Crankshaft postion Sensors: VR or Hall?

Hardware inside and outside of the ECU
Post Reply
E4ODnut
Posts: 207
Joined: Mon Feb 10, 2014 11:52 pm
Location: Gibsons, BC, Canada

Crankshaft postion Sensors: VR or Hall?

Post by E4ODnut »

I can understand the use of VR for crankshaft position sensing if you are trying to use an existing sytem, but for a new install, such as carb to EFI conversion, wouldn't hall be a better choice?

I've been doing a lot of research on this lately and I am becoming more curious all the time as to why VR with a 36-1 or 60-2 wheel seems to be the method of choice. I don't doubt that it can be made to work very well, especially witht the MAX9926 chip and circuitry, but wouldn't a hall effect gear tooth sensor such as the Honeywell 1GT101DC work just as well or better, be simpler to interface with, less susecptable to noise, give a dependable signal from very low speeds to more than anyone would probably need and be more tolerant of tooth to sensor clearances?
Robert
1995 Ford E-150, 300 CID I6 E4OD, Custom MS1-Extra
1992 Winnebago Elante 33 RQ, Ford 460 CID V8, E4OD, Custom MS1-Extra
1992 Bayliner 3288, Twin Ford 351CID Windsor V8s, Custom MS1-Extra
1995 Dodge Ram 2500 4x4 488 CID V10 5 spd. MS3 (in progress)
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: Crankshaft postion Sensors: VR or Hall?

Post by puff »

my guess: more expensive?
at 6k rpm on a 36 teeth trigger the duration of each tooth is 277µs, while the rise time for 1GT101DC is 150µs (just goggled the data sheet). probably people want to rev up their engines

btw, what advantages does MAX9926 have over LM1815? seems the same but from various vendors. all the stuff people write on automotive forums is often complete nonsense :D
DaWaN
Posts: 51
Joined: Sat Sep 20, 2014 6:54 pm
Location: Benschop, Netherlands

Re: Crankshaft postion Sensors: VR or Hall?

Post by DaWaN »

OEMs must have very good reason to stick to VR sensors: let me throw some arguments in favor of the VR sensor on the table:

Phase / offset / accuracy: the great thing of VR is the fact the middle of the teeth is always at the zero crossing of the signal, this makes VR immune to delays and/or phase errors.
If the VR signal is conditioned well inside the ECU you should not have a RPM dependent timing offset.

Electronics: a VR sensor is a complete passive component so very unlikely to fail. As these sensors are mounted in high temperature places reliability is a must. Any kind of Hall sensor must have built-in electronics for signal processing: this means the chance of a failing component is most likely bigger.
Also: the hall sensor signal is smaller in amplitude compared to a VR sensor signal, so tolerance on mounting Hall sensors might also be more strict compared to VR.
Obviously when it comes to wiring noise immunity a hall sensor is superior to a VR sensor: the wire carries a conditioned signal instead of the raw signal.

Last argument:
VW uses hall sensors on their 60-2 trigger wheels, it is interesting to see the wheel as the teeth are bent 90 degrees compared to a similar VR wheel.
This is a dedicated wheel for triggering and seems to be quite thick. The cost of a simpler wheel with VR sensor and shielded wires might just be cheaper compared to a Hall sensor setup.
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Crankshaft postion Sensors: VR or Hall?

Post by kb1gtt »

I've not owned a vehicle yet with a VR. I wonder if VR is more common on US vs foreign vehicles. Any how, this is how I see the differences.

MAX9926 vs LM1815. The LM1815 uses a shared GND for the signal wire. This forces a lower Common Mode Rejection (CRM) which causes the noise floor in the chip to go up, which decreases when it can accurately detect the triggering edge. This is most notable at the skipped tooth. I suggest the MAX9926 because it has a better CRM which allows it to properly decode the skipped tooth when the LM1815 fails.

