richfiles Posted August 4, 2017 Author Share Posted August 4, 2017 (edited) Yeah, I'm still around. Just a mix of stupidity at work, mixed with finding all 120 shrines in Zelda and... Minecraft... Dear Lord, I finally bit that bullet... Apparently, according to my stats, "richfiles has walked 171 km in Minecraft, and rode 25 km by mine cart, and played 144 hours..." You know... It's a lot, since I did build a rail station with pushbutton line selection, selectable 0-4 cargo cart deployment, and connected two villages with my super secret underground redstone powered base... Sadly, I discovered... a desert bordered by a savanna village... I want to find a village in a snowy area to the north west... Purely for the sole reason of having a rail line leading to "Borg Nordvestr"! Old Norse makes EVERYTHING sound cooler! It's funny that you got curious about progress today. After a couple months of no progress... Today I spoke to my boss at the machine shop about getting some time on the bigger mill to mill out the openings to the main instrument panel. I also looked into ordering female header sockets with flying leads... Only to realize I'm better off just soldering ribbon cable to the female header sockets I already have. Everyone sells individual pins for Arduinos... I want proper long sockets, and ebay is being stoopid™ (that or I just haven't found the magic (key)words to search by). I want to use them to attach the diode ROMS to the LED driver board, which will mount behind the main DSKY LED display board. I want to figure out a good LED driver as well. I ordered MOSFETs and transistors MONTHS ago. I need to stop by the screen printing shop and see if I can get an official purchase order... That is, if they are still willing to print the tape for my tape meter radar altimeter. I've unfortunately, had no replies regarding the images I submitted, so I'm not actually sure what is even going on with them. My biggest worry is the software... I STILL have no real grasp on programming C... I think I can cobble together the basics like analog meters and toggles and digital displays, since there is already base code to start from for KSP serial IO, but the prospect of coding the software to run my FDAI is daunting... I understand what needs to happen, just not how to implement the code to do it. *EDIT* One more thing... I had mentioned a while ago (in the simpit repository) that the structural rail for the control panel slide interferes with the leftmost edgewise analog meter. It's the entire reason I went and sold some of my single meters and bought more double meters. It let me eliminate some of the single edgewise meters, reducing my count to just four of the edgewise meters int he front instrument panel. One of my justifications for this was I didn't want to cut into the meter housings... I am now tempted to do this. It occurs to me, that if I cut into one of the meter housings, I can shift the four meters slightly to the left side, and net myself a hair bit more space. I've been concerned whether I have room for both the tape meter and the carriage meter. Moving the edgewise meters slightly over, even if it means cutting into a housing, gives me a bit more space to work with. Part of me is tempted to try to have a go at building the LED light boxes to direct the appropriate color LED illumination on the correct areas of the scales. I could check cut clearances then and see how that will turn out. Edited August 4, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
HighFructose Posted August 6, 2017 Share Posted August 6, 2017 Wow, most of this is over my head, but I am glad this project is still alive. I can't imagine how cool playing the game on this will be for you. One question, about how much will this project end up costing? (If you're comfortable saying of course) Quote Link to comment Share on other sites More sharing options...
richfiles Posted August 8, 2017 Author Share Posted August 8, 2017 LOLOLOL If I had half a clue what I've spent, I'd be fine sharing it. I'd have to figure it all out though... Hundreds, easily. Partly though, that's because I'm buying premium parts. You can get analog meters for a few bucks, but I've spent upwards of $10-40 a meter to get REALLY NICE ones. My FDAI cost a couple hundred bucks alone, and I still got it for an absolute steal. I spent way too much on LED displays alone, though to be fair, ebay refunded me for bad ones that came from China (skim the simpit repository I link to in the first message of my post to get the long term history of this project). I spent a lot on toggles with a fancy lever (and I'm not even sure if I like that style now...). I spent $70 on Futura typeface rubber stamps. I know I've spent $20-80 from ebay, dig-key, and other online parts sources as few times each... I spent $40 on aluminum panels once. Long story short... I have no idea anymore how much I've spent! Maybe that's a good thing! I have still kept my costs down from being even higher by salvaging and recycling many parts too. My enclosure is mostly recycled materials that I got for free. The important thing, is that you can spend as much or as little as you like, to be fair. Max7219 modules from China are DIRT CHEAP... If you get a good one, you can have an entire array of LED displays for about $10-20. Cheap toggles are in a never ending supply. An Arduino clone is all you need (still go with a Mega), so there's no need to go with some overly expensive kit or official branded model. My mega cost $9.99 from Micro Center. It's entirely possible to keep costs down, if you limit the amount spent on parts, go with cheaper parts, and maybe even do a little salvage. Quote Link to comment Share on other sites More sharing options...
