years ago in an effort to downsize.  (sad face emoji)

• Adding the Special Sauce

  KJoh • 09/20/2026 at 23:46 • 0 comments

  The objective was to do something different, so this is where the Arduino comes in.  At first, I worked with a nano and a 4-digit display, as that was what I had on hand.  Here's some shots of the breadboarding:

  And the schematic:

  What I came up with was a race pattern to provide some animation to what was displayed, with the resulting clock speed being a fixed 25 or 66.  The arduino code can be modified to do whatever the user would like.  For example, one enhancement someone suggested was making the race animation accelerate or deccelerate versus a single fast or slow speed.

  It's easier to view the linked youtube video than describe how it works.  

  The arduino code is in the attached files.  I implemented this as a state-machine, my go-to move for any project like this.  Interested to hear any comments on the approach.

• Messing with an Arduino

  KJoh • 09/19/2026 at 23:44 • 1 comment

  No, I didn't try to use a 555...

  My first approach is what I eventually went with.  I could read the state of the Turbo LED to determine the mode.  The arduino could then control a display accordingly.  Here's a sketch of the wiring:

  I fed the Turbo LED to an arduino nano as an experiment, with a 10K pulldown resistor to avoid any dimming of the LED brightness (see comment below).  An alternative was to sample the Turbo switch, but this was most convenient, worked the first time.  In hacker fashion: if it worked, keep moving forward.

  5v power for the arduino came from a spare molex power connector in the machine.

  This arrangement reliably determined the Turbo mode, without affecting the operation of the computer in any way.  Next, it was on to the display.  What should I do?  The next log will be published in a day or two.

• 66MHz and... 25MHz (????)

  KJoh • 09/19/2026 at 23:31 • 0 comments

  Landmark Speedtest gave a surprising result.  The fast clock speed of 66 MHz made sense given that the CPU was a 486DX 66Mhz.

  However, the slow speed result (Turbo OFF) was 25 MHz.  Did that make sense?  You'd think at first it would be, for example, half of the base clock speed, or some other predictable fraction.

  Looking into it further, I learned that slower machine speeds were often achieved by messing with RAM wait states and other such tricks.

  No matter, it was immaterial what the actual number was.  I could trust Landmark Speedtest to give me the effective speed, and I could always change the number later if I learned any different.

• Option B: Something with some pizzazz

  KJoh • 09/19/2026 at 23:25 • 0 comments

  I decided to go a slightly different route.  I wouldn't try to make the display a real-time measurement of the clock speed.  It would still be something hard-coded to make it simple, but with more than just a clock speed A versus clock speed B.

  The first step was to determine what the actual clock speeds *were*.  Here is a shot of the motherboard, extracted when I was first working on the machine.

  I tried the usual sources for identifying the motherboard, but came up empty.  If anyone can help, I'd be interested.

  (side note, one of the things to work on was to replace the old battery on the motherboard (seen in the upper right).  The replacement fitted was a CR2032 with a diode to prevent charging - another project, done many times by people all over the web.  There was some minor leakage from the old battery, and any degraded traces were easily repaired.).

  To determine the clock speeds, I therefore resorted to using Landmark Speedtest, a tool I've used in the past.  More info here:  https://dosbenchmark.wordpress.com/research/landmark/

• Option A: Conventional Clock speed display

  KJoh • 09/19/2026 at 23:14 • 0 comments

  As many will tell you, those old clock speed readouts didn't actually measure the clock speed, they just display whatever you decide to display by means of jumpers.  The turbo switch did the rest.

  You can read more here:

  https://forum.vcfed.org/index.php?threads/turbo-button-jumper-settings.38001/

  A typical example is shown below (source: Vogons):

  Is this what I wanted to do?  I figured it would be fun to try something different.

• First looks

  KJoh • 09/19/2026 at 23:07 • 0 comments

  This is the machine as I first got it, with the cover off.  Quite yellowed (I would later retrobright it), and missing any sort of clock speed display on the front panel.

  Below is a view from inside the case (apologies for the focus).  You can see the reset and turbo button header connections, the LED headers, with a blank space on the right where a display would be.  It would fit a two digit 7 segment display, as would be customary.

View all 8 project logs