Goodbye Linode

Recently, I’ve been bombarded with the notifications saying that my disk io rate treshold usage has exceeded certain amount up to 55819/1000. By that time, I only have one self-hosted wordpress blog.

I have been a Linode user since March 2013, therefore it has been more than a year an a half before I pulled the plug.

It has been a nice ride, learning how to set-up a hosting server using ubuntu, doing upgrades like a pro sysadmin, hacking through terminal and so on. I’ll miss that and probably I’ll come back later.

It was just me. At this point in life, I just want to focus on writing my blog. So, goodbye Linode, you have been great.

General Tuning Advice for Arducopter

I’m going to pick some important points in terms of tuning the arducopter. At the time of writing, the current version is AC 3.1.

My two main problems with my tricopter are these two: maintaining Alt-Hold and wobbling oscillation while quickly descending.

So, here goes:

via []

These two terms (Rate Control P, Throttle Rate P) go a long way to making your copter perform well, and are highly dependent on the thrust to weight ratio of your copter. More thrust = lower gains.

  • My copter oscillates quickly (smaller movements) when stabilized: Lower your gain in Rate Control P.
  • My copter oscillates slowly (larger movements) when stabilized: Lower your gain in Stabilize Control P.
  • My copter gets wobbly when descending quickly: Raise your gain in Rate Control P. You can tune most of this out, but it’s impossible to descend into your prop wash without some wobbles.
  • My copter is too sluggish : Raise your Stabilize Control P gain. These gains make ask the copter to respond faster to angle errors.
  • My copter yaws right or left 15° when I take off: Your Motors are not straight or your ESCs are not calibrated. Twist the motors until they are straight. Run the ESC calibration routine.
  • My copter angles back in the opposite direction I’m flying after forward flight: Raise your gain in Stabilize Control P.
  • After aggressive flying my copter leans to one side 10 – 30°: Do what ever you can to remove vibrations and isolate APM. You can even try and add a small weight to APM to keep it from vibrating. Also, just land for a few seconds, then continue flying, it will give APM a chance to correct without the vibration.
  • My copter won’t stay perfectly still in the air: Run the level command on a flat surface. (hold disarm for 15 seconds to invoke.) You can also fly in auto-trim mode in a windless (important!) environment. Any wind will cause the changes you make to work against you when the quad rotates 180°. You can use your radio pitch and roll trims, but remember to re-center them when you set up your radio with the config tools. Using the radio trims can have a negative effect in SIMPLE mode when yawing. Never Trim Yaw, Your copter may start spinning on it’s own. (Quads are also susceptible to drafts. They will need constant corrections unless you install an optical flow sensor. One day…)
  • My copter flies well, but then dips a motor arm in a fidgety manner while hovering: Your motor may be going bad or the bullets connectors between the motor and ESC may be at fault. Vibration from a bent shaft or unbalanced props can make bullet connections fail momentarily stopping the motor. A motor with bad bearings takes more power to spin. The ESC could be cutting out to protect itself. Or it maybe flying slow enough to stall the motor. Attach a current sensor between your battery and the motor and test the difference between a good motor and the bad one. If you are seeing higher draw in the bad one, replace or fix it.
  • In Loiter, my copter constantly overshoots: Try and increase your Nav_P term. You can also tune down your I terms because in some cases, the Nav_I or Loiter_I can causes overshoot. Making the Iterms 0 when in still wind is the best way to tune Nav_P.
  • My alt hold above 10 feet is only about 1-2m accurate: That’s actually the best you can achieve. The baro sensor is sensitive to light and wind. Use a felt cover to protect it from wind and make sure you don’t glue the hole and seal it up.
  • My copter increasingly swings up and down in alt hold. It eventually get’s down to the ground: Your THROTTLE_P is too high or low. You don’t need a lot of P to do alt hold. Think of how much you move the throttle to hold alt perfectly. Not much! That’s what you need P to do. I will ramp up as your battery goes lower to make up the difference.
  • I have a sonar installed but I can’t get it to work. Your sonar may be picking up electrical noise from the ESCs. Be sure to move it a few inches from any ESCs and try a shielded cable, if possible. If it’s cold outside, your sonar may stop working.
  • My quad loiters by rotating in a CW or CCW circle: Adjust your compass declination until it stops circling.

