counter over flow

[Fred_Flinstone]

Hello,

I hope some body can help me.

I'm trying to measure a speed of a shaft at around 0 - 7000rpm.
A hall-effect sensor picks up the signal from a gear wheel with 12 teeth, so the pulses per second should be around 1500Hz max.
The Sensor used is from allegro micro systems (link: http://www.allegromicro.com/sf/0673/index.asp) with a Schmidt-trigger.
The DAQ Card is an M series and I tried it with a 6036E PCMCIA card as well with LabView 8.
I try to measure continually but the buffer over flows with in a minute. I just use a basic setup with the DQAmax system the actual channel configuration with a graphical display and a numerical display. The settings in the Cannel configuration are Max frequency 3000Hz and Min 2Hz using a single counter and varying the Sample Rate between 1, 10 and 100 it makes no difference.
I cheque the signal form the sensor with a oscilloscope  it is a pretty clean square wave and the frequency reading form LabView and the oscilloscope are nearly equal -+100Hz  at 1300 Hz, so no noise is involved it does not show on the oscilloscope   (max frequency 100Mhz).
After this I used a signal generator to input a square wave signal into the DAQ it runs stable at 6400Hz at a sampling rate of 1 for 2 min than I stopped the DAQ.

Thank you for your help in advance

Regards,

Philipp

[Conseils]

Suggestions,
Have you got the decoupling capacitor close to the sensor device and linked in the test pin to ground?
Does your target,,, the gear meet the specifications in the data sheet?
"For TPOS to function as specified, the target must generate a
minimum of 120 G difference between the magnetic field over
a tooth and the field over a valley, at the maximum installation
air gap."

There is a pull up resistor that can be placed on the output of the device, this might help your triggering as well.
Are you running the device at 5v (basically is the device output (level) suitable to drive the input on the daq card?

Hope these help

[Fred_Flinstone]

Hello,

Thanks for the reply !

The signal generated is not so clean I get an average of around 7 Volts but when the pulse is triggered it over shoots to about 12 Volts and shortly drops to 5.5 Volts.
National Instruments told me to trigger a high the voltage needs to be between 2-5 Volts I built a voltage divider to try to stay in this range. They also told me that the max fall and rise time of the signal is 50 ns but that seems very short for me and the minimum pulse width of around 10 ns.
I don't have a decoupling capacitor can you explain to me what it does, please.

[Conseils]

The de-coupling capacitor stores a small amount of current and when the chip / integrated circuit demands some current (when switching for example) it ensures that there is sufficient to stop the voltage at the supply pins dropping. If you don't have one, you tend to get the sort of problems you are having and it also causes noise to other parts of a circuit. You find these capacitors physically close to chips especially digital / swithing types.

You should try to get a decoupling capacitor they are really very cheap, you could also readily obtain one from an old piece of scrap equipment, if you know how to do it or know someone.

Are you sure that the gear meets the technical requirements. They are quite specific for this device.

I think that there is a little confusion over the information on rise and fall times and minimum pulse widths. The following article might help.

Why Do I get Incorrect Count Values When I Use an External Signal as My Counter Source? http://digital.ni.com/public.nsf/allkb/E84361EAAF6C092686256AAA007ADA7B

[Fred_Flinstone]

Hello,

I had a look on the oscilloscope when the integrated circuit is switching the supply voltage does not drop so I think the current is ok. I tried it than with a capacitor (10000 micro Farad) but it maid no difference to the signal.

Than I tried to use a voltage divider (two resistors with 390kOhms) to reduce the average voltage of 6.5 volts to half of that to stay well within the TTL signal requirements, the problem I encountered was that the square wave started to be rounded off.

Than I tried to reduce the high spike voltage at the beginning of the signal with a capacitor but when I found a large enough capacitor to do so the logger did not read any of the signal and hat directly a buffer over run error.

So no luck until now I might have to use one of the low pass filters that were recommended in the like.

Tank you for the help.

Regards,

Philipp

[Conseils]

The size of the capacitor, in terms of capacitance is quite important and it is likely that the effect of such a large capacitor would have no effect at all. It is some considrable years since I did serious digital design work. But from memory..... its due to the internal inductance of what is likely to be a foil type capacitor (electrolytic). You must remember that the function being provided here applies to a very high frequency signal component during switching. The requirement is for a 0.1uF ceramic type probably (I can't look at the data sheet at the moment). This needs to be physically very close to the device, if you don't put it close to the device then the inductance of the connections / leads tends to swamp the effect of the capacitance and you don't get the desired effect of the decoupling capacitor.

Whilst you have considerd the TTL voltage requirements of the voltage input, have you considerd the effects of this network on the ouput from the device in question? Again I can't check the device specs at the moment but I seem to recall that this device has a low impedance output (18mA drive capability). The mismatch between the output and the potential divider due to the selection of the 600K Ohm load that you appear to be presenting to the device and the mismatch between this and the TTL input seem considerable (check this calculation but I think that your only allowing around 13 micro amps through the TTL gate). I suspect that this potential divider network has resistors way too large.

Again, I will ask the question about the gear profile, as I suspect that your gear profile does not meet the requirements and this could be the real source of your problems.

A description on decoupling from wikipedia
http://en.wikipedia.org/wiki/Decoupling_capacitor

Message Edité par Conseils le 09-23-2006 09:33 AM

[Fred_Flinstone]

Hello,

Sorry that I needed so long to replay.

I tried the to get the signal rise time to the specified maximum value of 50ns that National Instruments is referring to in the link that was posted with help of the LS7184 but that does not work, the pulse rise time measured on an oscilloscope is 200 micro seconds with the LS7184.

So I tried to use a Schmitt trigger 74hct14 this has a rising time of 20ns. But when the system is set up the rising time is 50 micro seconds.

The signal is well in the TTL range only the signal rising time I can’t achieve.

Any new ideas would be help full.

Regards,

Philipp

[Conseils]

It seems reasonable that if the signal generator works and your circuits don't with two different DAQ cards then it's probably an impedance mismatch between your circuit and the input DAQ card.

If the circuits aren't working, one thing that occurs to me is, have you adjusted / calibrated the scope probe correctly? Perhaps this is why you are not seeing the correct profile as at such fast rise times this adjustment will be important.

One thing that you have not mentioned is what hardware is between the DAQ card and your circuit.

I would indicate that I use exactly the same type of DAQ 6036E PCI cards and similar for almost identicle measurements, without issue. Although I have taken enourmous care over the conditioning circuits, power supplys, decoupling and screened connection wires. etc. In my instance timing errors would lead to measurement errors that would create an unacceptable noise floor or inability to correctly measure the signal. I sometimes use SCB68's with circuitry built onto the prototype areas http://sine.ni.com/nips/cds/view/p/lang/en/nid/4790.

I find that you have to be carefull what noise is pulled in through the connections and what earth loops can exists between any equipment and the computer. Earth loops are a particular problem as our equipment is often located in another room than that in which the test is occuring in. The equipment is often used off site as well, so we are not in control of earthing systems. To obviate these problems I often employ an opto coupler to opto isolate the test system from the computer system.

"The SCB-68 is a shielded I/O connector block for interfacing I/O signals to plug-in DAQ devices with 68-pin connectors. Combined with the shielded cables, the SCB-68 provides rugged, very low-noise signal termination"