MR60

Worlds Best ESC - Making it Better

That's a bold statement ... Worlds Best. But it's even larger than that. Not just Worlds Best, but best for most all applications less than 30 Amps (limit of the tests). That means:

- duration ships that only pull 2 to 8 amps per rotor

- most all 6S and smaller ships (exception of nano-ships)

- any-size FPV racer

- any other ship in between

Why almost any size? Shouldn't a small FPV racer use a smaller and lighter ESC for response? Yes, if it does better on a net-lift response test. In other words, when you penalize the ESC for it's weight, is it still better and faster? What i continue to see is ESC manufacturers downsizing critical components of the ESC at a net loss. They weight savings is lost because of greater thrust loss and response. In other words, this heavier ESC will out accelerate, in the real world, a smaller and lighter ESC.

Why post this? To move technology forward, we need to report to industry what works and what doesn't. For some reason (i don't know why), this ESC works better than all others tested:

- for generating maximum thrust from the motor***

- for net-lift efficiency or the grams of weight it can lift (after it lifts the rotor) per watt

- for response (how fast it can generate targeted lift)

These tests were conducted on multiple days on multiple rotors of highly variant size, always being immediately compared back to another DYS 40A multicopter test to ensure that the baseline wasn't changing.

The ESC that dominated is a DYS 40A OPTO Multicopter using SimonK. The photo is included because there are two others that carry a similar or same name.

- Not the white cover DYS BLHeli 40A

- Not the one that is says "Programmable" versus "Multicopter" in the blue/purple band across the front

3691280727?profile=original

Have i tested all ESCs? No, but if you are convinced you know of one that would work better, let me know. I've tested most all of the following and one or more of their variants:

- DYS

- Multistar

- Turnigy

- T-Motor

- Afro

- Motortron

- Quattro

- 3DR

- Spider

- KDE

- ZLW

- Aris

- EMAX

- AutoQuad

- Exceed

- HobbyWing

- Lumenier

To do a test like this, a highly repeatable and finite test stand is needed. It took a while to develop one but what works is one that:

- measures (at a minimum) volts, amps, thrust, motor temp (shoots IR up the aft end of the motor)

- eliminates harmonics between the rotor and load sensor (this proved difficult but achievable)

- is calibrated and proves repeatable within 1.5%

- controlled by a system that can precisely repeat a rotor test (uses a Audurino Mega)

- directly feeds the data into Excel for analysis (uses DATAQ)

- uses a test script that produces repeatable results

- uses a test procedure that minimizes repeatability error (used average of multiple tests)

How much better is this ESC?  On average:

- 4.4% higher net lift (after it lifts itself)

- 2.3% more net-lift efficient (usually the larger the better)

- from more than twice the response or the same response as other ESCs (usually the larger the better)

                                                       So how to make it better?

Step 1: Strip it naked. See photo below.

           ... remove the cover

           ... remove the heat plate (better to locate the ESC under prop wash to run cooler, see below)

3691280569?profile=original

Step 2: Right-Size the bullet connectors or wires (see above where heavy wires are replaced by 2mm bullets)

           ... remove the large bullet connectors or wires

           ... replace them with ones that are the most net-lift efficient (where heat loss = weight loss)

Step 3: Seal the ESC. Seal it with Electrical Sealant to protect from moisture and conductive dust

           ... tape or plug connectors and wires

           ... repeatedly spray each side from different angles

           ... a mistake i made was not sealing the bullet connectors and solder

               - don't tape them off like i did

               - insert a male connector into the end of bullets so sealant doesn't get inside them

3691280655?profile=original

Step 4: Locate ESCs under Prop Wash. See photos below. The turbulence generated by the prop does not adversely affect lift when the ESC is placed on edge to the prop wash.

           ... Use something non-conductive like hot glue to bond the ESCs to the motor mast or spar

           ... Face the FETS (the little square warehouses or Fire Emitting Transistors) to open air

           ... Protect the ESCs from below from ground contact (not needed here because of clearance)

back-side with hot glue

3691280584?profile=original

front-side with FETs completely exposed to open prop wash

3691280749?profile=original

Step 5: Tie up wiring. Use dental floss to secure wiring away from the prop.

***Note: The T-motor Air 40 in high-timing mode (an option) generated higher thrust, but at the sacrifice of efficiency and motor temp. Also, the T-Motor Air 40 was 2nd best and close in performance. If you are using an Air40, it probably isn't worth switching.

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –

Replies

  • I just read through all 21 pages of this discussion.

    First, I have a question that no one has asked. I get that the Aluminum heat sink quickly rises to the temperature of the FETs and may not be able to dissipate the heat to AIR very well with the heatshrink surrounding it and such.

