Jump to content
 

TurboSnail

Members
 View Profile  See their activity

  • Posts

    1,451

  • Joined

  • Last visited

 Content Type 

Blog Entries posted by TurboSnail

  2. TurboSnail
    For a current project, I need a tiny, but very high-ratio motor and gearbox, so modifying a servo to rotate continuously seems to be the perfect answer. The entire unit takes up less than 20x20x8mm and provides a motor, speed controller (I'm using it with R/C) and an approx. 120:1 ratio gearbox, while costing less than a fiver. But it does need some slightly fiddly modification first. I'm listing the process here, as much to aid my own memory next time I do this...
     

     
    Small but hopefully powerful...
     

     

     
    After levering off the base, you can see a tiny (4mm diameter) motor (the white circle on the right) and the circuit board. If you move the circuit board to one side, you find the potentiometer hiding underneath. This is what tells the servo what position it is in, so it can adjust correctly. All we need to do to make it rotate continuously is remove this feedback capability. To do this...
     

     

     
    Lever out the potentiometer so it doesn't have the connecting shaft attached (it has a shaft connecting the pot to the output gear which you can see in the cover to the left). File down or drill out the potentiometer shaft hole (the little white bit sticking out of the small black box in the photo) so the shaft cannot turn the potentiometer. This has removed the servo's ability to sense where it is - there is a better way of doing this for bigger servos by replacing the potentiometer with a two high resistance, low tolerance resistors, which will give you more stability, but I don't have room to do that in this tiny servo body.
     
    Flip the servo over and lever off the top cover to reveal the gearbox - be very careful with this or you will have gears everywhere and will have to spend 10,000 years trying to work out what order they went in. I speak from experience...
     
    The output gear (the one sticking out the top with a spline on it) will have a tab on the bottom to limit the range it can turn. Cut/file this off. Before you put the output gear back, measure the spline diameter - about 3mm is fairly common.
     

     
    To add a gear to the spline instead of one of the 'horns' that came with it (no idea why they're called that) drill out a gear to slightly less than the spline diameter - drilling this gear to 3mm (by drilling to 2.5mm, then 3mm to reduce the risk of decentreing the hole) results in a snug fit. I used a crown gear as the servo has to sit flat under the chassis so needs a right angle before reaching the wheels.
     

     
    Re-assemble the gearbox and cover, then test - you should have a continuous rotation servo! If it rotates when the control stick is in the centre position, adjust the potentiometer with a small screwdriver until it stays still. If you want, give it a blob of superglue to make sure.
     

     

     
    The potentiometer can then tuck back into the body as it is not physically connected to the output gear any more, making the installation nice and compact. Job done!
     

  3. TurboSnail

    Day 1

    By TurboSnail,

    I consider myself fortunate to have a job still, so I don't have much more time to myself than usual, although I can break off during the day to spend 10 minutes painting something or run the 3D printer. So this little series is going to be a short account of the modelling I get up to each day, just to help keep myself amused. Some entries may be very short!

    The first job today was trying to paint a grey sky backscene - and here, we're not off to a great start as I used spray paint on to plywood. It doesn't stick! Lesson learned, brush emulsion ordered...
     
    Job 2, making some changes to my forthcoming Howard Petrol shunter kit. It used to be 2 wheel drive, but getting the front pickup to work and still have the wheels turning was proving tricky. I've redesigned it to be 4 wheel drive, and use a cheap motor, assembling which was today's job. It currently looks like this... Hopefully it'll run by tomorrow.
     

     
    Stats:
    Pasta reserves: Fair
    Conference calls taken: 2
    Times cursed at slow WiFi: 3
    Cumulative pushups done: 60
  4. TurboSnail
    I've taken some brave pills and had a go at weathering my Sentinel. I'd just finished it when I had a look at it in a few different lights, and it was way too black, even for a heavily used, unkempt loco. Fortunately the wash was still wet enough that by adding more thinners and rubbing it back with a paper towel, I was able to get it back to a more reasonable state. Not convinced it's the best job ever, but I think it looks better than the unweathered version, so I suppose it's progress.
     
    I've also finished off the figures I was working on, just simple 3-colour jobs on most of them, with the help of a fine-tip Sharpie for facial hair. Again, I probably should have done more research into period colours before painting them, but oh well, they'll do for now.


     

     
    Stats:
    Number of WiFi dropouts yesterday: 16(ish) - coincidentally whenever I was uploading or downloading large PCB files, grr...
    Number of days taken to work up the courage to weather the Sentinel: 5
    Cumulative push-up counter: 825
  5. TurboSnail
    Right, that'll do. Enough of this blogging nonsense! Finishing the Bagnall (for now) feels like a good place to stop. It needs varnish and weathering, but those will wait until the name and numberplates turn up. Overall, I'm pretty pleased with it, the modified bits have worked well, so I'll probably do something similar for more locos in the future. 
     
