Hall sensors and steel parts

Does anyone know how steel parts influence magnetic field in a typical Magnet - Hall - Magnet setup? Is this even a valid concern? For reference, let's say I'm going to use a Cardanus coupling (joint) as the gimbal of my DIY stick and that linearity of the sensor output is the top priority.

One the first guy that use U-Join (Cardan) say that have no problem, since the Neodimium magnets is very strong and is placed very near the HALL (1-2mm).

Some guys use a circular metallic case around magnets/HALL creating a kind of Faraday cage. Industrial HALL "pot" have this kind of case.

But, why not use magnet resistive sensor - like the DIY MagRez (analog), MMsens (digital- need compatible controller) instead HALL?

- HALL require manually position the magnets relative HALL center position during assembly (and a 0,01mm is registered :) ) and manual calibration.

- Resistive sensor not require magnet precision centering and is auto calibrated every time that is plugged.

BTW - Since you are in Poland why not try buy Milan SIMM8RGe controller board, this work with digital Melexis sensor. In their Newsletter are mentioned the DIY market.

Thanks Sokol. Again! I'll note your contact down for future hiring of advisor if I start a business ;)

On to the meaty part. I know that in general the HALL-magnet assembly may be influenced by adjacent metal parts. But I'm looking for information that would help me understand the phenomenon as well as the magnitude of it with regard to the application in sim controls. For example I have some old notes on the subject containing figures like these

http://aussiemagnets.com.au/knowledgemanager/questions.php?questionid=23

but they hardly tell anything about the magnitude or the influence of irregular shapes as opposed to simple plates.

As for my selection of controller and sensors. I prefer to avoid obscure hardware of any kind. At least one I can't replicate easily myself. I couldn't find the Milan SIMM8RGe controller on his website and the MagRez would have to shine like 10 golden Suns to convince me to ditch simple and tried setup over that little undocumented gargoyle ;)

- HALL require manually position the magnets relative HALL center position during assembly (and a 0,01mm is registered ) and manual calibration.

Not if you're calibrating manually.

- Resistive sensor not require magnet precision centering and is auto calibrated every time that is plugged.

It's limited to symmetrical setups (forward and aft stick - same deflection) and I have nothing against manual calibration.

... and the MagRez would have to shine like 10 golden Suns to convince me to ditch simple and tried setup over that little undocumented gargoyle ;)

People are using this sensors in DIY boards at ~10 years. :thumbup:

The first revisions o MFG Crosswind (previous of the use of SIMM8RGe+Melexis MLX90333) use then. Bauer use in his pedal, stick.

They are document (for DIY controller use), just not in English.

GVL224 (no "busines advisor" in this), sold a assembled board read to use for ~8$.

But, deal with DIY HALL sensor (SS495A1, A1324) + magnets is very "fun". And work. :D

If a quick lookup doesn't return the following information in English (or at least a proper russian webpage):

- technical details

- where to purchase

I consider it obscure. Sorry for being so stuckup but I'm after getting my DIY stick and rudder pedals up and running, not adventures ;)

As for this "MagRez" sensor, it uses KMZ41, no?

Just entering KMZ41 in google spits out this pdf:

http://www.eetasia.com/ARTICLES/2002DEC/A/2002DEC20_AMD_AN18.PDF?SOURCES=DOWNLOAD

I have not yet finished reading yet, but it seems pretty interesting.

Sorry if OT.

- technical details

From DIY simmer POV, with scheme, part list: http://avia-sim.ru/forum/viewtopic.php?t=572&sid=6c87a4131c47d8cd94b9e42b0065098b

- where to purchase

DCS forum: http://forums.eagle.ru/showthread.php?t=145129

BTW - From today perspective both HALL and MagREZ can be considered "obsolete" as pot' because they work in analog mode.

Are affordable DIY digital options, MMSens+MMjoy2, or Hempstick + Melexis MLX90333 sensor (dont know if he finalize the implementation of support in the code), or commercial: Komaroff RAMS (+RedBaron controller)*

*Or the (difficult to buy) VKB MaRS (+ Njoy32).

:thumbup:

From DIY simmer POV, with scheme, part list: http://avia-sim.ru/forum/viewtopic.php?t=572&sid=6c87a4131c47d8cd94b9e42b0065098b

 

 

 

DCS forum: http://forums.eagle.ru/showthread.php?t=145129

 

BTW - From today perspective both HALL and MagREZ can be considered "obsolete" as pot' because they work in analog mode.

 

Are affordable DIY digital options, MMSens+MMjoy2, or Hempstick + Melexis MLX90333 sensor (dont know if he finalize the implementation of support in the code), or commercial: Komaroff RAMS (+RedBaron controller)*

 

*Or the (difficult to buy) VKB MaRS (+ Njoy32).

 

:

Your understanding of sufficient documentation is completely different from mine :) As you can see in the Hempstick thread here on the forums even though it's pretty well documented there's some crucial information missing.

It's also evident that you don't care about the costs. You can't seriously say the difference between the cost of Allegro-based solution and Melexis-based one is small. Especially for a 6+ axis project. Then there's the matter of simplicity! And the Allegro- based solution being close to perfect.

All that combined, the melexis would have to brew beer while idling while the fact that some circuit is "digital" would have to make me live longer for them to be worth pursuing :)

Simundza with his MFG supposedly went from "old gen" to "next gen" and how exactly did it translate to better performance?

As for this "MagRez" sensor, it uses KMZ41, no?

 

Just entering KMZ41 in google spits out this pdf:

http://www.eetasia.com/ARTICLES/2002DEC/A/2002DEC20_AMD_AN18.PDF?SOURCES=DOWNLOAD

 

I have not yet finished reading yet, but it seems pretty interesting.

Sorry if OT.

Get out of my thread.

Are you kidding me? Good find! ;)

Your understanding of sufficient documentation is completely different from mine :) As you can see in the Hempstick thread here on the forums even though it's pretty well documented there's some crucial information missing.

For build a MagRez I need the electrical scheme for part list and make the PCB.

This was on forums, as are PCB layout (easy to do in half hour with Laser printer method, see YT videos).

For DIY sim builder like you and me is only "documentation" needed.

Assembly, hook on USB controller like a pot' and test.

Hempstick project is well documented, but this project is more expensive, more complicated to write the firmware, configure, than a ~10-20$ MMjoy2.

It's also evident that you don't care about the costs. You can't seriously say the difference between the cost of Allegro-based solution and Melexis-based one is small. Especially for a 6+ axis project.

In absolute numbers (Digikey prices)

Allegro 1324 cost 2.2$.

MLX 90333 6.2$.

KMZ41 2.50$ (+MPC602, resistors, condenser add ~3$).

TLE5011 2.50$ (need 4 resistors 2 condenser ~0.2$).

So cost wise the 1324 win, but use the most expensive dont will hurt you bank account.

BTW - Thrustmaster use Melexis not because is "better", is because for then is more cost effective, this 6.2$ handle 2 axis. :thumbup:

Then there's the matter of simplicity! And the Allegro- based solution being close to perfect.

This simplicity is the enemy, you need deal in position a magnet 0,02 mm for left, or for up... for achieve a not "close to perfect" center position.

Experience from 3 joysticks, 2 pedals and 1 Suncom SFS retrofit with Honeywell/Allegro HALL sensor. But... work (despite the dozens of trial and error, tests needed). :D

All that combined, the melexis would have to brew beer while idling while the fact that some circuit is "digital" would have to make me live longer for them to be worth pursuing :)

For simmer's practical purpose the "digital" only mean less subject to external interference, possibility for use long cables...

Simundza with his MFG supposedly went from "old gen" to "next gen" and how exactly did it translate to better performance?

I don't know the reason, but I can guess:

Not depend on external source in other country - specially one in political unstable region.

Develop your own hardware and software - this can be used for other projects (I know the case of guy that buy his SIMM8... for use in other brand rudder pedal).

The "new" sensor require less parts to assemble (like TLE5011) relative to MagRez, in quantity production this save time and cost.

And from marketing POV the "new" in technology sold better (people buy new iPhone's every year). ;)

From simmer point of view I bet that in practice you will not notice difference in performance using one or other or eve a good brand potentiometer.

People still using the USB 1.1 8 bits (256 points) CH controllers + pot's and they are happy. :D

You know, I had to do some high resolution angular measurement for a project at work and found Austrian Microsystems to make some interesting parts.

