R/C Tech Forums

Go Back   R/C Tech Forums > General Forums > Electric On-Road

Like Tree12Likes
Reply
 
Thread Tools Search this Thread
Old 12-31-2017, 04:27 AM   #1
Tech Adept
 
Flygio's Avatar
 
Join Date: Dec 2010
Posts: 146
Default Imprecision on measuring sensor board deviation

What do you use for measuring sensor board deviation? I purchased a skyrc motor analyser just to discover that it has a 4% accuracy... It's pointless to look for matching sensor when you have that type of error... Means you could have +/-2 shift at 50... My sensor board could be 50-50-50 and reading 48-50-52 ...
Is there a more effective way for measuring sensors?
Flygio is offline   Reply With Quote
Old 12-31-2017, 07:23 AM   #2
Tech Adept
 
Join Date: Jan 2013
Posts: 175
Trader Rating: 3 (100%+)
Default

There is the motolyser, but I found looking for a very good sensor board a waste of time. To many variables that throw off the readings.
eazy70 is offline   Reply With Quote
Old 12-31-2017, 09:06 AM   #3
Tech Elite
 
howardcano's Avatar
 
Join Date: Jan 2012
Location: Olathe, KS
Posts: 3,656
Trader Rating: 36 (100%+)
Default

I've owned both of these devices. The more expensive one seems more repeatable, but it has absolutely no specifications I can find for accuracy in either the manual or on the web site.
__________________
Howard Cano
When race results are re-calculated using the IOF (Index Of Fun), I always win.
1993 ROAR 1/8 Pan National Champion
howardcano is offline   Reply With Quote
Old 12-31-2017, 09:28 AM   #4
Tech Fanatic
 
Airwave's Avatar
 
Join Date: Jul 2007
Posts: 824
Trader Rating: 1 (100%+)
Default

Are you aware that the precision range of the sensor itself is way worse than that?
Airwave is offline   Reply With Quote
Old 12-31-2017, 12:29 PM   #5
Tech Adept
 
Flygio's Avatar
 
Join Date: Dec 2010
Posts: 146
Default

Quote:
Originally Posted by Airwave View Post
Are you aware that the precision range of the sensor itself is way worse than that?
Really? The sensor itself or the electronic reading it?
Flygio is offline   Reply With Quote
Old 12-31-2017, 07:28 PM   #6
Tech Master
 
tbrymer's Avatar
 
Join Date: Feb 2011
Location: Arizona
Posts: 1,029
Trader Rating: 9 (100%+)
Default

I have a skyrc motor analyzer and it is completely repeatable. However, not sure what the accuracy is.
glennhl likes this.
__________________
CRC - C&C Motorsports - SpeedZone
tbrymer is offline   Reply With Quote
Old 01-01-2018, 12:35 PM   #7
Tech Fanatic
 
Airwave's Avatar
 
Join Date: Jul 2007
Posts: 824
Trader Rating: 1 (100%+)
Default

Pretty sure that the hall sensor cannot locate the position within a 15 range at least...
Airwave is offline   Reply With Quote
Old 01-01-2018, 02:21 PM   #8
Tech Elite
 
howardcano's Avatar
 
Join Date: Jan 2012
Location: Olathe, KS
Posts: 3,656
Trader Rating: 36 (100%+)
Default

Quote:
Originally Posted by Airwave View Post
Pretty sure that the hall sensor cannot locate the position within a 15 range at least...
Why? Do you mean accuracy of the sensor in locating the field, or deviation of the field position in relation to the mechanical angle of the rotor?
__________________
Howard Cano
When race results are re-calculated using the IOF (Index Of Fun), I always win.
1993 ROAR 1/8 Pan National Champion
howardcano is offline   Reply With Quote
Old 01-01-2018, 02:36 PM   #9
Tech Adept
 
two shoes's Avatar
R/C Tech Elite Subscriber
 
Join Date: Dec 2015
Location: Rockies
Posts: 113
Trader Rating: 15 (100%+)
Default

Sensors have about a 1% accuracy so for 360* would be 3.6* or +/- 1.8*ish...

