Criteria for selecting Low Pass filter for Hall Effect sensors.

[venkatgandham]

Hello,

I am trying to find the criteria for the Low pass filter selection for Hall effect sensor.

From my understanding If I know the frequency of Hall effect sensor then I can find the R and c values and Design a Low pass filters.

But there are two obstacle mentioned below.

There are two types of circuits before the input of hall effect sensor can be applied to CPU or ECM.
Type 1
Type 2

Type 1:

It has C1 (for ESD) and R1 resistors which is in parallel with C1 the output of both is given to Signal conditioning Circuit --> CPU or ECM.
Now I am confused where is the LPF filter in the type 1 circuit to restrict the high frequency inputs. My guess in signal conditioning circuit.
If that is true then how would I decide the vaules of C1 and R1. Why R1 is in parallel with C1. Which is missing in Type 2.

Type 2:-

It is open drain circuit, It has ESD Capcitor in parallel with the signal conditioning circuit. Now I would like to know why this Circuit doesn't have the R1 in Parallel with C1. Then comes the Signal conditioning circuit and output of the Signal conditioning Circuit is given to CPU or ECM.

Flow of Signals:-

Type 1: Hall Effect sensor o/p---> C1||R1 --> Signal Conditioning Circuit --> ECM or CPU.

Type 2: Hall Effect sensor o/p---> C1 --> Signal Conditioning Circuit --> ECM or CPU.

Please help me in identifying the criteria for designing the LPF for hall effect sensors. I appreciated your help

 

[BradtheRad]

Can't visualize the ESD configuration, nor the signal conditioning circuit.

In any case...

You have unknown resistances in neighboring devices, including the sensor and signal conditioning circuit. These will add their influence to the visible resistors you put near the capacitor. The effect is to throw off the time constant.

You will have to try various capacitors, to find one that yields the desired filter curve, in the arrangement you end up using.

 

[FvM]

Can't visualize the ESD configuration, nor the signal conditioning circuit.

So you do understand the nature of the intended hall sensor application and why it should use a low pass filter?

 

[venkatgandham]

Hello FvM,

Yes, Because the output of the hall effect sensor is square wave and you don't want to pass the square wave to the input of ADC of the microprocessor. This is because ADC won't be able to read it and it will be able to read the sine wave coming from the Low Pass Filter or Signal Conditioning Circuit. That's why I am trying to design the Low pass filter or Signal conditioning circuit.
F=1/2*Pi*R*C, I can find the value of the R and C once I know the frequency of tha Hall Effect Sensor. I am also trying to find the values of the ESD capcitor and resistance on the path to the Microprocessor.

 

[BradtheRad]

So you do understand the nature of the intended hall sensor application and why it should use a low pass filter?

Hmmm... Should I answer the question, or the questioner?

The hall effect sensor detects a magnetic field directly. It has 3 pins (the one I've experimented with).

It outputs a volt level in a range of 0 to 5 V (say), in proportion to the intensity of the magnetic field. Above 2.5V means the field is North, below 2.5V means it's South. (Maybe it's vice versa).

I must have missed any previous threads where the OP discussed his application.

I don't know why the poster specifically is adding electrostatic discharge protection. I can't be sure where he put a capacitor in parallel, and a resistor in parallel.

Hall sensors are not by definition required to go through a low pass filter.
So for lack of details I simply guessed that he's looking at a design something like the diagram below. It streamlines the discussion, since we can't always be sure someone will post the schematic.

I do know that whatever cutoff frequency he wants for the filter, will be influenced by other resistors besides the visible one next to the capacitor.

A situation is hardly ever so routine, that a simple formula (such as Freq_cutoff = 1/RC) is all the guideline anyone needs.

Whatever the formula predicts, the real value will no doubt need to be found by experimentation.

So out of an impulse to be helpful, I fashioned a brief reply along that line. It invites the OP to reply and to elaborate, if he wishes.

- - - Updated - - -

I am trying to find the criteria for the Low pass filter selection for Hall effect sensor.

Because the output of the hall effect sensor is square wave and you don't want to pass the square wave to the input of ADC of the microprocessor. This is because ADC won't be able to read it and it will be able to read the sine wave coming from the Low Pass Filter or Signal Conditioning Circuit.