richfiles Posted August 17, 2017 Author Share Posted August 17, 2017 Sometimes, you have to suck it up, and start on the more mundane tasks... I began cutting bits of copper sheet (actually, this particular sheet is VERY thin PC board material) and started work on the light boxes that will go into my five analog edgewise meters. Protoboard is cut and filed to fit the openings, and the edges of the copper traces soldered to secure the board to the light box. These light boxes will contain red green and yellow LEDs to illuminate the colored portions of the analog edgewise meters correctly. Outside the light boxes, the LED lighting will be the blueish-green LEDs that i previously acquired. This photo is not washed out, cause I had not selected the super bright LEDs like I needed... I'll have to desolder these and put the correct ones in. Red LEDs will go into the other slot, and the green LEDs will have no divider, as those color bars are only one color. I may need to do something different for air density... Possibly one long box with a bright blue LED shining at one end, and no LED at the other? Maybe? One down... 17 or 18 more to go... Like I said, sometimes you end up with some rather mundane tasks, that you just have to slog through. When all these are done, I'll also cut out side shields to stop light spill from the main illumination from BLASTING out the gap for the meter movement to pass through... Quote Link to comment Share on other sites More sharing options...
Shaun Posted August 18, 2017 Share Posted August 18, 2017 This idea is probably a bit too old fashioned, but what about a nice nixie tube countdown timer or something? Absolutely love the look of nixie tubes. Quote Link to comment Share on other sites More sharing options...
richfiles Posted August 18, 2017 Author Share Posted August 18, 2017 I absolutely love nixie tubes too, but they don't fit the Apollo style, unfortunately. I'm trying to stick with a fairly reasonably close NASA/Apollo theme. It's okay though, cause my PC case has an integrated nixie tube clock built into it! It also just happens to be the old wood cabinet of a 1939 Philco radio. Picture's a bit old. The PC is actually temporarily in an open chassis at the moment, cause it's on the bench being fitted with new upgrades and a reconfiguration from a slide out motherboard tray to a tilt out tray (that allows full length GPUs). One of the upgrades is a seconds display. Not shown, is a circuit I want to build that will take the signal from the drive activity indicator to spin a dekatron (ideally), or at least a ring of neon lamps. I'd prefer the dekatron tube, but I don't have parts for it yet. Quote Link to comment Share on other sites More sharing options...
Shaun Posted August 18, 2017 Share Posted August 18, 2017 So beautiful Quote Link to comment Share on other sites More sharing options...
richfiles Posted August 30, 2017 Author Share Posted August 30, 2017 Welp... I just spent money to order these keycaps... No idea if I'll even use them, but I couldn't say no... Guys... RIP JEB... Moar like RIP, MAH SIDES!!! Quote Link to comment Share on other sites More sharing options...
richfiles Posted September 5, 2017 Author Share Posted September 5, 2017 (edited) I'm a bonafide idjit... I'm glad I've been busy/lazy and not made any more of these stupid light boxes. I KNEW I'd designed the color bars to be that size for a reason... This is the part where the epic facepalm occurs. I'll need to order a couple of them. Specifically, the narrow versions of the short red and yellow, and the long green ones. Go figure, that I'd have green ones only in wide versions, and all my narrow yellows are long versions. I think I only have a single short, narrow red LED, so I'll need a few more. It's MUCH simpler though, than all the work necessary for those dumb light boxes! I get a lot of bleed through, thanks to the thickness of the plastic, so I might actually consider cutting out that space to mount the LEDs flush with the back of the scale. The secondary benefit of this, is since the LEDs would end up epoxied into place, I don't need a PC board to secure them, and I can run wires straight back, so there is no risk of anything shadowing the blue-green LED backlighting. On another note, I've started looking at viable parts to start constructing my digital readout carriage meter. I'm trying to decide whether to go with a stepper or a DC gear motor. To get a stepper with a threaded shaft is almost twice as much as a surplus DC motor that I found with a similar threaded shaft. I'd probably have an easier time controlling a stepper, but it's so bulky and expensive in comparison... The cheapest parts seem to be CNC and 3D printer parts, thanks to the volume those are seeing these days, but it's still approximately $35 on average for the stepper alone, with the threaded output shaft. It seems linear bearings are pretty cheap, and I have some steel rods already, for my linear guide. Man, it's SO overkill. Literal CNC parts to move a pointer with LED displays on it. Everything cheap though... It doesn't have the reach. I could try a setup with some manner of long arm, hinged further back, and get a different assembly with about 90-95 mm of travel, but it seems like so much extra mechanical work, when I know exactly how to do the whole job with the CNC parts. All I know, is I wanna order soon, cause I want to build the meter before it starts getting cold outside. I want time to make the meter (since it's the last instrument to be made or acquired), so I can know exactly how it'll turn out, and start placing parts and measuring to cut the main panel. I'll likely need to do that outside, so as to not bother the neighbors, thus outside. If it gets cold before I have all the instruments ready and measured... It probably won't happen again till next year, and likely no sooner than late April or May. And yes... I live in Minnesota, where the start of September is when we begin dreading the upcoming cold season. Actually, I once did that on June 21. Winter is coming... **EDIT** Ordered the stepper from Poland. Best size and overall length, and even with shipping, a very reasonable price. Shipping is a gamble, as it could arrive as early as the 13th, but it could also take till the 25th. Lets hope for the earlier date. I'll have to see what I have on hand for linear guides. If I can find an old scanner, I can salvage from it.