Max amps of a 18 FET 4115 controller, check my math.

Ok, I think I get it now. Phase A only 3 of the 6 FETs are driven by PWM while 3 in Phases B and C are either floating or grounded depending on the commutation.

Let's talk theory, ignore losses for now and hope I have the cycle right (you should get what I'm after though)

So say phase A is driven high, phase B is floating and phase C is grounded, Phase A has 3 FET's in parallel with a total of 3mOhm (9mohm each), phase C is providing the ground which also had 3 FETs in parallel with a RDSon of 3 mOhm. If our supply voltage is 100V we now have 100V passing through Phase A, 6 FETs with a sum of 6mOhm combined from the high and low side. So are we only dissipating power in the high side 3 FETs at 3mOhm or in all 6 equally at 6 mOhm?
 
Seriously guys, you are helping me understand this so much better. It's almost in my grasp. I just reread a post electroglide made and it started to click.

Alan B, that's what I thought myself but I've never seen it mentioned in all of my searches, thank you for confirming it. Now I can finally get somewhere.

Working through problems like this really help me understand WTF it is I'm doing. I had to go through the same learning process when dealing with installing an 80mm turbo where my muffler goes and work through tons of compressor and thermodynamics info with help of a few friendly people. Fortunately, this is much much cheaper and everyone here shares info and isn't thermo dynamics :lol:
 
liveforphysics said:
Whoa whoa. In an 18fet controller, all phase current passes through groups of 3 parallel fets, not 6.

:lol: :mrgreen: :lol:

omg.

i am looking for 9 FET controllers info, and this thread fell into my all seeing EYE (blame the search), now i am gonna rape this 1 as well, sorry, luke havn't finished ES wiki, he's fault i am here !

reading here now raised new questions along the way, scary polynomials, what's gate ringing, is that something that happens at WOT or low RPM ?

i don't understand why the battery phase ratio makes a difference on the motor, from some old thread i recall with full throttle or hyena and mac motors the motors were jerky till they played with the ratio and increased it above 2.5 for it to work smooth, if we limit 2 systems to 50a bat and 50a phase what does the motor cares in what way he gets his 50a pahse ?

if we limit bat to 10a and phase 100a, motor starved so controller chops the voltage and converts to phase amps
if needed ? (makes big power losses, but is the motor jerky or smooth)

and 1 i asked ellswhere, what side of FETs is under the most stress, the high side ?
logicaly the current must reach the cathode to empty anode, so similar amp flow to heat parts, but when it get's out through low side does it go through 1 bank of low FETs (if low FETs is where current exits, didnt find answer to that yet) or does it exits through 2 banks of low side FETs of the non active while 1st completes his 1st 1/3 cycle ?

zombiess i see you have a 9 FETy ! how many bat amps you feed it ?

got another 1 on the way, if FETs have such a low ass resistance why they heat so much powering continually only 12v 4a bulb through it ? does it has tiny miny contacts inside ?(laymen analogy would be aprriciated)

edit: forgot most important that tbothers me few days, can i use a clamp DMM to see those sneaky AC phase amps ?
that'l cancel many complications and contemplations if people could'v just measure what a groundless pit fall their motor is for amps.

on digi-key year+ ago saw 4110 FET's rated 180a, and usualy 4310 were 120a, now every 4110 i see anywhere is rated 120a, wtf ?
 
zombiess said:
, but I suspect most of the FET failures we are experiencing are from high burst current at very low commutation speeds.

Hey guys...I just read your thread. and I would just like to say Thanks for putting in a sentence I finally understand :lol:

Yes...I have cooked the MOSFETS on two controllers now...and both at low speeds, pulling off from a start..slowing down again..pulling off ...when I had block times set too high..up at 5 seconds or so 18 FET 4110 Lyen controller at 84volt...65 amp battery and 170 amp phase current
I am playing it safe now...up to 100volts with 65 amp...160phase and 05 second block time...want to up the figures a bit but and fed up with letting the smoke out of my FETS
 
Sorry for your losses Neil.

Block time is a dangerous thing - basically telling the controller to ignore the safety limits for a time - so the current limits are disabled. This is only safe for a setup that has enough resistance in the motor and wires to protect the controller. A low resistance setup will produce currents that are too high for the controller. Block time risks the controller.

In the BLDC controller we have one high side FET bank conducting and one low side FET bank conducting at a time (the third motor phase lead is open). It is up to the designer how to utilize that, but he (or she) needs to do motor Commutation and Pulse Width Modulation. They could PWM either or both sides, but the high side drive often needs to be AC coupled and can't handle DC. Since commutation always switches when the motor is turning and PWM may not, and for a number of other reasons, it is often set up so the high side is commutated and the low side is PWM'd.

There are switching losses in FETs, and PWM is higher frequency than commutation so it switches more often, so the PWM side has greater switching losses than the commutation side. Another loss is the flyback current. When the PWM switches off the collapsing magnetic field in the motor creates voltage that flows through the inherent body diode in a commutation side FET. This is the "current multiplication" that is often mentioned that happens in the motor.

