Electrinic Load

[smarbaga]

Hello All
I am about to put together an eletronic load.
This load will primarily be for testing my new
electric car motor controller design.
I would like the load to be variable.
my car motor controller is fed by 20,
80 amp, 12 volt gell cell batteries.
hooked up in series, parallel (120+ vdc).
a zillion short circuit amps, but 4, 100 amp fuses.
I would like to test this controller up to 350 amps.
I have 15 of these FUJI EVK31-050 IGBT's
these IGBT's being from Fanuc motor drives.
I was wondering if anyone out there would have any idea
how i can use both sides if these transistor modules.
perhaps to get double ( or so) the current out of them.
attached is a pdf of there layout ( specs) .

ps: when i get further along with this eload thingy, I will post
pics, schematics and software,
as it will be PIC controlled, with data logging, etc.
thanks for your help
reguards garry

[vger]

The EVK31-050's have the collector of the lower transistor and the emitter of the upper one connected together internally. So seperating them to connect them in parallel for double the current will probably not be possible. Those things were designed to be switched on and off, and not controlled in an analog fashion. If you have 120 V at 350 Amps you need to disipate 42,000 watts (35 1200 watt hair dryers worth). I seriously doubt those packages would handle that kind of heat production. You might be able to use half sections of the modules to switch a multiple fixed resistor load bank. Looks like the single transistor power disipation is 300 watts (absolute max) so you would need 140 of them to vary the load in an analog fashion. They can handle 50 amps each (absolute max) so to be safe, say 40 amps. If you build up some resistors (electric stove heating elements maybe) and switch them with the transistors, you could go in steps up to your required 350 amps. Suppose you have 6" elements that are 1500 watt (at 240 V) and you parallel 4 of them to give you 3000 watts at 120 V. Then you could go up to 350 amps in 14 steps of 25 Amps each. This would require 14 of the EVK31's for switching.

Steve

[smarbaga]

thanks for your reply.
i do have 5 of those heatsink blocks.
for a total of 15, evk31 moduels.
i guess it is a lot of power to get rid of.
i don't know of any other way of variably
bench testing a motor controller up to this
much current, or say even 200 amps.
i can not load the motor down, it is just to
powerfull.
i guess i will start out with one evk31 block
and see what i come up with.
i made a battery load tester in 1988 for black & decker
for there battery lawnmowers.
i think it was 60 irfz44's with op amps driving each of them.
there design engineer had to submerse the heat sinks in a cattle
drinking water trough and reflow cold water into it to keep the fets
cool.
when i went to there lab and he showed me the setup, it blew the
sub pannel right off the wall, lol, good thing it was a concrete wall.
if there are any other ideas or sugestions,
they would be greatly appreciated.
thanks, garry

[vger]

Garry,

If you want to use the motor as a test load, you could couple it to a hydraulic pump in a closed loop with a flow valve as a load control. For current monitoring just use a shunt...
http://www.deltecco.com/WB.html
and a simple meter. Or you could just get a dynomometer of the type used for engine testing.

Steve

[wildwestpat]

Hi Garry

Sorry did not see that you are asking about load switching not motor control switching which makes some of the points below redundant. Suggest you use an generator as the load as you may be able to use a 'spare' traction motor. No matter if it is hydraulic or electric the power will turn up as heat that has to be ditched. Hope you like hot baths. I have used a large industrial snail fan as a load in the past but noise was a problem.

There are several basic electronics problems in what you are thinking about parallel operation of discrete semiconductors.

1. You can not connect semiconductors in parallel to double the current capability. The safe working area graphs indicate that the devices you have are relatively low current capability. The voltage and current must remain in the lower quadrent of this graph under ALL conditions.

2. The load is inductive. This means that as the transistor starts to turn off it will be subject to an increasing voltage as the flux in the inductive winding produces a voltage as the flux collapses. This voltage will be many times higher than the battery voltage. This switching spike has to be captured and clamped so that it does not breakdown the transistor junction.

3. As the transistor turns off the instantaneous current and voltage will produce power and this produces heat in the junction of the transistor. Heat can not be transfered instantaneously to the external heat sink. The speed of heat transfer is known as secondary breakdown and is responsible for melting the semiconductor junction and is described in the data sheet on the graph headed secondary breakdown. You will soon need a good osciloscope with a current monitoring proble so that you can plot the switching voltage current curve. The first elbow region is the critical one here.

4. To operate the transistors in parallel it is necessary that they are matched for their dynamic current voltage characteristics to a high degree of precision to ensure even power sharing. Even transistors from the same parent wafer will not be sufficiently well matched. This is why the electronic control gets complicated.

5. For an electric vehicle the braking power complicates the electronics if the power is to be returned to the battery.

6. Bridge operation may be an option as this will double the power by increasing the voltage accross the motor.

7. To help save expensive accidents I suggest you make your own fuses as follows:- Use normal fuse wire but imerse it in water - a jar with a plastic lid and a pair of brass bolts used to terminate the test rig battery whilst ensuring that the fuse wire is fully in the water. The wire fuse wire heats up and the differential between carrying current and rupturing is a lot sharper than for a conventional fuse. This can save a lot of grief as the water cased fuse blows much faster than its air cousins. However you need to do some calibration to get sensible fusing currents.

8. You may care to use voltage accross the switching semiconductor as a control input element for the logic that drives the power control. This can help by preventing the next stage of the switching action starting befor the previous one has completed. The power transistors will not turn off quickly from the saturated on state.

Hope this helps even if it makes gloomy reading much of it gained by kiling very expensive transistors. When you have a dead one it is worth an attempt to try and inspect the base emitter junction of the output transistor. Paint stripper can help. The difference between a voltage kill and a power kill are then very obvious even to the naked eye. The power kill will have actually melted the semiconductor and showes as a crater. A voltage kill shows a a mark on part of the base emitter comb. It is useful to differentiate between the two types of death as the solutions lie in different circuit tweeks.

Regards

Pat

[wildwestpat]

Hi Garry

Just a further point you have 20 x 12 volt 80 amp batteries. These are of the gell type so the average discharge current would be 8 amps or 2 kilowatts as a conservative estimate.

If I am reading you right this means that you need to do some careful calculations on the absolute peak power you can dump into your load. Those cells can supply an awsome current but not for long! Suggest while you are in pic mode you might like to use cell temperature as a monitored parameter.

Good luck - Pat

[smarbaga]

thank you all for your advice.
i think i will just go to the junk yard and get 20
or so stove elements and wire them together as needed
on a bus bar.
perhaps suspending them on flat metal bars.
maybe even using them as shunt resistors, in the future.
i think i can get them for a $1. each.
this would be a quick, cheap and dirty fix.
i guess there would be a bit of heat.
i also think i will have to be aware of fire and
a deadly voltage.
maybe i can find some real shunts at the junk place as well.

[Pandinus]

Water heater elements might be safer If they are mounted in a tank of some sort.

[wildwestpat]

Hi Garry

Just a point about safety since you mentioned it. A shock from the DC supply will cause your muscles to lock unlike AC which causes a muscle spasm which will throw you clear of the accidental contact.

Stay safe regards Pat

[LZ1TWB]

Hi,

I think it's not right to test a motor controller with a resistive load, as the motor has some significant inductance and will affect the signal form. So the load should have the same electrical specifications as the motor if you have to be correct.

Anyway you say you cannot load your motor. Why not just put it in your car, bring the car's wheels in the air, drive the motor and press the brakes?? It is just as simple and dumb. Of course, you will sacrifice some break life, but whatever. The testing cannot get more real than that.

Todor

[smarbaga]

thanks for the ac/dc shock headsup.
i actually thought that a dc shock causes your muscles to contract,
making it impossible to let go of a wire if you are holding it.
reguardless, its the silect killer, if one is not always aware.
i will put sone heater elements on he controller.
if the controller can hold around 100 amp for a minuit or so
without getting to hot, i will put it in the car.
the inductive load thing is a good point.
so i don't want to spend a lot of time on a electronic load.
( more time than on the actual controller. )
a small e_load may be handy to have around, say 10 amps.
ps:
i have a .pdf schemaic but when i go to add the pdf the upload says
"upload failed" same with the zipped version.
???

[wildwestpat]

Hi Garry

The effect of the DC shock depends of course on which nerve paths the electricity tracks and hence which muscles are caused to contract the net effect is that you don't get thrown free as there is no violent muscle spasm as with Ac that can leave you wimpering against the opposite wall! I have had one or two near misses - in one I unwould a high voltage capacitor by holding one foil in one hand and the other in the opposite hand and pulled them appart to investigate the insulating film for tracking. Even with the capacitor shorted out with the dump stick there was enough residual DC charge to lock my arms in the extended position until a work mate saw what had happened and knocked the foil to the ground breaking the circuit.

On a happier note just been down the gym and saw the rowing machines. These use paddles in an enclosure that has ribs to stop the paddles rotating the water too fast. They appear to absorb a lot of energy for a long time. May be a quick mod to a washing machine might make a good dynamic load when you get to testing with the motor.

Regards

Pat