**EDIT 2** PRINTERS!!! I realized every old inkjet printer dumped on me probably has a good quality linear slide... Now if only I'd played around with belt drive instead of screw drive, I might have paid nothing for the carriage meter. Oh well. The screw drive will actually be easier to build than some complex pulley contraption. Edited September 5, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
richfiles Posted September 6, 2017 Author Share Posted September 6, 2017 Ugh... Why me... I just spent $24 on a stepper motor from Poland. Won't arrive for between a week to 2.5 weeks... And I just found the PERFECT mechanism in a cheap scanner to do THE WHOLE WORKS... It's a plastic rack and pinion gear setup, and it's actually ideal, cause it completely encloses the rear of the slide. Since it's plastic, I can EASILY cut it and shape it and glue it and whatever I need to do to make it work... So yeah... totally wasted my money, when i had something I could salvage all along. Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 9, 2017 Author Share Posted October 9, 2017 (edited) Oh dear... It's been a full month, and I've literally done nothing. I have realized I can cheat on ordering the yellow narrow bar LEDs... I can just cut them in half. They contain 4 separate LED dies with 8 pins to power them. I can literally split them in the middle. It might be necessary to sacrifice one side to have clearance to cut one to shape, but that is no big deal. I have several dozen of them. I'll still need to order the red and green ones, unless I cut down a few of my larger ones of those too. Not sure if I have enough of those though. Didn't order very many of the wide reds in particular. If I mount them carefully, I could probably mill out opposing corners, and get two red narrow/short LED bars from one wide/long LED bar. Honestly... It'll be easier to just order the correct size. With how busy work has been, I probably will get them before my next build session anyway. I decided to grab another electric typewriter I found at a local thrift shop. It has a long linear rack gear, and the stepper has a pinion that rolls along the length of that rack gear. I think It'll make an exceptionally quick mechanism, which will be nice for those very short burn maneuvers. The typewriter mechanism I already had is built with a pulley and cable mechanism, and I realized it might actually be well suited to linking up with my throttle lever, so i can implement automatic mechanically responding full/kill throttle buttons on the controller. Meanwhile... My Polish stepper motor has started collecting dust. That's $24 I'll never see again... Edited October 9, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 19, 2017 Author Share Posted October 19, 2017 (edited) Minor update... I decided to order the Red and Green LED bars. It's too much hassle to cut the wide LED bars into narrow ones, and those are pretty expensive anyway, for LEDs at least. I will still cut down the Yellow bars. I have narrow long ones, which will be super easy to mod into short ones. I can just mill the unnecessary portion off. The cut edge of the Yellow bar LED that will leave the translucent epoxy fill exposed won't matter, cause it'll be glued up against the opaque housing of the Red LED. I kinda wanna try to work on the analog meters this weekend. I think I have a full three day weekend free. It'd be nice to try to work on the linear portion of the carriage meter as well, but I'm not sure if I'll actually get to it. Just hoping Digikey gets my package to me by Saturday. Edited October 20, 2017 by richfiles GAH!!! Tracking says Monday! WHY!!! Why on my three day weekend! D: Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 22, 2017 Author Share Posted October 22, 2017 Great news! The LEDs showed up today (pays to have Digikey located in the same state where I live), and I was able to mill the long yellow LEDs to the correct size. Tested them all to make sure I didn't damage any of them, and super glued them to the red LEDs that just arrived. The Green LED needs no mods, since I chose to make the green marker a single long marker. Now I need to find my epoxy, so I can bond them to the analog meters. I'm actually very tempted to cut slots into the polycarbonate to reduce the gap between the scale and the LED, to reduce light spill. The angle was intentional, to help accommodate the arc of the analog meter. Not a huge thing, but I did it just cause I could. I'm hoping to have a fully completed analog meter, with full backlighting finished by tomorrow! Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 23, 2017 Author Share Posted October 23, 2017 (edited) I finally tweaked all the meter scales to be what I need them to be. In the end, i decided to put intake air on the main panel, simply cause It ultimately saves me a meter. I was gonna have Max Temp % be it's own meter, but decided I wanted to ditch the meter with the broken movement (my repair is not the best). I also wanted the normal single meter to be my atmospheric density meter, since I had more room for the graphic. I decided to give the diagonal line a try, but I have a copy without that line as well. Also, for Max Temp, I decided to double up on the red LED section. I wanted it to be a little more prominent. Depending on how the colors print out, I may tweak the atmosphere meter a bit I need to see how they illuminate. As for other progress, I cut the final modifications into the last analog meter that needed modding to prep it for a backlight. I only need to cut one more bit of polycarbonate, and bend it to the correct shape, for that meter, and trim the center ridge of the last unmodded factory backlit meter, and they'll all be ready for new scales! Hoping to be finished with that by the end of Monday. It would be pretty incredible to have all the analog meters officially done! One more thing... I know the atmosphere scale doesn't actually show percentage of actual atmosphere... I think it's more of an exponential or logarithmic scale or something. Google is REALLY failing at finding up to date details of how this actually works, and parts of the Kerbal wiki seem to be rather lackluster... If I could change the scale to something more like: .0001 .001 .01 .1 1 ATM |__,__|__,__|__,__|__,__| Something like that'd be my preferred scale. I'm not sure what my low cutoff should be though. I'm also not sure how the game differentiates atmosphere scaling between planets... I have to assume the scale always displays max density at whatever one local atmosphere of pressure at "sea" level is? Like if I land at the lowest surface altitude at Duna... The meter should be at max density, right? I mean, that makes sense. Man, it's been so long since I actually played... It's kinda crazy! Anyway, I'm open to suggestions and facts... Percent atmosphere kinda... It just doesn't feel right. Every other meter is Percent fuel, or percent electric charge, or even percent max temperature... Even the Air intake is a percentage... It'd be nice to have ONE thing read an actual value... Also debating if I ought to have the colored part of the atmosphere meter fill in the entire white space of the meter, or if I should leave some white space bordering it. I'll have to see how it looks, I guess. I'm slightly reluctant to print the scales on the Nekoosa paper till I have the Atmosphere scale finalized. I only have a limited number of sheets of the stuff. Anyone know what KSP considers to be the atmospheric density cutoff before space as a percentage of 1 Atmosphere? Is that percentage consistent between bodies? I guess the last thing to ask, is what is the actual units output by the plug-in? Is it unique to each planetary body with an atmosphere, or is it a fixed % of 1 atmosphere scale? I'd like the meter to appropriately show density, but I'm not quite sure how to show it. === Anyway, a quick update... I got the last polycarbonate panels cut and heat formed to the correct shape, and I even have printer sized sheets of the Nekoosa paper cut! I'll probably have all the scales installed once I have the Atmosphere scale's design 100% finalized! === Another quick update... Found out KSP Serial IO sends the raw value of atmospheric density in units of kg/m^3. That's fine, but it adds complexity to the programing. I'll need to have the controller know what the ASL atmospheric density for every single body is, so I can calculate the percentage of one local atmosphere ASL to output as an analog value for the meter. This includes all the OPM planets and moons that have atmospheres. If anything is ever changed int he stock game, or the mod worlds... I'll have to edit the program's constants and update it... That's gonna be fun... Still trying to find if atmosphere has a universal cutoff point at the transition from atmosphere to space. Also need to find out if the atmosphere gauge on the KSP UI is linear or logarithmic or something. Edited October 26, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 27, 2017 Author Share Posted October 27, 2017 So, looks like the onscreen meter can be divvied up into 9 equally divisible sections (there are more subticks, but they are not relevant)... I also found the info I need. KSP considers the atmosphere at atmospheric height to be .000001 of 1 ASL atmosphere... That means 18 ticks are being used to display 7 orders of magnitude (1, 0.1, 0.01, 1m, 0.1m, 10µ, and 1µ Atm)... Of course the game would handle atmosphere on 7 orders of magnitude. That's a nice, divisible number! Oh wait... let's divide it by NINE on screen! Welp... Now I'll have to have my physical analog meter use different tick divisions than the onscreen meter. I have definitely decided to use that particular scale though. I like the numeric values for the application. It just means with 7 ticks, there's only 6 primary divisions (and of course 13 total sub ticks / 12 subdivisions). KSP's atmosphere scale's color segments are not equally spaced either, so with different sub ticks, I'll be forced to alter the ratios of the color transitions to other places on the meter to make it match up with he ticks correlating to the order of magnitude changes. Not thrilled about that. I was really hoping to have one to one correlation with what was on screen, but it's just not possible putting numeric ticks that correspond to real numbers. Now, the only question that remains, is do I try to make a smooth gradient, to hide the mismatch, or do I retain 4 color divisions, and try to reshuffle them to match up with particular ticks on the scale? Quote Link to comment Share on other sites More sharing options...