    Another project that's come up every now and then over the course of this blog is the AEC shunter, which now runs! Needs a flywheel adding to smooth it out a bit more, but I'm happy with where it is now, and looking forward to weathering it so it doesn't look quite so toylike. See video at the end of this post.
     
    But I'm stopping the blog here, mainly because I have to remember to post something each day, and if I have a less motivated day, or too much goes wrong, it just turns into trying to talk about nothing. So I'll go back to posting in my various threads, thanks to those who have followed and commented on this blog, I've tried to take onboard as much of the useful info as possible! If you do want to keep up with any of the things I've been working on, see these links:
     
    Workbench thread
    Layout thread
     

     

     
     
  6. TurboSnail
    My modelling workbench chair is one of those cheap office chairs, with a gas strut providing the lifting mechanism. I've got a sinking feeling that the gas strut has a leak...
     
    I rediscovered a couple of old projects (while looking for bits for another project), the NER Class H (later Y7) and the AEC shunter. Cue some faffing about curing intermittent pickup issues, and a short while later, the H is running happily up and down the test track on a 9v battery (video below). Nice and slow and smooth, too.
     
    The AEC is a different story, the chassis sort-of works, but has nowhere near enough traction, being 2-wheel drive. The body is already very heavy, so I can't add more weight to help it. This means a chassis redesign to make it 4wd, not too hard, but will take time.
     

     
     
  7. TurboSnail
    As hinted at yesterday, I've been re-jigging the chassis of the AEC shunter. It's now 4 wheel drive and hopefully still very easy to assemble - I'll find that out at some point, as it's currently printing. The body I've already made will hopefully be reusable. I say hopefully, as it will need some weight removing to fit the motor, and that's what caused me to break the little Hunslet a few days ago. Wish me luck...
     

     
    Stats:
    Built: Southall works 1938
    Power: 78hp
    Status: Currently being restored, not sure if intended to run on a preserved line somewhere
  8. TurboSnail
    Last night was taken up with the build of the new prototype AEC chassis. It now runs, with 4 wheel drive, but needs some tweaking to get it a bit smoother and quieter. It's built with 3D printed gears and no bearings (due to issues of supply for both), so it is very loud, and not great at smooth starts. I think I'll have to make another one with better components once all this isolation jazz is over.
     
    I also livestreamed this chassis build, but I wasn't very organised so it's a bit of a messy video!
     
     
     
    Stats:
    Time between printing and building the chassis: 20mins
    Total build time: about 60 mins, allowing for much faffing about and dismantling the old chassis
    Cumulative pushup counter: 1170
  9. TurboSnail
    I had a Ruston 48DS body lying around after nicking its chassis for the Sentinel. So I've done some CAD for the prototype version of the class, built in 1936, which allows it to be used much earlier than the Hornby examples. Next job is to see if it still fits the chassis! I'll probably make it's modified version as well as used by Heinz.
     

     
    Stats:
    Number of attempts it took to get the desk I ordered from the completely different address the bloody courier delivered it to: 3
    Number of times profanity was uttered due to the above: lost count
    Cumulative push-up counter: 1295
  10. TurboSnail
    The prototype 48DS (though it wasn't called a 48DS until much later) fits the footplate! Some nice design from Hornby here means I can reuse the footplate, which is great as it has a lot of weight in it, so it should still run well. Lots of pressure at work at the moment so my time is given over to that rather than modelling, but hopefully will get primer on soon.
     

     
    Stats:
    Number of bolts/screws left over from the desk and chair I put together yesterday: 8
    Level of concern that gives me: none
    Cumulative pushup counter: 1425
  11. TurboSnail
    I managed to prime and paint the little Ruston yesterday evening, although it's not the neatest job I've ever done. A generous weathering will hopefully hide most of my indiscretions. 
     

     
    Stats:
    Number of Ruston chassis' I need to buy to run all the loco bodies I've designed or am working on: 4 (this may also increase in future)
    Cumulative mugs of tea counter: 64 (estimated)
    Cumulative pushup counter: 1555
  12. TurboSnail
    I am becoming slightly obsessed with the Hornby Ruston 48DS chassis, to the extent that I'm now looking at a number of other things I might be able to fit it under. Suggestions welcome! 
     
    Last night I was too tired to do any proper modelling, so I had a mess around with my Howard/Hibberd/Planet CAD, and it looks like the Ruston could be made to fit. The wheels are a little bit too small, but reasonably well hidden on this loco, and the bonnet would have to be made about 2mm wider. I'm still debating whether this is an acceptable compromise - it would run much better than its current chassis.
     
