Category: Uncategorized

After looking at the specifications of each actuator, here is a summary of the actuator characteristics:

1. 500mm actuator, 12V DC, 10mm/s, 750N load
2. 300mm actuator, 12V DC, 5mm/s, 5000N load

Two main conclusions can be drawn from this:

1. 300mm will extend more slowly than the 500mm actuator.
2. Our chosen relay and programmable logic controller must be able to handle supplying enough current for 4x 12V DC

 

For the control of the linear actuators, we can choose between relays, motor controllers, or a combination of both. We probably need a combination of both so we procured relays and motor controllers.

 

For relays, we have procured the Jayoyi 16 Channel Relay Board Module Optocoupler. It is well-suited for our project as its power supply satisfies the DC 12V range. Also, it is one of the best PLC relay expansion modules available as it comes with 16 channels and has an optocoupler. The perk of an PLC relay expansion module is that it increases the number of safe outputs of another safety relay. These are suitable in applications where several devices need to be stopped. The optocoupler enables electrical signals to be passed between low and high voltage circuits via short optical transmission paths while keeping them electrically isolated. This removes electrical noise from signals, and isolate and protect-voltage devices from high-voltage circuits. Essentially, this ensures that our lifting arm can work properly without erratic operation, which is certainly important for safety!

 

For motor controller, we have procured the Dual Channel H Bridge Motor Drive Module Controller. It satisfies the DC 12V range, has reversible control and PWM speed control. This means that there can be finer control over our actuators. The actuators should be able to be locked in certain position with the motor controller. This will be helpful for the control of the double actuator configuration.

A brand new week! What shall week 9 hold for us? We are as excited as you to find out!

First and foremost, if you still remember, last Friday, we tested multiple clamp-grips, and where necessary, redesigned and reprinted them. Now, they were all ready for another round of testing, together with the clamps.

The first clamp to be tested is the black one that would grip sideways. The video below shows the optimised clamp, which is a pull clamp, with the black grip.

Similar to the red sideway grip that locks onto the brown metal tubing, we designed an almost identical, mirror-image pink grip that locks onto the adjacent brown metal tubing, rotating inwards from the right, with the help of a pull clamp.

Thirdly, we also designed and built a new clamp that fits nicely onto one of the rectangular boxes on the pneumatic unit. As this clamp comes down vertically from the top of the unit, we were cautious of the clearance, and took care to make it as short as possible (please see the photo and video below).

You might also remember that our clamp grip with 3 holes did not fit into the unit last week. Our redesigned clamp grip now, does. 🙂 *Insert big smiley face*

First, you can see below that in the originally printed grip, the two holes that are supposed to fit into the nuts on the unit are just slightly off from the nuts (short of reach) and they cannot reach the nuts. We coupled the revised clamp with a push clamp, as seen in the video that follows.

All in all, most of today was spent re-optimising all clamps to ensure they can fully tighten and secure the unit. Other than the ones already shown above, this included adding reinforcements in the form of more heavy-duty angled brackets (circled with blue and red) into the aluminium profiles of the clamping mechanism.

Optimisation also included drilling and filing the profiles to ensure that whatever M6 screws we needed would go fully into the profiles and be flush with the outer edge of the profiles, without protruding out of the profile (similar to what we did last week with the short aluminium supports on the clamping mechanism) (please scroll down to the end of that post). 

Here is a neat video showing all the clamps working so far!!!

Other than the clamps, we also ordered more aluminium rods from the friendly Mr Raman from the SPMS workshop.

Lastly, we are starting to consider the control method for the actuators in the lifting arm. In the next post, we will elaborate in detail the options we considered and the method we eventually go with.

Today, we did the following!

We modified the base by shifting the position of the two long aluminium profiles within the base so that they can support two adjacent vertical aluminium profiles that would serve as the vertical support tower.

The below photo shows us “in the flow” as we adjust the two aluminium profiles within the base. The modular design of the aluminium profile assembly makes it relatively easy to make any adjustments we wish: we simply have to unscrew the brackets, shift the position of the aluminium profiles, and then rescrew the brackets. However, for longer and larger profiles like these, the brackets are larger, and also harder to be adjusted and screwed precisely, so this took some time.

