Friday, November 2, 2018

24-gage steel, bad choice

So I ordered some 24-gage (.025") galvanized steel.  Doesn't seem to be a common size, but found a person on etsy selling 8"x 8" sheets for $1 each ($6 shipping).  It's really stiff, I don't even have to build up a prototype to know it won't flex like I want it to (which means the ball will have to hit it pretty hard to register, which isn't what I want).  Well at least the good news is that 28-gage is readily available at home depot in 12x12 sheets, or I can acquire a 24" x 48" sheet for $18 and pop out 672 parts at 2 cents each.  Hopefully I can still re-use this metal for other parts in the future (ramp flaps, ramp side rails, short guides).

Tuesday, October 30, 2018

Next revision

Ok so I cost reduced it down even more by making better hardware choices and cheaper metal (I believe a thicker 26-gage galvanized sheet metal will be just fine).  I didn't realize how expensive spring steel is (VERY), and stainless isn't that much better. They make pipe clip buttons out of galvanized steel so I know it has some springback.  Also I had some 28-gage galvanized steel at work and although it's a bit on the thin side, it functions well.  I already updated my 3d models, made some new 3d printed parts, and built a whole new assembly.  It feels decent, the leverage arm is increased so the force to push down on the switch isn't bad at all:

So the new price per assembly?  59 cents in parts (not including labor) at prototype volumes.  Before I go off and start having metal brackets waterjetted, I really want to test this out.  I plan on making a life test rig to see how many cycles this can withstand before something breaks, or something gets bent where it stops contacting.  I have 2 ways I can build this:
1. Get some U-channel (so the ball is guided), add the target on one side, and add a shooter lane auto launch mech at the other side.  Only piece missing is a delay timer to keep firing over over.  Luckily someone at work is an electrical engineer who recently built such a box to run some solenoids.  Basically I just need something that will kick on 50vdc for about half a second to fire the coil, turn off for 3-5 seconds, repeat.  This could get noisy however (I've seen pinball test rigs in person).
2. Start with a u-channel again, but this time mount it on top of a half circle, and connect it to a crank mech (sort of like a crib that rocks).  If I made it say 2 feet long, there would be enough space for the ball to gain speed from gravity as it rocks back and forth.  This is more mechanical, but then it would just be a geared motor running continously.

Monday, October 29, 2018

super quick shipping :-0

So I ordered those switches yesterday, they were already in my mailbox today.  I'm pretty sure there's an electronics store on ebay that's very close to the central post office in my area.  So I went ahead and decided to adjust my standup target design and 3d print some parts (still needs some tweaks)

There's literally just a switch holder, the switch itself, a custom target face, a thin metal bracket (I made this one from galvanized metal but it needs to be spring steel), few screws, and a brass insert.  I did a BOM cost estimate and including 3d printed parts I'm at around 80 cents for a complete assembly.  The biggest cost drivers are the bracket and brass insert (I could possibly lower those prices if I bought in volume, not prototype quantities).  In production quantities, I could totally see getting this down to about 50 cents per assembly which is a huge cost reduction over the $8 retail stand-up targets.

Sunday, October 28, 2018

homebrew isn't cheap

So one thing me and Hugh from OPP can definitely agree on is parts aren't cheap, and playfield BOM's can really get out of hand quickly.  Individually parts don't seem expensive, but when you start adding them up (bulb sockets, any sub-assembly with coils) it can really add up.  Honestly many homebrew guys ends up spending 2-3 times as much money making their own game than what it would cost a pinball manufacturer to build one (because of volume of parts).  This is the reason why Hugh grabs populated playfields when he comes across them, the parts are worth more than the whole.  I mean really look at any sub-assembly on pinball life.  Each flipper assembly is $40, each sling assembly is $40, each drop target assembly is $50.  And this isn't a knock against Terry, far from it.  I think the idea that you can buy these assemblies is great for someone that doesn't want to scrounge parts from another game.  If you look at all the parts, there's a lot of cost in there (both in formed metal, assembly, etc).  Now it's not like Pinball life designed these and said "hey, that's the price".  No, he effectively had existing parts reverse engineered, bought them in volume, and his employees build them up as needed.  If you were to try to buy the individual parts and make them yourself, you'd actually spend more money than buying the assembly.  Not that it makes much sense (less assembly, cheaper), but that's often how things work out.