richfiles Posted October 29, 2017 Author Share Posted October 29, 2017 (edited) I'm liking it... Still debating whether or not to include the diagonal line in the Atmosphere scale. I have two saved revisions with and without the diagonal line. The onscreen Atmosphere meter color gradient scale ratios are 1:3:3:2. Mine is 1:4:4:3, to fit the tick marks on my scale. It's close enough. I really wanted numeric values other than percentages for at least ONE meter. Sadly, this day has already mostly gone by, so I probably won't have time to install these today. I need to be able to use the dremel to take down high spots where I trimmed the center ridge on the factory backlit scales, and to drill the holes for the mounting screws (the scales are inserted, and the holes drilled for the polycarbonate and the housing, all at the same time, so when the screws are tightened, the tabs on the end of the polycarbonate pieces pinch down the ends of the scales against the housing of the meter. Hmm... Decisions, decisions... I certainly like the numeric orders of magnitude scale better than the old percent scale. Hmm... I might also need to slide the word "Atmosphere" slightly to the left. I think I somehow left it slightly off centered. I've also decided it's going to be absolutely necessary to cut out a gap in the leftmost meter to make room for the steel beam that supports the horizontal control panel. I absolutely need as much room as I can muster, if I want to fit both the radar altimeter tape meter and the ∆V carriage meter into this thing. The DSKY's added width was not something I counted on, and I really have to conserve space... That steel beam takes about 0.65 inches (16 mm) away from my overall width, but honestly, I think I NEED every spare bit! Edited October 29, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
mountaineerman Posted November 20, 2017 Share Posted November 20, 2017 On 3/13/2017 at 12:59 AM, richfiles said: So, two thoughts for the day... First, 'The Legend of Zelda: Breath of the Wild' is AWESOME!!! The new Nintendo Switch is AWESOME! I'm awestruck that this level of graphics is even possible in a hand held these days. Now I'm kinda wondering if KSP will be ported tot he Switch. The Wii U version was canceled, but only cause the system failed to sell sufficiently. The Switch is fresh and new! It still has the potential to be a winner, and the initial sales numbers are excellent!Quite frankly, the Switch is quickly proving to be my favorite system. I have a PS4 and a new 3DS, and I already know given the choice, I'll buy titles for the Switch. The portability actually is kinda cool. I think a lot of people think of it as a handheld that hooks to a TV, but It honestly feels more like a console you can take on the go, to me. I really can't even explain it... I feel like it leans more console than handheld. To me, that's good. I can honestly say that I'd never really consider spending money on KSP on the PS4 to be really worth it for me. Why? I have it on my computer. I actually could see getting it for the switch as being reasonable, given the portability! Not only that, but since the Switch has keyboard support, why not hide full keyboard support in the game, along with the regular console controls! So then, my second thought of the day revolves my translational hand control. I have a Fairchild Channel F joystick that I plan to use for RCS control. It's a digital controller, but it's actually a genuine 4 axis controller... up and down, left and right, in and out, and rotate CW and CCW. I got the controller from ebay, only cause I remembered the unique controller, thanks to my owning a Channel F as a kid. It's about as close as any controller ever got to the real translation controls on the Apollo and Shuttle vessels. I have no intention of modding the controller... I just want to make an opening where it can insert through, and then have a soft clamping mechanism to secure it. It'll plug in to a DIN-9 socket on the inside of the panel. That lets me preserve a vintage video game artifact, while still making use of it. My actual second thought though, relates to where I am going to mount this. I'd assumed that I was going to find some manner of tiny corner of free space on the instrument panel to mount this. After all, I'd reduced my meter count to 4 meters, now that I have so many double meters, vs the old single meters. It occurs to me that there's really no obligation for the THC to actually be mounted INSIDE the panel at all... It does have a 9-pin D-sub connector, after all! I could honestly mount it ANYWHERE in the desk, and it would work! Potential mounting points are just under or beside the left side of the control panel surface, or where my concept drawing currently shows the CRT box. The CRT box isn't happening anytime soon, and probably won't even fit there, so that space would be reasonable to utilize. It's actually very similar to the location of the THC on the Space Shuttle, so I'm more than cool with that. I can simply mount the clamps to the outside of the control panel, and have a cutout int he desk to make space for the controller to pass through. One of these days, when i get more time, I'll try to update the concept art. In the meanwhile, I have a Digikey order to place for some parts. Hey @richfiles, Do you mind me asking what those joysticks you are using are called, or where you got them? They look like 2-axis potentiometer based. I've been having a hard time finding something good for my own project. Thanks! Quote Link to comment Share on other sites More sharing options...
richfiles Posted November 20, 2017 Author Share Posted November 20, 2017 (edited) 3 hours ago, mountaineerman said: Hey @richfiles, Do you mind me asking what those joysticks you are using are called, or where you got them? They look like 2-axis potentiometer based. I've been having a hard time finding something good for my own project. Thanks! The normal looking joysticks are a common style, mass produced and often sold on ebay. Just search for "2 axis joystick", "3 axis joystick", or "4 axis joystick". The two axis version is a simple left-right, up-down joystick. The three and four axis versions have a larger knob, and it actually twists clockwise-counterclockwise. The four axis version does that, and also has a simple pushbutton on the top of the stick. They're cheap too. Usually under $15-20 shipped. Generally, they have a 10KΩ potentiometer for each axis (though there are variants with other values, like 5KΩ). The ones I bought feel nice and solid. If you wire one side of the potentiometer to ground and the other to Vcc on your Arduino, then the middle pin will produce an output readable by an analog input pin. As for progress... Work's been kicking my butt. They fired someone in my area, and just decided they weren't going to bother to hire a replacement. Two people working a job with a three station workflow... Yeah... That's gonna be real efficient... That's how they wanna do it... Fine. The 10+ hours of overtime on this check will be appreciated. On a VERY positive note... I finally seem to have had the time (and reason) to learn KiCAD! My boss at my second job wanted me to design and build a hall effect sensor tester to test the commutation boards on the brushless motors he manufactures at his shop. I told him I could wire wrap something really quick or take the time to design an actual PC board. He said make it nice and professional, so I basically got paid to figure out KiCAD and design a really nice board. I'm only about half way done with it, and the autoroute is a bit derpy. Usually traces have to stay ON the board! It's a work in progress. I need to clean up the text so things aren't jumbled up, and I need to fix a few trace routes. I definitely need to ad a ground plane to the thing, at least around the switching power supply. I am actually going to make the board about 2/3 wider, now that I've confirmed the enclosure will support it. It'll let me add the power switch and spring post holes and pads to allow everything to directly attach to the board. Edited November 20, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
mountaineerman Posted November 20, 2017 Share Posted November 20, 2017 Lovely, thank you very much! I will have a look. Quote Link to comment Share on other sites More sharing options...