So in a 9 FET design there are three one FET switches and three two parallel FET switches, so the designer has a choice of where to deploy one FET and where to deploy the two parallel FETs. On the both sides there are on-resistance losses. The motor phase current is flowing all the time in the commutation FETs, and when the PWM switch is open the phase current is also flowing in the inherent body diode of a commutation FET (the one connected to the node that is being PWM'ed). This is the "multiplication current" from the flyback of the motor magnetic field collapsing. The PWM FET only flows battery current when it is closed. So the PWM FET doesn't see phase current, it sees only battery current!

From this I would expect the 9 FET designer to put two FETs in parallel on the commutation side. These FETs see phase current plus PWM multiplication current plus commutation switching losses. The PWM side FETs only see battery current plus PWM switching losses. It depends on how switching losses compare to forward current losses in the devices being used.

The only knob the controller has to reduce current is to reduce the PWM duty cycle. So enforcing any of the limits is done this way. Battery current limit, phase current limit, speed limit, cruise control, etc. So even when you whack the throttle open fully the PWM will still occur until the speed gets close to maximum. In low turn motor setups with high battery voltage the PWM may never stop, even at Wide Open Throttle and full speed.

Let me know if there is anything wrong in this analysis. Each time I go through this I realize a bit more detail. I am applying basic electronics and physics principles to the problem and trying to work through the effects.

On the jerky behavior when limits were set low, this is a feedback instability in the controller. It is seeing currents that are over limit so it is reducing the PWM to control them, but apparently the feedback parameters are not correct for the situation and are oscillating. It is reducing the PWM too much, and then releasing it, and then reducing it again. Nothing wrong with the low limits, it is the feedback parameters that are wrong for the situation.

A True RMS clamp current meter should measure phase amps okay as long as the frequency is within the ratings of the meter. This is not 60 hz and not sinusoidal so higher bandwidth is required.

4110 FETs come in different packages, perhaps your 180A units were in a larger package with heavier leads. TO220 packages can generally only handle about 75A per lead, so even the 120A number is not useful. Thermal factors dominate anyway and even 75A is not going to stay cool enough to last.
 
Alan B said:
Sorry for your losses Neil.

Thanks for the thought...but I take it all as par for the course when you start pushing the limits of these things without knowing where the limits are !!


Alan B said:
Block time is a dangerous thing - basically telling the controller to ignore the safety limits for a time - so the current limits are disabled. .


Yes, I did understand that from the moment i started adjusting that setting...so i did realise I was playing with fire..On quick start ups and rides it is fine, it is only where stop start riding came in to play that the issues occured
 
oh boy oh boy, i got a neuron migraine, but something did get in my brain ,thank you, you catch fast, all this in 15 hour tour !

commutation side stressed more, and is on the high side, but what is the high side, is it simply the FETs directly connected to batterys + with their middle leg ?

do you have any guesstimates by how much % the high side is more stressed ? as to figure how much more power can go in with additional support of FET.

at what you run your 9 FEt controller ? by the motors you use shouldn't you use the 9 FET on 8x8 motor with hight impedance and 12 FET on 2806 ?

"This is not 60 hz and not sinusoidal so higher bandwidth is required" is this the frequency of "electrical RPMs" ? if a controller has top 40,000 then 1/3 of that is what my meter should be able to measure ? even less because i wouldn't test this at max wheel RPM.

yes people said 75A, jeremy some time ago, though i didnt look what package they were because wasnt aware to package sizes and thought the 4 variants offered by digi-key were all TO-220, but if we burst 180a through 180a FET for 2-3 sec and same through 120A FET wouldnt they be impacted differently ?
how FETs die, like a fast blow fuse or slow blow ? i bursted 60a batt, i am sure it hit over 120a on phase in 6 FEt controller, if it's 120a can it take over 120a in bursts ? or maybe it took it because it is 180a.
not that i expect to 180a continues, just a marginal safety, usually i am out of the danger zone in 5 seconds (while experiment), usually i get to 20kmh and then it's ok to WOT with my motor impedance (what ever it is).

77a with 312a limit depends how fast you'l be out of the inefficient zone, if you'l go as neilp, go, stop, go, stop you can still blow yours.

damn it, while looking for more info about high/low side, stumbled upon more totaly different projects but related to murdering amps, now i need to jump on those guys, this is neverending stooooryyyyyyy laa la la, la la la
arlo is a nut put toroids just to choke phase amps after the commutation stage (holy damn i can talk like half a grown up, PWM stage, huh).

luke ! did you put any Frankenstein add ons for running your colosus ?

yesterday i was 6 hours after my bed time, and now 3.5, i thought i can be strait never slip down the sleepery slope, never lose count of days or pass lunch or almost late to work just for a little dose, i ma fkn info addict.

oh oh, you also might know what is the gate driver, people talk about it like it's some single part ("replace the driver", "i'll replace the driver" "check the driver"), all i see connected to the gate leg is resistors caps and transistors and then to the MCU.
 
scriewy said:
i am looking for 9 FET controllers info, and this thread fell into my all seeing EYE (blame the search), now i am gonna rape this 1 as well, sorry, luke havn't finished ES wiki, he's fault i am here !

zombiess i see you have a 9 FET! how many bat amps you feed it ?

At least this is a question I can answer LOL. This controller was running 24S, 99.4V hot off the charger, IRF4310 FETs, 35A battery, 88A phase. Ran it like this in 100F+ weather for over 100 miles, never failed, still have it, still works great but is now retired to the parts bin.
 
Subscribed.
 
Back
Top