Thursday, November 18, 2021

My experience with a grown man Electric Scooter - the Chinese 5600 Watt Laotie Ti30 - Part 1

For a couple of years I have been playing around with the popular Xiaomi M365 which I bought back then. In spite of a few nitpicks it is an okay scooter, and good value for money for those looking for a small vehicle that can cover moderate distances (about 20 km per charge) while still being carriable (it only weights about 12.5 Kg). 

It does have some known reliability issues, such as the tires getting flat quite easily, and the folding mechanism reportedly breaking while riding:

Depending on the speed and circumstances, the latter can produce a painful experience for the rider. It never happened to me personally, but I found that soon into having bought the scooter, creaking noises around that hinge appeared. I resolved it by putting a rubber spacer in the area around the hinge in the folding mechanism. I assumed that without these vibrations and noises the stress on this mechanism would also be lower, and fortunately never had a problem personally.

I tried buying a steel version of the hinge locking part (which is what usually breaks), only to be scammed by Chinese suppliers who would just send another aluminium part similar to the original one, in spite of advertising it as steel.

More recently I decided to up the game when I found that a different class of e-scooters began to emerge in the market at acceptable prices. And by a different class I mean with a 20x power rating compared to scooters such as the Xiaomi M365. This of course, at the eye of the vehicle regulations in most countries, can make it difficult to keep it as a license free electric vehicle comparable to the M365. For example in Portugal, an electric vehicle is only considered comparable to a bicycle in terms of unlicensed use, if its output power is limited to 250 Watts, and the maximum speed it can propel the vehicle is capped to 25 Km/h.

The Laotie Ti30 (as with many similar chinese e-Scooters - all seem to come from the same "foundry", in spite of the different brands) has switches and modes (i.e. "Eco" mode) which will limit its power so that the motors cannot propel it at above 25 Km/h unless the user takes the more deliberate action of pressing the switch.

Once in "Turbo" mode, the raw power of the motors is unleashed. With its two 2800 Watt (peak power) hub motors, a matching battery (a 16S8P 60 Volt 38 Ah LiPo battery made of 21700 size cells - the same format used in modern Tesla vehicles) and two 45 Amp (actually 37 Amp, but more into that later...) Electronic Speed Controllers (ESCs), users have reportedly reached about 85 Km/h without too much effort.

In my experience with the beast, I didn't dare to go that fast, but I could confirm that upon engaging "Turbo" mode, it will zip past 40 Km/h in a blink of an eye. 

It only takes a slight touch of the throttle for it to pull violently, as if the scooter would be trying to get rid of the rider. It is quite an intimidating experience. I found it slightly more manageable to only disable Eco mode together with having 2 WD disabled, i.e. run in Turbo mode only on the rear motor. Turbo mode and 2 Wheel Drive is like riding a rocket. Slight increments of throttle will produce massive acceleration. While riding at 25 Kh/h it will still try to pull a wheelie when pressing the throttle in this mode. Didn't even consider starting in this mode. It is probably not a good idea for someone with limited experience and relatively light weight (I have 75 Kg - eventually a 100 Kg dude will be safer here).

Ordering the scooter was simple and Banggood seems to be the exclusive seller of this particular brand. I was surprised with the quick delivery - shipped from Poland via UPS it took 7 days to arrive to Portugal. As the purchase was done within the EU, I didn't incur in additional custom fees.

Given its weight, it is not a carriable item like the M365. Its 50 Kg are far from being an easy load to carry by hand.

The construction is quite solid - with a mix of aluminium and steel parts. The "yellow pages phone book sized" battery (for those who know how these looked like back in the days) accounts for 20 % of that weight. The front suspension fork is very similar in robustness and size to that of some motorcycles. The single suspension damper at the rear is more frugal and similar to that of some bicycles. It is not protected from dust, which is somewhat of an issue if left that way.