 

Once the base was ready, we proceeded to cut the aluminium profiles for the support tower (2x 1.5m), and then attached the primary lifting arm with its actuator.

The below video shows us cutting the profile with the aluminium cutter.

This photo shows the heavy-duty brackets we used to secure the vertical profile with the horizontal profile on the base. We made sure to get high-qualtiy heavy-duty brackets from Prestech (even though they are significantly more expensive) as this joint bears the entire weight of the whole arm and it  is extremely important for the integrity of the arm.

 

After attaching the two vertical aluminium profiles, we had to attach the pillow blocks to serve as the freely rotatable joint between the primary lifting arm and the two vertical support profiles. This video shows the process.

The first linear actuator (attached to the vertical support tower using the actuator brackets that we bought) will then provide the control needed for the primary lifting arm. The below photo is the completed assembly of this stage, and the video below that shows the actuator in action – as the actuator extends, the primary lifting arm is raised higher, with the pillow block as the pivot.

Neat! The actuator is able to smoothly control the angle of the primary lifting arm without any problems, for now. As per our calculations and hopefully, it should retain this smoothness when the secondary lifting arm + clamping system + 18kg load is attached.

We decided we had enough of the main arm, and we continued working on the clamping mechanism. Today, we designed and 3D-printed customised clamp grips (2 completed 3D-printed grips, 1 currently printing) and then tested them out. Where needed, we also paired these grips with appropriate clamps.

1) This yellow clamp below will lock onto the rectangular reliefed part of the pneumatic unit using the two screws on the relief (highlighted below with a red box). This will be done by simply pushing in the yellow clamp into the reliefed part, which will then help to prevent any unwanted vertical motion of the pneumatic unit.

2) We had also 3d-printed another red grip (on 1st July) to lock onto the thick brown metal wires. However, we had yet to attach a clamp to the grip, so we did that today. We also optimised the relative distance of the clamp and the aluminium profile attached to the red grip to achieve tight and smooth clamping of the brown metal tubing.

This video shows the initial clamp, which was not tight and also suffered from a rough, disjointed motion. Thus, we adjusted the aluminium profile (on the left) attached to the clamp closer to the right, nearer to the profile (on the right) attached to the red grip. The second video shows a much smoother motion of clamping. Hoorah! #Optimisation

 

3) We also 3d-printed a black grip that would fit into the relief with 4 big screws on the unit. When done correctly, this will be a very good, tight grip that would prevent motion in all directions except the horizontal, forward direction.

To refresh your memory, this is how the clamp should go in.

<to show photo/video of the black grip going in with the assembly> 

4) Another good point to grip is shown below: Again, we chose to clamp from the side, rather than from the top. We set about designing the clamp, and it is currently printing! We shall test this grip with a clamp when it is ready. 

5)  Yet another good place for a grip is this: a pair of very solid-looking bolts that could be gripped from the vertical direction. We have to be careful though, as we well know by now, the vertical clearance is limited, so the height of this clamp system has to be as short as practically possible.

Below is the solidworks file of the clamp. The two holes at the end will fit into the two nuts, and the hole in the middle is for the M6 screw to secure this grip into an aluminium profile.

However, when we printed and tested this grip, the holes were not aligned (unfortunate!), which meant more precise measurements and careful adjustment of their positions in the file. Also, the protruding ledge was too long, and the grip could not fit into place. We shall redesign and reprint, and then test the revised grip tomorrow.

Another day, another us. We came in with a mentality to clear the clearance problem. Will we do it today? Let’s find out shortly..

Luckily, our creative juices were flowing quite fast. One initial idea was to try unclamping only partially (to reduce the need for clearance), and to try rotating the clamping mechanism with the pneumatic unit slowly out of the installation chamber. This seemed promising, I mean, we didn’t feel that we have many options at this point, and trying to slowly squeeze through the clearance seemed an option.

<try to insert photo?>

With some renewed confidence, we (confidently) walked up to Tony to pitch our grand idea.

Drum rolls please…in 3… 2… 1…

Badumtss.

“Cannot lah!”

Instant rejection.

Because the clamps are designed for full unclamping, a partial unclamping would make it difficult for the technicians. In the dark space, they would have to judge how much to partially unclamp so that there was just enough clearance for the unit to squeeze through.

