Most current put through your 18 FET controller?

[zombiess]

I'm talking TO-220 based, mainly Lyen/Keywin/Hua Tong etc. I say this because I know Arlo1 might pop in here and say 350A so I figured I'd beat him to it

If you could list your sustained during acceleration battery current (and your set phase amps) which will be close to your peak and if the controller lived I'd like to know. I'd also like to know if the controller got hot while doing it. I am asking because I'm playing around with an EB318 controller I built up using match IRFB4110's on my 18.6kV hub motor and it's working much better than I thought it would. It's barely getting warm and I'm beating on it and am always in PWM since I have the PWM set to a max of 99%.

Right now I'm at 126A peak and sustain around 115A from 0 to 40+ MPH before it drops down to around 85A at just over 50mph. Phase amps are programmed to 170A. My controller is slightly warm externally, probably around 80F on the case. I did not bother to put a temp sensor on the FETs as I have on my other controllers when I built this one because it was a blown up board I got for dirt cheap and fixed not knowing if it was going to actually work. I am running 18S LiPo with a fully charged voltage of 74V. I have modded the controller the same way I build all my controllers. Matched FETs, individual bergquist silpad insulators, sanded the heat sink bar on both sides and applied thermal paste between the case and the bar. I also made sure the bar is completely flush with the case when tightening up the screws.

Thanks for participating.

 

[Arlo1]

I had all the settings maxed and we had a nichrome shunt in parallel to the oem shunt and I soldered the original up a bit too This was with the old program before the xpd was around. Mine was a kit from methods. Best I ever got was ~160 battery amps as I accelerated this would drop off fast as I gained speed fast. That controller was the one luke road in my bmx the only reason it ever died was colossus.
The controller is the one used in this video![youtube]KCUGjwviiJE[/youtube] I can try to find what the settings were but trust me everything was maxed I even moded r12 to regen over 100v and I did I ran this on 24s charged to 100.8 and regen right away out of my old house strait down a hill.

I had so much good luck with that controller.... THAT IS HOW THE WHOLE EV THING STARTED FOR ME!

 

[MadRhino]

I have an 18 fet Infineon on my Demo 8 that has about 10 000 Km, seeing bursts of 150 A batt 24s Lipo, and another one exactly the same on my V 10 that will hit 7000 Km this week. Every try higher never lasted the first 1000 Km, I've fried a lot of them to find out. The small case 18 fet Infineon is the biggest one that I like to fit on my bikes, because it can be fitted in the triangle of most large DH frames and narrow enough to build stealth and neat. If you can make one that can last with higher power, I am a client and will even buy a 3rd Cromotor with it :wink:

 

[zombiess]

I have an 18 fet Infineon on my Demo 8 that has about 10 000 Km, seeing bursts of 150 A batt 24s Lipo, and another one exactly the same on my V 10 that will hit 7000 Km this week. Every try higher never lasted the first 1000 Km, I've fried a lot of them to find out. The small case 18 fet Infineon is the biggest one that I like to fit on my bikes, because it can be fitted in the triangle of most large DH frames and narrow enough to build stealth and neat. If you can make one that can last with higher power, I am a client and will even buy a 3rd Cromotor with it :wink:

I run my controllers with safety margin. That is typically 50-66A per paralleled FET for phase amps no matter if it's a IRFb3077 or an IRFB4115. I am experimenting with this salvaged 18 FET controller to see what it can do. I will then do the same with the IRFB4115 18 FET controllers I'm building. I need to take the controller off the bike and add another shunt plus a temperature probe so I have some comparison data between the IRFB4110 and IRFB4115 which I have not done yet.

Based on my experience so far, a few people I've talked to and those who posted here it appears I can bump the current up some more. I think I will stop at 150A battery, 200A phase if it lives at these levels. I hate blowing stuff up because it means more work and a dead controller, I dislike both.