    My question is- if ONE FET gets hot (compared to the others on the same ESC) and that heat is dissipated across the entire heatsink AND the other FETs, could that save that one FET from damage?

    Next, I saw pages and pages of arguments about new, faster ESC designs using newer FETs and other components. I saw a lot of arguments back and forth but no actual data. I saw Forrest offer to test any of these new ESCs and more. Can we get one of the racers who loves the new ESCs to work with Forrest and end the arguments with real data?

    Personally, I'd be interested to know if:
    1) The new, lighter ESCs can handle heat better and do not require a huge heatsink.
    2) The newer ESCs can handle crashes better.

    My application is a camera ship for film production that will:
    1) Fly for 30 minutes on one battery
    2) Be stable for slow pans and reveals
    3) Be fast and agile enough to follow boats, cars, and other fast moving objects

    cheers,

    Robert

  • MR60

    Rob - Thanks for jumping in. As always, excellent put. I do need to show the data. I also need to show the data in a format that is easily accessed and meaningful. I'm closer to the latter, but still not ready. But that doesn't mean i can't share the data i have so far. So thanks for the push.

    Let's start with the DYS 40A Multicopter. Since i'm claiming that they are the best, then they probably won't sue me if the data is for some reason misleading. I and others have given a lot of thought to this in trying to come up with a meaningful response test, and we think this format is the right presentation.

    Your insight is right on about the speed of response. It's much slower than we think. Keep in mind that what you are seeing below is an average of 299 up thrusts and 299 down thrust trying to target a 380 gram thrust change. The test analysis collects data on the first 250 ms (1/4 of a second) of response in increments of 1/100 of a second (10 ms). If then smooths the data using a 3rd order polynomial (rather than doing over a thousand samples).

    This is how it looks for a 12" prop using a high torque motor MN3508-29 380KV at 6S (so it's fairly quick at that voltage tossing around a small T-motor 12x40 prop).

    3702214838?profile=original

    Then from this data (and from looking at max thrust data), we can calculate the the critical components of rotor acceleration:

    - acceleration for the first 200 ms is 5.2 grams/ms^s up and 4.1 grams/ms^2 down

    - the rotor can achieve 200 grams of the requested thrust change at a rate of 20 Hz up and 17 Hz on a down throttle.

    - the maximum velocity, (the right most place on the acceleration curve where acceleration = 0 or maxes out), which is a function of maximum thrust, which in this case is 1047 grams.

    You probably know better than i on why this is so slow. Is it because, internally, the ESC is doing the following:

    - how much RPM do i need to generate to achieve the input thrust?

    - let's try give it this many amps and hope i don't overshoot (taking a best guess for internal PIDs)

    - small pause

    - what did that do?

    - OK i have this much left so let's repeat (where upon it does this again and again until it gets to the desired RPM)

    ... and motors can only generate so much torque when trying to rotate a prop with a lot of drag and mass out a ways from the hub.

    I owe more data like this. Using our new test procedure, i'll repeat this test for a FPV 250 setup. and i'm looking for a representative ESC to test against that has all of the latest "high-tech" stuff. Hopefully, the new test will show that the "high-tech" stuff is better (it focused on fundamentals like amp throughput first and then on the techy stuff). 

    • There is something that doesn't compute in my mind. Forrest, I do honestly appreciate that you are trying to build-up some kind of a measuring standard so we can compare apples to apples... But, here I have one problem. Actually more than one.

      1) Which component are we testing here - ESC, motor, props?

      2) Which combination are we testing here - ESC&motor, motor&props, ...?

      Of course your original theme is all about ESC, but what we have here are co-related components making up a system. A single component by itself is useless to us, right!? So, for example when testing an ESC alone (I guess here) you would connect some PWM signals on the input, some dummy load (RLC) and measure the response at output, right? But that is useless, I guess again.

      Now if you try to measure a system... that's another beast. Usually we are looking toward optimizing a system and it might happen that a system as a whole might work best with "average" components, not the best ones. So when you say that "TheBest-XYZ" ESC is better than "NotSoGood-XYZ" ESC that is true for a certain ESC & motor & propeller combination!! Change one of the other two components and the story is different altogether! So whenever you are trying to compare, or test one component against the other, other parts of the system shouldn't change. So when testing ESC's you have to use the same prop&motor for all tested ESC's. Right? Well, maybe not! Let's look at the props here. Change a pitch and the response is different again. And it might even happen that an excellent ESC with previous prop is below average with a new one.

      Well in automotive industry we use electric brake. So to properly test ESC&motor here, the good idea might be to use also a small electric brake instead of a propeller. This way a whole bunch of variables just vanish. Plus you can measure electromechanical efficiency and all other bunch of parameters for ESC&motor system.