More importantly, if they did not gauge correctly, the clamp could clash with the metal wires, and if the lifting system continued to pull the unit out of the chamber, the metal wires could be potentially damaged by the clamps. Not the best idea for now.

But of course, we were not ready to give up. Within minutes, we pushed out another idea into the pipeline. The existing clamps were unnecessarily long, and the clearance we thought we needed might have been unnecessary.

So, we tried to make the clamps as short/small as we could, and hopefully that would decrease the clearance needed. Tony said this was a feasible route, and told us to just get our hands working on it. And so we did.

<insert more photos? if any> 

In the meantime, in a bid to find new inspiration, we decided to revisit the photos we took during the depot visits. Unfortunately, for the worse, not better, we chanced across this one…

Terrible news for us. While Fadzuli’s very-kindly-done measurements (and somewhat misleading photos) seemed to suggest quite a bit of vertical clearance, this was clearly not the case! Here’s a recap on what he sent us a few days ago: He had labelled this at 25 cm. Since the width of the unit was 10 cm by our previous measurements, the vertical clearance, logically, would be 15 cm. Lo and behold, from the photo previously, we were doubtful if even a 2020 aluminium profile could slide through that narrow clearance. Bad, bad news.

When Tony heard this, the look on his face was priceless.

“Ok definitely need to change liao. This one confirm cannot”

So a complete overhaul was in order. But how?

The first idea was to try to eliminate the clamping system on the top side of the unit, somehow. It made some sense – since the clamps were the ones causing so much clearance problems, our problems would be cleared if we cleared the clamps! Right?

The idea went something like this: Of course, clamps were still necessary to keep the unit secure and bring it up to the installation point. But perhaps, just perhaps, we could get the technicians to unclamp the top-side clamps when the whole unit is just reaching the installation point (but not yet locked in). There would still be ample clearance to unclamp the top-side clamps.

After they are unclamped, some sort of rotating joint could allow these clamps to be rotated inwards and folded just above the unit, so that they won’t cause any clearance issues. Since it was just maybe 20-30 cm to the installation chamber, we reckoned that a gentle movement by the arm was all that was needed to move the unit into the installation point and the rest was for the technicians to manually slide and lock the unit.

It was a very appealing idea. If clamps caused the problem, remove the clamps, and problem is solved, right?

Of course, all ideas must pass through our God Tony, just as all bills must be passed by the parliament.

“Ehhh, maybe you can try. Eh actually… cannot lah. There is too much risk leh.”

After some intense debate, we admitted defeat. If the arm malfunctioned or lost balance just a little bit, there simply was not enough clamping force/security to catch the unit. We were imagining the ideal situation where the bottom side clamps and some minor locking profiles would be sufficient to keep the unit secured ALL THE TIME, under ALL SITUATIONS. This was, in an engineering context, simply asking for too much. One wrong push by the operator of the lifting system, or a slip by the technician, and the unit could be sent flying DOWNWARDS onto the floor, and possibly shatter into pieces, along with the sanity (and the jobs) of whoever witnessed that event.

It was time to put on our thinking caps again.

*thinking cap time*

*thinking cap time*

*thinking cap time*

*thinking cap time*

Oh, some additional arrivals today, by the way: 4 x 300mm Actuators (5000N ones); 100 x brackets for 2020 profile (clamping mechanism).

The actuators were indeed 5000N rated, a fact met with a big sigh of relief by the whole team. These were the main lifters of our primary arm, and we could allow for absolutely no leeway in the carrying capacity of these actuators.

Similar to the 750N ones, we did an informal test with the table. Here is a photo: (we’d like to emphasise that the person (whoever that was) in the photo was on the OTHER table, and there was NO ONE on the table that we were testing)

The table was lifted above the ground with absolutely no issues. Awesome! We still need to wait for the pillow blocks to arrive before we can build the arm, so we put that aside for now.

*thinking cap time*

*thinking cap time*

*I think we ate some food*

*Yea we definitely did. I ordered some Lei Cha from the vegetarian stall. Lei Cha is nice. I recommend Lei Cha.* 

*thinking cap time*

*thinking cap time*

Eureka! After some time, we think we got the RIGHT idea.