Zoysia Grass Appear On Day 7

Here in my place, the temperature were constantly hot at about 37-39 Celcius. Following some advices from the forum, I began the seeding process of the Zoysia grass that I obtained from SeedRanch via EBay.

I’ve made sure that the top soil was smooth and level, and evenly spreading the grass seed. Afterwards, I’ve used the back of the rake to softly mixing the seeds and the soil. The seeds were approximately resides at about an inch in the soft top soil.

Next, I’ve used a helicopter sprinkler to evenly water the soil. Since I’m always at work, I spent about 5 minutes watering the soil at 7.30 am, and later another 5 minutes at 5.30 pm.

I was a bit nervous, thinking whether the grass will survive or not. However most of the materials that I’ve read outlined that Zoysia grass is probably one of the type that is very robust and aggressive. The photo below shows the grass development at day-7.


Take note that the grass seeds started to germinate probably in between these 7 days. So no rush, and don’t worry if you’re just started.

Next, I’ll just keep on watering the grass morning and afternoon till the root network is strong then I can water it once a week.

Will post more on the development of the grass. Thanks for reading!

I bought a Flashforge Creator 3D Printer

I ordered a Flashforge Creator 3D printer on 30th Jun 2013 from AliExpress. The item was safely arrived from mainland China by FedEX on the 4th July. Pretty quick, and I’m very happy to report that the machine is doing very well!


As usual, I started off by calibrating the build platform. The Creator model was derived from Makerbot Replicator, therefore the LCD interface shows the similar guide on leveling the platform step by step.

Once done, I printed the calibration box that came together inside of the SD Card. The current model has a heated bed platform, therefore it is a good idea to print the model with raft support to ensure that the printed item stick on the platform well.


Flashforge Creator is much better than the Mbot3D printer. Creator operates much quiter, the filament guide is more well thought for and it comes with heated bed platform. The price is much cheaper than the Mbot3D as well, that might be depending on which model that you are getting. My Mbot3D is a dual extruder model (which I got wrong on more extruder is better, you’ll be happy with one, trust me) that cost me $1k++ with shipping, and the Creator is a single extruder with heated bed platform that cost me $800++ with shipping.

As for me, I think that I’m going to stick with Flashforge Creator 3D printer for a while. However I’m still having fun with the no-heat bed platform on my Mbot3D. So the journey continues…

Mengapa 3D designer memerlukan 3D printer

Seseorang designer memerlukan 3D printer untuk melihat bagaimana prototype yang telah direka berfungsi dan dipersembahkan, sama seperti seorang penulis memerlukan printer untuk menyebarkan penulisan mereka.

Mungkin ada yg akan argue, mengapa menggunakan 3D printer, walhal injection molding telah ada? Pada fikiran saya, situasi ini sama seperti ‘mengapa menggunakan desktop printer kalau di tempat kita ada syarikat percetakan?’.

Rasionalnya adalah, syarikat percetakan mencetak untuk skala besar, sama seperti injection molding, yang mana hasilnya memberikan kos yang rendah. Manakala desktop printer digunakan untuk mencetak satu atau dua salinan teks, sama seperti 3D printer, kosnya lebih rendah dan efektif untuk semakan.

Sekiranya sesebuah research group itu terlibat dengan prototype design, saya rasa 3D printer adalah sebuah indispensable tool. Bayangkan dari fasa idea, lukisan CAD 3D dan akhirnya functional prototype di dalam masa 4 jam di atas meja di pejabat.

Saya sedang merekabentuk self-stabilizing camera holder (gimbal) untuk quadrotor saya. Nanti kalau ada gambar baru, saya akan upload lagi. Till then, see you in the next update : )



Penyelesaian kepada botol Medela with Suction Cup yang tak stabil.

Medela adalah salah satu daripada pengeluar pam susu yang berkualiti tinggi. Bagi satu pakej standard, Medela membekalkan botol, pam susu dan suction set. Walaupun ia tergolong di dalam pakej premium bagi para ibu, saya diberitahu oleh isteri saya bahawa botol yang dipasang suction set biasanya tidak stabil. Medela ada bottle holder, tapi ia tak termasuk di dalam set yang biasanya dipasarkan. Selalunya kalau baru siap pump, susu akan tumpah walau tersentuh sedikit suction cup tu. Menangislah para ibu time tu.