I referred previously in this blog about Chinese products that tend to require technical attention or improvements by the user since day one for safety and reliability reasons. This one, in spite of the rich experience that is likely to provide, is no exception. I will keep that as the focus of this post, as there is much to be said and done in that respect for this scooter.

I knew what I was going for, as I read videos and posts from other users about this scooter, and it was useful because it gave me the opportunity to perform some preventive care before stumbling into problems.

One of the important things that should be done before riding, is to protect that rear damper, as I have explained. It comes unlubricated - which in part is good because it will not capture the sand and dust quite as much, but will cause wear due to friction. As such I started by disassembling the rear suspension and apply grease to all of its moving parts. The inside of the damper had to be lubricated as well. 

Because now the damper will be more prone to collect dust which will be permanently attached to its surface due to the grease, it is important to protect it. While a rubber bellow is the ideal solution to protect this part, it may be difficult to obtain the correct size and it will take time. There is however a quick, cheap and reportedly reliable solution, which consists of wrapping the damper with a few layers of kitchen plastic wrap (as can be seen in the photo above). It can then be secured with zip ties, and it is ready to go. I followed that approach and so far it is doing great.

The other important aspect is the tightening of all the screws. These should all be checked because are not guaranteed to be properly torqued, given the rush in which these products are assembled.

One important screw is only accessible from the bottom, and in order to minimize creaking sounds from the steering, it needs to be better tightened. It is part of the folding mechanism, and it helps keep the tension together with the locking lever:

Some report that 5 or 6 Nm of torque on this nut is sufficient to avoid noises. I personally tried to keep a balance between the locking lever still be movable, and no slack or creaking sounds being apparent.

One issue that I found right after unboxing, was with the turn signal switch: I could only engage the right blinker, and when trying to turn on the left blinker I noticed that the travel range of the switch was not complete, as if something was blocking it. At first I still considered if there would be an electrical problem, but upon opening the battery deck and checking the LED blinker, it seemed fine. The problem had to be either in the switch on the handlebar, or the cabling that goes to the blinker itself.

Decided to open up the switch, and found that some of the plastic of the casing that keeps the switch in place was broken, causing a  misalignment with the opening where the button slides.

The issue was solved by properly attaching the switch to the casing with epoxy, and adding a screw (visible in the photo on the left side of the casing) that acted as a detent for the rear of the turn signal switch (replacing the broken piece of plastic).

The blinker lights are relatively nice, offering a progressive LED sequence in the direction of the turn.

Not without a slight nuance however: in the front a constant blue light glows when the blinker is engaged:

While relatively subtle in intensity and size, it is not technically legal in many countries, where blue lights in civilian vehicles is strictly forbidden (these are reserved for authorities and first responders).

Speaking of legal aspects with the lights, there was another couple of compliance problems in this scooter to account for: first it came with a couple of RGB led stripes attached to each side of the case. Besides making the scooter look like a Christmas tree, these blink and present a range of colors which is also not legal in a vehicle (at least in Portugal and several European countries). I simply removed these and the corresponding constant current adapters, saving some space in the already extremely crowded deck:

One last element requiring changes was the headlights: these have a red LED ring, and internally there is also a red power LED. 

Not sure how much of a selling point these non-compliant lights are in China, but for me it's just about the additional work of having to remove these.

The ring was just a matter of cutting the power wires from these, but the internal red LED is slightly more difficult to remove, as it is an SMD LED soldered into the single-sided aluminium PCB (LED3 in the image):

For removing it I used two soldering irons, by using one on each leg of the LED. The soldering irons need to be quite hot as the heat will dissipate very quickly with this type of board. It is also important to remove the board from the aluminium casing, as this helps further cooling the PCB.

I didn't get to understand the purpose of the yellow wire going to this lamp, as it is not connected to anything in the deck.

My first glance of the battery deck after opening it for the first time, was of observing an absolute mess of wires crammed together in a tight space:

Cable management aficionados and people bearing OCD issues would have a hard time with this vision. In the middle of the spaghetti, the disconnected yellow wires without a known purpose.

