Hardware suggestions for a gas mass flow controller

[markmakeitso]

Saw your email address, is an option, but don't think we have anything particularly sensitive. Not a ton of LV content just yet, that might come later.

-Are Viton seals okay for both butane and nitrogen? One supplier we talked to was urging metal seals, although my research pointed towards good compatibility for both.

-According to process specs our target canister volumes (various sizes) determine a mass loading rate of butane, from which we can calculate a CCM flow rate, and that same CCM rate is then also used for nitrogen, and the two flows are mixed before entering a canister. This would generally require paired MFCs, one calibrated to butane, one to nitrogen, each with identical flow rate ranges. So a pair of MC 0-100 SCCM MFCs for example. Correct?

-Do you handle annual NIST recalibrations in house, or could you recommend a 3rd party shop to perform that? Idea on pricing if done in house? (feel free to opt out of those on the forum) Any suggestions towards making that process more convenient, besides making our physical system easy to disassemble?

That's what I've got for now. Appreciate the help,

Mark

[Kevin_Alicat]

Hi Mark,

Happy to help!

-Viton seals should be fine with both N2 and butane. I generally go through a list of several different elastomer compatibility charts from various companies (whatever I can find online) to get a good consensus since sometimes you will find variations where one company says it's OK but another states that it's just good or even poor. In this case, N2 is fairly inert and butane is listed everywhere I could find as being perfectly compatible with Viton (FKM).

-This sounds like a good way to go, though there's a small point that may be good to clarify. My only hesitation comes from a difference of units of measure, where you state CCM but are likely referring to SCCM. Small difference, but the SCCM (volumetric flow rate standardized what it would be at a given temperature and pressure) is more akin to counting the number of molecules passing through the system rather than the space that those molecules take up (which is what the purely volumetric flow rate, CCM, will measure). Two mass flow controllers one flowing 50 SCCM of butane and the other flowing 50 SCCM of N2 will produce a gas mix of 100 SCCM 50/50 N2/butane (the mass low percentages translate almost precisely to molar mix percentages). Mixes based on volumetric CCM flow rates will have a pressure and temperature dependence in the molar composition of the resulting gas. 

-Annual calibrations can be done on-site by the customer, through a 3rd party calibration house, or through the manufacturer's in-house calibration department (I believe most MFC manufacturers provide re-calibration services).

If done on-site, you will need to a reference that has an accuracy spec adequate to your needs, a means of communicating with the device, knowledge of how to alter a calibration (some MFC companies provide this info if requested), and lastly: to be OK with the calibration no longer being held to the specification provided by the calibration house/manufacturer (since you'd be taking it on yourself to ensure the calibration accuracy). This option is usually is done if there is no need for a specific certification of the flow (such as ISO 17025 certified, etc.) or if the accuracy needs to be just kept within a ballpark. The main benefit is the reduced calibration cost, as you would only need to keep a single portable meter up to calibration standards. The drawbacks are that the calibration is not certified and the device health cannot be monitored (if it goes back to the manufacturer, they can tell if the device is behaving oddly and pinpoint the reason - such as debris in the path/valve or damaged sensors.

3rd party calibration house option: Many calibration facilities have the tools to provide great calibrations and can usually certify them to specific standards if that's required for your documentation. We provide information on how to calibrate our devices to 3rd party calibration facilities if asked, so they should be able to tweak things if needed (some may perform a calibration check first and do no alterations if it is within the desired specification). Depending on their location, shipping may be easier.

Manufacturer's in-house calibrations: Offers additional bonuses, such as being able to determine if the device is behaving oddly (may need a large calibration tweak due to contamination or a damaged sensor, which can be found and brought to light before being fixed and sent back). They should know the best practices for modifying their calibrations. Our calibrations are NIST traceable and while I don't think I can provide specific numbers for pricing (it varies by accuracy requirement and flow range), I can say that they are fairly low. Feel free to email me if you require more specific pricing on various options, and I can get you in touch with our application engineers.

To make the calibrations go smoother, all I can think of are three things: installing in-line filters before the MFC's, potentially picking threads/ports that make life easier for you, and possibly getting a second set to swap out for a rotating calibration. In-line filters, such as a sintered brass filter, will help reduce the risk of debris entering the MFC and potentially clogging the valve or flow region, which can impact the operation or calibration. Threads/ports such as SAE or a Swagelok compression/VCR/VCO may be easier for you to install/uninstall if you end up needing to do so with in small spaces. Rotating calibrations are mainly useful if system downtime is more expensive for your operation than a second set of devices, and allows you to simply remove one set and replace it with another that is in-calibration. It is more costly, but I have seen some companies elect to go this route due to the requirement for 100% up-time being key.

Feel free to let me know if you have any additional questions.

[markmakeitso]

Hi again, thanks for the answers.

