Mass Flow Controller

Current images of fusor efforts, components, etc. Try to continuously update from your name, a current photo using edit function. Title post with your name once only. Change image and text as needed. See first posting for details.
Post Reply
Chris Mullins
Posts: 67
Joined: Sun Feb 19, 2017 11:32 am
Real name: Chris Mullins
Location: Shenandoah Valley, VA
Contact:

Mass Flow Controller

Post by Chris Mullins »

I had been using a needle valve to deliver hydrogen to my cyclotron, but even at best it was very difficult to get a repeatable flow. The valve was too high flow in the first place - the usable adjustment range on the 10 turn valve was 1/8" turn on the high end, down to hand tight on the low end. Gradually it shifted so "hand tight" was even too much flow. Here was my original gas system, with needle valve and 1/4" copper tubing:
needle2.png
needle2.png (121.45 KiB) Viewed 5280 times
I found a bargain 10 SCCM (oxygen) MFC on ebay for $35, and grabbed it. To keep things simple I bought a $13 supply to generate +/-15V (https://www.amazon.com/gp/product/B0752TRXDC), and soldered some leads directly to the MFC edge connector to test with a protoboard and a 10K pot. I connected it directly to my diffusion pump and vacuum gauges, left the input side of the MFC open to atmosphere, and varied the control input from 0-5V. I also mapped out the valve open/close lines (short to ground opens the internal valve, stopping the flow), and adjusted the MFC offset trimmer:
mfc1.jpg
It seems to work as expected, so next step was to try it on the cyclotron, still in protoboard stage:
mfc3.jpg
It was immediately better than the needle valve, and for the first time I could get repeatable cyclotron runs on different days! Time to move from protoboard to final design.

Instead of a simple pot to adjust the setpoint, the final design uses ... a 10 turn pot. I added a 5V regulator for some stability, and added a cheapo ($2.40) LED meter to give a visible reading on the setpoint voltage. I found an old Radio Shack 75 ohm TV A/B switch, which I emptied out. That barely fit the power supply digital pot, switch to toggle the MFC valve, and power connector:
mfc_control.jpg
Bench test of the MFC controller, attached to the MFC:
mfc_built.jpg
The MFC uses VCR connectors, so I ended up switching away from Swagelok and copper tubing. The flexible stainless steel tubing is much easier to work with since I'm still moving things around fairly often. Here's the new setup:
mfc_installed.jpg
I'm still waiting on one more length of flexible tubing to go between the chamber and the MFC. That will get the MFC off the lab jack, and permanently mounted onto the wood base.

Even 10 SCCM is too high. My operating setpoint is between 0.05 and 0.15 volts. The MFC stops regulating around 0.03V. I'll keep an eye out for a 1 SCCM unit - those seem to be pretty rare on ebay.

My bargain $35 MFC is working very well, but it ended up costing a lot more to acquired various VCR fittings and tubing!
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Mass Flow Controller

Post by Richard Hull »

Some times a great buy on one end of things leads to a fortune spent to implement that fabulous buy. This is typical and not unusual at all when seeking out stuff on the cheap. We can fall into great deals that soon fade from greatness as the wallet empties trying to get them to work and mate with our vacuum systems. You have joined many of us in this mistaken path to costly betterment. In amateur vacuum systems this is as common as the sea sickness. Getting one's "sea legs" in vacuum systems can cost, but the lessons learned are well and truly learned.....and paid for.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Mass Flow Controller

Post by Jerry Biehler »

You can actually change the fittings on most MFCs. The swagelok or vcr fitting unscrews to a 9/16" o-ring seal fitting which is a pretty standard block fitting for swagelok.

I was going to say 10cc was going to be pretty big when I read it. 5cc is about the smallest I normally see.
User avatar
Rich Feldman
Posts: 1470
Joined: Mon Dec 21, 2009 6:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Mass Flow Controller

Post by Rich Feldman »

My inexperienced understanding is that 5 or 10 sccm
is the ballpark for full scale flow in the actual sensor capillary of a typical MFC.
In other words, when you buy a unit with that value on the nameplate, 100% of the flow goes through the measuring element.

So I guess a 1 sccm MFC would employ special design or operation of the thermal element.
Or maybe just special parts in the flow control valve, for a more delicate touch in the not quite all the way closed region.
All models are wrong; some models are useful. -- George Box
Chris Mullins
Posts: 67
Joined: Sun Feb 19, 2017 11:32 am
Real name: Chris Mullins
Location: Shenandoah Valley, VA
Contact:

Re: Mass Flow Controller

Post by Chris Mullins »

Richard: Yep, the bargain MFC ended up costing a lot more than first planned, as one expensive piece led to another. I was originally going to adapt from VCR immediately back to Swagelok and stick with my copper tubing, but one idea led to another .... I'm still very happy with the end result as long as I don't think about the extra cost, which could have been better applied somewhere else in the system

Jerry: Interesting, I didn't know that!

Rich: That makes sense, and would explain why the 1 SCCM controllers I see on ebay are so much more expensive.
User avatar
Rich Feldman
Posts: 1470
Joined: Mon Dec 21, 2009 6:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Mass Flow Controller

Post by Rich Feldman »

Chris's report prompted me to take apart, when I finally found it, a partial MFC that Greg Courville gave me "to see what's inside". It's got body, sensor, and valve actuator, but no electronics. Chris, if you prefer I'll remove this from "your" thread and start a new one, but you already took the ideal subject line. :-)
DSCN1012.JPG
This specimen's nameplate says 0-10 mln/min for H2+D2, which might be no coincidence since Greg had recently finished a fusor project. Anybody know if R2-D2 exists? The n after symbol for milliliters stands for normal (0 degrees C and 1 atm pressure). The "standard" in symbols like sccm might be ambiguous, at least enough to matter in trade, since IUPAC changed their definition of standard pressure from 1 atm to 1 bar.*

Here the ribbon-cable strain relief and sensor cover are out of the way. Half of the yellow foam is missing, because it got lost and the picture was taken during re-assembly.
DSCN1023.JPG
Now with the extrusion and front piece of foam out of the way, we see the sensor capillary. There's also what looks like a resistor, and in back view a thermistor (?).
DSCN1018.JPG
DSCN1020.JPG
I didn't trace the internal wiring while it was exposed, but made some electrical measurements later at the 1x5 pin connector.
Pins 2, 3, and 4 behave like a series-connected pair of 1.05 k resistors, I assume the upstream and downstream sensing coils. Pins 1 and 5 behave like an isolated 1.29 k resistor, which could include heater coil and both discrete components, not necessarily all in series. After applying about 6 mA for about 15 seconds, the resistance was down to 1.27 k. After the whole unit spent about an hour in the freezer, the 1-5 resistance was up to 1.324 k and the others were down to 0.964 k. Now, after 50 minutes out of freezer but still in plastic bag, 1-5 resistance is 1.294 k and the others are 1.006 k.

I am not ashamed to show a picture from near the beginning of the shoot, as evidence that there was a piece of yellow foam in front of the capillary.
DSCN1015.JPG
.
Who, besides me, is curious about what's under the big solenoid? How big are the valve orifice and pintle (?) for 0 to 10 sccm H2?

Last time we discussed "mass flow" in detail here, I finally got one thing straight. Learned that the thermal mass of 1 scc doesn't depend a whole lot on the specific gas. IIRC, thermal sensor corrections for gas type might range +/- 20% from the middle, even if there's a 100:1 range of molecular weights. But "one part fits all" won't work for flow restricting valves. Other things being equal, flow through an orifice goes down (in volume units) and up (in mass mass units) as the square root of molecular weight.

[edit] Found a definition of "normal", and yet another definition of "standard", in MFC catalog from Bürkert GmbH & Co. ...
normalstandard2.JPG
My advice to newbs: don't try to remember a universal meaning of "normal" or "standard" when talking about gas volumes. It's different in different branches of science and industry. What are the temperature and pressure when contracting to buy or sell a billion cubic feet of natural gas?

On my pre-edit point: If the proportional valve in Chris's MFC has parts sized for 0-10 sccm O2, the same solenoid current range would let 0-40 sccm of H2 flow, IMHO, since hydrogen is 16 times lighter. I expect that factor would NOT apply to the external flowmeter output or setpoint input. Compared to 10 sccm O2, 10 sccm H2 carries about the same amount of heat (which is what we sense and what we set)
with about 1/16 of the mass, and can flow through an orifice 4 times smaller in area with the same pressures. That's why most giant electric generators are cooled with H2 gas.[\edit]
All models are wrong; some models are useful. -- George Box
Post Reply

Return to “Images du Jour”