When you see pinball manufacturers being conservative on mechs in the last 5-10 years, there's probably a good reason for that.  Other than Stern, nobody really has the volume buying power.  I think Stern is plagued by several overhead issues.  They have a giant building (110,000 square feet) that has to be powered, lit, heated/cooled, insured, maintained.  They have a TON of employees not only designing games but building them.  Then there's the licensing, and not just the ones they do something with (buying a license for future use to make sure nobody else makes that game is money out the window).  But just look at Beatles which is practically identical to Seawitch (add a magnet and a spinner which are mechs that already exist).  Unlike Seawitch (with basic coding for rules and numerical scoring) it will have deep rules and full color LCD animations that need programming.  Beyond that, it's rumored the Beatles license (including the measly 5 songs) cost them a million dollars.  So if they plan on selling a total of 1,964 (we'll see how well it sells), that's really only an additional $500 amortized into each machine's BOM.  I'm hearing numbers like $7500 for the bottom gold edition, $12.5k for the platinum edition, and $25k for the diamond edition (no extra features for upper tiers either) which is clearly a cash grab.  You might say "oh well, if there's not much offered in higher tiers, clearly they will sell mostly gold editions and the higher tiers just won't sell".  You'd be wrong.  Not only are they screwing their customers, they are also screwing their distributors if what I'm hearing is true.  Distributors have to buy packages.. IE if you want a few gold editions, you must also buy the higher tiers in order to get those.  It'd be like if you were a car dealer that just wants base model camry's, but toyota won't sell you just base models.  You MUST stock the high end camry's too.

So anyway, without sounding like a JPOP of re-inventing the wheel (he made things custom for the sake of being custom).  In my case, I want to cost reduce parts to make pinball more affordable (not more expensive by requiring buyers to order a custom sheet of glass).  Recently I started coming up with a concept to cost reduce standup targets.  I'm still contemplating the best way to do it, but once I get a design I like I'm absolutely going to pursue getting them made.  Whether it be just selling it to the homebrew community, or if a smaller pinball company likes the idea I'm happy to sell to whoever wants them.  Now standup targets aren't terribly expensive (about $6 each), but again.. Start throwing a couple rows of 5-bank and you're quickly up to $60.  There really isn't much to it, but because we are still using those antiquated brass leaf switches with tungsten contacts and fiberglass spacers they are still pricey.  When I was deep in my homebrew projects, I felt strongly about replacing the leaf switches on the flipper buttons with tactile switches.  It has a better feeling of that spring snap, and micro switches never need adjustment, plus today they are dirt cheap.  I actually did a cost estimate (not including labor), and I could easily get the cost of a standup target to around $1 depending on if I cut brackets from aluminum channel or 3d printed a bracket (which is still low cost, but less labor).

Cherry switches are REALLY cheap if you know where to shop around.  I bought some really tiny ones about a month ago to play around with that can handle 1A.  Though they are so small that they are hard to design around.  The holes themselves are like 1.8mm (which means you are really using tiny screws).

I recently made an order for some slightly larger ones rated at 5A that also include one of those roller wheels (which should help eliminate wear as the edge of the switch blade rubs against the back of the target).

The difference in price?  The first one can be bought for 50 cents each in quantity of 10 (US seller).  If you're buying from China, 7 cents each.  The second nicer switch is also 50 cents each in quantity 10 (US Seller), or 10 cents each if you order from China.  Since I plan to use a bunch of these I ordered 10 from the US seller so I can get them quicker, and 20 switches from China where I don't care if they take 3-4 weeks to get here.

So I started looking at the $50 drop target assembly.  You have a $10 coil, a $5 leaf switch, a bunch of zinc plated bent steel, it all starts to really add up.  Coils are used everywhere in pinball because that's what they started to use to make flippers flip back in 1948 and they just kept using them.  For things like flippers and slings, absolutely there's no other cost effective way that can get those to move quickly with a lot of torque.  I see no reason why drop targets need to be so quick (other than coils are what have always been used because they were available).  Even if they were half the speed of coils, you'd hardly notice the difference.  Also coiled drop targets are really loud when a large bank of them reset.  Whether that be individual inline banks, or a wide bank that is reset by a single coil because of the mechanism required to raise them all back up at the same time.

Now I believe the first pinball to use a servo (in the true sense of an off the shelf enclosed servo) is the blinder mech on the Data East Tommy (1994) followed by the turning head on Sega Frankenstein.  One could argue the rotating head on party zone is a servo, but in reality it's a 12v motor attached to a gear reduction gearbox.  So while it has properties of a servo, it's a motor and gearbox and not cheap by any means.  It's unclear if the homer head in simpsons pinball party is servo based from the manual.  However, it does seem like Stern has at least moved to cherry switches for their drop target banks (but still uses a coil to reset).

Looks like Pinball Life does sell a Stern 3-bank drop assembly, but it's not clear if it includes switches (and it's friggen expensive at $199!)

I mean if I had to estimate the cost of that assembly I'd say it probably costs about $20 in parts.