    I'm also having a go at semi-freelancing another... thing - I'm not going to reveal what it's based on yet, as there's a good chance it won't work. It is very small though!
     

     

     
    Stats:
    Time spent on CAD instead of sleep: too much
    Hours before I have to go back to work (hooray!): 100(ish)
    Cumulative pushup counter: 1690 (ow!)
  13. TurboSnail
    I've not really had the motivation to do proper modelling the last couple of days, so I'm just sat doing more CAD. I've also been trying to fix my 3D printer as that's been playing up again. However, this has resulted in the prototype 48DS CAD being finalised, and the Heinz version modelled as well. Here they are in the print supporting software.
     

     
    Stats:
    Hours left on current 3D print: 7hr30mins (ooh, exciting...)
    Number of Rustons bought today: 1 (it won't be a Ruston for long, their chassis' are so nice)
    Cumulative pushup counter: 1830 (when I started this, I wasn't expecting it to go on for more than about 2 weeks...)
  14. TurboSnail
    I've now finished the Ruston works no. 177530, the prototype that would eventually become the 48DS. The loco was built in 1936, then was sold to Heinz, during whose ownership it was slightly modified (hence the two versions in yesterday's post). I've put together a basic video showing the build and the removal of the DCC socket that was needed to make it fit. Hopefully my voice isn't too soporific!
     
    If you'd like a set of prints to do this conversion, have a look at ts3dmodels.blogspot.com where I've got more info and contact details.
     
     
    Stats:
    Total hours spent building this loco: 3 (roughly)
    No. of attempts needed to get the loco to fit the footplate: 1 (First time! This never happens!)
    Cumulative pushup counter: 1975
  15. TurboSnail
    Howard lives! Last night all the wiring (there's not much) was finished, and the wheels given a good clean. Testing with a 9v battery (I still haven't made a controller yet), it works pretty well, and the pickup is good for such a short wheelbase. It could probably do with a little more weight if I'm going to pull more than about 5 wagons at a time, but the plan for it is to be one of the shunters for my planned 6-wagon Inglenook layout, so it should be fine.
     
    It has wooden gears, I'm hoping to have found a way to make plastic ones by the time it becomes a proper kit. Should make it quieter than it is now!
     
     
     
    Stats:
    Miles jogged: 1 (slowly)
    Mugs of tea consumed: 5
    Cumulative pushups done: 125
  16. TurboSnail
    I've got my science hat on today (it glows, and there are many bells and whistles) and am having a stab at making a controller. I haven't got one here and it's irritating trying to test locos with just a 9V battery. So, the plan is to make something that I can use for testing, which means adding a degree of automation to shuttle a loco back and forth along a short length of track. That means Arduino.
     
    So I've had a play around with a new Nano Every on a breadboard with a PWM motor driver - total cost less than £20. If I bought the unbranded eBay equivalents it would be even cheaper. I've got it set up running a motor up to full speed, hold for 5 seconds, then ramp back down to stop. Then the same in reverse. I've tested a load of different motors and they all seem to be running ok. Next step is to mess around with the programming a bit more, add manual control and make a nice neat PCB for it all to sit on. No small task then!
     

     
    Stats:
    Total motor test hours (so far): 2
    No. of motor types tested: 5
    Cumulative pushup counter: 2125
  17. TurboSnail
    More progress on the controller, I've managed to put some code together to make the controller respond to a potentiometer, operating as a centre-off (as I prefer that to having a reversing switch). Experimenting with throttle curves and various amounts of dead zone in the centre too, so it buzzes less and has nice smooth control at lower speeds. 

    So the next part is to make this all into a nice neat PCB. I've started on the schematic in KiCAD (free PCB design software), and so far I have the following additional specification:
    12V aux connector (for powering lights etc.) Additional reverser (in case I wire up a motor the wrong way round) Two terminal block inputs for additional inputs (might run a shuttle in future) Additional switch input (for some sort of future expansion) erm... Tea maker? Heath Robinson sandwich constructor?  


     
    Stats:
    Projects I've started to avoid having to paint a backscene: erm... at least 3
    Adequate reasons for making a controller rather than buying one: Not enough... apart from fun 
    Cumulative pushup counter: 2275
  18. TurboSnail
    Well, I sort of enjoy it anyway. PCB design is a bit like fitting together a puzzle, only there's no one set solution so you can be more creative. Especially when trying to fit things onto as small a board as possible. Here's the controller solution I've come up with, it could be made smaller (currently 55x70mm) but at the expense of ease of assembly and clarity. Regulator in the middle, directional LED at the bottom, power on (with LED) in the top corner and shuttle and reverser switches also at the top. Input and outputs all on the back of the board, along with the motor driver and Arduino.
     