Now that I got a better idea what waveform you want, here are schematics showing three low-pass filters.



I'm guessing that a curled triangle wave is satisfactory, because it is not possible to turn a square wave into a sine using just a capacitor or inductor.

As you can see, the impedance of the previous stage contributes to the filter response. Likewise is is also affected by the impedance of the next stage. Notice that the maximum observed output is affected as well.

It's because these impedances are hard to determine, that makes it impossible to predict what microfarads you need.

You can calculate a ballpark figure, yet it is likely to be off by one or two orders of magnitude.

F=1/2*Pi*R*C

Right, I forgot.

 

[FvM]

I must have missed any previous threads where the OP discussed his application.

Of course you didn't as there are none.

I apologize for the witty comment. But seriously speaking, there isn't more validation for the low pass point in the original post than for the ESD one. The only difference it that you can imagine possible motivations for a low pass in some hall sensor applications. It may serve a reasonable purpose.

The poster is asking how to select the low pass filter. How can you really help without understanding the actual function of the low pass for the application under discussion? Without any substantial information, I prefer to doubt if the low pass has a purpose it all.

 

[godfreyl]

Seems like there are some questionable assumptions being made.

Because the output of the hall effect sensor is square wave...

Why do you think the output will be a square wave?

...and you don't want to pass the square wave to the input of ADC of the microprocessor.

Why don't you want to pass a square wave to the ADC?

This is because ADC won't be able to read it...

Why not? An ADC can read any waveform. It won't work if the frequency of the input signal is more than half the sampling frequency, but in that case a low pass filter won't help anyway.

 

[soppcb]

**broken link removed**

 

[venkatgandham]

Hello Godfreyl

The LPF or signal conditioning circuit is used just to clean up the square wave signal and remove any voltage offsets or slow transients. The ideal output from the LPF or signal conditioning is a clean square wave.

 

[BradtheRad]

there isn't more validation for the low pass point in the original post than for the ESD one. The only difference it that you can imagine possible motivations for a low pass in some hall sensor applications. It may serve a reasonable purpose.

The poster is asking how to select the low pass filter. How can you really help without understanding the actual function of the low pass for the application under discussion? Without any substantial information, I prefer to doubt if the low pass has a purpose it all.

Yes, all I could do was suppose the need for a LPF was real.

Sometimes we must 'fill in the dots'.

Or a poster may ask one question, but really needs for a different question to be addressed.

Anyway I noticed the OP used some advanced lingo. So I decided just to answer along the lines of his initial question.

Since he mentioned that he wants a sine wave (post 4), I realize we should ask if he is using (a) a hall effect switch which puts out highs and lows, or (b) a hall effect sensor which puts out an analog signal.

Maybe he wants to isolate some frequency. Maybe he would be better off using a bandpass filter.

 

[godfreyl]

Since he mentioned that he wants a sine wave (post 4)...

...and then he said he wants a clean square wave (post 9).

A low pass filter isn't going to fix DC offset or "slow transients". Maybe a comparator or Schmitt trigger would be more appropriate.

 

[venkatgandham]

Hello Brad,

I was busy last week that's why couldn't reply you. In past posts I was trying to define the criteria for Hall effect sensor interface, Initially I wasn't sure how this is going to work and described what I am doing. Now I know that there was no need of getting a sine wave at the output of Signal conditioning ckt. When it's already a square wave, the filter is used to remove the unwanted frequencies.

I came up with the following design Criteria for the interface.

Type 1:-
Hall effect input (it has divider consisting of a Rsource (1k) and Rsink (25 ohm)) ---> ESD cap parrallel to resistance --> signal conditioning circuit (LPF)--> Micro, for Type one hall effect sensor which is not an open drain circuit. I know the resistance and capacitance value for the interface to design it. The values for the R is around 47K and C= 10nf


For Open drain Circuit Type2.

Hall effect input---> divider consisting of a resistor and capacitor --> signal conditioning circuit (LPF)--> Micro.
for this we have capacitance and resistance values in C= 2.2 nf and R= 2.1k. I didn't find the filter values which can be found by F=2*Pi*R*C.

