Electronics > Projects, Designs, and Technical Stuff

pinout puzzle on big BLDC drivers (repurposed automotive motors)

(1/4) > >>

mtraven:


got 2 of these motors from a surplus store, they told me they had tried to test one and let out some smoke, but for $10ea, I bought the pair.  from the remains, I can tell they connected 12v to the one plug(pigtail) that had been left with the motors.  I am reasonably certain this was a signal line that wants to see 0-5v (maybe -5 to 5), but def not 12v.  Polarity may also have been reversed, hard to tell at this point.  These were power steering motors from a 2007 land rover and I don't have a specific purpose in mind just yet.  They seem pretty beefy, I would like to get them spinning and see what I have.  I think these were setup kind of like a computer fan-give it 12v and a signal and the driver takes care of the rest.

My goal is to use the destroyed driver to workout the pinout for the good one and then design a BLDC driver for the bad one.  So in response I am looking for any info on the motor, how it was originally implemented and suggestions on where to begin on a new BLDC drive.

full photo set here:  http://"https://photos.app.goo.gl/fFFZnPuG3wyVfVfp7

After ripping the driver apart & spending countless hours researching, here is everything I know:

-I see one chip blown up.  I can only read the last few digits on that chip (????9072c3) a few digits of the manufacturer are : ???ANXDTG4  and the G is underlined.  I cant get anywhere in IDing this chip. 

-The other 64pin chip is an infineon-TLE7189QK (3 phase BLDC driver)

-The largest IC, presumably a microprocessor, is d70f3485gj(a2).  Not Found any info on that.

-power steering (full asm) part #:   80-30278OR






PINOUT:
-main power connection is obvious (giant spades)

-from the single row header(left), the cable was still with it & only had connections on 4 & 5 (contacts in blue)

-from the photos of this motor in its asm., it looks like the 2 row header also has just 2 connections. One going to each row of the header, reasonably sure pin 6 is the one on the inside row.  The second one looks like pin #2, but very well could be #1.


edit: fixed broken link

jhpadjustable:

--- Quote ---I think these were setup kind of like a computer fan-give it 12v and a signal and the driver takes care of the rest.
--- End quote ---

1. You need to look at the whole picture, and not ignore the small stuff. For example, that SOIC-8 at the center right is a TJA1051 CAN transceiver which strongly suggests that this motor is meant to be a node on a CAN network, where status information and possibly control will be exchanged.
2. Reverse-engineering "in these uncertain times" has a lot more to do with "open source intelligence gathering" as the spies call it. A bit of web searching seems to confirm that some Land Rovers do indeed have electrically assisted power steering controlled via the CAN bus, and that the head unit plays a role in translating steering angle into torque commands to these motors.

Anyway, you're going to need some sort of CAN interface to work the unmodified device. An "ELM327" OBD-II reader might do. You'll also need some software to view and generate CAN messages, which is a matter of taste and budget. Then you'll need to identify the higher-layer protocol in use, then try to figure out how to send messages to the motor controller to get the desired results of torque. Here's a free ebook that might help you get started: http://opengarages.org/handbook/ebook/

Or, you could dispense with the existing controllers entirely as needless complications and build your own BLDC controller. If you want results, and not a crash course in CAN (see what I did there?), you'll probably prefer this option.

mtraven:

--- Quote ---1. You need to look at the whole picture, and not ignore the small stuff
--- End quote ---
  my knowledge base is rather limited, so its more like I am blind to the small stuff...I could have starred at that board for hours & never figured that out.  thats why I reached out for help, and thank you for that, makes the decision to just start from scratch easy.  I had hoped it would be as simple as a couple of analog input voltages, oh well.  And if that board is that specific, I wonder if its not purpose build for the sort of back-and-forth action it would see steering a car vs the more continuous manor I would probably use it in.  There has to be a PID loop in there right?  at very least, I would think that would be tuned differently.  All things I would have control over in a custom driver...alright I am sold.