 

[scriewy]

Zombiess or any 1 ells, i am dieing to know for a few years now why no 1 just takes a clamp meter and sticks it ! in the phases to eliminate all the guesswork ?

till now the only refference i found was fechters encounter with max battery to phase ratio 1:4 in a factory preset analogue controller dating befor 2010 (with unknown settings).

many poeple can buy a clamp meter, but, not every 1 gonna go for a scope, too scientific.

what ? Is the sampeling rate of a clamp meter not good enough as an osciloscope ?

i'm drulling to buy a clamp meter, but had no time to read on types and functions i better pass on so not to spend 300$ for extra options i'l never need.

ask weight of poeple, that's a second reason for blowjobs

didnt you have a scope zomb ?

 

[Arlo1]

I did put a clamp meter on a phase wire and I can tell you it does not read properly because its rated for 60 hz

 

[John in CR]

....Based on my experience so far, a few people I've talked to and those who posted here it appears I can bump the current up some more. I think I will stop at 150A battery, 200A phase if it lives at these levels. I hate blowing stuff up because it means more work and a dead controller, I dislike both.

Controller current limits are very dependent upon the motor. Lace your 2 turn Cromotor up and the controller is unlikely to survive at 150A.

I'd like to see you experts get your hands on the programming software for the Wuxi controller boards. They seem to handle more difficult to drive motors better than the Xie Chang controllers, which is evidenced by running noticeably cooler.

 

[zombiess]

Controller current limits are very dependent upon the motor. Lace your 2 turn Cromotor up and the controller is unlikely to survive at 150A.

I'd like to see you experts get your hands on the programming software for the Wuxi controller boards. They seem to handle more difficult to drive motors better than the Xie Chang controllers, which is evidenced by running noticeably cooler.

Pssst... John, I am running this into my 2 turn 30uH inductance Cromotor right now and it's only getting slightly warm, like 80F on the case after 20 mins of beating on it at 99% throttle while pulling peaks over 120A during acceleration and sustaining +80A at just over 50 MPH for a mile ( I was playing with cars from stop lights). I think I'm going to try turning it up to 200A phase (66.6A per FET in each 3 paralleled bank) 150A battery and call it done if it works OK at this power level which I think it has a good chance of doing.

The EB3xx boards manage phase current limiting differently than the EB2xx boards that are common in the 24/36 FET controllers. If I mod the shunt on my EB3xx boards it only effects battery amps and has no noticeable effect on the phase amp setting. I took an 18 FET controller and reduced the shunt from 1.33 mOhm (what the controller expects) to 0.50 mOhm and had to keep the phase amps the same unlike the EB2xx board where I would have had to reduce the phase amp setting the same percentage as the battery setting to keep it from putting out way too many phase amps. The EB3xx boards must have a way of making some sort of ball park measurement of the actual phase amps vs just guessing like the EB2xx boards did. I've noticed an inductor type trace leading off the phase output back into the driver section and then running back into the controller chip, they could be using the RDSon of the FET and using it as a shunt to get a close measurement of the current. I really haven't reverse engineered it much to tell for sure, but it's definitely different. I've posted about this many times.

I'll be testing out an 18 FET IRFB4115 controller I built next to see how it handles the current and compares to this one.

 

[AussieJester]

very interested *subscribed* best of luck :wink:

KiM

 

[electr0n]

I've been running 90 battery amps and 190 phase at 24s for a couple thousand km on my 18 fet. Good to know I can bump it up even higher, but I think I'll need to install a wheelie bar first :lol: Has anyone figured out how to smooth out the acceleration yet? I love the smoothness I get at say 18s, 24s is kinda touchy.

 

[MadRhino]

Every 4115 fet controller that I've pushed to 150A batt had fried within a few hundred Km, if not at first ride. For this reason I use only 4110 now, always running 24s Lipo.

Of course, motor winding, wheel size and total riding weight are important factors, along with terrain and riding style. My bikes are all between 68 and 79 Lbs, Cromotors and x54 in 24" and 26" wheels, I am 174 Lbs, I ride very hard on and off road, steep mountain trails everyday in the summer, snow and salted streets in the winter, 365 days a year rain or shine, avg 12 000 Km per year.

 

[zombiess]

Has anyone figured out how to smooth out the acceleration yet? I love the smoothness I get at say 18s, 24s is kinda touchy.

I have. I will soon be selling throttle tamers. Technically I already am but I only have 8 and they are sold. Have a batch of 70 PCBs coming in about 2 weeks.