richfiles Posted November 20, 2017 Author Share Posted November 20, 2017 Much better than before! This bodes well for when it's time to actually build the FDAI Synchro Emulator board. I belive there are 5 analog switch chips and 5 DAC chips, plus the sine wave generator, as well as 9 transformers, all being controlled by an Arduino. Honestly, with my newfound understanding of this software, making a board shouldn't be that terrible. It'll be better than wire wrap and a whole lot less brain twisting than trying to route that setup manually. Also, I'm starting to learn some of the hot keys... Good Lord, there are hotkeys everywhere... Quote Link to comment Share on other sites More sharing options...
richfiles Posted December 16, 2017 Author Share Posted December 16, 2017 (edited) You're looking at the final art for my first professionally designed PC board! Minor non instrument panel related update... Work's been brutal... Definitely a job meant for three people, but we've only had two people since October. It's been rough. Management sees nothing wrong, of course... They can just pad my hours (as a part timer) to full time, and keep paying me part time wages with no benefits... Thanks guys... Anyway, my other job (which has actually been suffering because of all this, because I have trouble finding enough time to go out to the machine shop now! ) has had this test device project to deal with, and I do belive I have a finalized design! It's been an interesting road, learning how to use KiCAD. I'm certain I will use it to do the FDAI synchro emulation board now. I really have no reason not to. I was considering going with OSH Park as supplier in the future for small boards, but man, my boss's (the machine shop) supplier is a place called Circuits West. Website ain't all that impressive, compared to OSH Park, but they offer a $31 special for up to 60 square inches (387 sq-cm)... That's 3x 6x10 inch (15x25 cm) PC boards for $31. That's actually pretty insane! Thing is... Their website makes it vague whether that's for all three boards or per board. They're not as fancy as the OSH Park boards, but it's a basic HASL (solder coating the copper instead of gold plating), green solder mask, and white silkscreen, so it's not that bad... except another weirdly worded bit on the website makes it sound like only one of your three copies is plated... Seriously, if they wanted my business, they could have put more than 20 minutes into making their website. Went with OSH Park... And that marks the first time i've ever designed a PC board from concept to manufacturing using CAD and a professional manufacturer! Here's the backside. KiCAD has very bad tools for via stitching (none), so it takes a few extra "hoops" to jump through to tie things like ground planes together over the board. Still, KiCAD is both free and open source. I'm alright with that! I guarantee i could have done a better ground plane if I had tried, but It's in the "good enough" range, and I really needed to get this manufactured... We already built our first batch of hall effect commutation sensor boards. These hall boards mount inside electric motors and tell the position of the rotor, based on sensing magnetic fields generated by a ring of magnets. A functioning board will always have on or two outputs the same, but never all three. It also has a pattern that changes based on direction of rotation. With this tester, we can hook it to a hall board to verify functionality (if an output is stuck on or off, eventually, as it switches through it's outputs, All three signals will end up being the same. This latches a fault detection circuit on, lights a fault LED, and requires a reset button be pushed to clear the fault. This lets us detect bad hall effect sensors or bad wiring or soldering (shorts or opens). In motors, if we spin the rotor by hand, and read the hall pattern, a motor wired incorrectly will show an incorrect pattern. Now, on to where this is relevant to Kerbal stuff... This marks a major change for me! This is literally the first time in my life that I've designed a PC board entirely using a computer, and sent the design off to be professionally manufactured. I've been meaning to learn this skill for the past 15 years, and just kept putting it off. Hehe... I guess being paid to learn it was a good fire under me to motivate me! I may still use wire wrap or hand made PC boards from time to time... Mind you, that this tiny board cost $40! It comes out to about $5 per square inch. I might contact Circuits West by phone or email and ask them for some clarification on things the website isn't so clear about... If I get an answer that I like, then i might go through them for the FDAI (navball) control board. That one wis gonna be sorta, kinda big, with 13 regular chips (minimum) and either 5 or 10 power amplifier chips (depending on if I can get clean signal separation on the stereo channels), plus 9 small transformers... Yeah, she's gonna be a big'un! If that 60 square inch board deal for $31 can qualify for a single board, then I'm good. If it doesn't... Well, the deal is you get three identical boards, and if each board can only be 20 square inches... that may actually still okay, cause the FDAI circuit is triple redundant. It literally will have three identical copies of everything except the input stage. I can split the design so that three identical boards can be tied together at the ends to perform the job. i'd only populate the input stage portion on one board, and jumper the signals across the other two boards. Data signals would pass from board to board, and each board would only manage a single axis, with three boards providing the three separate axes I need. It's just gonna be super annoying if they only do solder mask and HASL plating on one board, and the other two end up being bare copper. It's still better than nothing though, and i could conformal coat it after assembly, but it won't look as nice. Edited December 16, 2017 by richfiles Quote Link to comment Share on other sites More sharing options...