In looking at the OEM board, creating my own version is a bit intimidating, but something I have wanted to tackle for a while now.  I like the idea of having the logic separate from the power stage (it seems some IC's are both)...that may be the only option for a motor of this size anyways.  I have been playing with a samd21 dev board, that seems more than capable of timing and sensing and prototyping circuits is easy.    On the other hand, there is no shortage of purpose built BLDC driver chips, with all sorts of features...I don't even know where to start...and that pretty much requires a custom pcb.  which is fine once I come up with a final circuit, but I am going to want to play with/test  that circuit before having them printed.

then on the power size, how might I size the mosfets?  I have seen a method for setting up a virtual ground to measure the resistance of the coils then ohms law it from there?  and if I am building my own, if the motor will tolerate it, I would love to run that voltage up...maybe 24v? 48?


edit: a quick search showed breakout boards are available, so if a smt BLDC driver chip is the better option, I should be able to prototype on one of those boards just fine.

Are there any well documented guides or tutorials you would recommend on the subject?

jhpadjustable:
That's fair. Anyway, the moral of that story is don't forget to look up the little chips too, as they might hold critical clues to some aspect of the system under study. (Yes, sometimes even the little SOT23 as in the lower left might hold something important.) A couple of analog voltages would not have been the automotive way. Those guys like to see robustness against noise, local intelligence and diagnostics, and not running extra wires without a good reason. Silicon is cheaper than copper, especially by the hundred-thousand.


--- Quote ---I wonder if its not purpose build for the sort of back-and-forth action it would see steering a car vs the more continuous manor I would probably use it in.  There has to be a PID loop in there right?
--- End quote ---
Surely it is. Since they come in pairs, it might be reasonable to assume that each one pushes in its own single direction The lack of multiple power MOSFETs seems to confirm that the driver there only runs one way. Maybe that means you would be ill-advised to use it for propulsion, or should observe a duty cycle limitation and not try to go all the way to town and back on it. Or, maybe it'd be fine. Lots of possibly hidden mechanical details, which I don't claim as my wheelhouse (no pun intended).
The control is probably implemented by the big processor at lower left. There might be more to it than one simple PID loop, since road surfaces do vary. For that matter, the PID loop may be distributed over the network.


--- Quote ---I like the idea of having the logic separate from the power stage (it seems some IC's are both)...that may be the only option for a motor of this size anyways.
--- End quote ---
They sort of did. I speculate that inductor and MOSFET at the upper right is a constant-current driver they use to apply measured torque.


--- Quote ---then on the power size, how might I size the mosfets?
--- End quote ---
It appears they used a 55A, 40V device (upper right). A steering motor would need to maintain torque into a 0rpm condition, so I'd think they would size it generously. You might as well follow suit at minimum. MOSFETs are pretty cheap. But maybe it's not brushless after all, if there's only the one power driver transistor?


--- Quote ---I would love to run that voltage up...maybe 24v? 48?
--- End quote ---
You're on your own there. I'm more of an industrial networks guy who happens to sort of know a few dribs and drabs about motors. I can only say that, due to the presence of that buck converter-like arrangement in the corner, I wouldn't be 100% sure it's meant to run even at the full 12V for any great length of time. Again, motors aren't really my specialty, so hopefully someone else with more domain expertise can weigh in on that and on the finer points of drivers.

max_torque:
These EPAS motors are not that useful in traction or continuous operation because they are only rated for short term operation, for steering assist, which on a road car is a very part time duty cycle.  The driver is also fairly limited in terms of continous current.  The large mosfet will be used as a primary power switch, to enable the power to the motor to be shut off for fail safe no matter what happens to the 3 phase driver itself. As steering systems as "safety critical" they tend to include dual switching capability and often dul processors, with one operating as a equizzer / monitor to check the performance of the system is appropriate.

It will definitely be CAN controlled, but the majority of the functionality is within the embedded micro, and the assistance and inertia negation functionality is hard baked into that device.  EPAS racks typically include a column torque sensor which is an analogue or phase shift device to indicate to the control micro how much torque needs to be applies by the rack mounted motor.

The motors are pretty high performance and low inertia, but as mentioned have a poor thermal co-effiicient, but you may have more headroom in your application because these systems are validated to run in a 85 degC engine bay and still meet their performance targets.


In all cases, it will be far easier to just use a pre-existing universal 3 phase motor controller to drive the motor.  What you will need to check for is the motor position feedback signals to see how those are arranged.  Often these use a bipolar magnet on the end of the rotor shaft, and a small (typically SO-8) sensor ic on the pcb that reads the rotor position.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod