Paul & Sabrina's cheap DIY 144v motor controller

[mcudogs]

Earth shielding at one end is used to eliminate capacitive coupling into a low voltage circuit. The voltage induced into a circuit from a high current conductor is mainly transformer coupling. The only way to eliminate this is to create an opposing field source which cancels out the original. This is done by connecting both ends of the shield to earth. This creates an earth loop current in opposition to the original disturbing field. The value of the field is much smaller than the original but because it is much closer to the low voltage circuit the field tends to cancel out the main disturbing field.

We have been using this technique in telecommunications for years to reduce induced noise into telecommunications cables from power lines nearby. So not all earth loops are bad.

[squiggles]

Originally Posted by mcudogs

Earth shielding at one end is used to eliminate capacitive coupling into a low voltage circuit. The voltage induced into a circuit from a high current conductor is mainly transformer coupling. The only way to eliminate this is to create an opposing field source which cancels out the original. This is done by connecting both ends of the shield to earth. This creates an earth loop current in opposition to the original disturbing field. The value of the field is much smaller than the original but because it is much closer to the low voltage circuit the field tends to cancel out the main disturbing field.

We have been using this technique in telecommunications for years to reduce induced noise into telecommunications cables from power lines nearby. So not all earth loops are bad.

This could explain why Australian Telecommunication is so poor wherever it is on copper . Full of earth loops, particularly after rain. I would go as far as to say most of the induced noise into telecommunications cables in this country has very little to do with the proximity of power cables and a hell of a lot to do with earth loops often the result of water ingress. It is certainly a common occurrence in country areas.

In the 1980's much research was done into induced noise in sensor circuitry as the popularity of computerised SCADA systems increased. In particular the installation in electrical substations where we experienced substantial issues with induced noise for fairly obvious reasons. One thing we learned is the noise was worst when shielded cables were earthed at both ends, in some extreme cases (long cable runs) it was worse than unshielded cables.
The best solution was twisted individually screened pairs earthed at the computer terminals.
There is another bad thing to do as well (probably not significant in the EV but possibly in telecoms), don't leave unused wires in the same cable unterminated, they work as damn good antenna and amplify the noise level nicely.

Also important that the earth bonding is at the computer end, if it is done the other way nasty little voltages can appear at the terminals and blow things up, can be a little embarrassing.

[mcudogs]

Hey Squiggles, the earthing of telecommunications cable screens at both ends is not used on a regular basis. It is mainly used in AC rail traction environments and is very expensive to implement. The cables have aluminium sheaths and the sheaths are earthed at 1km intervals. I was involved extensively in the testing of the effectiveness of this screening and I can assure you that it works well.

I would not even consider using it in an environment like a sub station where there are significant earth currents running through the structure. You would be running part of the disturbing current through the shield and as you found out the effect on the interference would be to increase it.

However, in an EV the traction battery is generally only earthed to the chassis at one point and there should therefore be minimal earth paths for the traction current.

The design of electrical noise reduction in systems has always been a hotly debated topic and due to the complexity of the issue it is impossible to come up with a one approach fits all scenario. Let's just say that if you have an interference problem you should firstly shield the cable and earth it at one end only. If the problem still exists try earthing the shield at both ends, but only if there are no earth return currents going through the chassis.

BTW I agree completely about the state of the telecommunications system in Australia. I think you will find the majority of the issues arise from poor maintenance by our favourite Telecomms company. TXXSTRA

[squiggles]

QRail?

[mcudogs]

Picked it in one.

[NiHaoMike]

For those making water cooled controllers, instead of using an inverter with a fountain pump, what about a bilge pump with a simple PWM controller?

[mcudogs]

Originally Posted by NiHaoMike

For those making water cooled controllers, instead of using an inverter with a fountain pump, what about a bilge pump with a simple PWM controller?

We could use one of the spare outputs of the ATMEGA to operate another Mosfet. I don't think it would need to be PWM'd, just on or off when the temperature rises above a certain value. How much current do they pull?

The Mosfet could also be used to drive a small fan to cool the heatsink on non water cooled versions.

[NiHaoMike]

The small one I have (for a portable sink that I'm going to build) is rated at 1.8A maximum at 12v. It will run at lower voltages and will probably last longer as well. I have heard about overclockers running cheap bilge pumps at 5v for water cooling computers.

[Renny_D]

Now that this thread is 217 pages it would be nice to have the first page updated with where to find the final schematics and parts lists if there are ones. I'd love to give this a shot. Thanks for the hard work and for sharing it with us.

Renny

[MPaulHolmes]

good idea!

Right now, I'm doing a test run of the wiki. I'm ordering everything on there, and am going to take pictures of every step of the process. Then the wiki will need to be updated based on any errors. It needs a test run of a full controller. The assembly process has taken several practice runs. That's one reason Ben's controller is taking so long to be sent out. I need it as a model for the cleaned up pictures and video process of the next controller.

For now, check out:
ecomodder.com/wiki

It's not really complete, though.

I need to catalog screw sizes, drill locations on heat spreader, miscellaneous parts, etc...