6 - 9 - 12 FETs ... Why 9?

Ok I know what a hex bridge is ...3 phase drive ... So 6 FETs and if you double them or triple them you use 12 and 18 FETs but how are 9 FETs used? Is it in the free wheeling or what? Searching doesn't seem to find anything ANYWHERE ... I'm talking about motor controllers. Thanks in advance!

There's been a fair bit of discussion here on ES, but becuase the search engine doesn't look for antyhing less than 3 characters, you can't easily narrow down all the threads with "FET" in the title to find them.

Here's one I posted in, with a bit of info (but not complete)
https://endless-sphere.com/forums/viewtopic.php?f=2&t=86477


And this one
https://endless-sphere.com/forums/viewtopic.php?f=2&t=19642&hilit=*fet*+*odd*

They parallel one side of each bridge section because the heat load on one set of FETs is slightly higher than the other.

It the heat load higher because the internal diode is being used to handle FET off time current flow or just a difference in P and N FETs?

Thanks guys!

Wow that first link "not complete" discussion sure nailed down a few other things I had not thought about like low side only commutates while high side PMW's plus because they allow the high side off time to rely on the low side diode that adds a lot of diode drop wattage. Says a lot about the two stage controller design I have been working with where the hex bridge just switches the phases and another stage feeds the B+ to the bridge with a pair of high low FETs through a low loss high frequency inductor. No PMW to the motor windings so the motor runs much cooler and radiates only communtation frequencies and less noise!

Thanks guys! I have the Grinfineon 35amp arriving today in the mail ... My first order from Grin Technology.

Generally only N FETs are used, P FETs are generally not chosen due to increased cost.

Adding stages to the controller usually results in more cost and more overall loss.

It's not adding stages it's a two stage controller ... First stage is a buck regulator and the second is a 3 phase generator. How about Regen with 1 volt of bemf that doesn't pulse the motor windings with short circuits to induce voltage spikes up to the battery voltage instead it is done with a low loss high frequency inductor that is boosting the voltage from the synchronously rectified output of the motor windings to a capacitor Bank that is charged to the bemf of the motor as it spins.

Each stage adds loss. It is difficult to achieve the lowest loss this way.

Each stage adds cost. It is difficult to minimize cost this way.

Not saying it is not interesting, I've looked at the same ideas in the past. Go for it! Build a few.

I don't have to build a few ... I have a few ... yeah a little pricy but at 100-200 amps up to 60 volts they do pretty good at 1-2% loss. And the motor runs much cooler than with single stage controllers.

Good quality single stage controllers are very efficient. Don't compare to the cheap junk.

Very little motor heat comes from ripple.

60 volts is a bit low, but many ebikes are under 60V.

Demonstrate your stuff, love to see it.

Make it the size of the PhaseRunner and more efficient, may be a market for it.

Ebikes don't have a lot of space.

The motors these were designed for turn 6000+ rpm at 60 volts ( 2 turn 0.013ohm) ...no need to go higher voltage but it's a FET change if you need to and some were made with IGBTs. These are the first generation controllers, the newer ones are much smaller and are integrated into the motor cover and do 300amps. As for eBikes, you do know there are Federal speed limits if you want to still call them a bike. Having ridden motorcycles for many years I know you really don't want to be going too fast on a bicycle frame at least not without more body protection. Bad things can happen ... Been there, done that, and have the scars to prove it. I am running a RoboteQ 50 volt 150amp controller on a Chinese rear wheel hub motor direct drive with 35 amp phase limit just to see what these cheap hub motors can do and the Regen is insane. As it is I can feel the motor wires getting warm in my hand at that 35amp phase current limit ... the controller output wires are 8 awg ... yeah lots of fun making up reducing connectors.

My first Grinfineon 40amp controller and CA3 that arrived yesterday (yahoo!) powered up with a major throttle issue ... not to randomly working when going through the CA3 ... At one point the wheel jumped from stopped to maybe half speed so fast the torque compressed the rear suspension ... good thing I was running on a rear wheel stand ... apparently the phase current is controlled by the throttle adjustments when passing through the CA3 and I may have a connection issue in the wiring between the CA3 and controller so although the CA3 was ramping up the throttle with phase current limiting set to 35 amps, it connected to the controller at a "go really fast now" level causing it to kick. Max phase current is 90amps? Yeah that would not be good! And I thought the controller power connectors we're not supposed to spark when connected to the battery? What's up with that? The Andersons were already pitted out of the box! I had to upgrade my pre-charge connector to solve that problem. Next I start checking the crazy long leads!