     
    Stats:
    Approx. hours spent on this design: 2 (it's not that hard)
    Number of WiFi dropouts while trying to type this blog post: 3 (I don't type that slowly, my WiFi is just rubbish)
    Cumulative pushup counter: 2425 (I've got stuck at 150/day...)
  19. TurboSnail
    Erm... you may have noticed this post is a little later than usual... That's because I have done hardly any modelling today! Doing long hours at work is making me lose my motivation a bit, at a time when I really should be doing more modelling to distract myself and lower my stress levels (my work messaging app is are still beeping every few minutes - or would be if I hadn't just muted it). Oh well. 
     
    Anyway, I've put some writing on this PCB. Does that count as progress? I've tidied up the tracking a bit to make them less obvious and able to carry more current, and found sources for most of the components I want. The controller has also gained a name...
     
    Still to do on this: round off the corners/tidy up the measurements to nice round numbers (this will make enclosure and front panel design easier). Finalise component choice and order the bits. Then order the boards!
     

     
    Stats:
    Depth of material removed from the bottom of the barrel whilst writing this post: >3mm
    Cumulative pie-eating counter: 1 (how I've managed to go this long without one, I don't know - steak and kidney, since you ask)
    Cumulative pushup counter: 2575 (in a weak attempt to make up for the above)
  20. TurboSnail
    Well, I do have one other ongoing project... this semi-freelance version of a Planet 20hp. It's a bit larger, so maybe it's a 30hp? I don't know, it's just a bit of fun to keep my mind off other things. 
     

     
    Stats:
    No of my brains insufficiently active to generate stats: 1
    Cumulative pushup counter: 2725
  21. TurboSnail
    While I was messing around with controllers, I thought I'd make a thingy to help with some basic measurement jobs and to act as a reference for various things. No idea if it'll be any good, but worth a punt since it wasn't expensive to make.
     
    It's got an average OO person, back-to-back gauge, centre finder, scale ruler, actual ruler, track width gauge, some other handy measurements and wire/axle size identifier. Probably a bit small, but any bigger would have been more expensive.
     

     
    Stats:
    Total controller spend (so far): £50ish (but I nearly have enough components for two...)
    Cumulative pushup counter: 2875
  22. TurboSnail
    Following a request from a friend, I put together a print full of stuff that might be useful for scratchbuilding and bashing - lots of chimneys, springs, domes, a backhead etc.  Plus some steampunk bits... Hopefully it's all of some use! I've been thinking about providing the various bits of all the loco kits I make as spares, so if that's something you'd be interested in, do drop me a message.
     

     
    Stats:
    Successful controller automation tests: 5 (everything worked! Yay!)
    Cumulative pushup counter: 3025
  23. TurboSnail
    A more detailed look at supporting all the models I printed yesterday - I've been asked by several people for a bit more detail on this, so here it is... A bit of a marathon, this video!
     
     

     
    Stats:
    Success rate of these models: 90% (two small failures)
    Cumulative pushup counter: 3175
  24. TurboSnail
    A bit more development on the controller concept - PCBs have been ordered, so I'm doing some more work on the software. It's now set up to respond to the throttle using centre-off (so no clicking a reversing switch every few seconds while shunting) and it seems reasonably smooth. I had to incorporate a bit of a dead zone in the centre otherwise it was a bit twitchy - it seems pretty smooth now. I've also wired in a light-dependant resistor, which stops power to the track when it senses dark - fortunately this works, as the little Ruston would fall off the end of the desk otherwise!
     

     
    Stats:
    No. of lines of code needed to make this work: 55 (with extra spaces to make it easier to read)
    Correct resistor to use with a 20k to 2M LDR: 1k (this took considerable experimentation to get right)
    Cumulative pushup counter: 3325
  25. TurboSnail
    I should probably explain my choice of components for the controller a little bit. It's a mix of SMD and through-hole, normally you'd go one way or the other, but I felt in this instance it'd be better to compromise. As such all the resistors, caps and LEDs are SMD to reduce the number of holes in the PCB, hopefully making it look neater. The same goes for the 2x 15-pin headers holding the Arduino on. The rest (switches, terminal blocks etc.) are through-hole, as that's what is readily available and cheap. This hopefully confines most of the messiness to the outer edges of the PCB.
     
    The other thing I'm quite happy with is finding a throttle potentiometer with a centre-detent - on of the biggest issues I have with my current 'proper' controller (a Morley Vector, otherwise excellent) is that you can't tell by feel alone when the controller is centered, and therefore off. This leads to occasions when you think you've turned off the power, when in fact it's still on and your loco is sat on a dead spot, so that as soon as you knock the boards or move the loco (to uncouple it, for example), it takes off again. Having a simple notch so you can feel the off position should solve that nicely.
     

     
    Stats:
    Total component count: 23
    Cumulative pushup counter: 3475
×
×
  • Create New...