70mm monster Scorpion in the works

What we need is better controllers, not bigger motors. I do not know why manufacturers keep making bigger and bigger motors while there are not controllers out there to run them to their potential.

It looks coo, though. :wink:

Matt
 
ron van sommeren said:
.
Something like this Matt?
http://www.endless-sphere.com/forums/viewtopic.php?f=30&t=28841


Sadly, no, but a step in the right direction perhaps. Those single layer SMT fet controllers wouldn't last 20 seconds of low-speed operation on a motor with low resistance and inductance.

Perfect for motors that have a prop for a load though (where phase current increases at the square of RPM), or even wheeled devices that have extremely low weight (so the time spent at low RPM is nearly nothing).
 
ron van sommeren said:
Grrrrmbl&^%@(!!! :)
translate...www.sinusleistungssteller.de then ? Under References/Projects some very nice videos of traction applications (quad, shovel) and GeeBee racer. Sinusoidal commutation, not trapezium.


That controller is able to function, not because it's sinus, but because of that pair of shunts on 2 of the phase leg outputs. It's the only controller I've ever seen that can see phase currents, which is the only way to survive at low RPMs.

Does anyone have a picture of the board for these new ESCs? Perhaps they have something different than all the other ESC's we play with, Castle, Jeti, Hextronix, etc which all fail in seconds or minutes in an Ebike, yet perform fantastically in RC vehicles.
 
LFP I bow before you, humbly ask you to clarify, the RC ESCs all die within seconds or minutes?? I thought there were folks here using them medium - long term? Not even the CCMM? The HV160? Are none of them going to work? I'm still hoping my turnigy 90/100 will deliver the juice to the drive I'm building...

Eric
 
Hi LFP..
Has anyone looked at or tried one of these, I thought they looked promising,
What do you think?? http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=8937
Regards John.
 
Hmmm looks interesting... Anybody try it yet?
 
liveforphysics said:
... Those single layer SMT fet controllers wouldn't last 20 seconds of low-speed operation on a motor with low resistance and inductance. ...
Bring bike up to a certain speed and then 'engage' the controlle? Possibly with a slow ramp-up in speed too? Disable controller below a certain speed?
 
ron van sommeren said:
liveforphysics said:
... Those single layer SMT fet controllers wouldn't last 20 seconds of low-speed operation on a motor with low resistance and inductance. ...
Bring bike up to a certain speed and then 'engage' the controlle? Possible with a slow ramp-up in speed too? Disable controller below a certain speed?


Yes, if you can get the motor up to speed, then an RC controller can work, yet in my experience only the Castle HV160 is able to consistently maintain sync at speed.
We've done various things, pedaling first dragging the motor up to speed, slipper clutches (Recompence) to prevent the motor from stalling (and take shock load off the drivetrain), 2-spd transmissions to enable starting from a stop to have the least load against the motor as possible, etc etc.

The end result for all of those experiments (after $1000's of dollars in blown RC controllers in my case) is the person wanting/wishing they had sensors and/or adding sensors to the motor and running it from a controller that uses large through-hole components and has a real thermal path from the FET junctions to the sink (which no SMT fets are able to have, even if the sync is held at ambient temp magically, the Rth between the junction and the back of the epoxy case makes the sync little more than a gesture.)

Once you go sensored with proper controller, you have a setup that works, and you're happy to not need to carry around a spare RC controller or two in your bike tool kit to try to make it somewhere and back.
 
The only issue I have with sensored is the timing problem. They either have to be timed for high RPM or low RPM.

Assuming you can work around the sensorless startup issues and stay under 8kw, (which I have) then they work fine.

I wish there was a way around the timing problem.

It sure seems like everything has its drawbacks.

Matt
 
recumpence said:
The only issue I have with sensored is the timing problem. They either have to be timed for high RPM or low RPM.

Assuming you can work around the sensorless startup issues and stay under 8kw, (which I have) then they work fine.

I wish there was a way around the timing problem.

It sure seems like everything has its drawbacks.

Matt


Matt- Have you seen Burtie's magic timing box?
 
LFP - a couple of the "sync"s in your post I presume should have read "sink", i.e. heat sink.

OK, I'm hoping I've still got a situation where an RC controller will survive. We'll see.