      But we still need a propeller, right? So, this is an equation with three main variables with many "sub-variables". Let me just name a few: 3S, 4S, 6S, KV, No. of poles, pitch,...

      Mathematicians found a solution for such equations - "per partes" partial equations.

      My point here is, that it is impossible to set a standard measuring/comparing test unless we settle for the common ground.One would be no props (when testing ESC&motor combination). The other might be to use one very common and widely available motor for ALL measurements (for ESCs).

      And then, when we find the "Best in The World" ESC&motor combination we can test different props with it.

      I'm very nervous about this issue, because for a long time I'm trying to find a proper method to "measure", what is the best props&motor&ESC combination. And sincerely ESC being the lest important. I know, wrong thinking as ESC are the less reliable of the three.

      • MR60

        You definitely understand the problems I've encountered in the past year in trying to measure response. You are absolutely right. But, what I found is that the problem is less complicated than I thought.

        Background:

        - decided to use your recommendation of "standard" rotors

        - used six common superb rotors (Tiger Motors & KDE Direct) 

        - created 5 applications

          ... 250 Racing                      6" prop with the TM2206-13 2000KV

          ... 500 FPV Camera Ship    12" prop with the TM3508-29 380KV

          ... 700 Camera Ship            17" prop with the TM3508-29 380KV

          ... Endurance Ship              30" prop with the TM U8 100KV (only 6S ESCs could do the latter)

          ... Moderate Heavy Lift        30" prop with the TM U10 heavy lift (ditto)

          ... High Heavy Lift               30" prop with the KDE 8218XF 120 KV heavy lift (ditto)

        - found the most efficient and responsive prop for each application (tested 90 props)

        - then tested response

        My expectation was that three or so ESCs would raise to the top of those various applications (within the limit of only testing up to 6S ESCs).

        What I found surprised me. One ESC did the best in all the applications even on a weight adjusted basis.

        Note: Yes ... a real heavy lifter would require a +12ish S ESC 

        That surprise is why I chose the words "Worlds Best 3S to 6S ESC". From FPV racing to a 6S Heavy Lifter, it alone rose to the top of all the applications.

        I also found an interesting correlation. By simply testing max net-thrust, you have a good idea about the quality of the ESC in allowing amps through to the rotor ... or what the response will be.

        In hindsight, it sort of makes sense. If an ESC allows more Amps to the motor (less resistance, etc), it will allow the rotor system to have faster response.

        On the impossible part.  On efficiency of rotors for a multicopter, there is no doubt that it should be measured in terms of net-lift efficiency (it's efficiency after it lifts itself). But for response, the statistic is too application based. So the numbers need to be given in terms of raw acceleration (grams per microsecond squared) and Hz along with the mass of the rotor. Then it's up to the designer to look at ship mass and it's location to figure out which rotor has the optimal acceleration of its mass-distances.

        On "per Partes". When i can generate actual data, that is my preference. When not ...

        On What Are We Testing.  For efficiency, it's always the entire rotor system (electrons, ESC, motor, prop). For response, i tried to come up with a standard statistic that could be used, but it's not possible given that mass is not a subtract from force but a ratio. So the mass will just be reported along with the data so the designer can calculate force differences of various ESC, props, and motors.

        Great questions. Insightful and provoking ideas that you raised. Thanks.

        • Forrest, some time ago I bought hobby wing 20A esc to lightweight my copter but I don't change them yet it's have now 40A rcTimer esc octo, I read in some part that they are the same as you show here, perhaps I'm lucky to have the better esc and I didn't know :O  is there something that I can test at home to know if they are giving better performance that the lighter ones (hobbywing) to avoid changing with no better results? my target in that quad is long time performance 

          • MR60

            The Hobby Wing is different, but it's still a good ESC (on a 30 minutes flight the best ESC will only deliver a half to one minute more time. Not enough to discard working ESCs. Use the ESCs you have. But if you need to buy a new ESC, consider the DYS 40A Multicopter.

            • Thank's Forrest, perhaps the lightweight that I have in mind give me more than one min, including wires change, remove connectors, shorten wires, etc. winter task :)

              I still have little customers difficults here to buy were I want and find special materials or components, I'm doing ingeniering to try to buy a silly pixracer LOL, I ask for 3M paint to some electronic specialized friends here and anyone know it UFF!! but I still have possibilities to find yet, I suposse the technique it's usefull for any esc to lightweight it.

            • @Forrest

              "

              Is the Al block large enough to absorb all the heat transferred from the ESC in a 10 minute flight? 

              "

              I can give you correct answer.

              There is no need for Al block to absorb all the heat transferred from the ESC in 10 a minute flight, since Al block is active element

              radiating the same thermal energy into plastic shrink wrapping and into air via air inlets and outlets.