Surprise, surprise. An idea so brutally simple that it was outright insulting to our collective intelligence.

Instead of clamping the pneumatic unit from the top, why not clamp it from the side? Why not?

No one really knows why we did not think of this idea. It was just an assumption that all of us were working with, and failed to challenge when all assumptions needed to be broken.

Here are the photos of the rough assembly that we got out by the end of Friday. As you can see, the clamping system now approaches the unit from one side of the base of the unit, rather than from its top.

Very beautiful pictures. Very beautiful clamps. Very beautiful project.

Here is a random easter-egg photo of 1 x Min Htoo drilling holes into the 2020 aluminium profiles so that they can be secured tightly onto the 3060 main aluminium profile and act as supports and locking profiles for the clamping mechanism.

 

Well, after a lot of ideating, trial and error, assembling, and some drilling and cutting, we were done for the day, with a much better outlook of what’s to come, and much healthier sanity levels. See you all!

Yet another day, we rise to the gentle embrace of the morning sun and the ever-so-familiar song of the “uwu” bird. And of course, the first thought on our minds is that damned clamping mechanism.

Luckily, Qian Xin had delivered us some great news the previous week that we had just been assigned a mentor, by the name of Savior Jie Huang, who’s a CN Yang alumni in Mechanical Engineering and is currently doing PhD studies in the same field. Will he deliver us from the pain of the clearance problem? We walked (or took the bus) with great anticipation for this mysterious hero, expecting him to disseminate nuggets of infinite knowledge to us like a tree dispersing its seeds.

Unfortunately, he was as equally stuck as us on the problem. He suggested some ideas on telescoped arm (basically a retractable arm system), but we had already explored that idea and it’s too commercially difficult to source, not to mention it will struggle with the load. It won’t really help us with the clearance method either.

LUCKILY, he does have some working knowledge of mechanical simulations and SolidWorks, so we passed our assembly files to him for him to tinker around a bit. Our attempts at SolidWorks (while very valiant) had always ended up in the whole arm system disintegrating into many parts due to a SolidWorks glitch whenever we put in the actuator sub-assemblies (they need a length limit so they don’t overextend or overcontract). Not so valiant, I guess.

As for the clamping problem, what we really needed was a new way to clamp the whole unit and avoid the metal wires, the metal handle and the roof of the train…

After Jie Huang left, some slightly better news: the heavy-weights of our lifting system, the linear actuators arrived! These were the 3 x 500mm Linear Actuators (+ 8 Actuator brackets to attach these to the 3060 aluminium profiles)

Wonderful news! We could finally toy around with these actuators…Nope. How could the day go so well? Of course not. The actuators were actually rated 750N, not the 1500N we had selected.

HOWEVER, we also knew that these were for the secondary arm and not the main arm, and the load they would have to support is just that of the pneumatic unit and of course of the secondary lifting system itself. Furthermore, we’ll be using multiple actuators, so their combined carrying ability should be well within safe limits.

As a rough test, we also tried to lift one of the tables in the MnT lab with the actuator ALONE, and guess what, it managed to do it without any problems. That’s awesome news that we didn’t get a faulty actuator.

In some other news (while still thinking about the clearance problem), Tony instructed us to source for more nuts, screws and brackets for the aluminium profiles, as he reckoned that what we bought might not be enough (given that some of them were bound to be slightly defected as well). Thankfully, Terence was around in SPMS, and he readily agreed to us dissembling his prototype, and mind you, that was the best craftsmanship prototype for last year! How gracious of him and his group.

Also, not much new inspiration hit us, so we proceeded with the 2x heavy duty ball joints, which also just arrived. We needed a way to secure the heavy duty ball joints to the whole clamping system. To ensure maximal fit, we decided to custom-design the ball socket instead of buying one. Ideally, this double-ball-joint system would allow us to rotate the clamping mechanism quite freely and also allow the manual sliding motion for the technicians to lock the pneumatic unit into the train.

Our first prototype for the socket looked something like this! The important thing was to make sure that the inner curve of our 3D print fits the ball as much as possible, which it did! Thank you to the Ones above. Our day is slowly getting better.

