Kerbal Instrument Panel: In-Desk Apollo Themed Hardware Controller

[richfiles]

Got the small single outlet mounted. I'll secure the wire into the corner later.
It's a polarized outlet, and neutral/hot have been double and triple checked as correct.
You can see a simple lamp plug end inserted into it below. Simple as.

On a side note, after cleaning most of what was on the old desk, and doing a little cleaning in the kitchen, I opted to do another project.
I gutted a pair of Gen 4 iPads for their LCDs, and made a pair of nice monitors from them. One has the housing flipped, so the controls are on opposing sides.
The plan is to mirror the pair of iPad panels from my workbench to this new second set, so I have the same screens available for my workshop's desk.
You can see one of the two screens powered up. They both are tested, and work great! I'll mount the vesa mounts to a pair of rails, and mount that to an arm.
The cases are really nice, though I have no plans to redo my homemade one... I like how homemade it is. 

Again, I streamed it, so if curious, you can see just how a pair of iPads are turned into a pair of monitors.

As a final note... I saw the post about the status of the KSP forums...
I hope that they continue to be supported for a long time, but I still have my worries.
I have just backed up this entire build thread, and I hope to back up more of my posts.
I really do hope that this corner of the internet doesn't just up and disappear, but...
Take Two is... Not a company I have ever trusted. 

Edited July 8 by richfiles

[starship26]

Wow!

[richfiles]

Small update.

I built the pedestal to raise my monitor arm by 2 inches (close to 5 cm).
This is simply so that I can mount the monitor to the desk, and know the arm is high enough to not collide with the console.
It's made from scraps leftover from building the desk, so the wood grain, thicknesses, etc all match up with the rest of the desk.
For all of six seconds, I thought of cutting plain rectangles and trimming the sides... Then I decided to take the extra steps to make it fancy! 
The taper of the lower section kinda gives me the vibes of the stepped desktop sections, and the rounded bit is the same as the front edges.

I think it looks quite nice, and I look forward to mounting it!

I have also extracted the CRT from the old Brother Word Processor. I'm realizing that the time to mount it was last year, when I was cutting the first desk pieces.
Basically, I need to fit it into place, and I need to do it the hard way, or the even harder way... I really only have two realistic options...
I could fully disassemble the left desk segment (additive construction is a real kraken if you need to modify literally part number one).
That would allow me to pull the part I need to modify off and work on it incredibly easily. The other option is to remove the lower brace and cut an opening in situ.
This is harder, cause I need to make precise cuts in awkward, tight spaces and angles. If a tool slips, I risk marring the rest of the desk. 

One thing that I did do, was to cut the bezel out of the original Word Processor housing. The mounting holes for the CRT are on the back too.
I have an *awful* router that's not up to the task, but I also know where I can borrow a better router... If I use a follower bit, I could mirror the opening.
Those bits have a bearing the same diameter as the cutting path, and they will roll along a template... Like this bezel, and perfectly mimic the edge.
If I screw the bezel to the wood, I can use a straight follower bit to cut the opening, and then use tapered follower bits to follow the opening and cut the bezel edge.
The big catch 22 is that i need to screw the bezel down... and the inside of the desk doesn't have clearance for the router...

I'd have to have screw holes on the outside surface...

I do not like that one bit... My one option might be to find a follower router that has the bearing at the tip of the cutter, and not the base.
I could screw the bezel to the inside, then plunge the bit and cut till the bearing meets the bezel on the inside.
It might work, but I also risk mangling everything... It makes me nervous...

I absolutely adore the crazy form factor of this tiny ultrawide CRT! There's honestly a funny reason for it's specific size too! It's got some crazy roots!
This little widescreen CRT will fit right in with my 49 inch LG 5120x1440 ultrawide and my 3840x1100 mini-ultrawide on the desk step on the right side.

So I mentioned roots... Both this CRT's dimensions and the Intel's success  have roots in the CTC Datapoint 2200 "smart" terminal.
The Computer Terminal Corporation (CTC) Datapoint 2200 was a serial "smart" terminal, but could function as a small stand alone computer as well.
The 2200's CPU wasn't one single chip, but rather many small logic gates tied together on a motherboard to create a multi-chip processor. 
CTC contracted Intel and Texas Instruments to try and develop their CPU architecture into as few chips as they could, ideally, one chip.
TI tried but ultimately failed to create a viable chip. Intel succeeded, but were delayed...
Datapoint went on with a multi-chip processor for their next model, not using Intel's chip.
Intel and CTC still were in an awkward situation. While delayed, Intel did succeed in creating the one chip CPU... But CTC no longer needed it.