Both the interface circuit are very similar to the one you posted below for square and Triangular wave.

In type1 you have resistance and capacitance in parallel and then going to the LPF.

In Type 2 which is open drain --> Just have a Pull down capacitance to the resistance and the output of that is going to LPF.

If it becomes difficult to visualize let me know I will paste the Circuit, but I am trying to avoid that because of the Project.

I want to know which type is used very often and why? From my understanding it should be the type2.

Why there are two types of Hall effect sensors, why didn't they go for the Open drain only.

I know the working of hall effect sensor but don't know why they have adapted two ways of designing it.


Thank you for your help and I am learning from your responses.

 

[1206.sachin]

Hello All
If the O/P is square wave you can choose RC low pass filter cut off = 10 * Fo/p max ( Fo/p max is Max O/P frequency )
This is thumb rule.

If your O/P is having pull up resistor use that as R & calculate C for desired cut off.

You can add schmidt trigger after LP filter. It will improve square wave shape.

Open drain sensors are used for ORing two sensor O/P.

 

[BradtheRad]

If it becomes difficult to visualize let me know I will paste the Circuit, but I am trying to avoid that because of the Project.

Certainly you do not need to reveal any secret aspects of the project. Your question appears to be about non-proprietary signal conditioning.

For you to get real answers you will need to post a diagram, so we can visualize what kind of signal you are getting, and what conditioning it will need.

Just click the 'Add an Image' button to bring up a window to choose a picture, and upload it from your computer. Highlight the link that comes up, and paste it in your post to display it.

 

[venkatgandham]

I have uploaded the Image of the Circuit I was talking about. Below is the Url for that image.

 

[BradtheRad]

Yes, that tells us a little more.

I suppose the supply wires to the sensor are implied? Because the sensor normally has 3 terminals.

The sensor output conveys a certain amount of current. It has a certain output impedance. The neighboring resistors be high enough value that they do not drain away your signal. You want your sensor output to be dominant.

The diodes are okay where they are.

I believe you will need only one capacitor at most.

Right now it is important for you to spend a session examining the waveform coming from the sensor... and only from the sensor... while it detects the magnetic item (whatever that is), doing whatever it does.

You will observe whether the sensor puts out a waveform that is usable, in the sense that it (a) has sufficient amplitude, and (b) swings within a suitable volt range, and (c) gives a stable signal both when idle or when detecting.

There may be other criteria as well.

After determining that, you must then discover whether the microcontroller can handle this signal as is, or whether you first need to condition it.

The conditioning may turn out to be one of the other suggestions given above, perhaps in addition to a low-pass filter. You may end up needing to add a schmitt trigger, or maybe an amplifying device, etc.

If you want square waves, then that suggests you want to detect the pulse frequency, or time between events, etc.

If you want sine waves, then that suggests you want to detect intensity, or presence of more than one signal, etc.

 

[FvM]

If you want square waves, then that suggests you want to detect the pulse frequency, or time between events, etc.

If you want sine waves, then that suggests you want to detect intensity, or presence of more than one signal, etc.

In fact, we don't know much about the input signal. It's said to be a square wave, but which frequency, fixed or a larger range? Venkatgandham told in post #4 that he wants to get a sine after the filter, in post #9 it's a clean square wave. Looks like a helpless case.

 

[godfreyl]

In fact, we don't know much about the input signal. It's said to be a square wave...

IMHO, that's unlikely unless someone's invented a way to make magnetic fields with square edges.

It's possible the sensor has built-in circuitry that gives a binary on/off output, but then the signal shouldn't need conditioning.

 

[venkatgandham]

sensor will need a conditioning circuit even though it produces a digital output, how about when you touch a sensor with the static field of 8kv that will destroy the input of ECM hence you have a ESD cap and also to remove some of the unwanted frequencies or glitches from the sensor input.

IMHO, that's unlikely unless someone's invented a way to make magnetic fields with square edges.

It's possible the sensor has built-in circuitry that gives a binary on/off output, but then the signal shouldn't need conditioning.