RLGaming.com

I am trying a different approach to figuring out what is accurate and what is not. I am attempting to mathematically calculate the unit scaling of the imported models and then measure parts and compare them to known dimensions to see if any of them are accurate or if it's just some of them that are inaccurate. There is guy on my OpenHarrier Facebook group that works on the SJU-4 seats and has access to them who is going to get me some good fresh measurements.

So the math didn't quite work out. What I did discover is if I scaled the important model exactly to 1050x, the follow dimensions become correct in the Razbam model:
* width between the insides of the lower part of the seat: 18" (measurement verified by a guy that works on these)
* width between the inside walls of the left and right consoles: 22.06"
* width of the console instrument plates for the left and right consoles: 5.75" is the correct width, and these are within 1/8" of this.

The following become incorrect:
* width of the MFD's are supposed to be 6.5" but the Razbam model comes out as 7" (which is incorrect)
* width of the UFC and center pedestal come out slightly incorrect (too big compared to the real thing)

I am still working on getting other measurements to determine exactly how big the front instrument panel should be. What I can surmise is that the Razbam model is just plain incorrect on some dimensions. You can never scale the model and get everything to match up with the correct dimensions because there are errors in the Razbam model.

Speaking of which, I have discovered that the real left and right consoles have a downward slant to them from front to back that is NOT modeled by Razbam. This slant is about 10 degrees down slope relative to the floor of the cockpit. The razbam model has these 100% horizontal with zero slant - which is absolutely, verifiably incorrect. I confirmed this with some folks that work on Harriers just this afternoon. Not sure why Razbam got that wrong. Truthfully, this detail really doesn't make much of a difference and I doubt people would notice. I can easily replicate this tilt - I'm still trying to determine whether I will do that. It's something to think about. Keeping the consoles flat works fine, too, and is what the other Harrier builders have all done. Nobody has tilted them downwards like they are supposed to be. I guess I'd be a trailblazer if I go forward with that.

Anywho, I'm kind of on a construction pause anyway and won't be able to physically work on the cockpit for at least a couple of weeks for unrelated reasons. But the good news is, I don't believe I have to rebuilt my seat anymore. The only thing I need to recut are the center pedestal and front instrument panels - and I already was planning on cutting the instrument panels again anyway because I did a sloppy job cutting them out the first time.

Rebuilt the side consoles taking into account the 10 degree downward slope of the consoles towards the rear. The center pedestal and instrument panel have not been rebuilt yet correcting for known scaling issues.

Man, that looks awesome.....

Can I stop by sometime to...drink that Tyskie?

Looks great!

v6,
boNes

Sure, man! How about this weekend? My wife and kids are leaving for Florida tonight and I'll have the whole house to myself for over a week.

Oh cool! Right now Sunday is fairly free...

v6,
boNes

Cool, how about Sunday evening? I want to maximize build time in the garage with my wife not being here.

Got a few days of solid building in this past weekend and made a bunch of progress. It's really starting to take shape!

You and Creeper put my handyman/building skills to shame!

Ohh man thats a beauty...

You have any plans to wrap and fibreglass it... sand it down and paint it with a partial livery??

It will be sanded, skinned, filled, and painted, but not with fiberglass. That much fiberglass would actually be pretty expensive and not worth the effort. I do have some fiberglass experience. Actually the headrest is fiberglass, and I still have do the seatback in fiberglass. But the cockpit outer skins will be 1/8" sheets of plastic most likely, which will be easy to cut and install, and will bend easily to the curves I need. Same with the glare shield. Same with the inside skins. Originally I had designed the outside structure to be an approximate scale of the real fuselage, but it would have also made the support structure too weak not being made from metal - especially with regard to supporting the canopy ring. I made the outside radius of the canopy ring greater than the real thing (resulting in a thicker canopy ring) for purposes of strength. If I had made it the accurate outer dimension, it would be too thin for MDF and would easily snap if somebody grabbed it and tried to pull themselves up by the canopy ring. Anyway, to support the canopy ring, I also extended the outside walls by an additional 1.5 inches on each side. That is again not accurate. I'm more concerned with making the inside cockpit dimensions and inside walls accurate - which should be pretty close to the real thing when it's all said and done. (The Inside skins will require some minor additional interior framing that I haven't done yet). If I want to make it look like more of the shape of the Harrier fuselage, I would also need to make the nose narrow in, which I have not accounted for - but that could be done later.