It's a stretch to call motor current "controlled" when talking about a pwm-throttle controller. Most of those controllers don't even measure motor current. They measure battery current and with a motor at low speed the motor current can easily be 3x the battery current. It drops quickly as the motor accelerates.

There are many adjustments in the CAV3 to get the results you want. Depending on what you are setting up, testing with no load on the motor can be confusing.

The controller capacitor bank is connected directly to the input wires, big arcs are normal unless precharging is used.

The PhaseRunner is a different animal entirely. Much more sophisticated FOC control in a much smaller controller, it measures and controls motor current directly, and implements real torque throttle like an electric car or motorcycle. Much improved control.

JanComputerman said:

As for eBikes, you do know there are Federal speed limits if you want to still call them a bike.

Click to expand...

No, there aren't. There's lots of discussion about it here on ES if you want to dig up the threads, but the CPSC regulation that "defines" an ebike is only for the purposes of what a manufacturer can sell as one.

Doesnt' have anything to do with what states actually define anything as (as all that stuff is up to each state to deal with).

JanComputerman said:

I am running a RoboteQ 50 volt 150amp controller on a Chinese rear wheel hub motor direct drive with 35 amp phase limit

Click to expand...

Unless that controller is unusual, then as AlanB noted it won't have any way of actually knowing what the phase current is, only the battery current, and then it "guesses" via math on the battery shunt reading what phase current *might* be.

My first Grinfineon 40amp controller and CA3 that arrived yesterday (yahoo!) powered up with a major throttle issue ... not to randomly working when going through the CA3 ...

Click to expand...

If you haven't gone thru and completely configured the CA3, then it's behavior might not be anything like what you want until you do, as there are many interdependent settings in it. Go to the CA3 page on Grin's site, and follow the links there to the Teklektik UUG / etc, then follow all the steps in that guide to setting all of the menus up the way you want them to be, to ensure predictable behavior.

If you had Grin set it all up for you to start with then it'll at least have all the basic stuff done, but if not it'll all be at the defaults which may or may not be useful for you.

apparently the phase current is controlled by the throttle adjustments when passing through the CA3

Click to expand...

No, it's not.

First, the Phase current isn't controlled at all in any of the Grinfineons I'm aware of--they simply dont' have a phase current sensor, so they don't even know what the phase current actually is, only the battery current.

Second, the CA has no control over anything other than what the throttle voltage is, and switching the ebrake line on and off. The only thing it knows about what the controller is doing is what the battery current is, via either the standalone shunt or the controller's shunt, in the battery negative wire. How the CA interprets that and then sends control signals via throttle and brake can be very simple or fairly complicated, depending on how you've setup the CA, but it still cant' control phase current as there's no mechanism to do so.

And I thought the controller power connectors we're not supposed to spark when connected to the battery? What's up with that?

Click to expand...

Unless you have installed a precharge setup, or one comes preinstalled, then that's completely normal behavior. You can look up the many threads about it via "spark" or "precharge" to find ways to fix or minimize this.

Ahh yeah I wiggle the wires and the wheel starts and stops ...I think I have a bad connection somewhere ... Been through the configuration before it even arrived and I am familiar with ramps but I will admit there were a few gain settings in there I didn't see in the documents I read. I push buttons for a living so I figured out most of it but yes ...twist the throttle and nothing happens ... Wiggle the wires and twist the throttle and it runs ... Riding it the same would happen when I hit a bump, Regen would work sometimes then no throttle so I would pedal then hit a bump and it would slow to 4 mph by the Regen working. The only thing I did not do is tie the ebrake input high or low going into the CA3 but when it was connected to a N.O. brake lever switch the screen throttle icon changed and nothing worked again. I monitored the throttle input and output with the CA3 screen and it outputs a varing throttle signal to the controller but sometimes the wheel spins and sometimes it doesn't.

As for the phase current in the RoboteQ controller yes it knows the bemf of the motor as well as it's rpm and battery voltage and current and it calculates the resulting current in the phase and therefore knows the phase current ... It doesn't have anywhere else to go! And yes the end result of a 35amp phase current is really tame acceleration and Regen braking along with an 800rpm/second ramp. At 120 amps I climbed a 16 degree dirt driveway up a mountain so I know my phase amps also skidded down the same driveway with Regen.