Another problem with the headlights which I immediately spotted was the absence of one of the support screws:

Worse than that (and probably the ultimate reason why there was no screw), the hole for the screw was not threaded. So besides having to find an appropriate screw, I had to thread the mount in order to be able to tighten the screw into it.

Back to the battery deck and the wire spaghetti, I found that in my unit the ESC's were not covered by a thermal pad. Based on videos from other users it seems that they used to add thermal pads, but for some reason in this particular unit they didn't. I ordered a 2 mm thermal pad which I plan to add, as soon as it arrives.This is important because it will allow the heat to be more directly transferred to the metal case, instead of building up further in the ESCs. As this is not a ventilated space, the only way the heat escapes is by conduction in the surfaces, some convection in the air contained in the volume, and finally to the outside, via the external surfaces of the battery deck.

While there is not much leeway for proper cable management given the very limited available space, I found important to group the cables with zip ties, and identify and label the connections. Also added some heat shrink material around some of the cable bundles in order to better protect and isolate these. The result, albeit not ideal, is slightly tidier:

In a next step I plan to replace the large yellow 3 phase motor connectors with higher quality gold plated 6 mm bullet connects. Besides the superior electrical contact, those are also more compact than these large boxes. And as you may have figured out, space is at a premium inside this battery deck.

Out of the box, the battery does not have any particular preparation against water ingress. There is some foam adhesive around the wires leaving the box, but nothing else. In order to provide a more substantial protection, I added some black temperature resistant silicone of a type normally used in the automotive industry. While it does not have an explicit designation, I believe it is equivalent to RTV Silicone:

Applied it to the surfaces of the deck that contact the lid, in order to form a joint, and to the cable openings:

Regarding the brakes, not much had to be done. These are of the relatively popular brand Zoom, and the only thing I found important was to adjust the calipers by loosening its screws, let the caliper bite the disc, and tightening back the screws while pressure is applied on the brake.

This ensures that both brake pads are at an even distance from the disc. The rear brake disc also required some straightening, which I did with the help of a kitchen fork by correcting the bent section in multiple iterations until it would no longer hit the pads.

My major issue came after just a couple of rides in the scooter: on the second ride and as the battery gauge dropped to two bars, I realized that I no longer had front traction. At first I thought it could be related to the battery level. Some form of limp mode. But after letting the battery charge overnight, on the next morning, to my surprise, the front motor would no longer work, and the scooter would always run in "Eco" mode, regardless of the position of the buttons and gear modes.

At first I thought the first ESC would be fried or something. But after opening up the front motor ESC and making some basic checks to the MOSFETs, I considered that at least the transistors were probably not blown.

Looking closer I realized that there was a large resistor that seemed to have had a very rough time:

There was even discoloration of the color codes, which made it a challenge to figure out what value it was supposed to be. I removed it and it measured about 500 K. Then I decided to open the rear ESC (which was working well - and by the way in order to rule out motor problems I connected it to the front motor and it worked normally), and found that it seemed to be exactly the same hardware, with the exception of a small SMD capacitor which was not present in the front ESC. The same resistor in this healty ESC measured 100 Ohm - quite a big difference from what the other one was measuring. The resistor was clearly burnt in a somewhat catastrophic way.

I tried to do some reverse engineering to this section of the circuit and was able to understand that this resistor seemed to have some kind of a current limiting role in the somewhat rudimentary switching mode DC/DC converter that this circuit consisted of:

Basically this converter has to transform the 60 Volts from the battery into the 5 Volts required by the micro-controller. In order to achieve this in an efficient way, a switching mode power supply design is used. On the secondary side there is a 78L05  linear regulator that further regulates and drops the DC voltage to the required 5 Volts.