Good article here: Hall Effect

If you look at the sensor on the board and can read the numbers, you can look up the datasheet online (Mouser, DigiKey...) DATASHEET

It appears circuit design, magnetic thresholds and distance all play a part.
Nerobro likes this.
two shoes is offline   Reply With Quote
Old 01-01-2018, 02:56 PM   #10
Tech Master
 
gigaplex's Avatar
 
Join Date: Dec 2015
Location: Melbourne, VIC
Posts: 1,659
Trader Rating: 2 (100%+)
Default

Quote:
Originally Posted by two shoes View Post
Sensors have about a 1% accuracy so for 360* would be 3.6* or +/- 1.8*ish...

Good article here: Hall Effect

If you look at the sensor on the board and can read the numbers, you can look up the datasheet online (Mouser, DigiKey...) DATASHEET

It appears circuit design, magnetic thresholds and distance all play a part.
You're misunderstanding what the accuracy in the data sheets refers to. Hall sensors measure a voltage induced by a magnetic field. Their accuracy rating refers to the magnitude of that induced voltage. That's quite different to the accuracy of timing measurements in a dynamic environment.
Nerobro likes this.
gigaplex is offline   Reply With Quote
Old 01-01-2018, 04:11 PM   #11
Tech Adept
 
two shoes's Avatar
R/C Tech Elite Subscriber
 
Join Date: Dec 2015
Location: Rockies
Posts: 113
Trader Rating: 15 (100%+)
Default

Quote:
Originally Posted by gigaplex View Post
You're misunderstanding what the accuracy in the data sheets refers to. Hall sensors measure a voltage induced by a magnetic field. Their accuracy rating refers to the magnitude of that induced voltage. That's quite different to the accuracy of timing measurements in a dynamic environment.
Please enlighten me as to the accuracy and how to determine it (real question, not being snarky). The RVDTs and LVDTs I deal with on a daily basis have an accuracy of 0.2% repeatable. These range from 90* rotational to 4"-40" straight line for position, not a very "dynamic" environment.

My basic understanding is that these are used as switches, once a voltage threshold has been crossed, it switches state, indicating pole swap. The timing is determined by how far ahead or behind each phase fires in relationship to where the sensor picks up the position of the pole... Where/how is accuracy lost or gained? Can you saturate the sensor at a high enough speed?
two shoes is offline   Reply With Quote
Old 01-01-2018, 07:33 PM   #12
Tech Master
 
gigaplex's Avatar
 
Join Date: Dec 2015
Location: Melbourne, VIC
Posts: 1,659
Trader Rating: 2 (100%+)
Default

Quote:
Originally Posted by two shoes View Post
Please enlighten me as to the accuracy and how to determine it (real question, not being snarky). The RVDTs and LVDTs I deal with on a daily basis have an accuracy of 0.2% repeatable. These range from 90* rotational to 4"-40" straight line for position, not a very "dynamic" environment.

My basic understanding is that these are used as switches, once a voltage threshold has been crossed, it switches state, indicating pole swap. The timing is determined by how far ahead or behind each phase fires in relationship to where the sensor picks up the position of the pole... Where/how is accuracy lost or gained? Can you saturate the sensor at a high enough speed?
Well to take the data sheet you linked to, it has two different options for sample rate. The fastest is 20Hz. When measuring RPMs in the 20k range (a little over 300Hz), you're not going to get any remotely useful data to determine the position of the rotor.

To get a precision of 1% positional accuracy, then you need to sample at a frequency at least 100 times higher than the highest RPM the motor will reach. This also ignores any effects of latency of sampling, or other sources of inaccuracy.
Nerobro likes this.
gigaplex is offline   Reply With Quote
Old 01-01-2018, 07:51 PM   #13
Tech Addict
 
Jorge T's Avatar
 
Join Date: Mar 2011
Location: At a chemical plant in TX
Posts: 555
Trader Rating: 13 (100%+)
Default

Sensor selection for brushless motors
Attached Files
File Type: pdf Selecting-Hall-Effect-for-DC-Brushless-Motors.pdf (379.2 KB, 55 views)
Nerobro likes this.
__________________
RSD
Jorge T is offline   Reply With Quote
Old 01-01-2018, 08:24 PM   #14
Tech Adept
 
two shoes's Avatar
R/C Tech Elite Subscriber
 
Join Date: Dec 2015
Location: Rockies
Posts: 113
Trader Rating: 15 (100%+)
Default

Anyone have a PN for one used in an RC motor? I do not have a PN visible on the sensor boards I have laying around.
two shoes is offline   Reply With Quote
Old 01-01-2018, 09:49 PM   #15
Tech Addict
 
Join Date: Sep 2010
Location: Chicagoland
Posts: 544
Send a message via AIM to Nerobro
Default

Quote:
Originally Posted by two shoes View Post
Sensors have about a 1% accuracy so for 360* would be 3.6* or +/- 1.8*ish...