Walaupun Medela ada bottle holder mereka, ia tak mudah didapati melalui pembelian online atau di kedai-kedai. Mungkin mereka tak perasan bahawa masalah susu tumpah ini adalah masalah yang real. Oleh yang demikian, sebagai seorang designer, saya perlu mengambil langkah pro-aktif untuk menyelesaikan masalah ini. Lagipun 3D printer saya masih lagi di dalam calibration stage.

2013-06-28 09.26.00

Mula-mula, saya menggunakan ProE (CAD software) untuk merekabentuk bottle holder. Saya buat ukuran diameter standard Medela bottle dan menjarakkan sedikit supaya suction set ada sedikit clearance.

Kemudian, model tersebut di save di dalam format .STL supaya boleh dibaca oleh ReplicatorG. ReplicatorG diperlukan untuk produce G-Code yang boleh difahami oleh 3D printer saya. Setelah siap, G-Code tersebut dihantar ke 3D printer untuk menjalani proses printing. Oh, saya letak engrave nama di atas produk ini untuk melihat sama ada 3D printer saya boleh mencetak nama di atas produk yang dihasilkan.

Keseluruhan proses design mengambil masa selama kurang dari 10 minit (saya memang dah biasa dengan ProE sejak 2004). Manakalau proses printing pula mengambil masa selama 4 jam. Saya menggunakan infill sebanyak 5% kerana model ini tak bergerak dan tak perlu terlalu lasak. Feedrate yang digunakan adalah 30mm/s, dengan suhu 230 C bagi menghasilkan lekatan plastik ABS yang stabil.

Hasilnya, sebuah platform bottle holder telah berjaya disiapkan. Saya dan isteri membuat beberapa cubaan menyiku suction cup untuk melihat sama ada botol ini mudah tumbang. Nampak gayanya platform ini sangat stabil bagi penggunaan biasa, yang mana ibu-ibu tertumpah susu sebab tersentuh suction cup tersebut.

3D printer extruder making the ‘tick tick’ or ‘clunk clunk’ noise

For the people who are having problem with the extruder making the ‘tick tick’ or ‘clunk clunk’ noise, I locate below some of the solutions that might help you:

1. Look for response from Ben:
2. Right Extruder problem:

Buat maintenance Mbot 3D Printer (Nozzle Tersumbat/Clogged)

Selepas 50 jam ber-3D printing, hari ini saya melakukan sedikit maintenance terhadap nozzle MK8 extruder. Apa yang saya perasan, nozzle di sebelah kanan ni tak berapa smooth flow ABS dia. Masa print tu, ada fenomena motor skipping, yang mana filamen masuk dengan ada bunyi ‘tap tap tap’. Secara normal, filamen sepatutnya masuk secara smooth dan steady.


Tool yang digunakan untuk membuka:
1- allen-key saiz 2.5
2- spanner/wrench saiz 7mm


Saya buka extruder dengan allen key, dan meletakkannya di atas printer. Kemudian, saya pre-heat nozzle kanan sehingga 230 C supaya ia mudah dibuka. Kemudian nozzle tu saya rendam dalam acetone, kerana print material saya adalah ABS.


Sekejap je habuk ABS dah cair dan larut. Okay lepas ini saya akan dapatkan tolak keluar ABS yang tersekat di dalam nozzle, dan pasang semula unit ini. Harap-harapnya printing lepas ini akan smooth.

2013-06-25 09.52.10

Gambar di atas menunjukkan nozzle yang sudah direndam selama 24 jam. Saya tarik/tolak keluar ABS yang tertinggal di dalam nozzle, kemudian akhir sekali saya flos dengan wayar 0.25 mm.

Last sekali, lihat gambar di bawah, kita boleh nampak cahaya melepasi nozzle yang sudah bersih.

2013-06-25 10.00.28

Victory! Print saya sekarang boleh kata as smooth as silk!

Sebagai rujukan, saya lampirkan link-link berkait di bawah, semoga dapat membantu:
1 – Tips dan respon rasmi oleh MakerBot
2 – Cara-cara membuka extruder MakerBot
3 – Rendam dan flos nozzle