Solved 6 FET E-Crazyman, 120% & current limiting observation

[NeilP]

Just been playing with some E-Crazyman 6 FET 4110 controllers I bought on E-bay.
EB306 boards

Current limit set to 30 amps, phase to 80.
running at 50 volts

Current monitoring with a Stand alone CA with the supplied e-bike.ca CA remote shunt

16 inch wheel with unknown motor re configured as Delta terminated

Block time zero

When at 98 % it runs up to about 22 mph...current is limited to 30 all the way up to speed, apart from a slight blip over 30 on start up.
At 22 mph (480rpm) is where it tops out for speed, current down to about 14 amps.

But here is where it gets weird.

go up to 120% speed setting and the current limit jsut seems to get ignore totally.

Fromm 22 mph, speed accelerates slowly till somewhere between 23 and 24 mph (500 and 525 rpm) and then the bike gets a kick in the pants and current shoots up to 56 amps and more

I assume this kick in the pants happens when the controller goes over the 100% point and drops out of PWM..as mentioned here
http://endless-sphere.com/forums/viewtopic.php?f=2&t=49504&hilit=myths

But the fact that it ignores current limiting is worrying.

Not really looking for any solution as I intend to run at 98 %, but just putting this up for information/interest really

Comments of course welcome

 

[zombiess]

That is pretty interesting that the current shoots up that much, but I have an idea why, but I'd have to scope it to check. I personally don't like running a controller over 100% as I feel the gains just aren't worth it when I've tried it.

I've noticed there are several differences between the EB2xx and EB3xx boards as I have noted before. One of the major differences is that modding the shunt on an EB2xx board changes both the phase and battery amp values, but on the EB3xx boards it appears to only effect the battery amp values. The new controller chips might have some way of doing a crude phase amp measurement vs just guessing based on my experimentation. It's just a theory though, but I do know for a fact they act totally different. My reason for believing this that I modded an EB318 shunt from 1.25 ohms down to 0.3x mOhms and when I corrected my phase amp setting along with my battery amp setting the bike barely ran. I figured out the battery amps scaled correctly based on what the controller expected for a shunt value vs the much lower one, but the phase amps needed to be set like normal for the bike to run correctly. I also noticed that my regen current never increased with the lowered shunt. On the EB2xx if you mod the shunt you get more regen current, but on the eb3xx the regen current stayed the same no matter what I did to the shunt.

 

[NeilP]

I could try scoping it my end if you had time (via PM ?) to give me some pointers on scope setup.

I have a Tektronix 475 Dual-Trace 200 MHz Oscilloscope I was given a few months ago. I last used a scope maybe 25 years ago, so it is just a bunch of knobs and a screen at the moment.

I have 4 x of these 6 FET EB306 boards here, and had planned to mod them with 2 shunts instead of just the 1. I believe Lyen is now sending them out like that now, to reduce heating and resistance rise as current increases and shunt heats up and reduce resistance as seen by the FETS. His advice is after modding, just set the software to a 2 shunt board setting..so a 9 or 12 FET profile in XPD,

 

[zombiess]

Try setting your block time to zero and see if it still stays over the limit. If so I wouldn't use any value over 100 percent.

 

[NeilP]

Try setting your block time to zero and see if it still stays over the limit. If so I wouldn't use any value over 100 percent.

From my first post
:wink:

16 inch wheel with unknown motor re configured as Delta terminated

Block time zero

I have it set in the 4 speed (pulse x2 to ground) mode and am just playing with settings really. It is for a compact folder so I plan to remove the speed switch and get one more bit of clutter off the bars . CA will go too, and just relying on controller LVC...minimal wiring. Planned on leaving it at 98 % and around 25 amp battery, just playing with settings now for interest. It was originally a Wye termination, but even at 100 volts in a 24 inch wheel it barely made 32mph. A 16 inch wheel really slowed it down, but ridiculous torque. so Delta and 50 volt is giving more realistic performance

 

[migueralliart]

I've had the same experience with the lyen 6 fet which is an infineon controller. Thing is I gave up on the 120% setting a long time ago. It just seems that it is overly trying to achieve that extra speed. My controllers all get hotter at 120% compared to 98%. IE my trike on 50amps and less than 100% never gets too hot with a lot of stop and go. For some reason cruising at 110% or more it get warm pretty fast.
Just my 2 cents

Also how much did you pay for your controller. Does ecrazyman has it in stock ready to buy or did you email him? I can't get ahold of him thru ebay.