Before we knew it, it was already 6+ pm, so we decided to call it a day. The clearance problem continued to linger over us, and it was quite clear (pun intended) from our facial expressions. We decided to heed Tony’s advice, and just wait for that “shower” moment. Good bye, for today!

As we overhauled our initial vertical clamping method, we had to rethink all the clamps (gg so much effort).

 

However, with our collective genius (!!!), we managed to create 3 new clamps using toggle clamps and some 3D printed parts. In the creation of these 3 new clamps, we definitely kept some crucial concepts in mind:

1. Secure the stability of the unit with the maximum number of points of contact to stabilize the unit in a 3D space. The unit will likely be rotated 270 degrees about its longest axis and may be subject to some rotation about its short axis across the unit, perpendicular to the unit, due to the rotation of the long unit to fit through the train door.
2. Do not cross the top most bunch of wires as this will cause our clamps to hit the roof of the working space.
3. Clamps must be able to be slid in from the base of the unit without requiring vertical motion. This will enable us to better unclamp and slide our the unit in the working space.
4. Clamps used are restricted to toggle and push clamp connected to aluminum profiles directly or via brackets
5. Power supply housing and wires should be avoided and not directly clamped. Too much force should not be exerted on these sensitive electrical components as they may be damaged.

 

Here is what we have done so far:

 

From right to left, Clamp 1 and Clamp 2 are pictured below:

 

 

As can be seen, Clamp 1 has 2 points of contact. Using the current pictured orientation of the unit as our reference, clamp 1 will prevent motion upwards and towards us.

 

Similarly, Clamp 2 has 3 points of contact. Using the current pictured orientation of the unit as our reference, clamp 2 will prevent motion upwards, towards us and to the right.

 

Clamp 3 is pictured below and it has 3 points of contact too. 2 points of contact to prevent the unit from moving towards us and 1 to prevent the unit from moving downwards.

upload images

 

There is a direct relationship between the points of contact and the stability of the unit when being lifted. Hence, we are aiming to have at least 10 points of contact for clamps in total when the whole clamping system is done. Since there are already 8 points of contact with these 3 clamps to the right of the CG of the unit, our progress with the clamps seems very promising!!

 

Also, we made progress with solidworks. We created a holder for the long L-shape metal tubing, to be coupled with a clamp.

We first thought of this design below, but realised it was not the best idea to approach and clamp the metal tubing from the top, since we want to avoid any kind of clamping from the top as much as possible given the limited vertical clearance.

Furthermore, it would also not be able to prevent any downward vertical movement of the metal tubing (away from the grip), which makes the unit more prone to becoming loose from this grip.

A much better idea was to clamp the metal tubing from the side using a rotatable aluminium profile. (Please note that the photos were taken from 2nd July’s post, purely for illustration purposes of why we chose this design. On 1st July, we had not yet printed this grip.)

This is how the locked red 3d grip would look like after the profile was rotated clockwise (using an appropriate clamp). This way, the metal tubing cannot drop loose in the downward vertical direction (unlike the previous design). The only can it can become loose out of the groove is if it moves horizontally to the right. Thus, we settled with this second design.

This design will be subject to some adjustments, but the addition of this grip will increase in the number of points of contact for the unit, which is, of course, good news.

Also, we 3D printed and refined the design of the ball joint holder”

Here are our 2 ball joints on top of 3 2020 aluminium profiles. This will make up our double ball joint configuration in between our third arm and clamping system for great adjustability of the clamping system. It works by allowing the ball joints on each side to rotate individually. The combined rotation of the 2 ball joints will allow the whole double ball joint configuration to be especially adjustable.

upload images

 

Hello all!

We have built the aluminium framework that forms the base of our entire lifting system. The dimensions were informed by our prior calculations on Monday, 24th June, which involved finding the centre of gravity (CG) of our pneumatic unit.

<to be updated>

In case you forgot already, our actual pneumatic unit from the SMRT Tuas Depot was supposed to arrive yesterday (20th June). Unfortunately, a puncture tyre poked a hole in that plan.

Finally this afternoon, our friendly LTA engineer secured another van to deliver the actual, new pneumatic unit to the Making & Tinkering Lab for us to work on. (unit labelled with a big red circle in case you can’t see it amongst the organised mess of the bench)

As a comparison, here is the old unit (now sadly dumped into a corner of the MnT lab store)

Now, we were in for a slightly nasty surprise – there were a few key differences between the old and the new unit, which increases the obstacles for our clamping system and makes our life significantly harder.