Good article here: Hall Effect

If you look at the sensor on the board and can read the numbers, you can look up the datasheet online (Mouser, DigiKey...) DATASHEET

It appears circuit design, magnetic thresholds and distance all play a part.
Yey, data sheets, boo for misunderstanding how they relate to this application.

Though I"m not sitting here and doing a long read of that data sheet, that particular hall sensor is "somewhat atypical" of what you'll find in a r/c car. That one has some logic built in to it, so it's outputs are in the "hz" range, when we really need tens of khz.

That hall effect sensor would not be useful in a r/c car motor.

Also, it's important, that you're relating degrees of rotor position to the parts per whatever accuracy of that sensor, and those two things do not relate like that. The sensor is measuring the magnetic field. That field is exponential in shape, and also "mostly in weird angles" when it emerges from the rotor. There may be some very, very, small degree of angle that the hall sensor goes from "no" to "yes" depending on the shape of the magnetic field.

This is a much more likely candidate: http://toshiba.semicon-storage.com/i...dName=TCS40DLR

it's rated at 25hz, but it's also dumb so doesn't have any internal logic to mess us up.

While we're at this, we should talk about how hall effect sensors work. What they do is they setup a stream of electrons. The magnetic field pushes that stream around. They switch "on" when that stream is pushed into a bucket somewhere else on the die. Depending on where this bucket is, or buckets are, you can setup a hall sensor to trigger on north, south, or both poles!

Quote:
Originally Posted by gigaplex View Post
Hall sensors measure a voltage induced by a magnetic field. Their accuracy rating refers to the magnitude of that induced voltage. That's quite different to the accuracy of timing measurements in a dynamic environment.
Pretty close, usefully close unless you're designing an ESC. They measure the strength of a magnetic field by how much they divert a stream of electrons. The design of the "catch" on the die can make the resultant output very nearly digital, or a nice smooth sine.

Quote:
Originally Posted by two shoes View Post
My basic understanding is that these are used as switches, once a voltage threshold has been crossed, it switches state, indicating pole swap. The timing is determined by how far ahead or behind each phase fires in relationship to where the sensor picks up the position of the pole... Where/how is accuracy lost or gained? Can you saturate the sensor at a high enough speed?
Well, they more indicate where the pole is in relation to the stator. More than "indicate pole swap". Timing is how far ahead, or behind, the rotor that the stator switches states. That's right.

Hall sensors are crazy repeatable. Now how they're being listened to...well that's more something we should talk to ROAR about. Speaking of which:

well crap, that PDF isn't coming up easily.

Another way to determine it... is searching digikey, and looking for analog output. All the sot23 pieces with analog out, are $0.60+ each. These things are not likely giving analog voltage out.

So the sensor harness just provides "checking" of where the stator is. (that's why the motors can start from a standstill, or low rpms when back EMF isn't doing the job for them.)

The honeywell PDF also indicates that the hall sensors are being used as switches, as opposed to absolute rotor indication.

Back to the original question, how can accuracy be collected from this "on/off" data. Well.. you just look for the rising, or falling edge of the signal. Between the exponential nature of the switch, and the magnetic field, I'd be surprised if there was more than a few tenths of a degree variance in trigger position. I think the effects of reflow and surface tension on the final position of the switch would be of greater importance.

Also... now I'm wondering if sensored ESCs turn blind after a certian RPM< and just depend on back EMF. Some of my ESCs will drive both sensored and sensorless. ROAR allows both. So... I wonder.
gigaplex, two shoes and moistAF like this.
Nerobro is online now   Reply With Quote
Reply



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off



All times are GMT -7. It is currently 12:56 AM.


We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.
Powered By: vBulletin v3.9.2.1
Privacy Policy | Terms of Use | Advertise Content © 2001-2011 RCTech.net