Pro's of VR, it's reasonably cheap and rugged for potentially high temperature environments. This separation of high temp components in the engine bay with lower temp components in the ECU was originally a much lower cost solution than hall. So OEM's used it and those designs are often still hanging around. As well the bandwidth of a VR is much lower, than with a sharp edge of a hall. So it makes it easier to pass EMC testing that's required for OEMs before they can sell the product to the public. So OEM's have some benefits in using it. The accuracy of the signal is based on RPM and what electronics you have to decode the signal.

Con's of a VR, you have to hit a min RPM before you can detect your crank angle. Think cold starting condition with a soft battery. You may not get enough RPM's to detect and start the engine. As well it adds delay to ECU sync times. If you have to wait for say 5 teeth to pass before the ECU can start to decode the wheel, you have just wasted that time of cranking. Also at high RPM you voltages increase which means you need better wiring to make sure you don't fail to keep sync.

Pro's of hall, you can test it at 0RPM, which can be handy when trying to figure out if you have a proper polarity. AKA is the top of the tooth a 1 or a 0 in software. Hall can do this very easily, as you can rotate the engine by hand and check polarity with a multi-meter. This low RPM also allows the ECU to decode the wheel quicker, often if you save the last known crank angle, you can use that as your starting point and get crank sync on the first tooth edge. AKA you can keep sync even if you turn the engine off for a week.

Cons of hall, your accuracy is determined by the internal circuitry found in the hall. They have circuits that hunt for the proper magnetic field, this can add a bit of jitter to the output signal. Also the sharp square wave transitions add a certain level of harmonics to the signal, which can cause EMC problems which can be expensive for OEM's who have to under go that testing. Often they need to slew limit the signal, and use special wire to wire the hall, which has shielding and such. Yes it's only like $0.10 per install, but for OEM's with large quantities that can be significant.

Both VR and Hall are used in high RPM, both can either succeed or fail at high RPM and high tooth counts. I don't see a significant difference there. Hall's jitter tendencies requires some extra effort in software to ensure you have proper decoding. It's easy for the signal to produce say +/- .2 degree crank wheel sensing, but if your software doesn't address jitter properly, you can easily get that +/- far above the desired .1 or .2 tolerances.

I would generally suggest hall for most people if they are installing something new. If you have an OEM VR, then use it, but if you can go hall, that's far less problematic for most people who use rusEFI, as the software is capable of decoding with small levels of induced jitter.
Welcome to the friendlier side of internet crazy :)
E4ODnut
Posts: 207
Joined: Mon Feb 10, 2014 11:52 pm
Location: Gibsons, BC, Canada

Re: Crankshaft postion Sensors: VR or Hall?

Post by E4ODnut »

I don't think cost is an issue. Digikey offers a 1GT101DC for $Cdn 88.69. Diy Autotune offers a 1GT101DC (or perhaps a clone) for ~$Cdn 42.00. CY Sensors offers what looks like a 1GT101DC clone for ~$Cdn 19.40. I didn’t have much success finding new aftermarket VR sensors but Digikey has nothing less than ones at ~$Cdn 114. I’m sure that wrecking yard parts can be had for much less though.

Here are a few interesting links:

http://www.farnell.com/datasheets/1680391.pdf
http://www.cy-sensors.com/
http://fullfunctioneng.com/info/Hall%20vs%20VR.pdf
http://www.electronicproducts.com/Electromechanical_Components/Hall_vs_VR_Which_speed_sensor_is_wrong_for_you.aspx
Robert
1995 Ford E-150, 300 CID I6 E4OD, Custom MS1-Extra
1992 Winnebago Elante 33 RQ, Ford 460 CID V8, E4OD, Custom MS1-Extra
1992 Bayliner 3288, Twin Ford 351CID Windsor V8s, Custom MS1-Extra
1995 Dodge Ram 2500 4x4 488 CID V10 5 spd. MS3 (in progress)
puff
contributor
contributor
Posts: 2961
Joined: Mon Nov 11, 2013 11:28 am
Location: Moskau

Re: Crankshaft postion Sensors: VR or Hall?