So where does that leave me?  Well designing my own mech of course.  I played around in solidworks this weekend on and off and came up with the mech below:

It consists of a custom 3d printed target, a custom 3d printed sleeve/bracket (which allows the cheap roller switch to be mounted), and a servo.  The cheapest servo I could find on hobby king is $2.35:

So my BOM consists of:
$2.35 - servo
$.10 - cherry switch
$.04 - two screws for switch
$.08 - two screws to mount servo
$.10 - Target (cost of PLA to print)
$.26 - mounting bracket (cost of PLA to print)
$.10 - servo arm
$3.03 per assembly

If we're doing math, that's approximately 16 times cheaper than an equivalent single drop target.  What does that mean for homebrew?  Well if it's robust enough (time will tell), it means if I'm only budgeting for 3 inline drop targets ($150), I can now put nearly 50 drop targets in my game for the same cost and not compromise the design because parts are too expensive.

Tuesday, October 16, 2018

Me and Hugh (OPP) finally meet!

So last week I was on vacation.  Flew into Boston, drove up to Maine.  Last minute I'm like "Who do I know in the area?", and I realized Hugh was like 40 minutes from the airport.  I didn't have his phone number so I emailed him early in the morning hoping he would check it.  I felt bad not giving earlier notice, but I couldn't pass up this opportunity.  Turns out he was going to be home that afternoon, so me and my wife drove up barely going out of our way.  He gave me a tour of his basement, and the opportunity to finally play SharpeShooter3 pinball.  Even though I've watched videos of him playing, it's more bizarre when you're playing the physical game and you hear your own voice on it.  We chatted a bit, and I think our biggest negative takeaway with both projects was much it annoyed us that the inserts were off.  It's most likely a combination of bad measurement / Photoshop and Microsoft ICE not doing a great job of merging scans, and the vinyl stretching during application.  The last can be fixed by printing art to something solid like thin polycarbonate sheet (same way outside edge does their hardtop overlays for existing machines):

Photoshop merge.. not much you can do about it.  Ideally you want to scan a playfield in one pass.  I do wonder if there was some sort of registration (like drawing out a grid pattern every inch) would help the software align better.  I should do a test on some art I don't care about (like a poster) and see if it improves the accuracy.

Since my day job has finally slowed down a bit, I want to start getting back to some ideas for the next playfield layout from scratch.  Before I start cutting full playfields, I think I want to test out some ideas on smaller chunks of wood as proof of concept.  Having multiple 3D printers I can quickly make real parts from 3D models.

Sunday, March 18, 2018

van halen delivered

So end of February, the van halen machine that Hugh from openpinballproject worked so hard on arrived.  We had used to get a quote which I thought was fairly reasonable, especially since it was a van used to transport.  Nothing against large semi-trailers, but I've seen many games get destroyed by a forklift driver that just doesn't give a crap about merchandise getting damaged.

So right away, I had to pack my Back to the future game temporarily (there's only room for 3 games upstairs, and I'm at capacity).  I may still stick Van Halen in the basement, but I want to do some tidying up first.  So right away I took the horribly rusty feet off and replaced them with new ones.  The legs themselves are a little rusty, in fact one has a stripped thread (so I used a nut on each side to keep the height adjusted correctly).  I have about 16 used legs I'm trying to get powder coated black from a co-worker that will eventually replace these.  I don't feel it needs to have the period correct Bally legs since this is a custom homebrew.

Now I have to preface before I talk about the following that none of this is a knock on Hugh.  The fact that he was able to build up a custom game in such a short period of time (I think it was 6 months?) starting with what I imagine was a very beat up Dolly Parton is amazing.  He stripped the playfield (twice I believe), scanned all the existing plastics and sanded the playfield and scanned it so I could get artwork done.  I came up the rules and he implemented them by writing the code to get it done.

So I did make a couple improvements tonight.  Nothing bugs me more than a dirty and rusty lockdown bar mech.  Even though it's not seen while the game is in normal play, when you do take off the glass (get a stuck ball, fix something under the playfield), it's the first thing that sticks out like a sore thumb.  So I went to the local harbor freight and bought a few wire wheels for the drill and went to town.  I also hand sanded the lockbar with 500 grit sandpaper to get that a little smoother.

I also added the panama captive ball that I 3d printed.  It doesn't get in the way of the ball path coming off the spinner, it doesn't get in the way of the shooter lane, and it's shootable from the left flipper.  Sometimes the left sling will fire it up into it, or sometimes if you konk it just right from the side it'll make the ball travel up to hit the switch.  I never liked that odd opening that Dolly Parton had where the ball goes back into the shooter lane.  I felt like I was committing a sin driving that screw into the playfield, but sometimes you have to take a chance.

So the more glaring issue in general is the playfield overlay lineup.  I don't think this was an installation issue, but rather it's very difficult to get accurate measurements of inserts.  Not sure if I'll ever go through the trouble of trying to fix it or not.

The inline drop targets need decals, as does the spinner.  Both have the original from the Dolly Parton.  Speaking of which, the spinner never got code for a sound.  Hugh has sent me the files and instructions to update it, just haven't gotten around to it yet.