Presence of wiring system and additional metal rod on the top side of the unit

• • The original unit. (the top side is to the right) As you can see, it is extremely bare around the parameter
  • The new unit:
  • These wires and the rod present a significant obstacle to our clamping system. as we realised that any of the gaps between the original grey metal bar of the pneumatic unit and the wires/bar (see below images) is too small for the aluminium profiles of our clamping mechanism to fit
    (the 2020 aluminium profile is too thick to fit through the first gap)
    
    (the 2020 aluminium profile is also too thick to fit through the second gap – although it can squeeze through the two metal bars, it is blocked by the base of the unit):
  • As such, our clamp will have to come from all the way above these wires.
  • Luckily, the design of our clamp uses angle brackets which allow easy adjustment of the width of the clamp, so all it needed was to unscrew the angle brackets, adjust the width of the clamp, and to screw the brackets in again

Additional enlarged metal handle in middle of unit

• • The old unit:
    
  • The new unit:
  • This metal handle is significantly larger than that on the old unit, and it even overhangs beyond the base of the new unit. Thus, our clamping system will have to avoid this metal handle.

Well, time to work on revising our clamping system so that we can overcome these challenges! Adios, for now.

Today, we hoped to finally get the pneumatic unit to work on. However, unfortunate circumstances and a nasty punctured tyre poked a hole in that plan. 🙁

Tomorrow, may be a better day.

 

On the bright side, we refined our clamping system by adding a long (adjustable) clamp that can reach to under some of the attachments to secure the unit in place while it is rotated upside down.

 

 

Also, we were linked up with our mentor, Jie Huang, who has some experience in Solidworks! He gave us some valuable advice – focus on a design that can lift our payload first, then look into stress/strain analysis. We will also be meeting him next week. 🙂

 

Finally, we procured some pillow blocks with 30mm bore diameter for use as rotating joints to connect two the two support aluminium profiles with the main lifting arm (rough sketch below). The pillow block allows 360 degree free rotation ability via the aluminium rod (30mm diameter) that fits through the 3 pillow blocks. To install control, actuators will be installed (as per our prototype designs) between the two support aluminium profile and the main lifting arm (not drawn below for simplicity).

We were in talks with the LTA engineers to get a pneumatic unit to work with in lab. After a few days of waiting, they informed us that there was a pneumatic unit in the SMRT-NTU Smart Urban Rail Corporate Lab and linked us up with a researcher working in that lab! Finally, we could get our hands on the pneumatic unit and start visualizing and materializing our solution. 🙂

 

We went over to the SMRT-NTU Corp Lab with a trolley and brought the unit to the Making and Tinkering Lab. However, we soon realized that this pneumatic unit (which was on display in the SMRT-NTU Corp Lab) is not the same as the unit currently used in the trains. The pneumatic unit we got from SMRT-NTU Corp Lab is pictured below on the left and the pneumatic unit used for repair and replacement works in Tuas Depot is pictured on the right.

 

 

We won’t be able to get the actual unit anytime soon too.

 

This meant that we would be working with different dimensions and different attachments on the metal base of the pneumatic unit whilst brainstorming and creating our clamp system. Our crafted clamp system had to be adjustable to allow for these tweaks in dimension and attachments. Also, this probably meant that we will not be able to custom-build a grip that can hold onto any of the attachments.

 

We had to think of a way to work around this… Thinking caps on!

 

Also, since we will visit the Tuas Depot again tomorrow, we brainstormed some of the train and depot measurements to take:

1. Height between train floor and top of door
2. Height between depot floor and train floor
3. Distance between pneumatic unit and LED screen cover
4. Height between bottom of pneumatic unit and top of LED screen cover (this area will be known as the Working Space)
5. Angle of pneumatic unit with respect to normal from ground
6. Length by which to slide pneumatic unit sideways
7. Distance between two doors in one car
8. Distance between the door edge and vertical grab bar (if present)
9. Distance between the glass barrier beside the seats on both sides of the door
10. Maximum height of the train car
11. Length, breadth and height of pneumatic unit