 

[NeilP]

Yes, it will get hot running 50 plus amps!! I don't plan to run it at those settings.it is a 16 inch wheel folding bike with a wobbly handlebar and a frame that feels like it will fold while riding it!!

http://www.ebay.com/itm/72V-1500W-brushless-controller-for-E-bike-scooter-/300813212040?pt=LH_DefaultDomain_0&hash=item4609dd5988

same as from Ed Lyen pretty much..$75 as opposed to $79 from Ed...but Ed's postage would have probably been cheaper, so pretty much the same price.
I do not actually know why I did not buy from Ed. I have always had good service form him .

I am assuming that of course they are the same controller..as Ed sells.

The wiring is not quite as I would like..the CA plug is not quite right..it works, but the negative wire for shunt and other functions are combined..should be two separate wires to give accurate monitoring of current , so I am having to change that and add wires myself for the 3 speed or 4 speed switch, add the solder bridge to enable regen braking etc. Lyens come down already done.

Any way E-crazymans e-mail ecrazyman@gmail.com
Ed Lyens e-mail address lyen@hotmail.com

I e-mailed him via e-bay ..no idea if he has them in stock ready to go or if he has to make them up. But if you have already bought from Edward...just carry on with Ed and save yourself the work.

 

[migueralliart]

Yup I bought from lyen and ecrazyman before.

Communicating with lyen is better. Ecrazyman replied to emails like once every couple of days.

Although I gotta say I bought a 12fet from ecrazyman that to this day has seen LOTS of abuse. 3077fets baby 50 amps ALL day long on 48V.

 

[NeilP]

Lyen can be slow with e-mail too..I think the reason is that they are not sad buggers like some of use that check e-maisl every 10 minutes

 

[zombiess]

I've had the same experience with the lyen 6 fet which is an infineon controller. Thing is I gave up on the 120% setting a long time ago.

Just FYI the controllers are not made by infineon, they are made by Xie Chang. The 120% is not the best solution for more speed. It's normally better to increase the voltage. I have read of several guys on here pushing modded 6 fet controllers to the 80amp range which is very impressive. Personally I would suggest not running above 60 amps, but take that with a grain of salt as I personally have not abused a 6 fet myself yet.

I have 2 eb306c ( optional sensor less mode)controllers, one my wife built for her bike and another I built I need to finish. I do love these eb3xx boards and the options of switching between 4 speeds. I previously scoped the eb3xx boards and they run 100% pwm like the eb2xx series. I have not verified 99% keeping them in pwm yet.
I encourage others to join me in finding out more info by testing.

I haven't dealt with cell man, but I still buy parts and send customers to Ed Lyen who has always responds quickly.


Anyways just wanted to chime in with my 2 cents since I have been doing a lot of research on them lately.

 

[NeilP]

The 120% is not the best solution for more speed. It's normally better to increase the voltage.

Yep fully get that, but my GF's bike runs at 12s, (50volts) and I had 6 spare 4s hardcase packs, and the same extruded alloy section to make the box the same as the GF's bike...same pack design, same charger, so not voltage increase for this bike. it is a small folder for the nose locker of the aircraft,(PA31-350) so can't go with a bigger pack.
Motor was in a 24 inch wheel initially, now fitted in 16 inch..so that slowed it down even more. The swap to Delta config brought the speed up higher than before...a beefy little motor.
It has appeared on here before in this thread

http://endless-sphere.com/forums/viewtopic.php?f=3&t=25399&hilit=thompson&start=25

 

[NeilP]

I found out why it APPEARED that current limiting was not working.

the controllers have had a mod done, un requested, fitting a second 4 milliohm shunt an SMD shunt on the underside of the board.

So the profile chosen when programming the controller needs to be a profile for a 2 shunt controller..so a 12 FET controller profile has to be used .

I was setting 35 amps with the 6 FET profile..so it was setting a current limit of 70 amps and not 35 that i was expecting.