CTC and Intel came to an agreement. CTC would keep their money, and Intel would get access to the CTC Datapoint 2200 instruction set.

The Datapoint 2200's instruction set was the model for the Intel 8008 CPU, and it's instruction set was expanded into the Intel 8080 CPU.
The 8080 instruction set was then further expanded to support 16-bits and became the x86 instruction set used in the 8086 and all of it's successors.

This is the instruction set that forms the core foundation of every x86 or x64 CPU in use in modern PCs today! CTC GAVE it to Intel!
The 8008, 8080, 8086, 80186, 80286, 80386, 80486, the Pentium, II, III, 4, Core series, Xeon... They ALL share those roots in the CTC Datapoint 2200 ISA

While that's a pretty neat bit of computing history... That still doesn't answer the glaring question...

What is with the funky widescreen CRT? 
Why was the screen for a terminal built in 1970-71, in an era where most CRTs were far more boxy and not wide... Wide?  
Turns out the reason was simple...

1. It needed to be the width of a sheet of paper, so you could see a whole line at once.
2. Since memory was so expensive and limited then, not displaying the whole height of a page saves video RAM. Just scroll and pull in lines from main RAM as needed.
3. It was modeled after a very familiar shape in the computing world... For the era... The screen is sized to the dimensions of a standard IBM paper punchcard!

This squat CRT proved popular on a few machines, including some of Hewlett Packard's PCs, economy data terminals, and word processors, such as this Brother model.
Long after the IBM punchcard had gone the way of the dodo, this CRT form factor carried on in niche applications, until LCDs truly replaced the CRT for good.
While this tech might be well and thoroughly obsolete, I absolutely adore the old look, and want to reuse this tech for my Kerbal controller project.

I mean, look at how narrow this skinny lil' guy is! I just love this CRT!
It'll pair nicely with my little 3.5 inch color RCA CRT. In portrait mode, it's width is nearly the same as this CRT's height. Perfect!

 <-- skinny 

[richfiles]

PROGRESS...
I will split this into two posts...
Right Desk Segment and Left Desk Segment...

The monitor arm pedestal has been mounted! I'm really glad I decided to fancy it up a wee bit! 
If you look closely at both the image above and below, you will see a metal "sleeve" in the central monitor mounting hole.
The monitor arm uses an M10 bolt to mount it to a surface. The included bolt was not long enough to pass through the desk surface and the pedestal.
I got a longer bolt, which is fine, but I don't know if over time, the added bolt length might warp the bolt.
Because of this, I purchased a 10mm steel capillary tube with a 1mm wall. This gives me 12mm diameter of steel vs just 10mm.
The 10mm bolt fits snugly enough through the tube to not fall out by gravity, but can easily still be hand spun. A truly perfect fit!

It's a kinda rough cut, but I wanted better access to the space under the desk, and blocked by the cross brace mount.
The reason it is so rough, is I did not have access to my weak router or a jig saw... I roughed it out with a table saw. Jeb would approve. 

Also notched the cross frame brace. This just gives me better access underneath.

I'm realizing I should maybe add another screw or two, since I cut down that particular board. 
I have no idea why it cut it through the second screw hole. Could have moved it an inch over... 

Important thing, is I can access the monitor arm mounting bolt now!
This makes mounting monitors far easier! Who knew! 

Next up is the mini-ultrawide monitor that I want to fit onto the right side of the desk.

I had a pair of old cheap drawer slides that I will never, ever use... So I used them as not drawer slides.
Specifically, I trimmed the roller off the ends with tin-snips, and then mounted these internally threaded 10mm steel ball bearings to the ends.

Bet the dude that filed patent 5785400 never expected his idea to be used like this! 
Thing about this screen, is that it is a touchscreen. I honestly want to be able to take advantage of that.
I kinda sorta wanted a way to lay this screen flat, in case I ever wanted to use it in that orientation.

Just checking my mounts didn't break anything. Also, lying flat. I'll likely find some rubber feet for it at some point.

Here is the monitor mounted to the desk! How, you say?

Magnets, of course! They worked so well for my keyboard, so I'll continue the theme of magnetically detachable things here!

