You should ask Tesla Motors for some tips.
http://www.teslamotors.com/performance/acceleration_and_torque.php
200KW.
Is this still considered a "green" vehicle?
http://www.teslamotors.com/performance/acceleration_and_torque.php
200KW.
Is this still considered a "green" vehicle?
Nah. They use battery pack that weights 450 kg and outputs only 200kW.theAnonymous1 said:You should ask Tesla Motors for some tips.
Besides, manufacturing that battery pack uses awfully lots of energy. And battery pack alone cost something between 10000$-30000$
15000$ and 100000 miles lifetime of battery pack doesnt sound so cheap&green anymore, conventional mid-sized(by european standards) gasoline or diesel car can do better than that.
PS. I deny any negative comments on EV technology when MZZJ-motorcompany starts marketing full line of green dragster go-karts
Very interesting topic!
Although I have been working with electronics all of my life my hands on experience with power MOSFETS , IGBT`s and PWM circuits is new and only something I started studying and playing with in the last year as I suddenly had the need for a PWM DC motor speed control for a DIY EV I`m building. Considering my limited experience here I question the value of using such a high rep switching frequency. I think you would eliminate issues if you went to a more moderate frequency as I have done at around 6 KHz. I started higher but went down because of Miller and gate drive issues. I found a quiet area around 6 KHz where motor noise is not bothersome to me. At 4 KHz it was terribly annoying. I also question the need for the electrolytic caps that you are beating up so hard. Why do you need these at all? I see that it increases peak circulating currents by a huge factor and adds complexity and engineering problems. I use the term `circulating currents` in the same context as one might say `eddy currents` in a magnetics system, i.e. unhelpful power being pushed around, which costs power to push around. Does it result in a real and practical power increase at the rear wheels?
I would encourage going to at least 2X your present battery voltage ASAP. Would increasing battery voltage not eliminate whatever need you have for all those capacitors? I think IGBT`s would then be quite acceptable. IGBT`s are what I use. 600 volt jobs so I don`t fry them with what the snubber diodes cannot deal with. I am operating at 48 volts.
Although I have been working with electronics all of my life my hands on experience with power MOSFETS , IGBT`s and PWM circuits is new and only something I started studying and playing with in the last year as I suddenly had the need for a PWM DC motor speed control for a DIY EV I`m building. Considering my limited experience here I question the value of using such a high rep switching frequency. I think you would eliminate issues if you went to a more moderate frequency as I have done at around 6 KHz. I started higher but went down because of Miller and gate drive issues. I found a quiet area around 6 KHz where motor noise is not bothersome to me. At 4 KHz it was terribly annoying. I also question the need for the electrolytic caps that you are beating up so hard. Why do you need these at all? I see that it increases peak circulating currents by a huge factor and adds complexity and engineering problems. I use the term `circulating currents` in the same context as one might say `eddy currents` in a magnetics system, i.e. unhelpful power being pushed around, which costs power to push around. Does it result in a real and practical power increase at the rear wheels?
I would encourage going to at least 2X your present battery voltage ASAP. Would increasing battery voltage not eliminate whatever need you have for all those capacitors? I think IGBT`s would then be quite acceptable. IGBT`s are what I use. 600 volt jobs so I don`t fry them with what the snubber diodes cannot deal with. I am operating at 48 volts.
rcavictim said:
Those all are input side filtering caps only. Input side wire inductance is the reason for that huge cap bank. apprx 4uH wiring inductance stores 2 joules at 1kA, and if switching at 10khz that would result 20kW losses _somewhere_ You can snubber it, or slow down switching speed (and burn more power in switches), or see if your mosfets can handle that much in avalance.
mzzj said:
You lost me when you said slower switching speed would eat up more energy in the switches. Since the transition time from off to on is fixed by the device gate parameters (shunt capacitance) and the drive circuit, switching at a slower rep rate gives more time per switch cycle, or a higher percentage of total time, with the pass switch device in the on state. How can this increase switch losses? In my explanation, higher switching speed increases switch losses, it does not decrease them. This is opposite of your claim. What am I missing?
BTW, that 20 kW you are afraid of losing in your example is reactive power. It will go back to the motor or the batteries. If it was being used up it would be making something pretty darned HOT
and you would notice that. If you snubber it, then those diodes will get hot and the reactive power will be converted into heat energy in the diodes and related circuit wiring, radiated and indeed actually lost.
mzzj is right. The rest are misunderstanding the basics of a buck converter when parasitistic inductances are considered.
The capacitor bank is mandatory in order to avoid a few dozen KW of switching losses. Very fast voltage rise and fall times are also mandatory to avoid another few KW of switching losses. The operating frequency could be made lower, but a higher freq neither imposes any limit on maximum duty cycle nor results in higher switching losses provided that rise and fall times are reduced accordingly.