The as5311 has 12 bit resolution for a pole pair - that is 2mm of magnet strip. Or you can use rings, which are roughly 60mm in diameter, with a pole pair being about 12°. You get a ridiculus resolution (yes yes, in action it is actually way less) even for extremely small angles. Just cut or break the ring apart to make the sensor assembly smaller.

http://ams.com/eng/Products/Position-Sensors/Linear-Magnetic-Position-Sensors/AS5311

I even came up with an idea for knowing which pole-pair the sensor was at, but turns out someone already patented it:

http://www.google.com/patents/US20120217956

Interface is either I2C or SPI. In the end, we used optical encoders though.

edit: It seems the IC itself is not available at the moments, the demoboards are though:

http://www.mouser.at/ProductDetail/ams/AS5311-TS_EK_AB/?qs=sGAEpiMZZMvc81WFyF5Edo8DuZ7L%2fX8v2R71wvld6w4%3d

When it comes to Melexis, can those sensors be used without being "burnt in" first? Because that programmer is kind of expensive (2000€). I had the idea of buying a T16000m (30€) and taking the board and magnet, transplant it into a better gimbal. You would have to align them perfectly like in the T16000m though, since this is what the IC was burnt in for.

@Sokol

I did one project based on Allegro HALL. I have lost almost all of my notes on the project but I'm sure I would have remembered any problems with magnet-HALL placement. I have used a controller board from an old Logitech Wingman Extreme Digital 3D Pro (I don't know which is worse - the stick or its name) and at 1024 steps I did not notice any problems. I even placed the magnets directly on a steel axle. Am I missing something?

Russian

I told you I could deal with Russian "documentation" but only as the (truly) last resort. Not only my Russian is poor, the info is scattered, mostly forum based and mostly "experienced for experienced", and "and it works great with any sensor, as shown in, you know, that other thread Dimitri has started on 1C forums. Or simhq maybe..." Calling the information incomplete is a courtesy in this case. I'm not referring to MagRez. Just to Russian "resources" in general.

Prices:

1. I live in Poland. Correct the prices by a factor of 1.5 at least for the parts and a factor of 3+ for the shipping. Recently I've done a simple PWM fan controller. 10 different basic parts. I had to buy them in 3 separate internet stores.

1a. Only popular ICs and HALL sensors available.

OR assign weeks for parts and bargain hunting.

2. 8 axes gives times 8 the cost for sensor >assemblies<.

Other:

MLX90333 and TLE5011:

I wouldn't even know where to begin with them. Most probably not an option for me, even remotely.

EDIT: For later reference: http://simhq.com/forum/ubbthreads.php/topics/4097757/Re:_MMJoy_-_Build_your_own_USB#Post4097757

interference:

Have you actually seen interference noise on any output? With what controller? What cable length? As you could see I'm not an expert in electronics but I would be really surprised if it was an issue with any kind of basic filtering in place.

Arduino Micro and MMjoy2:

You may have convinced me on this one at least. But take a look at this. The very first question. What resolution it's capable of? I visit the webpage (fancy already in this circumstances) and...

https://translate.google.com/translate?hl=en&sl=ru&tl=en&u=https%3A%2F%2Fsites.google.com%2Fsite%2Fmmjoyproject%2Fnovosti-i-obnovlenia&sandbox=1

Nothing. I have found the info here (your post also)

http://simhq.com/forum/ubbthreads.php/topics/3994326/1

I don't like that it's not 12 bit but perhaps it wouldn't be wise for me to go for a 60 Euro board.

I have used BU0836 in this "upgrade" - I do this for other guys. Now use MMjoy2.

For cable use CD-ROM audio cables for short ones, and for long, e.g. hook a pedal in joy base, microphone cables, they have the shielding mesh around power and data cables.

I see people saying that "pig tail cables are better, but don't find this locally.

In my limited experience with DIY HALL:

SS495A1 result problematic, you see "micro-spikes" even in Windows Game Controller test.

Are a trick with 0.1 uF condenser between signal and +5V wires but this induce lag in response.

A1301 is less, but more complicated with magnet placement. Now have there 10 A1324 to try...

With MMjoy2 I can set filter in firmware for deal with SS495A1 problem, but in the last use I have to use a high value, dont like do this. At least I have this option to use with MMjoy2, with BU0836, no.

BTW - In Crosswind electronics upgrade in YT you see that is used filters too.