The process that led the resistor to get blown, was not obvious however. The first thing I tried (wrongly or not) was to replace with a new resistor and try the circuit as is. The behavior changed slightly: now instead of nothing happening when engaging "Turbo" mode, the rear motor would operate erratically, as if randomly alternating between Eco and Turbo mode. This led me to believe that I had either screwed up something further, or that with this pseudo-repair, the front motor ESC was trying to wake up but was not stable enough. If this was the case, the rear (and main) motor ESC would probably have some kind of logic where it would allow Turbo mode only if the front motor ESC would be properly detected.
I went back to the bench with the two ESC's, and almost randomly found that the zener diode that is next to the 13003 switching transistor didn't seem to be ok. It was a 1N4744A,  a 1 Watt 15 Volt Zener diode. Removed it and found that it was not acting as a diode, leaking current in both directions well below its reverse bias voltage of 15 Volts.

After some time waiting to source a new diode like this one (didn't have one with the same specs), I finally was able to replace it. 

Thinking in a bit of a "prophylactic" way, I decided to replace the 100 Ohm power resistor in both ESCs by a couple of 5 Watt 200 Ohm resistors placed in parallel in order to obtain the same 100 Ohm but with a 10 Watt rating:

The rationale for the replacement was the possibility that the original resistor might not have the appropriate power rating for the application. Still the actual cause was a mystery, and unknown if it was the Zener diode or the resistor what blew first.

After putting everything back together and securing the resistors and a capacitor next to these with epoxy, I went straight for a test run of the ESCs by mounting these back in the scooter with the wheels lifted. To my moderate surprise, the scooter was now working normally: both motors spinning, turbo mode reaching the expected max RPM, everything normal.

And as of this writing I still don't understand exactly what could have caused such failure. Apart from the Zener diode, the other semiconductors in the primary side of this power supply were not affected.

So far I did about 60 Km with the scooter after the repair without any issue.

With all the wire spaghetti, blown discrete components, and apparent lack of over-current protection for the small devices behind the battery, this made me consider the need (eventually urgent) of adding a fuse to protect these. While a big short on the main power rails that drive the motors would likely be protected by the Battery Management System (BMS), it is more questionable that the same would happen for a short happening in smaller wires which will burn before sufficient current is detected by the BMS. And having a fire next to a LiPo battery is the perfect recipe for disaster. 

There is one first switching element which controls the 60 Volts going to the control part of the ESC, the lights and other loads. This is the alarm box:

It is therefore necessary to add a fuse to its input wire, allowing the rest of the circuit (including the alarm itself) to be protected by the fuse. I measured that with the blinkers, the headlights on and the motors running, the current does not go above 500 mA (motor current does not go through this circuit, only ESC micro-controller power). So a 2 Amp fuse should be decent protection, while also not bringing exaggerate resistive losses.


JankeS said...

Thank you very much for your blogs. Did you mention the controllers are 37A instead of 45A? Everywhere in the specifications of the Laotie Ti30, it says that it has two 45A controllers just like the Boyueda.

Creation Factory said...

Hi @JankeS, thank you for your comment. I deliberately mentioned 37A / 45A because a few users learned that this controller (as sold for example in Aliexpress) is exactly the same hardware for both markings - same FETs in same quantity, and same microntroller. Given this fact I prefer to assume that the controller is actually 37A rated and not 45A (better be conservative with the expectations, for longevity and to avoid the white magic smoke). There is some leg room in the values the manufacturers decide to share, because it depends not only in the ratings of the FETs, but also on how efficient the cooling, at what frequency the transistors are switch, rise/fall times of the signal applied to the gates, etc. I wouldn't be suprised if they inflate the stated values just to satisfy some deal and push whatever they have in stock or are able to readily manufacture just in order to satisfy e.g. some high volume production of scooters with higher specs..

Unknown said...

Merci pour tous cette revue du Laotie TI30.
Voici un lien pour des contoleurs sinuosidaux pour ce modèle, finies les accélérations brutales et autonomie améliorée !