-Seal material: if you're happy with Viton I'm happy too.

-I've been sloppy with SCCM/CCM terminology, although I do largely understand the difference. Our system keeps the 2 source tanks nearly side by side, at relatively stable blue-collar lab conditions. The hope is that they would have nearly identical temps. I could monitor MFC-reported temps during operation and set faults if outside an acceptable window. An ambient temp sensor and additional inline pressure and temp sensors are also possible.

-The site manager definitely prefers 3rd party or manufacturer calibration over in-house, and the technicians know better than to fight for more work. Annual NIST is the requirement by spec. Our normal calibration shop doesn't appear to handle MFCs, so we might have to dig around a bit. I rather doubt we'd buy a second set of valves to swap, as reasonable downtime is probably acceptable. 

All good news here so far. Hoping we can order at least one MFC soon to start working with, and get the whole rig together soon after. Now that we're talking about an automated system dissatisfaction with the fiddly manual setup is rather increasing.

Have fun,

Mark

[Kevin_Alicat]

Hi Mark,

-Viton should suffice. It really only runs into troubles at low temperatures (somewhere around -10°C or so), if there are incompatible chemicals in the gas stream, or if you have a purity constraint that requires all stainless steel or metal seals (limits absorption of some chemicals, for cases like a gas analyzer that needs to pass through the full gas composition without loss).

-No worries! I've seen people use them interchangeably. The good news is that most MFC's will operate on a mass flow (SCCM) setpoint, which is the one you want to go with. The mass flow rates provided by the MFC's should already account for temperature and pressure effects. In the case of Alicat devices, we monitor the pressure and temperature of the gas (both output via serial as well) and convert a volumetric flow rate signal to the equivalent standardized flow rate (SCCM). Unless you perform calculations to go from SCCM to CCM or you select volumetric flow control on an Alicat device (offers both options), you will likely be controlling the SCCM mass flow rate by default on most MFC's.

-Glad to hear that down time is not so constrained for this project. That generally leads to a lot more stress in the event of needing to do any work/calibrations/repairs on the parts. Either 3rd party or manufacturer are good for you, and you may want to check on the estimated turnaround time for a recalibration from any of the places you find (as well as costs).

Automation for these types of applications is definitely nice! I'll be here as well for any questions that may pop up once you get the parts in. My experience with other devices is more limited, but if they're communicating over RS-232/485 via ASCII or Modbus commands I would assume the basics carry over fairly easily (and analog should further reduce any differences with LabVIEW interfacing).

One last thing to note: calibration and gas selection. Some MFC's are calibrated for a single gas, some allow a gas select change through a PC program, and some allow for gas selection to be done on the device directly (or through software commands without a PC program to load in new info/tables). N2 is usally the default calibrated gas on many devices, but the use of butane is something that you want to be sure that you have some way of either altering the gas select on the device or correcting for the calibration differences between the calibrated gas and the gas you are flowing.

[markmakeitso]

The quotes I received had some variation on gas selection. At least one manufacturer had an add on charge for anything not on the default gas list, which didn't include butane.

We had settled on Alicat MCs for this project, although I hesitate to turn this into a sales brochure post. The low end commandable flow rate, standard included screens, and multidrop RS232 were all pluses. Active LabVIEW support on the forum didn't hurt either honestly, and I certainly wouldn't expect such support if we had chosen another manufacturer.

Alright, good deal. I'm working on some UI issues with technicians today, and putting together a state machine framework. Rolling forward.

Thanks,

Mark

[Kevin_Alicat]

Glad to hear! I know other devices also work well for people, but I'm obviously a little biased toward Alicat. I'll be here for any added LabVIEW/Alicat/MFC support. Once you get to the point of working on the LabVIEW code, feel free to let me know if there are any suggestions you have for making any parts of it easier. It initially came about from a gracious customer who shared their work with us, and has since been evolving little by little to either add functionality or make life easier.

Cheers,

[markmakeitso]

Hi Kevin,

We're really close to ordering our first MFC. A few questions first though, if you don't mind.

-Our quote lists "Gas selection initially set to N2" and "Gas selection initially set to C4H8" (the latter I'm hoping a typo from butane's C4H10) for 2 0-200 SCCM valves. Just wanted to verify that the difference between 2 MC series MFCs, with the initial gas selection differing can be freely substituted for each other in our system, by simply using either the gas select command over RS232 or on the screen.

-Is there a price or functionality difference for DB9 ports instead of the 8 pin mini-DIN? I'm leaning towards DB9 instead of a BB9 setup, partly because we'll have several other 24VDC devices anyway, and could easily support the MFCs too. Also, the BB9 doesn't appear to support additional analog in or out, which might be useful later.

-Why the various DB9 pinouts on page 92 here? Are all available for MC series, or only certain ones? 

