Question about active sin/cos (resolver) to hall-effect conversion module

[absence209]

Hey guys!

Quick question about my current setup.

I currently have a motor that has 5 poles and outputs a sin/cos resolver signal using an RLS RMB29 module. For the controller, I have a far driver 72v model 72260 controller, which runs off of 3 hall effect signals.

For the sake of money and my mental well-being, I decided to stay away from using a sevcon for this (even though that's the appropriate controller) because of the 1000$ programmer and the rigorous tuning.

So for now, I've observed two possible options for getting this setup working.

The first is to use a circuit that actively converts the sin/cos signal into a 6-step commutation signal that repeats 5 times per revolution.

The second option is to replace the RLS RMB29 chip with an RMB29 variant that outputs a 5-pole commutation signal (no hacky workarounds required, everything is done in the RLS chip itself).

Now since both of these options lead to the same result (a replica of what would be 5-pole hall effect signals) would they work at all?

My main concern is that real hall-effect sensors provide angle data along with exact magnet position relative to the stator, whereas a sin/cos encoder only provides a rotor position (albeit a very precise one). Does the controller need this "extra" information in order to rotate properly? If so, how can get this setup to work with this? Has anyone here done this before? And lastly, if the resolver does somehow line up exactly with the magnets on the rotor (by chance) does that mean this setup would then be identical to what a typical hall-effect sensor would output?

Any help is appreciated! Thanks guys.

 

[mxlemming]

Hey guys!

Quick question about my current setup.

I currently have a motor that has 5 poles and outputs a sin/cos resolver signal using an RLS RMB29 module. For the controller, I have a far driver 72v model 72260 controller, which runs off of 3 hall effect signals.

For the sake of money and my mental well-being, I decided to stay away from using a sevcon for this (even though that's the appropriate controller) because of the 1000$ programmer and the rigorous tuning.

So for now, I've observed two possible options for getting this setup working.

The first is to use a circuit that actively converts the sin/cos signal into a 6-step commutation signal that repeats 5 times per revolution.

The second option is to replace the RLS RMB29 chip with an RMB29 variant that outputs a 5-pole commutation signal (no hacky workarounds required, everything is done in the RLS chip itself).

Now since both of these options lead to the same result (a replica of what would be 5-pole hall effect signals) would they work at all?

My main concern is that real hall-effect sensors provide angle data along with exact magnet position relative to the stator, whereas a sin/cos encoder only provides a rotor position (albeit a very precise one). Does the controller need this "extra" information in order to rotate properly? If so, how can get this setup to work with this? Has anyone here done this before? And lastly, if the resolver does somehow line up exactly with the magnets on the rotor (by chance) does that mean this setup would then be identical to what a typical hall-effect sensor would output?

Any help is appreciated! Thanks guys.

A TLE5012B E3005 soldered to a soic8 breakout board with a 100nF decoupling capacitor on VCC to ground would give you 5pp hall emulation.

And if you got the TLE5012B E1000 version you'd get an ABI encoder for your efforts.

 

[absence209]

and just to confirm that hall emulation would work right?

 

[mxlemming]

and just to confirm that hall emulation would work right?

Errrr

Isn't that what I said in the first line? Just pick the version of the chip according to what you need.

You still have to work out the right orientation if your controller doesn't auto correct it.

I mentioned encoder because fardriver have encoder options and if it falls within the right pulse per second range then encoder is better.