Voltage Surge Suppression from Hall Effect Sensor Transient Response 1

[ngio]

Hello all,

I am working with a "Hamlin 55100" Hall effect sensor (actually two of them with overlapping output signals 5V->~0V->5V), whenever I trigger the two hall sensors with a magnet I get a really ugly pulse (as seen in photo). At the end of the pulse there is a transient response that shoots the voltage up to 8V then resonates for a bit at 6V. I'm trying to remove this voltage surge and the resonating effect that happens afterwards. I looked into using a MOV, but as far as I can tell the MOV would not survive the amount of cycles it will experience (I'm planning on exposing it to over 250,000pulses).

The Hamlin 55100 spec sheet says to use a capacitor between the positive and negative wires of the sensor (but they don't specify a size), I've already blown up one capacitor (470uF) in my face trying this but with little to no success. ---I'd prefer not to experience the exploding capacitor again ---

please help if you can.

Thanks.

 

[Skogsgurra]

What is your load? A relay coil or similar? What you see is normal inductive kick-back and the best you can do is to use a free-wheeling diode across the coil (or whatever the inductive part is).

You cannot possibly blow up a capacitor with that voltage - unless you had the wrong polarity.

A MOV is a bad choice here. For many reasons.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[ngio]

Thanks for the quick response,

The hall sensor is attached to an electronic ignition system(EIS) that takes ~10mA (open collector input pretty sure) and a National Instruments DAQ system that requires minimal amperage if any.

It's interesting you say an inductive kick-back, the ignition system is controlled by a microprocessor, is it possible that there's too much current going to the ignition system currently? I did add three 1K resistors in parallel (to make a 333ohm resistor since I had none), to limit the current going to the ignition system. At 12V and 333ohms, the current should be 0.036A going to the EIS. If the EIS needs more current could it induce a "kick-back"?

 

[IRstuff]

When is the ignition coil firing relative to this spike? You could have:
> EMI noise feeding into your scope
> ground bounce from the firing the spark plug

See:

http://mgaguru.com/mgtech/ignition/ig108.htm

for similar voltage waveform.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[VEBill]

Are you using shielded, twisted pair wires?

The Hall effect sensor itself is probably very small (physical size) and would thereby be relatively insensitive to EMI itself. But the wiring running from the sensor to the input of the processor needs to be protected using basic E3 design techniques (twisted pair, shields, isolated signal return path).

 

[ngio]

Hi VE1BLL, I'm currently using twisted (shielded) pairs running from the hall sensor to the ignition system(w/ internal processor) and to the DAQ system to reduce noise. With this talk of twisted pair shielding I remembered that the actual ignition wires are not in twisted pairs or shielded (probably a very large source of noise right now, although I don't think it's the source of the post signal spike). I'll replace this and let you know if anything major changes.

Hi IRstuff, Thanks for the link to the file its interesting because the ignition system sparks right at the rising edge of the hall sensor pulse (therefore probably the reason for the effects seen). When you mention ground bounce... is there a way to prevent this? I'll increase the diameter of wire going to ground, maybe it will help.

I should also let people know that I am seeing low voltage spikes in other measurements (pressure transducer and optical encoder).

 

[IRstuff]

You can minimize ground bounce, and you can try doing differential measurements. Measuring signals in that type of environment is tricky.

The test would be when you disconnect the ignition coil, you get a clean signal. How far from the ignition coil are the Hall sensors?

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[ngio]

Currently I'm using differential signals for the DAQ system but not into the ignition system. The ignition coil is ~8inches from the hall sensors although theres a few metallic components between the two (covers and whatnot). Disconnecting the coil and checking the signals will be among the next steps (thanks for the recommendation), I guess it will tell me if the source of noise is in the ignition wires and the coil, or if its from the ignition system to the power supply (battery).

Unfortunately I'm off-site today, but tomorrow I will try it all out and let you know what ends up happening. Any last suggestions for stuff to try before tomorrow? I have already removed the following wiring issues: removed extra current loops, wire loops, wire separation minimum of 1ft between switching and analog signal wires. And I've added the following to the circuit: twisted pairs, shields, ground isolation. One item I haven't added (for reasons seen in my initial post) is the capacitor between the circuits and the ground, or between the positive/negative terminals of the loads. I have heard a few times that it helps in relieving transient responses, although my knowledge in sizing these is minimal.

I guess for some reason I was convinced earlier it had to do with the hall sensor but I'm starting to believe that it could be due to the ignition system.

 

[iop95]

Is sensor mechanicaly simetric regarding sensing moving part? I thing sensing moving part is approach to sensor at slightly high distance than to which departs; this may be reason that aproach pulse (low part) is very clean and depart pulse (+8V) have same oscilations. Also a small capacitor (about 10nF) between sensor signal and ground may solve this problem.

 

[ngio]

Thanks for the input, I've done my best to try and make the sensors as symmetric as possible to each other and to the magnet on the flywheel. I can double check to make sure the mounts have not moved, but they were fairly secure and rigid.

The 10nF capacitor should be placed between the signal and ground or the power and ground? The spec sheet recommends one between the Power and Ground wires (specsheet attached), but they do not give a size.

 

[IRstuff]

A top-level schematic showing grounding and power connections between the various parts of the system would probably be more useful.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[OperaHouse]

I find......

(actually two of them with overlapping output signals 5V->~0V->5V)

And your waveforms troubling. I suspect you are doing something with these sensors that we can't even imagine. Could you show how these sensors are hooked up?