I don't suggest pushing it much harder. It could work or it could blow. I do some mods when I build controllers to try and help them live under higher power. I then rate them at levels I feel they will survive at under as wide an operating range as possible. This thread is me gathering info because I don't know how far people have pushed an 18fet controller. My previous best was 110a batt 135 phase. an 18fet Lyen controller at 125v.

 

[scriewy]

I did put a clamp meter on a phase wire and I can tell you it does not read properly because its rated for 60 hz

yes, some 1 mentioned it's not the house 60hz, but did not say how many, so when i was at a store many months ago i asked EE clerk if he knows the frequency range of clamp meter he had, he didn't.

so there are no clamp meters with broader measuring AC frequency range ?
dont kill my drooling.

what numbers are we talking about 1khz+ ?

it's not the pwm frequency, right ?
and the PWM in our boards is what 10-20khz ?

we know sustained current of FETs, their peak current like irfb4110 around 600A for a few milliseconds, and only the phase current is a mystery to us for so many years.

like IF ! the total response time of the MCU and other bugs to overcurrent is 5ms and the FET peak capable 580A for 10ms then we can set phase to 450A and fly to the moon !
this screwy on the edge numbers derived from load on single phase is 1/3 of 450A, so thats like 150A per FET, meaning 1st 5ms through 1 FET 150A will flow, and by 10ms the MCU will get the current under control, after 2 seconds you'll be at speed above 20kmh, so the motors BEMF will be slowing the 150A current down and pretty fast drop down to safety TO-220 package 75A, i just didn't have space to write this from my phone, but that's what i theoretically meant.
when i peaked 80A with a 12 FET controller i think into phases at least 2.5 times more was going, if those were the common settings in crystalyte analogue controllers.

I think the reason RC motors so small and rated so high 1-3kw and high KV dont blow controllers is because they spin out of danger zone realy fast because they are never stalled under the weight of cars which do a free spin on every wot start unless the wheel gets stuck between rocks/planes, boat models proly struggle the most. In ebikes we stall the motor with our weight till our mass is accelerated, so controllers blow if not set properly with fatass marginal headroom.

no 1 with a scope wants to probe some phases or with advanced clamp meter ?

 

[scriewy]

I

 

[migueralliart]

I'll receive one of the 18fet controllers this Friday... thinking on running 80a battery and possibly 150+amps phase? 4110fets

Dunno yet we'll see.

 

[zombiess]

what numbers are we talking about 1khz+ ?

it's not the pwm frequency, right ?
and the PWM in our boards is what 10-20khz ?

we know sustained current of FETs, their peak current like irfb4110 around 600A for a few miliseconds, and only the phase current is a mystery to us for so many years.

like IF ! the total respons time of the MCU and other bugs to overcurrent is 5ms and the FET peak capable 580A for 10ms then we can set phase to 450A and fly to the moon !
this screwy on the edge numbers derived from load on single phase is 1/3 of 450A, so thats like 150A per FET

In PWM the frequency is about 16khz, at 100% throttle it's no PWM so the frequency drops to RPS * the number of pole pairs. PWM is in effect from 0-99% throttle. At least on the high side it is, I need to scope the low side and see if there is any PWM in the EB3 boards, probably not.

I don't push the TO-220 package FETs over 75A regardless of what the spec sheets say when I use the fat part of the leads I start derating the more FETs I parallel. I don't know if I'm comfortable pushing the phase amps more than around 66A per FET with 3 in parallel and that is with them matched for equal turn on times. You always need to leave headroom in components to account for variation such as heat. I believe everyone knows that as the FETs heat up (internally Tj is the most important value to look at) the current they can handle goes down. If you are already near the maximum values when the controller is cool, ride around for a bit and then say you start to climb a hill at low throttle, no more head room left and you get to enjoy your own personal miniature fireworks show.

Anyways, everyone can do what they want with their own property, I'm just sharing my thoughts and wanted to see how extreme others have been pushing their controllers so I have a better Idea of what I might be able to get away with safely on mine. I hate blowing up a perfectly good controller if I can avoid it, but I'm feeling a bit greedy with this one and am tempted to push it hard just to see how it works out / lives. I need to install a temp probe so I can monitor what's going on, I stop abusing my controllers if the FET case temp reaches 80C.