I created a template by placing the magnets on the steel balls and applying some paint to the backs, and pressing it onto paper.
Ran the screws through the paper, and got perfect alignment. The magnets are small cup shapes that the balls secure nicely to.

Cringe free video of the magnetic action... https://i.imgur.com/a6Cc2mN.mp4
I am so sorry, I was experimenting to see if people really do click on the cringe...
I feel dirty...

Anyway, it's getting so late that it's early...
I gotta work tomorrow, and the site 502/503'd so frequently, I wasn't sure I'd be able to post at all...

This marks the total completion of the Right Desk Half of my KSP instrument panel desk!

Work has been insanely busy (both jobs), but I'll try to post the results of the left segment tomorrow... 
Spoiler alert... It's at the same level of completion as this half of the desk!    

Edited September 9 by richfiles

[richfiles]

And now for the Left Desk Segment!

The big dilemma for me was removing the front panel of the left desk segment, so that I could mount CRTs into it.
Since the entire desk build was done as an additive process, old screws can easily be covered by new constuction.
To say I was concerned would be an understatement... The top desk and front panel were the first assemblies made!

The crossbrace obviously needed to go, as it covered the screws in the lower desktop that screw into the front panel.
At least those screws are really easy to access.

Here is the inside view... and... Oh thank the Lord.... All my worries could have been mitigated by just peeking inside...
EVERY screw is accessible!

Why did I ever worry about this! Front panel is removed! Simple as! Now I can easily mod it.

It's a clear shot inside, and what I'll do is mount a pair of drawer slides to either side, angled ever so slightly up.
The upward angle will ensure the CRT drawer naturally stays closed, and will keep the CRT drawer from scraping along the lower desk. 

There is a problem though... That front panel was 100% structural!
it was the majority of what was stabilizing the main lower desk! This NEEDS bracing!

Paper templates to the rescue again! I just need to cut out a board to support the lower desk at both ends.

Just a random shot of the removed front panel.
Again, this will be made into a CRT drawer, but till then, it can just rest snuggly in place to cover the opening.

Well... My choices were use the scraps I had, or go out to the storage unit. Scraps it is then!

Like a glove! A perfect fit! I am using the old front panel to re-secure the lower desk exactly where it needs to be as I secure it.

Screwed everything together and stained & sealed all the exposed wood.

MUCH more secure now!

Cross brace reattached.
I am going to plane that wood block to the same gap as exists between the lower desk and the cross brace, and add additional reinforcing.
This is an easy mod that I can add at any time, including after the desk is assembled.

It's a tight fit, but it is possible to access from underneath, though not really needed, since the CRTs will now be serviceable by way of a drawer.

Speaking of CRTs, I have extracted the small 3.5 inch color CRT from it's housing.

Kinda crazy how compact they made these old CRT TVs... considering the parts really aren't that compact.

Here is how these will be laid out on the front panel of the Left Desk segment.
The small color CRT is in portrait mode so that I can use it for IVA/KSC Kerbal portrait animations.
As I've mentioned before, the left CRT is from a Brother word processor.
It's basically a "Punchcard CRT"... An old aspect ratio/size that conforms to the old IBM punchcard standard.
They were used on early terminals and some early Hewlett Packard computers to make the machines compact.
Still gives you 80 columns, but doesn't waste valuable VRAM showing a whole page at once...
That's what scrolling is for!

As you can see, this barely fits vertically! The more I look at this desk, the more I wonder if I ever properly measured things!
spoiler... This entire desk was built off a single sketch with three actual measurements recorded on paper. That's it...
Everything after was subsequently pulled out of my... brain... You have no idea how truly off the cuff this "design" was...
Sorry for the blurry image. Took several shots, and my phone refused to focus at this angle...

Adding any angle to the shots let the camera focus, and you can better see just how closely things line up.

This image is flipped for your convenience. 
I just wanted to see how the desk would look with the CRTs more or less in place! I love it!

I'm thinking more and more about it, and I think the best way to cut the bezel for these will be to toss my cheap router into the trash,
and just mount the router bits into my mill, and carefully mill out the opening with the appropriate tapered bits. I think that'll work best.
Now that the front panel is no longer IN the actual desk, I can mount it and do whatever I need to finish it appropriately.

With that, the left desk segment is complete!
I still need to stain and seal the cable suspension pillar which will attach to the left desk segment.
All it does is string wires over a high traffic area, but this is just an attachment.
I can now FINALLY disassemble the old desk and reassemble the new desk in it's place!