I only found this out when I opened the box to do the mod myself,,only to see the SMD shunts already there

 

[gwhy!1]

Yup I bought from lyen and ecrazyman before.

Communicating with lyen is better. Ecrazyman replied to emails like once every couple of days.

Although I gotta say I bought a 12fet from ecrazyman that to this day has seen LOTS of abuse. 3077fets baby 50 amps ALL day long on 48V.

50A is peanuts .. i have been running one of my 12fet 3077 at 140A and tis fine, but i have the phase only set to 220A and blocktime set at zero, it makes for very smooth throttle control :wink: and dont get hot with peaks of 7.2kw .. happy days

Edit.. It was me that used to run a 6fet 3077 at 80A , did so for ages, without issue.

 

[NeilP]

3077 are higher current than the 4110's though aren't they?

 

[gwhy!1]

3077 are higher current than the 4110's though aren't they?

yes higher rated current but lower rated voltage, The rds is also lower on the 3077's so they don't tend to run so hot

 

[NeilP]

What about the gate driver side, can you just swap fets, so take out 4110's and fit 3077's?

Got a. 50 volt pack. That is 50 v hot off charge.

 

[zombiess]

3077 are higher current than the 4110's though aren't they?

yes higher rated current but lower rated voltage, The rds is also lower on the 3077's so they don't tend to run so hot

Your current is really limited by the TO-220 package. You can pulse current through any FET as long as the time period is short. Realistically the max any TO-220 packaged FET should be used at in a controller that you want to be reliable is around 75A. When in parallel and matched I derate them to a max of 50-65A each.

The RDSon heating is typically low compared to the losses these controllers have while switching and body diode losses under most riding conditions with 100% throttle being the exception. Personally I always keep my controllers programmed to 99% throttle because I don't like them dropping out of PWM mode, the mph gain in speed it gives me is not worth it. 100% throttle puts the controller into full block commutation so there is no PWM. It's OK for the overwhelming majority of ebike motors to run 100%, but I prefer to take the bit of extra heat that PWM causes vs the chance of popping an FET.

Good news is that if you want to use the IRFB3077's vs IRFB4110's is that they have just a slightly higher QG of 220nC vs 210nC for the IRFB4110. This means that they are about equal in switching performance given the same gate driver setup. Personally I have two 6 FET controllers built with IRFB4110's. IRFB4115's would switch slightly faster due to their lower Qg of 120nC, but these controllers will never run over 75V and with just a single FET they should switch decently by Xie Chang standards :lol: Take that with a grain of salt as I have not scoped a 6FET gate driver yet so I don't know exactly how slow it is.

 

[migueralliart]

Yup I bought from lyen and ecrazyman before.

Communicating with lyen is better. Ecrazyman replied to emails like once every couple of days.

Although I gotta say I bought a 12fet from ecrazyman that to this day has seen LOTS of abuse. 3077fets baby 50 amps ALL day long on 48V.

50A is peanuts .. i have been running one of my 12fet 3077 at 140A and tis fine, but i have the phase only set to 220A and blocktime set at zero, it makes for very smooth throttle control :wink: and dont get hot with peaks of 7.2kw .. happy days

Edit.. It was me that used to run a 6fet 3077 at 80A , did so for ages, without issue.

Been running 50amps due to the fact that the 9c can't take more without modifying the phase cabled.

But I agree my 12 get never got warm to the touch.

 

[gwhy!1]

3077 are higher current than the 4110's though aren't they?

yes higher rated current but lower rated voltage, The rds is also lower on the 3077's so they don't tend to run so hot

Your current is really limited by the TO-220 package. You can pulse current through any FET as long as the time period is short. Realistically the max any TO-220 packaged FET should be used at in a controller that you want to be reliable is around 75A. When in parallel and matched I derate them to a max of 50-65A each.

Yes 75A for the TO-220 package, so the 80A I was pushing them is maybe the max 'safe' current as through board lead length is minimum. The 6 fet controllers run slightly cooler for the same power as a 12 fet controller with the same fets fitted ( a 80A 6fet runs cooler than a 80A 12fet )

 

[NeilP]

This is only on a folding bike, with a wobbly steering column and 'flexible' hinge joint, and very basic brakes..so 25 to 30 mph and 30 amp is plenty !!