That makes sense. I like the idea of using plastic sheets. You going with a grey or going to have to paint it yourself afterwards?

Trichome wrote: ↑

21 Feb 2023, 19:47

That makes sense. I like the idea of using plastic sheets. You going with a grey or going to have to paint it yourself afterwards?

I've just assumed that I am going to have to paint. Everything inside gets painted the same shade of gray with the exception of the panel plates which need to be painted a flat black. Outside skins can be different shades obviously - no idea what I'm doing for that. Probably a long way off on that anyway.

That is some awesome detailed woodworking!

MFD's test fitted.

Started building out the support framing for where the UFC will go. The HUD components are just placeholders for now and not actually attached. Also added the box that will eventually hold the standby instruments (analog gauges).

Looking great Xpen. Its really starting to come along now.....I cant wait to see this all done.

Started work on some electronics for the OBU (option selection unit). These are 5 buttons next to 5 display units that have 4 x 16 segment alphanumeric characters plus 2 additional segments to represent a "selection" colon.

This breadboard has the the stuff for 1 of these displays. The displays themselves are custom manufactured in China for a guy in Australia that makes and sells high end sim equipment for us hobbyists. These happen to also be used on the F-18 UFC.

Unfortunately, I haven't gotten this to work yet. I had to drag out the oscilloscope and confirm I am getting data on the i2c data bus. I have a hunch that the LED driver chip I am using is not supplying enough voltage to light the segments, or something is wrong with the chip itself. The displays themselves work as I can light the segments manually.

I just need a little more time to figure it out.

Technical jargon / nerd alert...

So it turns out the 16 segment character display units I am going to use are commode anode. That means that the positive side (anode) for each character is connected to a different pin and it is the positive for every segment in that character. The negative side (cathode) is on 16 different pins, each pin corresponding to the same segment on every character. For example, the top left segment of all the characters are all connected together to the same cathode pin.

These displays have 4 characters with 16 segments each, plus 2 additional segments to represent a colon on the left side. 16x4+2 = 66 segments. I need to drive 5 of these displays from a single microcontroller, so that's 66x5, or 330 segments. Think each segment as a separate LED. No micontroller has 330 digital outputs. So what you do is use some sort of IC (integrated chip) to drive these displays. This allows you to control each segment of each digit using way fewer digital output pins.

The chip I planned to use was the HT16K33 which can control a matrix of 16x8 of LED's (128 LED's total). It uses the i2c (also called 2Wire) protocol which allows you to control the chip using only 2 pins from the microcontroller. 1 pin is used for a clock signal and another pin is used for the data signal. The data signal pulses between high and low (think on and off) rapidly at precisely timed intervals to represent data. The micontroller sends data to the chip, and the chip "remembers" the data and powers each segment on and off based on the last data sent. Here's the cool part. You can actually daisy chain up to 8 of these chips together using just the same 2 pins on the microcontroller. Part of the data includes a chip address to tell the chips which chip the data is for. The chips have an address that is physically set by using 3 pins with resistors connected to ground or left not connected (called floating).

The displays have only 22 pins on them. How does the chip control all 66 segments if the chip only has 28 pins (24 for connections to the displays) and if the display has 22 pins? Well, what the chip does is only power 1 character at a time by. It selects which digit to power and turns on the segments based on the data it has stored. Then it disconnects power to that digit and powers the next digit and the segments as per the data for that character. In short, it does this over and over again really fast. Your eyes can't tell only 1 character is energized at a time.

So basically this chip has 16 pins (1 for each segment, tied to the same segment on each character) and 8 pins to power each of up to 8 characters. The chip simply powers up each position on the character display by turning on one of 8 character pins and turning on or off the 16 segment pins based on the data it has stored for that position. As I said, it does this really, really fast.

Since one of these chips can power 8 characters, I had planned to power the 4 characters from display 1 and 2 on 1 chip, the 4 characters from displays 3 and 4 from another chip, and the 4 characters from the 5th display from a 3rd chip. That 3rd chip would also be used to control the 10 addition segments for the colon characters on all 5 displays.

Anyway, as it turns out, my displays from Tek Creations are not common cathode, they are common anode. Which means I can't use these chips because the polarity is wrong for a common anode display. Common anodes are not actually that (ahem) common. I will have to switch to a TM1640 chip. The concept is the same but it will require a separate 2 pins for each chip. And these chips are only available in China for some reason. So I have some ordered but it will be about 2 1/2 weeks before I see them.