I'm gearing down 100:1, so top speed of about 3 or 4 km/h (2 or 3mph), for hillclimbing only. The plan (and instructions to the user) are to engage the drive while approaching the hill, so that the speed sits down into the drive's range, rather than having the drive bring the whole thing up to speed.

The other failsafe I have (admittedly very inelegant, like most of my ideas such as sacrificial FETs etc), is that the driven wheel will have only limited traction, as most of the load it is pushing against is not bearing down through it. So I imagine if things get too much it will just slip.

We shall see. Soon, hopefuly!

Eric
 
ron van sommeren said:
liveforphysics said:
... Those single layer SMT fet controllers wouldn't last 20 seconds of low-speed operation on a motor with low resistance and inductance. ...
Bring bike up to a certain speed and then 'engage' the controlle? Possible with a slow ramp-up in speed too? Disable controller below a certain speed?

I have been using exactly what you describe and it has been faultless so far. No sync issues. No burnt ESCs. No excess ESC temperature. Only times I have blown controllers are as a result of running a motor as the windings were melting together. :oops:

Erogo said:
I'm gearing down 100:1, so top speed of about 3 or 4 km/h (2 or 3mph), for hillclimbing only. The plan (and instructions to the user) are to engage the drive while approaching the hill, so that the speed sits down into the drive's range, rather than having the drive bring the whole thing up to speed.

At those speed the user will have time to read the instructions as they approach the hill. :lol:
 
Erogo said:
LFP - a couple of the "sync"s in your post I presume should have read "sink", i.e. heat sink.

OK, I'm hoping I've still got a situation where an RC controller will survive. We'll see.

I'm gearing down 100:1, so top speed of about 3 or 4 km/h (2 or 3mph), for hillclimbing only. The plan (and instructions to the user) are to engage the drive while approaching the hill, so that the speed sits down into the drive's range, rather than having the drive bring the whole thing up to speed.

The other failsafe I have (admittedly very inelegant, like most of my ideas such as sacrificial FETs etc), is that the driven wheel will have only limited traction, as most of the load it is pushing against is not bearing down through it. So I imagine if things get too much it will just slip.

We shall see. Soon, hopefuly!

Eric


No, I mean sync, as in synchronization.

All sensorless drives (short of flux-vector type, which there are zero RC types that are able to do) just have to play a pre-defined series of pulses into the motor, called the starting algorythm, and hope they get rotor motion that generates BEMF that they controller can use to know rotor position. When this pre-defined series of pulses starts off, it's equally likely to be jerking the wheel backwards as it is forwards, but after that the average amount of jerks the rotor takes should be the desired direction of rotation. If the rotor doesn't rotate, then it has no way to know if it's applying torque in the right or wrong direction. If the rotor is stalled or not able to jerk into motion, then it has no way to continue to run the motor, and it just has to play the starting algorythm again and hope for the best. This is the nature of all sensorless controllers (excluding flux-vector drives).

So, this is how it gains sync to begin with, so it has an idea of rotor position to know which coil to energize next to continue making torque in the desired direction.

When you have a motor with a lot of poles spinning at a high speed, the pulses of BEMF it has to sense the peaks of (the transition between the rising and falling voltage in each phase) gets to be a very short amount of time, and motors with a BEMF that doesn't have a clear peak, or mostly rounded nose of the BEMF waveform can be very difficult to accurately position the rotor. When it looses rotor position and faults, this is called a loss of sync. Normally you get an awful high frequency squeal or scream from the motor, sometimes the motor tries to lock a given position, sometimes the controller just explodes into plasma, and most of the time you can just let off throttle, and gently re-apply and it finds sync again and continues to work for another bit.



The idea of fusing the phase leads is a definite fail, even if it's mosfets. The inductive spike flyback when the fuse pops will still kill the controller as the energy stored in the magnetic field in the motor tries to reach infinite voltage the instant the circuit opens.

Current sensors in the phase legs going to a fast uController that intercepts the throttle signal is the way to do it. Shunts in the phase legs with an isolated A to D on an isolated DC/DC and an opto-coupler to a uController is the other way to do it.
 
liveforphysics said:
..No, I mean sync, as in synchronization. ..