As for the CA3 yes it controls the phase current by regulating the throttle ...if it measures too much battery current it cuts back on the throttle ... The net result is control of the phase current. Now if you mean it only regulates the battery current and does not calculate the duty cycle effect on the phase current then that may explain the exceptional low end torque I just experienced for about an hour and 53 minutes ... It had some of the worse drivability I have ever experienced on an ev. Current throttle was nice as was power throttle but speed throttle ... Too much function in a 1/4 of a turn ...Regen, dead zone, throttle, upper limit dead zone. And speed throttle is the only one that give variable Regen and we have hills here so I use it 18% Regen in 16 miles today.
I ohmed out the throttle wire between the controller connector pin and the circuit board solder connection..it was good.

Next?

The controller can estimate phase current with a calculation. I haven't seen a low end controller that knew the coefficients of the motor that it would take to calculate back EMF, or the hardware to measure it, but I'm not familiar with the Roboteq you have (which is not really a low end cheap controller). Since the motor current changes over the RPM range by a factor of at least 3:1, and possibly a lot more, it makes a control loop like the CAV3 not behave the same at the ends of the control range. In particular they generally don't control very well at low speed (since they are adjusted to work well at high speed), so the phase current shoots up very high very fast when there is no back EMF.

If you test ride a Zero motorcycle you'll see the difference a torque throttle makes, where they actually measure motor current and do closed loop control on that. It comes out much smoother than PWM control (or the improvements that one can make with a CAV3), and the Zero throttle controls precisely from zero to a top speed well over that of ebikes. The PhaseRunner also has this, as do most any of the true FOC sinewave controllers.

Sorry that you are having intermittent problems. Those are not fun. That can lead to destruction of the controller, I would immediately remove power and go through every wire and connector and attempt to find the mechanical problem with the wiring/connectors before it causes an electronics failure. It is probably a bad crimp, bent pin, broken wire or bad solder joint.

Best of luck with your troubleshooting.

Arg lost the entire post ...darn wifi dropouts!

16 miles of riding around today having the motor cut out when hitting bumps. Definitely a loose connection somewhere and not the connectors. Throttle lag is a nightmare but I spent a few hours tonight doing the PID tuning, it was a lot easier on my quad copter flight controller! Everything seemed to be working well tonight until I turned the handlebars to the right about 45 degrees and then the motor stops. When I turn the handlebars back the motor started with a bang! Since the throttle was about half on, it went immediately to half speed!

The RoboteQ controller was $425 plus a few connectors and a cable for Communications. Definitely overkill for an under powered ebike motor but the throttle control is a dream! You just twisted and it goes faster and slower as smooth as can be.

As for the phase current ...it is the B+ current that flows through the high side FETs when they are on and when they turn off it continues to flow through the low side fet or the body diode of the low side fet. Completely measurable! This is standard switching power supply technology that I studied and worked with back in the mid 1970's.

For most controllers they shunt the use to measure battery current is between the battery and the capacitor bank (in the ground return), so they cannot see the circulating current in the motor at all, nor peak currents. They see only the average filtered by the large capacitance. That's just the way they are made. It doesn't have to be that way, but that's the most copied topology in ebike controllers.

Sounds like they did a better job on the Roboteq than the usual ebike controller.

Sounds like the same guy that did the shunt on boats ... The RoboteQ circuit board is surprisingly cleanly laid out and from the looks must bring a nice profit. The other two stage controllers I have use a shunt between the stages on the low side and since it doesn't pmw the motor it is the actual phase current there. As for the "filtered" current (battery current at that point) if you know the duty cycle you can calculate the phase current and from what I see it's a good bit higher than the battery current when I see 1000watts from my battery flowing and I'm going 5mph! Probably why the motors get hot and loose efficiency ... too high a phase current. I really wonder about my battery resistance measurement when I have 3 ft of wire running to the battery ... Bet that will change when I trim the leads!

#10 is only a milliohm per foot, so not too bad.

The controller can do a good estimate, some do while others don't seem to. Must be firmware or poor ADCs or filtering. The CAV3 doesn't have the benefit of knowing the PWM duty cycle so it is not in as good a position to do that estimate, so they just offer battery current or power options.

I looked at the RoboteQ at some point, but didn't try one. The PhaseRunner is very small and has a lot of features, the only limitation is the maximum currents are a bit lower than some might want to use. However for a 2WD I get to use two of them, and that's very adequate power for an ebike.

