Converting a Voltage signal to Vacuum reading

I have a vacuum instrument. The analog out is linear "with respect to the log of pressure".
It is a 0-10v signal being sent to a Compactlogix IF8 module.
The attached file shows the calculation for Torr, which is what I want to convert it to.
Ideas?? Could I just use a CPT instruction?

what the hell is wrong with PSI and inches of mercury?
/rant off...sorry.

EDIT: that rant was not aimed at you!...I reread that you want Torr, then yes, you should be able to calculate it with scaling math. You will also want to be able to adjust it if necessary when you first set up and verify/test the instrument.

Just scale your input to 0-10 as a REAL. Subtract 2.0 (for Pascal) or 4.0 (for Torr). Then plug the 'result' into an XPY instruction - XPY 10.0 'Result' Destination (REAL)

OkiePC said:

what the hell is wrong with PSI and inches of mercury?
/rant off...sorry.

EDIT: that rant was not aimed at you!...I reread that you want Torr, then yes, you should be able to calculate it with scaling math. You will also want to be able to adjust it if necessary when you first set up and verify/test the instrument.

Click to expand...

In regards to PSI or inches of mercury. This unit of pressure measurement is not really used in this industry (Vacuum heat treating) for processing aircraft engine components.
IE..
Torr (or Millibar overseas) has a far greater readability.
1 ATM = 760 Torr (1013 Millibar),
1 ATM = 14.69 PSIA
As you can see, there is far more readability compared to PSI.

In the industry of Vacuum heat treating, heat treatment is commonly done at 10-1 Torr range and lower.
Vacuum brazing is done in the 10-4 to 10-5 Torr range. (Above 8x10-4 Torr and the braze joint is no good)

This gauge (2 types used, known as Convectron or Pirani) is the UPPER level in pressure.
The LOWER Level of pressure reading is an ION gauge (2 types used, known as Cold Cathode or Hot filament type) and the range of that gauge is 10-2 to 10-10 Torr.
Capacitance Manometers are also used and they are VERY accurate but only cover 4 decades of Torr and are the most expensive. They do not read very well below 10-5 Torr and are VERY expensive in this range.
The best feature about Capacitance Manometers is they are "true pressure gauges", meaning they are not affected by the type of "residual" gas (Argon, Nitrogen.....etc) or water vapor in the system.

Even the pumping systems used to attain these vacuum levels are dependent on certain ranges of vacuum. It usually takes 3 types of pumps to get to the 10-5 range.

I could go on and on and on....

bernie_carlton said:

Just scale your input to 0-10 as a REAL. Subtract 2.0 (for Pascal) or 4.0 (for Torr). Then plug the 'result' into an XPY instruction - XPY 10.0 'Result' Destination (REAL)

Click to expand...

You may be close...but...the -4 is a decade measurement (exponent) so I believe I should multiply the voltage reading by .0001 (10-4) and that "result" becomes the exponent for the XPY instruction?

I just went by the equation in the digram.

Take the voltage and subtract and subtract either 4 (for Torr) or 2 (for Pascals).

Then use the XPY to raise to to the power of the resultant number.

E.g. 0 volts becomes -4 (assuming Torr) so the result is 10 raised to the -4 power

bernie_carlton said:

I just went by the equation in the digram.

Take the voltage and subtract and subtract either 4 (for Torr) or 2 (for Pascals).

Then use the XPY to raise to to the power of the resultant number.

E.g. 0 volts becomes -4 (assuming Torr) so the result is 10 raised to the -4 power

Click to expand...

Yep....
I think you are right.
Thanks..

Universal formula attached.

Tom - universal for linear but this has a log function also. So the extra XPY instruction is needed.

Bernie has provided the correct solution for this particular transducer. That is the same way we we do it with some of our sensors.