The capacitor bank is mandatory in order to avoid a few dozen KW of switching losses. Very fast voltage rise and fall times are also mandatory to avoid another few KW of switching losses. The operating frequency could be made lower, but a higher freq neither imposes any limit on maximum duty cycle nor results in higher switching losses provided that rise and fall times are reduced accordingly.
Well thanx for trying to edumicate me . I`m clearly not up to speed yet on this technology and not afraid to admit it. I only just began experimenting with PWM and IGBT power control this year.
Question. Do the higher power commercial PWM controllers made for the EV marketplace such as (picking a well known name for example only) the Curtiss, contain a bank of such electrolytic caps?
Question. Do the higher power commercial PWM controllers made for the EV marketplace such as (picking a well known name for example only) the Curtiss, contain a bank of such electrolytic caps?
here's a video of an electrical superbike
http://video.google.com/videoplay?docid=-2496640399999826485
maybe u can get those batteries and motor somewhere....
(btw, this is the bike i ride , but i ride with a normal engine in it)
when u convert HP to KW ,this bike is making 131KW from combustion engine (less with that electrical engine).
http://video.google.com/videoplay?docid=-2496640399999826485
maybe u can get those batteries and motor somewhere....
(btw, this is the bike i ride , but i ride with a normal engine in it
)when u convert HP to KW ,this bike is making 131KW from combustion engine (less with that electrical engine).
What do you mean by PAR?Eva said:
Our objectives so far have been complete opposite of limiting current.
If something breaks up that just need to be more stiff.
Last weeked we glued current limit circuitry to other control electronics, adjustable from 0A to 2000A.(battery current)
1500A setting proved to keep motors in one piece, even though motor stall current is roughly ~3000A.
HTFS-800 by LEM was a nice supprise, hall current transducer capable of measuring 1200A and costing only 22 euros(farnell)
Last time I was looking for current transducers the cost was something like 300 euros and those required bunch of auxilary supplies etc.
Sorry, PAR is the Spanish term, I meant torque.
Then what magnitude are you exactly adjusting with the "gas" pedal? Duty cycle directly? Non-linear duty cycle through a lookup table? Average PWM output voltage? Speed through a feedback loop?
BTW: You should use current limiting anyway, with short term and long term limits, maybe even temperature dependent limits, to get maximum performance without smoke.
Then what magnitude are you exactly adjusting with the "gas" pedal? Duty cycle directly? Non-linear duty cycle through a lookup table? Average PWM output voltage? Speed through a feedback loop?
BTW: You should use current limiting anyway, with short term and long term limits, maybe even temperature dependent limits, to get maximum performance without smoke.
In series wound motor torque is not directly proportional to current but proportional to I^2Eva said:
http://www.engineersedge.com/motors/dc_series_wound_motor.htm
Throttle potentiometer controls duty cycle directly.
Re: Current limits: Naah, where is all the exitement if FET doesnt stand for "Fire Emitting Transistor" anymore?
What reaally amazes me how well 24x MBR3045 schottky diodes are doing considering that we push them over the limits, they have to take 60A per leg with 50% duty cycle during accelerations. There could be all kinds of problems paralleing them and their negative tempcos, but no. With the two previous "not-so-optimal" layout power stages we blew 50 or so mosfets but never a schottky.
A system with torque proportional to i^2 and i increasing quite sloppy with duty cycle seems to ask for a more elaborate control scheme. I guess that your current approach results in quite jerky response as you press the gas pedal.
Also, are you doing the math of the schottkys right? I mean that 60*2*24=2880A !!
On the other hand, at 60A per leg Vd is around 0.9V so you would benefit a lot from synchonous rectification (with something having 0.004 ohms or so...) Also, you could remove completely the discontinuous-inductor-current to continuous-inductor-current (or even get a better implementation of regenerative braking) with properly controlled s.r.
BTW: What mosfets are you using?
Also, are you doing the math of the schottkys right? I mean that 60*2*24=2880A !!
On the other hand, at 60A per leg Vd is around 0.9V so you would benefit a lot from synchonous rectification (with something having 0.004 ohms or so...) Also, you could remove completely the discontinuous-inductor-current to continuous-inductor-current (or even get a better implementation of regenerative braking) with properly controlled s.r.
BTW: What mosfets are you using?
Actually throttle response is real nice, way better than any gasoline operated counterpart. Square or Log response seems appropriate, just like volume pot on amp.Eva said:
Yeap, that 2880A is pretty right. Only for first 5-10 meter of acceleration, one second or so.
Syncronous rectification would be nice during that one second, otherwise its not that helpfull in our case. (Full throttle 95% of the time)
Regerative braking is difficult with series-wound motor, with any easier motor type we would be using it already. For same reason we dont have reverse either. Reverse would be nice when you have underestimated your turnung radius and get stuck on the sidewall of the road.
Mosfets are FDP8441, but there are some better candidates rated for 100A or 120A from infineon and ST. Main criteria was next-day availlability
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.