Current sensing using proportional valve 3

[markus506]

Hi everyone. First post here, and not even sure if I'm in the correct forum.

I am currrently working on a new product, whereas I am controlling a proportional valve using a PWM output. The PWM is controlled by a dsPIC. The actual PWM driver is an ST VND600SP. I am using the current sensing ability of the VND to determine output current. The problem I am having is how do I calculate the actual output current to the coil based on the PWM signal? Basically, without knowing an exact coil resistance, how do I figure out my output current in order to compare it to the sensed current to determine if any compensation is required?

Best regards,

Mark

 

[itsmoked]

I think you should consider not trying to read the current. Just run the valve open loop on a timed basis.

Why?
1) Because measuring the current in that heavily inductive RLC circuit is going to be a monumentally big headache.
2) Aren't you actually trying to control flow of something? Well the flow is going to change at a different rate than the openness/closedness of the valve so trying to find the actual valve position is not going to help you control the flow much. (Certainly not worth the hassle you are embarking on!)

I would suggest that you profile the valve in your system by setting a bunch of fixed PWMs into the valve and actually measuring the steady-state flows that occur. Then graph this and finally tabularize it. Then you can quickly (very quickly with that dsPIC) interpolate from your look-up table what PWM you need to control your fluid.

Take lots of measurements near the closed position.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[markus506]

Thanks for the response Keith. One thing I didn't mention is that the user will have a device with an analog joystick which he will use in order to vary the speed of the system, or 'openness' of the valve. Also, there is not one valve that will be used, as customers may use different valves. The coils will likely have very similar characteristics, but since they can be from different manufacturers, there will be a difference.

I feel the only way I can control this valve is to read back the current and compensate. At higher currents (over 1 amp), the resistivity of the coil changes fairly quickly, thus the current through the coil drops fairly quickly. The monitoring needs to be fairly precise, since these valves tend to be fairly sensitive to current changes (i.e. large flow changes).

 

[itsmoked]

Hi markus.
I still think you are not going to have the control you are hoping for. Different valves just makes the 'current' method more complicated yet!

Is there any way you could put an orifice and a differential pressure transducer in the line? This would give very accurate flow and not care about the valve style or drive vulgarities at all.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[markus506]

Hmm...interesting. Are you talking of reading the pressure back to the micro and compensating that way? i.e. if the differential was a certain size, compensate by increasing the PWM duty cycle. So, where abouts would the 2 sense lines from the transducer be connected? (forgive my ignorance regarding these transducers)

 

[markus506]

I probably should have read your post better. You would put the orifice after the valve, then make one connection before the valve and one connection after the orifice?

Mark

 

[waross]

Hi markus506 and itsmoked.
How about a low value series resistor and measure the voltage drop? Feed the output through a bridge rectifier circuit to a resistor-capacitor circuit with a time constant calculated to smooth the pwm frequency.
If you need isolation wind a small C/T.
respectfully

 

[itsmoked]

Hi again markus.

Yes! That is a very common "cheap" method used for flow measurement. Any orifice will cause a pressure drop exactly like a resistor in an electrical circuit. And just like a resistor/orifice, the greater the current/flow the greater the voltage/pressure drop. Mind you it doesn't have to be a large pressure drop to get you some nice feedback. Sort of like waross' suggestion except measure the actual thing you want to control not something that only marginally affects what you are controlling.

Measuring the actual flow can let you report totals and other things that are not available by looking at the coil current. Things like, "is there any flow at all? Or is a tank empty questions?"

You can either use two different flow sensors one on each side of the orifice or you can use a differential sensor (which has two ports one on each side of its diaphragm). The individual sensors can tell you the actual system pressure whereas the differential sensor cannot.

In bigger pipes they call the orifice a restrictor plate. They use this method in a whole lot of places.

If your system can utilize this technique you will be far happier with your control results. (dare I say Happy as compared to Not Happy)

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[markus506]

Hmm..seems like a neat idea...however, from some of the prices of differential transducers, that option is likely out! I will look into individual flow sensors, but it looks like I may have to balance out system accuracy versus price at this point.

Thank you Keith and waross for your input. Most helpful.

Mark

 

[benta]

If you want to measure the current, why not measure the DC current supplied to the whole circuit? I can't imagine that the current consumed by the PIC and the internal VDN600 current can skew this a lot.

Benta.

 

[markus506]

No, it probably wouldn't Benta. The problem I have with measuring current is the ever changing resistance of the coil. For instance, the user immediately uses full output and has 1.5 A flowing through the coil. If this is held on for 1 minute, the current would conceivably reduce to 1.25 A. If the user then left the system alone for 1 minute, then used the system full on, the coil would not be fully cooled and the output current might be 1.35 A. Thus, I am always going to be compensating against new current outputs versus 'true' current output (through a cold coil). Thus, without being able to record the cold coil current output for each PWM output, I don't see how I can reliably compensate the system using current feedback so the user can expect to achieve the same solenoid position with a certain joystick position time and again.

