[[4. If yes to #2, what size should I make the PCB to fit into most cases -- ex: GK-01 style? 2 digit, 3 digit?]]

GK-01 fits in this type of case:

But for myself I would prefer a full three digit display in order not to be limited to 199 Mhz - like this:

but it would likely be too large to fit in 286/386/486 cases without some modding.

I'd be interested in the PCB or plans. The thought occurs to me I could do, for example, a blue-themed case or a red-themed case instead of the traditional green and have the turbo display match.

I have a couple of cases with the window but no LEDs.

wiretap wrote:

I was thinking about undertaking a small project in my spare time (which I don't have a lot of) to create a new turbo display. Y […]
Show full quote

I was thinking about undertaking a small project in my spare time (which I don't have a lot of) to create a new turbo display. Yes, I have plenty of old ones laying around, but I wanted to create something from scratch. I wanted to make this a more universal application so that I could share the design that fits in most AT cases (max dimensions, hole mount pattern ideas?) -- and I also want to design a 5.25" bay slot cover that can be 3D printed that has a turbo display cutout, turbo button cutout, turbo LED cutout, and other various cutouts that could be fit into an AT/ATX style case that doesn't already have an integrated turbo display hole. All parts would be easily obtainable off-the-shelf minus the PCB and optional 5.25" 3D printed bracket.

TLDR version:
1. Designing a PIC processor-based turbo display. (in DIP socket for easy install/removal)
2. Is there any interest in this if a design was shared for download, and/or if PCB's were offered for DIY kits?
3. If yes to #2, which basic design would you like to see? -- i.e. how should I make the pins hook up for max compatibility?
4. If yes to #2, what size should I make the PCB to fit into most cases -- ex: GK-01 style? 2 digit, 3 digit?
5. Any other random suggestions for the design would be helpful.

Thanks!

As of 1) - may I suggest designing for an Arduino rather than PIC . IDE and programmer for Arduino is so more easily available. Also it comes on a board for less than $5, ready to rock n' roll ..

I considered the Arduino, but they seem to cost is a lot more (the PIC is $4 in single quantity + $8 for a programmer = $12 total), and to have the pin count I need would require a much larger board that might not fit in the footprint needed. If there's a tiny Arduino option that's comparable, please let me know and I can modify the design. I haven't used Arduino's for any projects yet, but I'd be willing to do it.

- 2 pins for power (+5V and GND)
- 9 pins to drive 7-segment display (2 for driving transistors, 7 for the display segments)
- 1 pin for turbo switch status input
- 1 pin for driving turbo status LED
- 2 pin output for letting the motherboard know H/L turbo status

Anyhow, here's a very basic view of the PCB without traces, and not drawn with proper edges/sizing. (just the components laid out from schematic) -- I'm using Kicad for this. I still have to draw the actual board, mounting holes, and rearrange some components.

wiretap wrote:

I considered the Arduino, but they seem to cost is a lot more (the PIC is $4 in single quantity + $8 for a programmer = $12 tota […]
Show full quote

I considered the Arduino, but they seem to cost is a lot more (the PIC is $4 in single quantity + $8 for a programmer = $12 total), and to have the pin count I need would require a much larger board that might not fit in the footprint needed. If there's a tiny Arduino option that's comparable, please let me know and I can modify the design. I haven't used Arduino's for any projects yet, but I'd be willing to do it.

- 2 pins for power (+5V and GND)
- 9 pins to drive 7-segment display (2 for driving transistors, 7 for the display segments)
- 1 pin for turbo switch status input
- 1 pin for driving turbo status LED
- 2 pin output for letting the motherboard know H/L turbo status

Wouldn't you want 10 pins for display, just to be able to get the 3rd digit also? 😉

Even the Arduino Pro Mini has 14 digital IO pins beside power and 4 Analog inputs (which could be used for digital inputs pretty easily.) The same goes for Arduino nano, which has own programmer/bootloader to connect with USB.