Vacuum instruments use a log linear output because of the large range over which they may read. A range of 100 million counts or more is not unusual, and typically higher precision is desired at extremely low pressures. So 1000 to 100 torr covers on volt of range, 10 to 100 is one volt of range, 1 to 10 is one volt of range. .1 to 1 is one volt of range, .001 to .01, etc.

Typically 1 TORR is normalized to some mid-range voltage, in this case it is 4 volts. You need to refer to the manufacturers specs because not all of them normalize to the same value. You subtract the normalized voltage from the actual voltage to get your true exponent. As an alternative to normalizing the exponent by subtracting 4 you can multiply the result by .0001, (.0001*10^V) it has the same mathematical effect.

OkiePC said:

what the hell is wrong with PSI and inches of mercury?
/rant off...sorry.

Click to expand...

Inches of vacuum is completely useless at high vacuum levels. Its fine for things like vacuum conveyors and vacuum lifters, but those operate at pressures that are still a million times higher than many vacuum applications. For processes that have diffusion pumps, turbo-molecular pumps, and cryopumps and vacuum levels approaching those of deep space it just doesn't provide any kind of useful information. (milli)Pascals and (milli)Torr are the preferred units of the high vacuum industry.

Our stupid silly intransigence over converting to SI units is seriously damaging the USA, especially in the sciences and technologies. [/my rant off] It even affects mechanics: they have to buy twice as many tools.

My mistake, Bernie - thanks!

TConnolly said:

Bernie has provided the correct solution for this particular transducer. That is the same way we we do it with some of our sensors.

Vacuum instruments use a log linear output because of the large range over which they may read. A range of 100 million counts or more is not unusual, and typically higher precision is desired at extremely low pressures. So 1000 to 100 torr covers on volt of range, 10 to 100 is one volt of range, 1 to 10 is one volt of range. .1 to 1 is one volt of range, .001 to .01, etc.

Typically 1 TORR is normalized to some mid-range voltage, in this case it is 4 volts. You need to refer to the manufacturers specs because not all of them normalize to the same value. You subtract the normalized voltage from the actual voltage to get your true exponent. As an alternative to normalizing the exponent by subtracting 4 you can multiply the result by .0001, (.0001*10^V) it has the same mathematical effect.

Click to expand...

You got it also.
I also have to multiply the Torr reading by 1000 to be able to show Microns (Millitorr). The customer wants to see it both ways??
The big issue is if you look at the chart it is "log" between the decades, so it is slightly off.
It REALLY ticks me off that vacuum instruments with such capabilities cannot give me an EXACT "true" linear analog out across the range.

ONLY Televac has a "true" linear analog output for their gauges, but even then, it is only across 3 decades.

Capacitance Manometers are also "true" linear out but they are expensive (VERY expensive if required readibility is below 10-4 Torr) and they only cover 4 decades, so you usually need more than 1.

When we put together a control system for the vacuum furnace we have, I decided to go with a sensor that had communications to get around this issue. We are using an MKS 910 as it covered the range that we needed.

The only issue I had with the sensor was that it seemed to lock up after a period of time. I'm not sure if it was because I was hammering it with comms constantly or what, but simply power cycling the sensor got it back online. I had the power to the sensor already controlled by relay, so I just wrote my program to watch for a locked up sensor and cycle the power to it when needed. This was several years ago, so they may have improved things since then.

Brian

I mainly use capacitance manometers when I have a process gas that I need to control precisely. The cap. manometer is not dependent on the gas molecular weight so its an ideal fit for that application but less than ideal for straight vacuum. For hi-vac I use convectron + ion gauges. Lesker has a new instrument that combines a convectron and ion gauge in a single gauge/controller combination - haven't used it yet but it has potential. I've used Granville-Phillips gauges with device-net interface and been pleased with the results. G-P is now a part of MKS. I've only used the Televac multichannel controllers but their stand alone gauges are intriguing.

I still haven't seen any manufacturers producing gauges that talk ethernet/IP. End users are tired of not being able to easily and accurately log vacuum data, but for some reason the vacuum industry has been ridiculously slow to clue in.