Clutch - leave in or not?

Have you made or bought a converted vehicle if so this is for you
User avatar
EVguru
Posts: 305
Joined: Fri May 25, 2007 5:17 pm
Location: Luton
Contact:

Postby EVguru » Mon Aug 18, 2008 3:51 pm

I've seen more than one Series motor blown up due to popping out of gear, which is more likely when changing gear without a clutch. The added roatational inertia of a flywheel/clutch helps if this happens giving you just enough time to get off the throttle. A rev limiter would help protect the motor in these circumstances. It would also help with higher voltage vehicle where it is possible to over-rev in gear.

I had a stuck throttle on the Scirocco because of a damaged liner in the original cable. There was no way the brakes would hold the motor back and my normal instinct was to clutch out. This stopped the vehicle accelerating, but now the motor (the commutator at least) was in danger of bursting. The inertia of the clutch and flywheel gave me enough time to feather the clutch to control the motor speed whilst still being able to hold the car on the brakes. I was then able to go for the ignition switch.

A clutch also made control at low speeds much easier. It's difficult to get really fine control on a high voltage system an I would use a little slip when parking.

The aditional weight was fairly small compared to the weight of a series motor, but the standard clutch did struggle at times. I could get the clutch to slip accelerating hard in third gear.

As for loads on the bearing, a typical motor output bearing would be something like a 6006 deep grove ball race; 30mm I/D, 55mm O/D and 13mm wide. It's rated for a dynamic radial load of 11,900 Newtons @ 100,000 hour life expectancy. The axial rating would be about 25% of that.
Paul

http://www.compton.vispa.com/scirocco/
http://www.morini-mania.co.uk
http://www.compton.vispa.com/the_named

GregsGarage
Posts: 870
Joined: Tue Apr 01, 2008 5:27 pm
Location: Galashiels, Scottish Borders
Contact:

Postby GregsGarage » Mon Aug 18, 2008 3:57 pm

Pulling away from 3rd feels like my batteries are flat, but it will do it. The controller is a Curtis 1231C 144v 500amp. I can't tell you the exact current draw, depends on gradient, number of passengers etc. I will say that a ratio between 2nd and 3rd would probably suit my needs up to around 50-55mph, although I have been known to use 1st gear to get up a quite steep hill to my drive with the car loaded with 2 adults, 3 children and low batteries. :shock:

I believe you are going to use 2 small motors so you can always series/parallel shift them. The trick is going to be choosing the right ratio, figuring out how fast the motors can spin and what your speed requirements are.
Greg Fordyce

Daewoo Matiz
http://www.evalbum.com/4191

MalcolmB
Posts: 423
Joined: Sat Jun 23, 2007 8:07 pm

Postby MalcolmB » Mon Aug 18, 2008 5:00 pm

Thanks Greg. Doing sums is all well and good but nothing beats first-hand experience. Still wavering.

EVguru wrote:A clutch also made control at low speeds much easier. It's difficult to get really fine control on a high voltage system and I would use a little slip when parking

That's a good point. A clutch could also be useful on a fixed ratio system, although the added complexity runs counter to the KISS approach. User-friendliness is an important factor for me though.

User avatar
ChrisB
Posts: 4657
Joined: Fri Jul 14, 2006 11:33 am
Location: Hampshire on the Southcoast
Contact:

Postby ChrisB » Mon Aug 18, 2008 6:21 pm

EVguru wrote:A clutch also made control at low speeds much easier. It's difficult to get really fine control on a high voltage system an I would use a little slip when parking.........


Yep I'd add that the fiesta would have been easier to park etc if it had a clutch, you did end up having to be VERY gentle on the throttle during parking

ChrisB
I reject reality and substitute my own !!!!!!

User avatar
EVguru
Posts: 305
Joined: Fri May 25, 2007 5:17 pm
Location: Luton
Contact:

Postby EVguru » Mon Aug 18, 2008 6:30 pm

The difficulty with high voltage systems (above 144 volt) is of getting a suficciently PWM short pulses whilst making sure you turn the power device fully on to minimise heat generation. You start having to do other tricks like a lower switching frequency (Curtis 1221/31 - C) or random pulse skipping.

With lower voltage systems it's more likely to simply be throttle response. Having a 'low speed' switch that shunts the throttle input with a fixed resistance, so the pedal becomes say 0-1K full travel, would make the throttle less sensitive and would probably be a good idea for reverse anyway.
Paul

http://www.compton.vispa.com/scirocco/
http://www.morini-mania.co.uk
http://www.compton.vispa.com/the_named

User avatar
geekygrilli
Posts: 325
Joined: Sun Jun 24, 2007 6:56 pm
Location: London
Contact:

Postby geekygrilli » Mon Aug 18, 2008 8:04 pm

For what its worth I think I will leave the clutch in the next build so i have a clear comparison.

There is another cinquecento on gumtree with a popped engine for £160...looks tempting

User avatar
Jeremy
Posts: 472
Joined: Thu Jul 31, 2008 6:35 pm
Location: Salisbury

Postby Jeremy » Mon Aug 18, 2008 8:08 pm

EVguru wrote:The difficulty with high voltage systems (above 144 volt) is of getting a suficciently PWM short pulses whilst making sure you turn the power device fully on to minimise heat generation. You start having to do other tricks like a lower switching frequency (Curtis 1221/31 - C) or random pulse skipping.