richfiles Posted December 27, 2017 Author Share Posted December 27, 2017 It arrived, and I have it mostly assembled! This is officially my first professionally manufactured PC board! I made a couple mistakes that I will DEFINITELY aim to avoid in the future. A few ground solder pads on my ground plane had multiple connection points... enough that they were thermally bound to the ground pad. That was a huge pain in the retrograde... I also forgot to up the hole size for the pushbutton's alignment peg hole, and the two switch leads, since they were a bit thicker. I was able to make it fit by filing the square leads round, but again... huge pain! Good news, is making minor mistakes like this now means I'm far less likely to make them in when designing future boards! As for future boards... I've pretty much latched onto the plan to do the FDAI Syncro emulators professionally. Since most board manufacturers offer prototype boards in triplicate, I can simply do like I said, and make each board be 1/3 of the circuit. This will mean a few things... I need to decide whether each board will handle the DACs and amplifiers identically at the cost of using more components, or if I will conserve components with more crossover links between the three boards. Since each DAC and each Amplifier handle two signals, and I need three signals per axis (for three axes), it means I need nine DACS and Amps. In an ideal circuit, I'd use five DACS and five Amps, wasting only a single circuit. Going with three separate boards means I can't share one part across a pair of separate axes, without additional links between boards. The part count isn't significant. I can add six DACs and six Amps without an issue, be it space, complexity, or cost wise. The main consideration is communication. Over a serial bus, do I want to communicate with six devices instead of five? Do I have time to waste on an extra frame of data, when the micro running the show is busy with sine lookup tables and multiplications? Alternately, does it simply become easier to split it up the task? If each axis is entirely separate, and entirely identical, then rather than have a single micro handle those calculations for all three axes, I could have three small micros handle the identical functions on one axis each, and only communicate with two DACs, rather than five or six. I guess I'm erring on the side of six DACs and Amps, with three small micros, so each axis is 100% independent from the other two. The only thing shared between them, is an inverted and a non-inverted 400 Hz sinusoidal reference frequency coming off a bit of vintage retro hardware, and the general serial bus for the main Arduino to retransmit received values to other devices. Each of the three micros would only need to sniff for it's specific axis value, and only process that, and nothing else. Hmm... I was pretty broke when I ordered the DACs... I might need to order one more, if I went super cheap and just ordered five... Here's the blank board, as it came from the manufacturer. Gotta say, while it's not specifically Kerbal hardware related, is HAS been an amazing learning experience! Quote Link to comment Share on other sites More sharing options...
hugopeeters Posted January 4, 2018 Share Posted January 4, 2018 Dude, that is awesome! That's a great skill to have for professional level hobby projects Quote Link to comment Share on other sites More sharing options...
richfiles Posted January 9, 2018 Author Share Posted January 9, 2018 (edited) Welp... Final update on my diversion from Kerbal controllering. Here it is, during final assembly, and below, finished, minus any labels that get added later. I need to trace out one issue, and make another adjustment. The green LEDs are almost painfully blinding, so I'm just gonna pop a bigger value resistor in to dim them. The red fault LED in the reset switch is kinda dim, so I'm gonna boost that one by reducing the resistance to it. I've already run the circuit on a simulation, so I know it works, but I have one of two fault conditions not triggering the fault latch. I figure I either made a dumb mistake in the layout, and have to bodge a wire in, or I have a bad solder joint. Nothing too critical. Minus that one problem, nothing to serious. This is for one of my employers. It's used to test hall effect (magnetic) commutation sensors used in brushless motors. I'm particularly happy that the power supply worked, first try. As I said, it's my very first switching supply. I definitely want to move away from 9 volt batteries on toward either AAs or LiPos on future projects. Don't worry, this is probably the last post about this bit of side tracking. It DEFINITELY has taught me a lot, and it specifically has left an endgame in sight for making the navball controller boards! My main job seems to have let up a bit on the relentless torrent that's been keeping me late, and now that this project is done, I can finish up some other work I need to do for that same job, and then maybe get back to building my controller! **EDIT** My goodness, those green LEDs I selected are EFFICIENT! Even with a 910Ω resistor, they are still incredibly bright! Yikes! I was gonna measure the current, but forgot. Oh well. Box is closed, and I never want to open it again! In the end, there turned out to actually be SEVEN issues! 1: I failed to consider orientation of the reset button/fault indicator, and as a result, had to use spacers when mounting the board to leave enough room for the part. 2: The spacers meant it was no longer possible to electrically mount the board directly to the push post terminal's threaded posts. 3: This turned out to actually be a good thing, since I managed to wire the mounting pads in the wrong order. I had to cut the traces to the two outer push post mounting pads, or they'd short to the wrong post. I was forced to manually wire the posts using ring lugs and wires running from each post to the secondary input header (I'm glad I implemented that). 