Constant Current Power Supply

[Brian2000]

A friend of mine worked for a company that went under and I was able to get my hands on this accelerometer:

http://www.omega.com/pptst/ACC786A.html

I want to use it to characterize a motor that might have a bad bearing and to in general play around with it. I have a simple voltage DAQ with a high enough sampling rate to pick up the frequency range I'm expecting the bearing to show in...but I'm having issues trying to find out how to make a simple power source for this thing...

The full description is in the link, but it was it requires a 2 to 10 mA current at 18 to 30V. I've seen some descriptions of using OP-Amps, but I'm not sure exactly what my power requirements might be so I don't know how to select alternative components to run at the voltage and currents stipulated.

Any help or advice would be welcomed.

 

[BobK]

Removed.

Bob

 

[Brian2000]

Removed.

Bob

Ye giveth and then ye taketh away =(

What was thoust wisdom?

 

[(*steve*)]

Ye giveth and then ye taketh away =(

What was thoust wisdom?

He probably gaveth you some bad information, then tooketh it away before you were misinformed and were lead on ye wilde goose chase

 

[(*steve*)]

From the specs, I would say you need a constant current source backed with a voltage between 18 and 30 V and set to between 2 and 10mA.

The output signal is variations in the voltage across the device (nominally 12V) and you would typically AC couple it to your ADC.

The sensitivity is 100mV/g, with a max of 80g, s the voltage across the device varies by up to +/- 8V (4V to 20V).

 

[BobK]

He probably gaveth you some bad information, then tooketh it away before you were misinformed and were lead on ye wilde goose chase

You nailed it Steve.

After posting something useless (I thought maybe the 2 to 10ma was the output signal), I looked around the website for the device.

They sell a "power supply" for $450 for this sensor. Power supply in quotes, because it connects to just 2 wire on the sensor and provides an output that os -5 to 5V based on the acceleration. How they do this is kept mysterious, and any speculation I might have would be worthless.

Bob

 

[Brian2000]

I've mostly been looking at two references to how the circuitry and reading might work:

http://www.new.mmf.de/iepe_standard.htm
The above is the only reference I could find for that class of accelerometers.

http://www.omega.com/manuals/manualpdf/M5097.pdf
Page 11 of that pdf shows (what I hope is) a similar accelerometer's innards.

Again I'm not %100 sure that this material is relevant, but I think it has put me in the mindset of what I need to be looking for...

Thank you again for your assistance in this matter!

 

[BobK]

So page 5 of the second document shows you the circuit you need. You need a 24V regulated supply, a constant current diode and a capacitor. I think you can make this for less than $450!

Here are some constant current diodes that should work:

http://www.mouser.com/Semiconductor.../Current-Regulator-Diodes/_/N-ax1ml?P=1yzrlyu

Bob

 

[KrisBlueNZ]

Nice work Bob! Literally a valuable post!

 

[Brian2000]

Thank you bob, I'm not quite sure how I overlooked that example...

So $450? The diode you linked is cheap, as would be the capacitor...I already have the sensor...

So would I be correct in guessing that I can't just rig up 16 AA batteries in series to make the '24v regulated power supply'?

 

[BobK]

Yes, it actually reccommended using batteries because they would be noise free. And at 2ma draw, they will last a while.

Bob

 

[Brian2000]

I guess more specifically...the $450 number...where did that come from? Kinda scared me.

If the diode is a few bucks, capacitors are cheap, the accelerometer I already have gotten for free and now I can use some AA batteries...

What comprises the hundreds of dollars?

 

[KrisBlueNZ]

What comprises the hundreds of dollars?

You could ask them that question. I'd be interested to see their reply.

That company does seem to be very focussed on marketing and hype.

Perhaps the extra cost is in the pure gold core inside the company logo on the front of the power supply?

 

[(*steve*)]

the $450 number...where did that come from?

The marketing department.

 

[Brian2000]

Ahh...so the $450 was Omega's offering...

I thought that Bobk was implying that the cost in parts would be in the ballpark of $450, well thank you all for your help. I'll let you know if this works out for me probably sometime tomorrow or Friday.

 

[BobK]

Ahh...so the $450 was Omega's offering...

I thought that Bobk was implying that the cost in parts would be in the ballpark of $450, well thank you all for your help. I'll let you know if this works out for me probably sometime tomorrow or Friday.

Yep,

And I did say "a lot less than $450."

Bob

 

[Brian2000]

Just an update, I've got the sensor hooked up and collecting data. Lots of noise...time to start finding the sources, but since I'm looking at the frequency domain I'm not sure I can just look at a high or low pass filter since I want all frequencies from ~0-1000Hz which is my limit based on m max sampling rate anyway.

Thanks again for your help!