I'm not so sure that this is true now we have exceedingly fast, reasonably high voltage FETs and almost as fast, very high voltage IGBTs, is it?

A fast, low Rdson, 180A max, 75A average, 100V FET, like the IRFB4110, will turn on in around 25nS and off in around 80nS, so can be used with very short pulse widths even at switching frequencies that are compatible with a low inductance PM motor, like one of Cedric's designs. Assuming a 15kHz switching frequency and a minimum practical pulse width of around 110nS gives this device the ability to take the power down to about 0.18% of full power, which should be fine for good low speed control I would have thought. The higher voltage devices in the same series are just as fast, but have a higher Rdson (the device above is one I have a datasheet on, as I acquired a small box of them for my new controller).

High voltage IGBTs are a bit worse, with turn on times of around 50nS and turn off times of around 400nS, but will still allow pulse widths down to the equivalent of about 0.7% of full power, which again I would have thought was adequate for good low speed control.

The Curtis, with it's noisy low speed, 1.5kHz, switch frequency mode of operation was designed way back when power device switching speeds were very much slower. IIRC it uses some pretty ancient, slow, high Rdson, low max current, IR HEXFETs in a massively parallel configuration, which means that the poor old FET gate driver is having to try and dump current quickly into an extremely high combined gate capacitance to turn them on, hence the switching speed problem. They got around this by switching the PWM frequency down for low duty cycle operation, but this is a bodge that isn't needed with the more capable devices available today.

There's no real need to go much above 15kHz or so based on some modelling of motor current ripple I've done - even the lowest inductance motors that are currently available can smooth the motor current to an acceptable level at that frequency. Series wound motors, with their much greater inductance, would probably be happy at lower switching frequencies, although anything much lower might start to get obviously audible (but perhaps not to ageing aviators like myself who have already lost a fair bit of their ability to hear anything over 10kHz or so!).

Jeremy

MalcolmB
Posts: 423
Joined: Sat Jun 23, 2007 8:07 pm

Postby MalcolmB » Mon Aug 18, 2008 8:32 pm

Most of the electronics goes over my head, but from what I've gathered most of the controllers used in road-going EVs are still fairly primitive compared to say the RC brushless controllers. Which is the most sophisticated in terms of component use and control systems?

Regarding the sensitivity of throttles, or lack of it, are they generally based on linear or logarithmic potentiometers? I definitely find the throttle response of my Magura throttle very 'twitchy' at low speed even with the throttle ramp set to the most progressive setting.

User avatar
EVguru
Posts: 305
Joined: Fri May 25, 2007 5:17 pm
Location: Luton
Contact:

Postby EVguru » Mon Aug 18, 2008 8:33 pm

A fast, low Rdson, 180A max, 75A average, 100V FET, like the IRFB4110


Giving the specs for a 100 volt FET kind of misses the point about high voltage systems.

Turn on times are usually limited by the reverse recovery current of the rectifiers (unless running synchronous rectification) and turn off times by the parasitic inductance.

1000A/us = 10 volt per nH

Curtis didn't drive their FETs (IRF640 or BUZ31) anywhere near as hard as they could have, becuase of the difficulties they would have had complying with EMC regs.
Paul

http://www.compton.vispa.com/scirocco/
http://www.morini-mania.co.uk
http://www.compton.vispa.com/the_named

User avatar
Jeremy
Posts: 472
Joined: Thu Jul 31, 2008 6:35 pm
Location: Salisbury

Postby Jeremy » Mon Aug 18, 2008 9:16 pm

I only quoted that device because I happened to have the datasheet for it in front of me. I've just checked, and just as I thought, the higher voltage members of the same family have similar fast switching times, so my hypothesis still holds true. Anyway, the IGBT times I gave were for a 500V device!

Fast recovery diodes are readily available that will switch pretty much as fast as is needed, but a decent controller would use synchronous rectification anyway, with devices that switch at the same speeds as the main switches, so this is a moot point. All that's needed is a small dead band of a few nS to prevent shoot-through.

Other controllers manage to drive their output devices quickly and still comply with EMC requirements, most of which can be met with decent layout and close coupling of the main capacitors.

I had a feeling that the Curtis used those old IRF640s (about 35 of them I think, all in parallel), a device that IR released at least 25 years ago, as it's in my original copy of the HEXFET data book from 1982. Not exactly cutting edge technology................

Their big problem was as I've already said, trying to drive all that gate capacitance quickly. They may well have used slow diodes as well which added to the problem, as you've said. Either way, there's no reason why a modern controller has to be so limited. Control down to less than 1% seems possible as far as I can see.

On the topic of throttle linearity it seems that most are roughly linear. Whether this is ideal or not I'm not sure. I played about with altering the throttle curve on my electric bike throttle (using a small microcontroller between the throttle and controller) but found that it didn't make a startling difference to the way the bike felt to ride, even with some fairly extreme control laws. I'm tempted to try motor current control, rather than straight power control, if only to get a feel for what torque control feels like. There's something that seems more intuitive about using torque control for the throttle, but I could be completely wrong.

Jeremy


Return to “Conversions”

Who is online

Users browsing this forum: No registered users and 8 guests