-We'd like to monitor our supply pressures (at least 1 each butane and nitrogen) before starting a process, and the pressure sensor readings over 232 would probably be adequate for this task versus a discrete sensor. We could have shut off solenoids either before or after (or both) on each MFC, to serve as positive cutoff for gas flow, but also to allow supply pressure readings before flow is initiatied. What's your recommendation for that setup? Upstream vs. downstream MFC valve positioning can also be changed to enable this?

Thanks,

Mark

[Kevin_Alicat]

Hi Mark,

-Sounds like the C4H8 vs C4H10 was a butene/butane dyslexic moment on the part of the person who made the quote. The good news is that all of the Alicat devices with selectable gases are actually calibrated on clean dry air (the gas selection loads different gas properties and pressure/temperature surface maps for the device to use). The gas selection is more of a convenience preset, as it is user-selectable through either the front panel or the RS-232 communication interface.

-DB9 allows a locking connection, while the 8 pin mini-DIN does not. Outside of that, the DB9 has a second pin wired to the common ground which might be useful in certain situations (analog outputs, power, and RS-232/485 grounds are all connected internally, and are also tied to the flow body/casing by default). You are correct that the BB9 does not allow for analog functionality. If you go the route of the DB9 style connections and are comfortable making your own cables, you can wire them up to share a single COM line with the analog I/O still accessible (something like page 12 of this manual: https://www.alicat.com/documents/manuals/Gas_Flow_Controller_Manual.pdf, but with the analog being wired separately).

-The various DB9 pinouts are mainly for allowing drop-in replacement functionality in systems that are currently set up for a different manufacturer's MFC's. They try to replicate the more common pinouts that a different type of MFC may use to limit the changes required by a customer when changing to an Alicat MFC. The male/female versions of the Alicat default DB9 pinout are also dependent on user preference.

-The standard Alicat MFC's have a default pressure output of 0-160 PSIA, and the measurement is tied to the opposite side of the valve. This means that if you want to monitor the inlet pressure, I would recommend having the device set up with a downstream valve position. There are a couple potential drawbacks to this option: 1) if the inlet pressure is too high, you could risk damaging the pressure sensor of the MFC, and 2) if the inlet pressure is fluctuating or changes suddenly (such as opening a solenoid to go from 0 PSIG to 100 PSIG very rapidly), the sensor is exposed to the pressure changes and can lead to either slight fluctuations in the flow (in the case of fluctuating inlet pressures) or potential damage (in the event of a pressure change going from 0-100 PSI very rapidly, due to the high differential pressure generated by this abrupt flow). For an inlet pressure that has a fluctuation caused by something like a compressor generating pressure pulses, you can damp out some of the fluctuations by installing a buffer tank between the MFC and the pressure source. Shutoff valves would be a good idea if you want to be ensure no-flow conditions or to act as a shutoff to allow for removing parts without needing to fully depressurize the entire system (the valves used on the low flow devices have a leak spec from the manufacturer of about 0.2 SCCM or lower of He at 90 PSIG, and we do test every valve to ensure that it is below this threshold).

For the pressure readout, would you be OK with absolute pressure or do you require a gauge pressure reading?

Best regards,

[markmakeitso]

Hi Kevin,

Thanks for the continued answers.

-Alright, that's the best possible answer for gas setup.

-Definitely sounds like DB9 is the way to go for our project then. BB9 would be super quick to set up, but some advantages are wasted here, and some limitations are fairly undesirable.

-Sounds good, I'll keep those recommendations in mind when we're mapping out our physical system. Probably at least a pair of the 5 MFCs should have downstream internal valves then. Our quote lists "custom calibrated full scale range of 60 PSIA," likely from our stated 16 PSIG operating point. It's a pressure regulated feed off of a tank, so hopefully it avoids several issues from more challenging gas sources. Gauge would be preferable, but we could probably get by with absolute measurements.

Thanks again,

Mark

[Kevin_Alicat]

Hi Mark,

No problem!

I would agree with the statement on DB9 connectors, as the only drawback is that you have to make custom wires but it provides more potential functionality.

Sounds good on the pressures, especially since 16 PSIG is not going to be anywhere near high enough to cause any damage to a standard controller. The custom range of 60 PSIA means that the 160 PSIA sensor will be gained up and recalibrated to offer a little more resolution on the lower end. This is optional, and if you would prefer to have a device that you can take above 45 PSIG at some point in the future, I would recommend letting them know so that it can be left at a higher range. For gauge pressure, there should be an additional option to add a barometric pressure sensor to the circuit board to allow for baro and gauge pressure readout (absolute pressure - barometric pressure). If approximate gauge pressures are OK, you could try subtracting an average barometric pressure from the absolute pressure reading, but if a storm comes in and shifts your local pressure you will not be accounting for this.