              So Al block is not acting as thermal bank.

              I would suggest you to study theory of heat pipes and heat pipe

              cooling systems in PC computers, laptops, tablets.

              Passive heat pipe element can cool hot processor satisfactory to prevent overheating, elimiting the need for electric fan to do that job.

              I can test every ESC model, type at my

              Open Technology Park

              Open Fabrication Lab

              providing you and any interest person with study on thermal dissipation efficiency of any suggested cooling solution

              (study with thermal camera).

              But keep in mind, you can get relevant data on low thermal dissipation for modern high power FETs by Fairchild.

              Since all data on themal dissipation, thermal conductivity are known,

              professional designers of ESC do excellent job to design Ms of

              FET to meet standards.

              you are free to contact me any time for advice on design of

              cooling systems in power engineering, IT sector.

              Open Technology Park

              Open Fabrication Lab

              manta103g@gmail.com

              • Darius, my question to you as a knowledgable person is: So what kind of heat sink would be the best for this application (ESC - obviously)? Or I should say the most optimal. I know for fact that a "porcupine" type would be better than a solid block of Al, but space is a limiting factor. The same would apply to heat-pipe. Heat-pipes are excellent thermal exchanger. In a matter of fact the speed of transferring heat from one end of the pipe to another is second to none. But, but it does require some "volume". And in drones usually we do not have space nor weight to waste.

                What would you recommend??

                • T3

                  @Dorjano, under no circumstances listen to advice from Darius.  He has no idea what he is talking about.  Zero practical experience, just Bing.  I don't think he has ever flown a multirotor.

                  In my actual experience, I recommend you check out Castle ESCs for ideas.  Castle is a very high end brand.  The small finned heat sink is very effective, especially with ambient airflow over it.  The most important advantage they have is not having shrink wrap over the heat radiating surface.  You'll notice that some of their lightweight ESCs do have shrink wrap there, and if you look in the specs you will see that these ESCs require greater airflow to stay cool.

This reply was deleted.

Activity

gotham liked gotham's profile
2 hours ago
DIY Robocars via Twitter
RT @RoboticMasters: Monaco GP Circuit in the Donkey Sim (coming soon). Including buildings, tunnel and all! https://github.com/robotics-masters/sim-donkeycar-f1/tree/f1-tracks @diyr…
yesterday
DIY Robocars via Twitter
RT @breadcentric: Here are the details of #AWSDeepRacer finals: https://blog.deepracing.io/2020/12/01/aws-deepracer-league-finals-2020-round-1-schedule/ #awsreinvent2020 #AWSreInvent https://t.co/ovqsjp8V…
yesterday
DIY Robocars via Twitter
RT @breadcentric: #AWSDeepRacer League #awsreinvent2020 Open race is on Dec 1st - Dec 31st in three categories, 15 DeepRacer Evo (with LIDA…
yesterday
DIY Robocars via Twitter
RT @a1k0n: @SmallpixelCar @diyrobocars It's just something that's easy to track with chroma keying. I ended up using different colors on th…
Monday
DIY Robocars via Twitter
Monday
DIY Robocars via Twitter
RT @TinkerGen_: "The Tinkergen MARK ($199) is my new favorite starter robocar. It’s got everything — computer vision, deep learning, sensor…
Nov 23
DIY Robocars via Twitter
Nov 23
DIY Robocars via Twitter
RT @roboton_io: Join our FREE Sumo Competition 🤖🏆 👉 https://roboton.io/ranking/vsc2020 #sumo #robot #edtech #competition #games4ed https://t.co/WOx…
Nov 16
DIY Drones via Twitter
First impressions of Tinkergen MARK robocar https://ift.tt/36IeZHc
Nov 16
DIY Robocars via Twitter
Our review of the @TinkerGen_ MARK robocar, which is the best on the market right now https://diyrobocars.com/2020/11/15/first-impressions-of-tinkergen-mark-robocar/ https://t.co/ENIlU5SfZ2
Nov 15
DIY Robocars via Twitter
RT @Ingmar_Stapel: I have now explained the OpenBot project in great detail on my blog with 12 articles step by step. I hope you enjoy read…
Nov 15
DIY Robocars via Twitter
RT @DAVGtech: This is a must attend. Click the link, follow link to read the story, sign up. #chaos2020 #digitalconnection #digitalworld ht…
Nov 15
DIY Robocars via Twitter
RT @a1k0n: Got a new chassis for outdoor races (hobbyking Quantum Vandal) but I totally didn't expect that it might cause problems for my g…
Nov 11
DIY Drones via Twitter
First impressions of the Intel OpenBot https://ift.tt/36qkVV4
Nov 10
DIY Robocars via Twitter
Nov 9
More…