As I try explain bringing together all this scattered information in SimHQ (since no one did it until now in any place) ATMEGA32u4 have 10 bits ADC converter, but MMJoy2 firmware allow "oversample" - like Hempstick do from 12 to 14, so the real resolution can be more if use external ADC, e.g. 14 bits with build-in TLE5010/11 ADC. This was explained to me be firmware author.

The big problem in all this is, the guys that do this things dont know English, is difficult explain. As side note, they do "this things" because dont have affordable "Warthogs" in local Walmart's or Amazon's. ;)

The only inconvenient of TLE5010/11 is the need to make a small PCB (in that thread in SimHQ the guy sold ready to use for ~10$), and the use of 5 wires instead 3 on pot or (analog) HALL.

To end this: I am not trying convince you "this or that is the best", the intention was show options - I assume that you don't know the others, since you know but don't trust, I am wrong. Sorry.

If you have not read, don't miss this Hempstisk PDF:

I have used BU0836 in this "upgrade" - I do this for other guys. Now use MMjoy2.

 

For cable use CD-ROM audio cables for short ones, and for long, e.g. hook a pedal in joy base, microphone cables, they have the shielding mesh around power and data cables.

I see people saying that "pig tail cables are better, but don't find this locally.

 

In my limited experience with DIY HALL:

 

SS495A1 result problematic, you see "micro-spikes" even in Windows Game Controller test.

Are a trick with 0.1 uF condenser between signal and +5V wires but this induce lag in response.

A1301 is less, but more complicated with magnet placement. Now have there 10 A1324 to try...

 

With MMjoy2 I can set filter in firmware for deal with SS495A1 problem, but in the last use I have to use a high value, dont like do this. At least I have this option to use with MMjoy2, with BU0836, no.

 

BTW - In Crosswind electronics upgrade in YT you see that is used filters too.

 

As I try explain bringing together all this scattered information in SimHQ (since no one did it until now in any place) ATMEGA32u4 have 10 bits ADC converter, but MMJoy2 firmware allow "oversample" - like Hempstick do from 12 to 14, so the real resolution can be more if use external ADC, e.g. 14 bits with build-in TLE5010/11 ADC. This was explained to me be firmware author.

 

The big problem in all this is, the guys that do this things dont know English, is difficult explain. As side note, they do "this things" because dont have affordable "Warthogs" in local Walmart's or Amazon's. ;)

 

The only inconvenient of TLE5010/11 is the need to make a small PCB (in that thread in SimHQ the guy sold ready to use for ~10$), and the use of 5 wires instead 3 on pot or (analog) HALL.

 

To end this: I am not trying convince you "this or that is the best", the intention was show options - I assume that you don't know the others, since you know but don't trust, I am wrong. Sorry.

 

If you have not read, don't miss this Hempstisk PDF:

 

 

 

http://www.hempstick.org/download/articles/OnResolution.pdf

 

What do you mean by CD cables? Those? http://www.amazon.com/gp/aw/d/B00066GYXK/ref=mp_s_a_1_1?qid=1448528150&sr=8-1&pi=SY200_QL40&keywords=cd+audio+cable&dpPl=1&dpID=41XF9gSg9nL&ref=plSrch

Please don't get me wrong. The knowledge you have shared here and on other forums is invaluable. Plus I don't see anyone else being a competition to you in this respect. The reason I sounded like a nope-nope-meh-naah primadonna is I need to avoid specific solutions not to fall into a vicious circle. And that may happen easily due to my very limited experience with such modding and especially with electronics. The PWM project I have mentioned... Its purpose was for me to stop being electronics illiterate.

MFG:

MMJoy2 is for non-commercial use and yet Simundza uses it. Did he strike an agreement with mega_mozg or am I getting it wrong?

And does he use a custom built USB controller incorporating the Melexis Hall IC?

Yes, this type of cables - they are shielded.

Mega_MOZG said that this cable (he call "pig tail") is better:

The "pot" in the picture is TLE5011 over original pot case - used for turn the ring magnet.

But here "pig tail" is a kind of antenna cable... :huh:

MMJoy2 is for non-commercial use and yet Simundza uses it.

Impossible, MMjoy2 is based on ATMEGA mc, Simundza MFG SiMM8Rge is based on Microchip mc. ;)

Picture from their newsletter:

gif upload

I supposed that the circled chip is the MLX90363 for rudder axis.