I believe Ergo was refering to this ??...
uses large through-hole components and has a real thermal path from the FET junctions to the sink (which no SMT fets are able to have, even if the sync is held at ambient temp magically, the Rth between the junction and the back of the epoxy case makes the sync little more than a gesture.)

Has anyone looked at or tried one of these, I thought they looked promising,
What do you think?? http://www.hobbyking.com/hobbyking/stor ... oduct=8937
Regards John.
Well, if it did work and managed to hold sync etc, it would undoubtedly need a pumped water cooling circuit ( complication and cost ) , and its limited to 6s ( 22v).
 
Sensored controller.
I think that timming is fixed on those just because of simplicity.
In cars we have variable ignition with fixed input timing. it's the controller that change timming.

So It is just a programming thing. It could be even with sensor for startup and then switch to sensorless mode after with variable and optimal timming.

Why nobody used a simple centrifugal cart clutch with rc setup?
 
Joe90 said:
Sensored controller.
I think that timming is fixed on those just because of simplicity.
In cars we have variable ignition with fixed input timing. it's the controller that change timming.

So It is just a programming thing. It could be even with sensor for startup and then switch to sensorless mode after with variable and optimal timming.

Why nobody used a simple centrifugal cart clutch with rc setup?

I have been looking into centrifugal clutches myself. There are reasons I, personally, have never gone that route. For one thing, starting from a dead stop has never been an issue for me. Second, the Castle ESCs stay in sync very well. They are also easy to make survive in a setup under 6kw. Even up to 9kw, if you do not run them at very low throttle settings while pulling huge amperage, they survive fine.

I have heard more than one person mention sensored startup with sensorless kicking in once good sync and decent RPM is established.

Ultimately, it is really only us power mongers that need something better than RC controllers. :mrgreen:

Matt
 
Someone with the knowledge should make us one then, castle ??
Not big enough market I presume.
 
Joe90 said:
Someone with the knowledge should make us one then, castle ??
Not big enough market I presume.

Definately! Spend $50,00 on product development and then sell 10 units to the ebikers wanting sensorless controller. :lol:

I have been running my Castle HV110 on my sensorless WE hub motor lately with no problems at all. It does stutter at low speed, but once over 5 mph it works quite well. I have taken it up some fairly steep hills and it gets warm but not hot. All these RC controllers need is a little push off the line, proper gearing, and taking thier amp rating with a grain of salt in our application. Like Matt says, it's the guys pushing them to the limits and applying low speed, high amp loads on them that causes reliability issues. I also ran the same controller on my DR bike and it ran great but the bike was mechanically unsound which led to its failure. The controller never really got hot after repeated acceleration runs.
 
Joe90 said:
Someone with the knowledge should make us one then, castle ??
Not big enough market I presume.


They made some prototypes of a 20s 400amp ESC. I got 2 answers on why they stopped the project. Answer #1, the voltage is high enough to shock people (big whoop). Answer #2, which came from an engineer there, "they couldn't work the bugs out to make them reliable"
 
Joe90 said:
... it's the guys pushing them to the limits and applying low speed, high amp loads on them that causes reliability issues. ...
About de-rating controllers and motors, starting at 'for everyone else':
http://www.rcgroups.com/forums/showthread.php?p=11476940#post11476940
 
ron van sommeren said:
Joe90 said:
... it's the guys pushing them to the limits and applying low speed, high amp loads on them that causes reliability issues. ...
About de-rating controllers and motors, starting at 'for everyone else':
http://www.rcgroups.com/forums/showthread.php?p=11476940#post11476940


All useless info, and battery current protection on a controller offers nothing. Likewise, 100% duty cycle is the most efficient a controller can operate, making way way less heat than 80%. What he meant to say is not to exceed 80% of a components rated ability. This is also nonsense, most SMT gets in an ESC need to be defeated to 10% or less if you work the TDP numbers with the Rth values between the junction and sink/PCB.

A controllers FETs never feel battery current. Only the caps and some simple power traces feel batter current. People monitor and control current from the battery because its extrememly simple to do vs measuring it where it matters, the phase current, which is the loading the FETs see. A given controller may run xx motor all day long pulling 30amps of battery current. Now hook up a motor with lower resistance and inductance, have it batter current limiting to the exact same 30amps input, and your controller could smoke in seconds, as the phase current could be 10x higher now (or more, or less obviously).
 
Back
Top