Ugh. This is becoming more problematic the more I think of it. I also don't know if I can use the transducer approcach because of cost/lack of inputs to the system.

Mark

 

[waross]

Check out omega.com
they have a wide range of transducers at good prices.
respectfully

 

[markus506]

Thanks for the tip waross.

I'd like to say that I am extremely impressed and grateful for the help I have received here. Being a fairly new engineer (almost 2 years field experience) I am always looking for tips/opinions from more experienced people. Once again, thank you to all the contributed to my first thread!

 

[itsmoked]

Waross I beg to differ with you. Omega's prices are truly horrendous. BUT! They do have a nice selection. They only re-brand then stack on large mark up. You can always find the same product with the original maker's markings for less. (Well 'always' may be the wrong word as sometimes you just can't figure out who the maker is.)

markus; I re-read my last post. I said, [blue]"You can either use two different[/blue][red] flow[/red][blue] sensors one on each side of the orifice....."[/blue]. [red]Flow[/red] should have been pressure! Was thinking about flow so I slipped up there.

Questions:
What is the fluid you are expecting to be controlled?
What is the pressure it will be at?

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[waross]

Hi itsmoked;
Horrendous is relative. I am usually comparing Omega prices to industrial instrumentation. I am usually comparing with outfits like Foxboro, Rosemount, Honewell, Partlow and similar companies.
I look at their prices and then look at the Omega price and usually stop there.
Thanks for your comments. The next time I need instrumentation I will consider searching a little more.
And I bet you a cup of coffee that you have the Omega library on your shelf. If not, get it, it's free.
Respectfully

 

[itsmoked]

Yeah my back is healing from the last time I hoisted part of it.

And if your basis is the likes of Foxboro, Rosemount, Honewell, Partlow... I guess I see your point!!

Omega isn't a bad place to go if you just need "one". It does save time hunting around...time is money.. etc...

I'm still ticked at them for a shipping blunder. I needed something for a trade show in the middle of the UPS strike. They made it very clear that things shipped overnite were not subject to delays. So I made it very clear I wanted it shipped overnite. I would pay umpteen extra bucks.. etc.. Then they promptly thru it in the ground bin. Took weeks to get it. Then the sales guy was utterly unapologetic.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[markus506]

I will be controlling hydraulic fluid. Probably in the range of 2000 - 2500 psi.

I noticed that slip of the keyboard regarding flow and assumed you meant pressure. Thanks for clarifying though.

Mark

 

[itsmoked]

Ouch 2000-2500psi...
That does run the price up. ~ $70 times two. Probably too much money..?

Low pressure sensors for air and water can be really cheap $5 kinda things.

If $70 won't break the bank and the system is pretty static you could probably make assumptions and then use just one pressure transducer and no restriction.

Here's a suggestion that is wacky but may actually work pretty well depending on your system.

You could put a strain gauge or gauges on or around a hose and measure the hose diameter which will change with pressure. I keep grasping for something because valve position is just so poor a relation to flow and a novel approach to something, can win you a lot.

Still brainstorming... I watch hydraulic hoses squirm and change position according to what's going on all the time. What about a tilt sensor on a hose that relates to pressure..

Or back to the valve position. Any way you can observe the valve position? Use an armature follower or something? Then you could just control to the actual position.


Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[waross]

I like the valve positioning suggestion. If there is any way that the operating position of the valve can be accessed you may be able to fit a throttle position sensor to measure valve position.

Still brainstorming... I watch hydraulic hoses squirm and change position according to what's going on all the time. What about a tilt sensor on a hose that relates to pressure.. It worked for Bourdon, it should work for us.

Another thought.
Impedance versus resistance. At even moderate frequencies the inductive reactance will probably be high in proportion to the resistance. The higher the ratio the less effect the coil resistance will have on the current.
A transformer will serve the double duty of providing AC and increasing the voltage to the higher level that will be needed by the higher impedance.

By using a higher voltage and frequency AC current you can reduce the effect of the change in resistance of the coil with heat.
Whether you use PWM, duty cycle or a combination, a given output should give good repeatability on valve position.
Comments Keith?
respectfully

 

[itsmoked]

Hmmmm Not sure waross, how much that would help.. I agree if you are PWMing fast then the inductive component should dominate.. But the R part will remain..

I think looking at that part markus.. Its Current Sense outputs are horrendous.. +/- 10%!! Varying with time of day and position of moon.

You might have more luck using a +/- 0.5%:

http://pdfserv.maxim-ic.com/en/ds/MAX4172.pdf

Then either average the current reading in hardware or software.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>