4: The resistors I had selected for the Hall Effect Sensor LEDs were ridiculously undersized... (180 Ω to 910 Ω 2.2 KΩ) Rather the LED I selected was rated far too bright for what I needed. (I went back and replaced the resistors again, after finishing up at my regular job... They were still too bright. Current was about 1 mA @ 2.2 KΩ. Tried 3.2 KΩ, and got 800 µA, but decided to stick with 2.2 KΩ, just to play it safe. Didn't know how my power supply would handle low batteries, or whether the LEDs might dim more at the tail end of the battery life. I might yet change it again in the future). 5: The resistor I had selected for the Fault LED needed to be lowered a bit, as the fault indicator was a little dim. (150 Ω to 100 Ω) 6: The application note for the MAX756 Step Up DC-DC converter chip assumed that the user would simply connect their own signal to the standby/shutdown pin of the chip... Since the schematic showed it as not connected, I did the same. I never even considered it during the design phase. I couldn't figure out why the circuit was only getting 3 volts (straight off the pair of AA batteries), instead of the 5 volts it was supposed to generate. Discovered the problem because it would go away if I simply touched the shutdown pin. The residual electrical charge in my finger was enough to activate the chip. Fixed it with a small resistor jumping it to the power source, to activate it whenever the power switch is turned on... That was a dumb mistake... But not the king of derp... 7: Drumroll please... I managed to entirely miswire HALF THE FAULT DETECTION CIRCUIT'S INPUT! I am STILL uncertain HOW I managed to screw this one up. I had meticulously copied the schematic from my circuit simulator on my phone (ain't technology grand ), where I had PROVEN the circuit was validly functional, over to the schematic editor of KiCAD, the tool I use for PC board design. Somehow, when I decided to use the spare inverter gates to buffer the input, I SOMEHOW managed to leave half the fault detection circuit's inputs on the input side of the inverter gates... but took the other half and moved them over to the output side. What that meant, was that the circuit that was meant to detect if all three sensor inputs were logic high (5 volts) at the same time worked, but the circuit that detected if all three Hall sensor inputs were logic low (0 volts) would instead see the opposite, and so it would ONLY trigger when the other circuit was already triggering, and even then, it'd give the wrong output anyway. It was SUPER DUMB. The good news, is I only had to cut three board traces and jumper them with wire to the correct inputs. I STILL can't belive I made such a drastically severe mistake... Or even SEVEN mistakes in general. Sheesh... Good news, is it works 100% now! Bad news... My one job is still utterly monopolizing my time, and it's severely impacted my ability to get out of town to the machine shop and do my work there. I'm closing in on being nearly a month and a half behind on where I'd like to actually be with my work there. That's bad... I don't like it at all. Granted, I tend to only get 8-24 hours in a month period. It's not full time. Still, I'd say that I feel like I need to do about 50+ hours by the end of the month to catch up to where I'd like to be... And the month is drawing quickly to a close. I've already stated the problem with my current job. Having two people work a job that's clearly a three person job is stressful! They just don't grasp the concept! At the end of the day, I am sore, I am tired, I have no energy left. I AM most definitely burning out, and losing my patience. At the end of the day, I have to make a choice... Do I go in to the other job, drive out there, work a couple hours, and leave at 9 or 10 PM (21-22:00 for those who use a 24 hour clock), head back into town, and then get up in the morning to work at 8 or 9, and have to repeat that for 3-4 days, cause I can't get enough hours free to just go in and power through everything in two days... Or do I just go home and fall asleep early, cause I got nothing left in me? Sure, I could power through. I've done all nighters before. I've done low sleep... But it's BRUTAL. I'm not getting younger, and that is STRESSFUL! I can't keep punishing myself like that. I just end up like December... Sick for the entire month, never getting better. The other issue with going into the shop tired, is if I screw up, it could cost quite a lot. If a cable is $10 a foot for the raw material, and you have a 10 foot cable... Maybe I get the cables assembled, only to discover that I miswired them, cause I was tired and swapped the yellow and white wires. Gotta redo an end then! Say I'm so tired, I screw up something else up and the repair ends up leaving the cable short for it's minimum tolerance... That would be a loss of $100 on the spot. Boom! Gone! That doesn't even count parts I couldn't reuse in assembling either end. If I am tired and I drop a brake assembly, for a different product... Oh bummer, there goes the smooth bore of the bearings. Gotta tear 'em out and replace them. If I cut down a wire on the assembly too short, or I solder every connection wrong cause I was tired and rotated a part 180°... Yeah, there's SERIOUS potential to screw up when tired, and at that job, screw ups can be expensive. I'm not keen on doing that job tired. Supposedly, we're gonna have a massive streamlining soon, with some all new methods of preparing media at the lab... But dang it! It's taking so long to work out the kinks! Even if it saves time in the long run, I strongly suspect we'll initially be slower at it... I just want this stupidity to end. I want my free time back, and I want enough time left in any given day that I don't feel pressured to even fit time in for my second job! Ugh... Enough ranting. I need to sleep, or i'll be too tired for my regular job of +33 overworking.**EDIT 2** Got a raise at the machine shop job! Aww yeah! Edited January 30, 2018 by richfiles Quote Link to comment Share on other sites More sharing options...
hugopeeters Posted January 10, 2018 Share Posted January 10, 2018 Nice result! PS: did you run out of caliper batteries? Running it off the bench power supply? Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.