The LED strips in the backbox have hotspots, and don't cover it all the way.  This is an easy fix, just have to get to it.  Also there is a screw to hold the light panel together, it would be preferred if there was some sort of knob to make it easier to turn.  I'll probably get a hex version of this screw and 3d print a knob.  I also want to replace all the rubber caps used to hold the plastics on.  I NEVER liked these, my space shuttle even had them and right away I replaced them with locknuts because I can ensure the plastics are tight.  I like the colored nuts that mezel mods has started to sell, might grab some red ones to go with the theme.

The last thing I'm considering is moving the boards and transformer to the lower cabinet.  Right now there are velcro straps for strain relief for the heavy cables.  In my mind, the boards being the backbox never made much sense to me even from pinball manufacturers.  I'm sure initially boards ran hotter and needed venting.  I also get that it's easier to troubleshoot board issues in a backbox (standing upright instead of crouching over a box), but with modern hardware I don't think there's nearly the unreliability there once was.  Reason I bring this up is that if I ever decide to take it to a show, I worry about connectors being jostled around.  If most everything was in the bottom cabinet (at least the connectors for the cabinet plus playfield), the only wires going up into the backbox would be:
* power cable for monitor
* VGA cable for monitor
*12VDC for the backbox lights
* 2 pairs of speaker wire for the boombox topper

Tuesday, March 28, 2017

So much for ramping up, random thoughts

So I thought I was going to start working on this project again after the holidays.. Then my wife got a terrible earache infection that we're still dealing with (multiple visits to ear nose throat doctor, ear specialist).  Then to top it off my dog just got done with another spout of a pulled back muscle, thankfully this time it was just meds.  Then my 3d printer crapped out, thankfully my new prusa i3 mk2 came at the same time, and spent an entire weekend with my wife building it.  It still prints, but even with my fixes it consistently prints these weird artifacts on the sides that I can only attribute to some issue in the z-axis (or the board is garbling the g-code somehow?).  Here's what it looks like:

Speaking of which I must say, the new printer is awesome.  I read all the reviews, watched all the videos, they don't lie.  This thing prints flawless clean prints every time which is amazing considering how many 3d printed components it has.  The secret sauce as they say is in the auto calibration (in every axis).  Even when we were doing the initial calibration, it said (XY axis very skewed, but I can still adjust for it).  I said no way, let's get this as close to square as possible.  Loosened the nuts underneath, and kept bumping the one side until the x-axis bars were parallel with the grid pattern on the printer bed.  Re-did the calibration, said it was excellent.  Starting doing some generic solid test prints, came out great.  Then I went to re-print a job for a customer that had thin walls, super curling warpage on the bottom.  I noticed that the bed was jumping to 100C every time, I couldn't understand it.  Luckily prusa has a live chat support, and he walked me through some things.  Finally on my own I discovered I had a couple wires reversed, and basically the power supply was feeding uncontrolled voltage to the bed.  After swapping it out, accurate temperatures every time.  I've got my old printer up for sale locally on ebay, so far it's up to $102 with a day left.  Honestly if I threw it away I would still come out ahead.  It's made me some decent money over the past year.

So onto pinball.  In my last post I showed how I was updating my layout, and that I was going to have the shooter lane double as an outlane.  How innovative I am right?!  Wrong, turns out that SEGA did this on starship troopers.  I was listening to some podacst (coinbox?) and they talked about how that did terrible on route because if the auto shooter got weak or something obstructed it, you were basically draining balls.  I mean I could still do it and just realize that I need code to ensure that the player MUST score a minimum amount of points before it counts a drained ball.. Maybe I can come up with something even more innovative.

On a side note, I'm helping Hugh with another re-theme.  I know he hates that term when there's new programming involved, but mechanically it's the same layout.  So he's taking the Dolly Parton that's been sitting around, not changing the wiring, but re-theming it to Van Halen (per my suggestion since I'm so passionate about wanting a Van Halen pinball).  So far I've got the backglass nearly done, but still laying out the playfield and plastics art.  I'm going to try throwing as much at it as possible, including not only david and hagar, but every van halen reference in pop culture.

Lastly, I've been thinking a lot how I want to approach this next build.  Originally I was trying to 3d print the lower half so I could quickly toss them onto a blank piece of wood (which I may still, at least most of it), but I'm thinking I want to somewhat borrow Ben Heck's concept of splitting the playfield.  Many Pinside users had talked of such a concept in the "what's the minimum parts to build a whitewood" thread, but he actually did it.  I do NOT like foamcore.  I realize it's easy to cut, but it costs just as much for a sheet of foam as it does for 1/2" birch.  I would far rather have real material I can drill into than trying to work on something so flimsy.  Here's his rotisserie setup:
I'm going to try to do something very similar, but replace the foam with another piece of wood.