All thogh, if what you are planning is a "complete" design with the microcontroller onboard, I see the idea of the PIC - what type of programmer is available? Maybe a lot has happened since I worked with PICs the last time, but back then it took a lot of tool-chain installation etc., to get it up and running, where the Arduino is like a walk in the park to install tool chains. Also the Arduino boards (and programmers) are very easy available on ebay, Amazon, AliExpress etc.

Why do you need so many pins ? Nowadays you can get 7segment displays to which you can write in the number in serial. This way you can do this with a simple ATTiny chip !

I can easily add to the current design for a 3-digit display. I will create a variant for that.

For PIC programming dongles, they are easily found on ebay, Amazon, aliexpress as well.

I will continue to look into using an arduino. Just need to find the right footprint to fit, with the right pins available.

For the serial addressable 7-segment displays, those usually have a daughter board attached which makes the size or thickness too big depending on where it is installed in someone's case. Unless you can recommend a 2 or 3 digit serial display that just has through hole mounting and isn't almost an inch thick.

The design I currently have is about the cheapest/simplest I could make it with common components, not doing anything with 555 timers, led drivers, shift registers, etc. Keep in mind, this needs to fit behind a front case panel, so size is limited in many situations.

I looked at doing this as well, but quickly came to the realization that no two cases in my collection had exactly the same footprint. Also, 3-digit LED modules are really really hard to find -- at least ones that will fit in the space that most 2-digit AT cases have available. The digits need to be more compact than most generic displays with fixed pitch regardless of digit count.

FWIW, I did end up designing one module to re-use a 3-digit display from an existing case. The display worked fine, but the module used a proprietary controller chip (epoxy blob) that appeared to be DOA. I used an AVR (nekkid, not Arduino) and a cheap DIP switch for setting the digits:

I sent off for the PCBs, placed an order from Digikey for some of the parts that I didn't have on hand, and spent a day coding the ASM source. I then happened to find a better case for that particular build, so by the time this PCB came back from Timbuktu, it had fallen way way down the list of priorities, and now sits on a stack of other temporarily abandoned electronics projects on my workbench.

I took a middle-of-the-road approach to wiring. I figured, it's not really important if "Slow" and "Fast" modes get mixed up, so you can just connect the input (with an internal pull-up resistor) to the motherboard's Turbo LED output, or tap the turbo switch's Common pin en route to the motherboard. (Especially nice when the turbo switch is 2-pole, so you can bridge the center pins for each pole.)

I got the PIC-based design pretty much laid out, just have to route traces. I chose to make it the same form factor as the SFI-501. The top left front side mounting hole lines up with 3 different cases I have, and 4 other turbo displays, so that should be a decent universal fit. This design is also small enough to fit in a 5.25" bay cover.

SFI-501 for reference:

My design:

PCB layout:

Schematic:

Design changes that need to be made:
- Switch the mounting holes to the opposite sizes.. accidentally did them in reverse since I overlooked that I was working on the back side layer of the PCB, not the front side layer. 😵
- I'm going to use some smaller wattage resistors laid out horizontally instead of vertically.
- Use physically smaller transistors.

Next up is to test it on the breadboard. I'm leaving for a business trip today, so that will have to wait until next week.

Then of course I'll route the traces on the PCB in Kicad, and order some boards for a first run test.

Next iteration will be a 3-digit, and possibly Arduino based. If Arduino based, I think I would make an interposer board with breakout cable of some kind to save space -- this way you could have the small turbo display PCB mounted in the case, then attach an Arduino of any size, as long as the pins matched up. (i.e. just an i2c cable with power and ground going to the display portion)

You guys do know that, if you have an existing Arduino, you also have an AVR programmer -- right? You can just buy bare chips and flash them with the AVRISP sketch, or whatever it's called. No reason to use full-fledged Arduino dev boards -- even the tiny ones.