I think RoboteQ had a dual output controller, but for some reason it didn't appeal to me for this application. I don't think it had all the capabilities I needed, and it wasn't small, perhaps. I don't recall now. They are also not common here on ES, likely due to the cost.

PWM duty cycle is ideal but rpm and voltage could be used to estimate the duty cycle but anyway I'm not building a controller that way ...yet.

RoboteQ has a 60+ volt very small controller in the $250 price range about 1.2 inch thick by 3-4 inches but they overload it with communication options that drive the price up out of ebiker budgets. Most are available in dual outputs too. They have basic language programming which can be used to impliment any special control functions which I have played with a little.

I had a bit of a breakthrough this afternoon with my Grinfineon 35amp controller throttle problem ... the throttle contact in the ca3 cables was not making reliable contact because some of the stranded wires from the crimp on the female contact had protruded into the spring contact area preventing the contact from pressing on the mating pin from the car side. Moved the wire strands out of the way, bent the spring a little bit and it's working fine!

There you go. Good find. That is about what I was expecting.

My problem on the 2WD bike is the motor wind I chose (my first ebike) is a higher voltage unit, so 60V is not sufficient. The PhaseRunner is rated over 90V and is better for higher voltage lower current since it is a 6 FET design, so suits that particular setup nicely since I didn't want to throw away a good motor/wheel build. So that one runs 20S and about 84V full charge. The rear 9C is a bit anemic by itself on steep gradients, but with the front BMC gearmotor the combination is quick and sure footed on the dirt and gravel fire roads, and handles gradients that would melt the rear motor alone.

For my newer ebikes 60V is turning out to be a nice value, the 14S setups like the BBSHD, BBS02, and the BikeE's BMC as well as the Swift's dual speed geared hubs all work very well at just under 60V.

The 18S Borg needs a bit more to feed a big hungry Cromotor, but the Sabvoton FOC controller has proven to work well there, after many failures of more conventional controllers due to the high phase current peaks. With the precise adjustability and control of motor current I can set it to almost lift the front wheel, when I pedal hard on a hill the front wheel just bounces a little, so it is right at the edge. That way I don't have to deal with front wheel lifting when the throttle is opened hard. At least for my purposes I don't really need the front wheel to lift. It gets up and scoots very nicely though, and with variable electric braking it also decelerates and regens with nice touch and control, a great pavement eating commuting machine.

I would like to post a photo of my ride for the past ? 13 years I think ...it's been quite a while ... 225amp max Current but usually about 100amps and 36 volts was what I ran that would get about 28-30 mph but only slow about 1 or 2 mph up hills. The idea being low turn motors with less inductance so you can push the amps at higher rpm. I started with AGM 23lb Hawker 12v25ah batteries X3 and a few X4 for 40mph top speed on a 46lb 10 inch tire BladeZ scooter yeah 3x was 70 lbs of batteries plus the 46lbs of steel scooter and the 15lbs controller I sat on ...it was that big! Used NiMh F cell packs for a while then finally now with the lipo multistar and nanotech packs where I can carry the 30 Mile range battery packs in one hand and RoboteQ controller at 3.5lbs things are much lighter. I used to pick up that scooter and put it in my car trunk! I set up a ramp on the basement stairs with electric winch to bring it up and down. Now these Chinese hub motors do ok, have no where near what I am used to for power but at least I have SUSPENSION and can ride it anywhere a bike can go and the entire setup is 50-52 lbs that I can lift with one hand. It is a bit of a pain fitting in my 06xB but I managed to lower the handlebar so it can stand up in the car.

Yeah finding that bad connection for the throttle really gives the bike a whole new feeling when you twist the grip! It "goes" instead of "will it go this time?" I even let a buddy ride it down the street today ... So funny as he had not been on a bike in a long time and was all over the road totally oblivious to the SUV that was following him. It's a Dead End Street and the bottom is a steep hill but he had no trouble going down it and back up again.

Just an update ... Did a 37.1 Mile ride yesterday and no throttle problems ! Man that cruise control is great once you get used it turning on when you Don't want it! Good thing I fixed that ebrake switch before my ride. Btw it was a slow ride 11.0 mph average but do you guys get like 7.2 wh/mile? At one point it was as low as 6.5 but the last leg is up 360 feet elevation. Used about half my battery so it could go 70-80 miles per charge.

7 wh/mile is a low value, quite efficient, but 11 mph is low speed so that may be normal, especially if providing human power. You could compare it to the motor or trip simulators at ebikes.ca to see.