Post by puff »

probably you should take a look at Russian marketplace ;-)
e.g. I bought mine for less than $10 last year (brand-new vr sensor! from VAZ lada - probably same as from Cossack in Canada?). I guess, due to currency fluctuations it's now less than $5: delivery from Russia to Canada would cost you another $5.
http://electro-stavr.ru/?tovar=1&id_producer=334
the only difficulty, you'd probably need an individual to do so, as there are some customs-related difficulties for exporting goods from legal entities.

hm. may be I'd start a business exporting automotive parts from Russia? i'd love to! ;-)
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Crankshaft postion Sensors: VR or Hall?

Post by kb1gtt »

Welcome to the friendlier side of internet crazy :)
User avatar
abecedarian
Posts: 386
Joined: Fri Nov 15, 2013 10:49 am

Re: Crankshaft postion Sensors: VR or Hall?

Post by abecedarian »

Did you really think out your reply?
kb1gtt wrote:I've not owned a vehicle yet with a VR. I wonder if VR is more common on US vs foreign vehicles.
What does that matter?
You can lead the horticulture but you can't make them think.
User avatar
abecedarian
Posts: 386
Joined: Fri Nov 15, 2013 10:49 am

Re: Crankshaft postion Sensors: VR or Hall?

Post by abecedarian »

"This low RPM also allows the ECU to decode the wheel quicker, often if you save the last known crank angle, you can use that as your starting point and get crank sync on the first tooth edge. AKA you can keep sync even if you turn the engine off for a week."

No you can't.
You can lead the horticulture but you can't make them think.
User avatar
kb1gtt
contributor
contributor
Posts: 3758
Joined: Tue Sep 10, 2013 1:42 am
Location: ME of USA

Re: Crankshaft postion Sensors: VR or Hall?

Post by kb1gtt »

abecedarian wrote:What does that matter?
I have only owned hall, so it may make sense that I prefer hall because I've had it at arms length. Often people prefer what they are used to. AKA when you have a hammer, every problem looks like a nail. While I have worked with VR before, it was not my equipment, so I was limited in what I could learn. I may prefer hall just because I have had access to it. However being able to do basic checks with a DMM is really handy especially for someone who has less experience with electronics.
Welcome to the friendlier side of internet crazy :)
E4ODnut
Posts: 207
Joined: Mon Feb 10, 2014 11:52 pm
Location: Gibsons, BC, Canada

Re: Crankshaft postion Sensors: VR or Hall?

Post by E4ODnut »

Just an update on my research.

So far it looks like VR can produce be made to produce an accurate signal that occurs mid point of the tooth that is crossing the face of the sensor. That's the good news. The bad news is that it needs very good zero crossing hardware to do this, the trigger wheel must be matched to the sensor diameter, and noise has to be filtered out. All doable, but perhaps not easy.

At first blush, Hall seems like it should be a winner. Edge detection, little or no conditioning hardware required, good resistance to noise, but:

The operate point and the release point do not occur precisely when the tooth edge passes under the centre line of the sensor. What I have been able to determine so far is that they occur at some point before, but just how much before I have yet to find hard data on. It seems to be dependent on tooth to sensor gap and other factors as well. So far data sheets do not give me clear answers.

If the offset is constant over speed and temperature than software can compensate for it, but I don't know that it is.

Just to add to the complexity, there are at least two flavors of hall effect gear tooth sensors. The relatively simple magnetically biased hall effect gear tooth sensor, or the magnetically biased differential hall effect gear tooth sensor.

Research continues.
Robert
1995 Ford E-150, 300 CID I6 E4OD, Custom MS1-Extra
1992 Winnebago Elante 33 RQ, Ford 460 CID V8, E4OD, Custom MS1-Extra
1992 Bayliner 3288, Twin Ford 351CID Windsor V8s, Custom MS1-Extra
1995 Dodge Ram 2500 4x4 488 CID V10 5 spd. MS3 (in progress)
Post Reply