Lots of single resistors... Can't you use something like this? https://www.electrodragon.com/wp-content/uplo … k-resistors.jpg

Also, if you are going to have a full blown microcontroller there, make it controllable through some software. 😉 Start maybe with COM/serial connection and maybe in the end you have part of it running to ISA slot? 😉

And well, then you have the turbo button as input, but you can set each state to show whatever you want, from software.

If somebody asks me about one single most useless thing in the retro computers - it has to be the LED clock display. If taking such effort - why not make it actually usefull and show real CPU clock speed for example. Simple arduino powered frequency counter will measure up to 66Mhz with ease - which is good for everything upto socket 7 era... and you can always use it for different things like temperature display too...

I started going down that rabbit hole. Also had planned to attach the keylock switch, so little twitches from unlocked to locked would cycle through four settings of loudness on the PC Speaker via a resistor ladder and some FETs. By the time I had a quarter of the features I wanted, the module's PCB was already overflowing with pin headers and TH parts.

An entire generation survived by thinking the "8" or "16" of their non-turbo mode was anything even resembling accurate, while in reality, nobody had any idea what actually happened inside the box to make it slower.

Eh, it's just a novelty. It looks cool, it doesn't have to do anything useful. Let it be. 😉

Alrighty, I suppose I'll just make a few of these for a personal project since there doesn't seem to be much interest. Thanks for the feedback. 😀

As to using a resistor network, no, because each individual control line has to have its own resistor going to each segment --- a resistor network has a single common pin. If you were to use a single resistor on the common of the 7 segment display, you get varying brightness per segment.

wiretap wrote:

As to using a resistor network, no, because each individual control line has to have its own resistor going to each segment --- a resistor network has a single common pin. If you were to use a single resistor on the common of the 7 segment display, you get varying brightness per segment.

Really? You don't have to do it like that. Reverse the signal from mC itself or something, but (without going super deep into project) it should be possible to put all the resistors for each segment on a resistor network with a common ground or 5v.

But maybe i'm missing something right now quickly glancing over the pictures here...

GigAHerZ wrote:

wiretap wrote:

As to using a resistor network, no, because each individual control line has to have its own resistor going to each segment --- a resistor network has a single common pin. If you were to use a single resistor on the common of the 7 segment display, you get varying brightness per segment.

Really? You don't have to do it like that. Reverse the signal from mC itself or something, but (without going super deep into project) it should be possible to put all the resistors for each segment on a resistor network with a common ground or 5v.

But maybe i'm missing something right now quickly glancing over the pictures here...

Feel free to draw a quick schematic showing a better way. (You still need a current limiting resistor per segment.)

Yeah, you can't use common-pin packs for an LED segment with a common anode or cathode. You need per-segment control, current-limited. You would need external transistors, external individual resistors, or compromise and limit the current at the LED's common pin -- which will affect brightness depending on number of lit segments. (Doesn't look good, IMO, unless your speed is 33/66 or something else where both digits will be equally bright.)

They do make resistor packs that are individual resistors with in and out pins, though.

Update for those interested.. used Freerouter to run the traces, and it turned out pretty decent. There's only two 3 vias on this dual layer board. I am a little bit nervous about the top traces being so close to the PCB edge and the others by the upper mounting hole, so I may manually route those next.

I also corrected some schematic errors, labeled the jumpers (probably should add polarity next), and fixed the mounting holes (swapped left/right side).

I did a design rules check and it passed in Kicad, and also test uploaded it to Oshpark -- passed there as well. This is my first circuit board design, so I'm pretty surprised -- I'm sure something is wrong. 🤣

Next up is putting it on the breadboard and writing some simple C code.

Looks pretty good. I think I would move the 3 headers next to the mounting hole a millimeter or so closer to R5. Do you have enough room for the mounting screws' heads? That is one classic "oops". Also consider big clearance around mounting holes on the side, where the screws head is, so you won't cut traces when mounting. What clearance do you have between edge of board and the traces?