It's done!

Edited September 10 by richfiles

[richfiles]

Work is gonna be extremely busy for the next two weeks, so, I probably won't get around to tearing down the old desk right away.
Just figured I'd add that. Got some hard deadlines I need to meet, and so I'll simply be busy. I'll obviously update once I get things put together.

[richfiles]

Pfft... 2 weeks. Goodness... At one point, I was covering for someone's 2 week vacation! It's been kinda crazy at work... Anyway, I have a solid 7 day break coming up over US Thanksgiving. I'm hoping to get back to this project then.

I am hopefully optimistic about this upcoming break, and about this apparent sale of Private Division and KSP by Take Two... I sure hope a good developer bought it. Maybe the Kitten guys? Who knows. Just hope it's anyone better than Take Two!

Edited November 9 by richfiles

[richfiles]

On 1/6/2022 at 6:12 AM, richfiles said:

I may have found a suitable means to control the LEDs for my rate of rotation meters... The PCA9685.
The outputs are persistent, so you can set an output LED or meter, and forget it, as it retains the output till the output value is updated.
Because the outputs are also PWM, this chip can not only drive varying brightness LEDs, but also control analog meters!
The PCA9685 is a 16 Channel, 12 bit, PWM Driver, with a highly addressable I2C Interface. Unfortunately the chips are also "semi-unobtanium".
Unfortunately, most stocks are depleted for the chip alone. Fortunately, there are still many circuit boards available using the chip.
Fortunately, you can still order servo controller boards using these chips, but it's one big board plus an inflated price, per chip.
The boards are also not ideal for me, as I still need to design a compact board to drive the LEDs in the bezel that will surround the Navball.
I'd have to desolder the chips, and at that point, I may as well just buy the one weird package in stock and do a board around that.
16 channels is a fair bit, but it'd still take...

Heh... I never noticed I got cut off here... What I would have said is that it would take 15-18 of those boards.
How many depends on whether I separate red and green driver channels and whether or not I combine segments of different bars together or not.
What's important though, is I just ordered 20 of these boards. That should be enough to handle everything... I hope...
If I need to, I'll desolder the IC and use them on a custom board, or I'll just wire these boards to a custom board by way of either a backplane or wire bundle.

The important thing is that I need to control 240 LEDs, and for some reason, the Soviets decided Green is common cathode, Red is common anode. WHY!?!? 
Anyway, looks like I snagged these from Ukraine just in time, back in early 2021. I absolutely love the segment size and shape. Nothing like common LED bar graphs.

The green LEDs will be easy to drive, but the red ones will be... also easy! These chips are PWM, and they are set and forget. 
All I have to do is address the correct I2C address and set the PWM. For thee Red LEDs, I need only tie the common anode to positive, and invert the PWM.
0% PWM would be full brightness, and 100% would turn the segment off. When the PWM is in the high state, that's essentially off to a common anode device.
The PWM being low means it's at ground, meaning current can flow between the low PWM state and the positive anode. It's actually really simple!
This is the same technique I use to drive the bidirectional Red/Green LEDs in the DSKY keypad. (+) to (- ) vs (-) to (+) can be achieved with only two IO lines.

I have a 6 day vacation coming up over US Thanksgiving holiday. I also have a root canal the Monday before Thanksgiving... 
Honestly, I'll be relieved. Yes, it's gonna destroy my wallet, but I'll be able to think clearly without an ever present pain distracting me.
The highest priority though... I want to install the desk. It's been done for months now, but between work and life, there's always something...

If I have to unplug my router to keep from being distracted... I'm gonna get this DONE!

[richfiles]

I mocked up a few iterations of the Rate of Rotation meters.
For size reference, the 80 LEDs are spaced 1mm on centers, so this is about 8cm long.

Since the segments are only 0.5x1.6mm, I think it might be a good look to light them in pairs.
I want to light up scale ticks at half-ish brightness to create a pleasing set of graduations to define the center and overall size of the bar. 
I'm also considering dimly illuminating segments adjacent to the illuminated cursor to help draw attention to those highlighted segments.

80 LEDs per scale... Three scales... 240 total LEDs!  

It'll be a lot, but I definitely want to incorporate this scale, not simply for the aesthetics of it, but for it's usefulness.
With nothing but these three scales, I could get out of a free spin relying on instruments only! Worth it! 

Edited 9 hours ago by richfiles