Super Shielded Cable For EV Automotive Resolver??........

[Smokey]

I'm looking for a good cable for an electric vehicle motor that has a resolver.  There are always noise problems with this sort of thing, so I figured I would skip to the end and get a fully shielded cable (ideally with both foil and braid) and with independently shielded twisted pairs for the excitation, sine, and cosine windings.

But I'm a little surprised to not find more stuff either marketed directly for resolver use, or that meets the above shielding. 

Google turned up this, but I don't need a whole spool.  It would be great if this stuff was available somewhere by the foot.
https://www.tpcwire.com/products/wire-cable/control-cables-instrumentation/trex-onics-individually-shielded-resolver-cable

It sort of looks like CAT7 or CAT8 has the independently shielded pairs, but the outer shield looks like it's just foil.  I wonder if that would work?

What have you guys used for highly shielded automotive/industrial resolver applications??... or ... do you know of any material that's readily available that has all the above shielding?.... or... how else do you deal with resolver noise?

(None of the components are my design, so I can't add more filtering to the pcb or anything.  It's all off the shelf motor/controller).

[TERRA Operative]

My suggestion was going to be the types of cables you already found, but can you add ferrites to each end of the cable too?

[spostma]

I would try to run the existing cable through
 - corrugated tubing, or
 - a Tubular shielding braid (like https://tressimex.be/tubular-shielding-braids/ ) , or
 - if your grounding wire is braided, run it through that ground braid.

This will provide you with an extra shield layer.
Ground both ends of the outer shield firmly as you would do a RF ground connection.
Run the cable as close as possible over the chassis frame to minimize ground loops.

[jonpaul]

Auto, Av vehicle cables are readily available eg for CAN bus.


Knowing the resolver PHY TX, RX ands expected noise environment should lead to many cables, see Belden.com

Finally add shielding beads, CM and DM chokes and TAZ protection as needed.

j

[Smokey]

I would try to run the existing cable through
 - corrugated tubing, or
 - a Tubular shielding braid (like https://tressimex.be/tubular-shielding-braids/ ) , or
 - if your grounding wire is braided, run it through that ground braid.

This will provide you with an extra shield layer.
Ground both ends of the outer shield firmly as you would do a RF ground connection.
Run the cable as close as possible over the chassis frame to minimize ground loops.

Funny.  The motor manufacturer makes a BIG point about only grounding one end of the shield.  Grounding black magic. 

[jonpaul]

Hello:

The technique of EMI shiled and noise reduction is well known for decades, see SYNAUD COM and Noise Reduction by Henry OTT. Also Belden Cables.

In broadcast and audio we usually break the shield to chassis at one end of a cabole, eg floating RCA or BNC.

5 min of reading or researched would solve your questions.

j

[spostma]

Ground currents tend to take the the path of lowest resistance, which would be the metal frame.

IMO, the outer shield (braid) can carry injected common mode currents,
the inner shield would work as static shield, and could be grounded at the receiver side only.

But, just try out all grounding options, and see which combination gives the best immunity!

[Smokey]

...
But, just try out all grounding options, and see which combination gives the best immunity!

Yes, that would be optimal.  But in this case the resolver is deep inside an engine compartment of a vehicle, and the test condition is 1000A peak current output from the inverter (hooked to a 400V battery) which requires the car to be driving.  Instrumentation is a little hard when your whole lab bench is the back seat of a car moving quickly down the street closed track with professional drivers. 

[JeffK]

You mentioned "off the shelf motor/controller". I suspect the OEM of this equipment should have the recommended cable type, shielding arrangement and connection information already specified for satisfactory operation. But this may not be easily available for a one-off project from an high-volume automotive supplier.

I'd suggest starting with your signal of interest and defining the operating characteristics. If you don't have the specs on the resolver, https://www.tamagawa-seiki.com/ is a good place to review the basics of electrical signals produced by a resolver that is widely used in the industry. The excitation signal injected into the resolver should be in the range of 1 to 7vac rms, a load impedance of 70 to 130 ohms and a frequency range of 1khz to 10 khz. With the Sine/Cosine feedback voltage transformation ratio of 0.5. Or about 0.5 to 3.5 vac rms. Nothing here that requires cat6 or above performance.

The far more important interface is at the receive end where the resolver to digital conversion takes place. The ideal case is to lookup the specifications on the R/D conversion chip in your "off the shelf motor/controller". For example, Analog Devices has the AD2S80A (81A/82A/83A) series and the ancient AS2S90 series. Several good app notes are available regarding the interaction between the physical cable characteristics (impedance and cap./ft) and the effect on accuracy and signal/noise.

Tamagawa also offers some very good scope shots of magnetic and voltage noise interference examples in their user manual for the AU6805 R/D converter.


You can download it from here: https://www.tamagawa-seiki.com/products/resolver-synchro/smartcoder.html
   

Finally getting to the cable question: In my experience from the Industrial Automation design world, the environmental conditions and resolver connector termination is of prime consideration. Any decent braided shielded cable should be satisfactory. Finding a suitable cable that meets your temperature, chemical resistance, abrasion resistance, bend radius and termination requirements should be the work of a few minutes once you have the electrical interface and connector requirements defined.

For comparison, AMCI offers a H25 packaged resolver and a complete cable assembly on their website. Here are the resolver cables: https://www.amci.com/plc-automation-products/cables/

Nothing all that special as they appear to utilize Belden 9873. "Audio, Control & Instrument Cable, 18 AWG stranded (19x30) tinned copper conductors, polypropylene insulation, 3 twisted pairs, individually shielded with Beldfoil® (100% coverage), 20 AWG stranded tinned copper drain wire, PVC jacket."  Definitely not something that I could recommend for use in a harsh environment but gives you a starting point.