Based on experiments with N2O a while back, we found that the
temperature inside a tank with a cooled liquid - in this case not
cryogenic in the real sense, just very cold - did not change a lot
between the vapor phase and the gas phase. Its going to be cold inside
the tank no matter what, and the walls will also work to equalize
temperature between your optocouplers. In a tank made of e.g. aluminum
everything is pretty much same low temperature.
For the temperature shift to be noticeable in a short time frame, i.e.
without so much delay that it skews the calculated predictions, the
sensors must obtain heat from somewhere fast. If you think of old air
speed sensors, they deliberately added heat to the sensor by running
current through the sensing wire.
Further, some liquids will be in indeterminable state, notoriously for
N2O which has a soft boundary between liquid and gas phases. Add to this
sloshing, etc. and your optocouplers may not be so accurate after all.
The capacitive sensor has the advantage of being a "straw" which cancels
out some of these effects. We never got to test it, but we had a
suspicion that it would in fact work for N2O at the critical gas-liquid
transformation stage, since the capacitance is a function of the
dielectric constant, and that this would likely correlate to the N2O
density in some function. Capacitive sensors are not hard to make.
My personal afternoon thought would be to measure the acoustic resonance
of the tank. Resonance frequency would be a function of volume, density
and pressure/temperature. Temperature and pressure can be measured
accurately, and density derived from that, leaving free volume to be
determined. With some empirical testing this could conceivably also be
made to work in zero gravity.
Lastly, some people use ultrasonic distance sensing from the top of the
tank down to the surface. There are Arduino kits for this used in
robotics, but they are probably not cryo rated ;-)
/H
------ Original Message ------
From: "Pierce Nichols" <piercenichols@xxxxxxxxx>
To: arocket@xxxxxxxxxxxxx
Sent: 9/14/2015 5:16:20 PM
Subject: [AR] Re: Semiconductor cryo level sensing
I like the idea of connecting them all together to get a single reading -- much simpler wiring. However, it seems to me that it would be better to wire them in parallel, so one or a couple of dead diodes don't kill it dead. That also reduces the peak voltage from ~n*Vf to ~Vf.
-p
On Mon, Sep 14, 2015 at 5:08 PM, Ben Brockert <wikkit@xxxxxxxxx> wrote:
On Monday, September 14, 2015, Pierce Nichols <piercenichols@xxxxxxxxx> wrote:Nifty idea! My understanding is that all plain LEDs shift their wavelength with temperature because the size of the band-gap is temperature dependent. It seems to me like the same physics would affect the receive side of the optocoupler as well... so it's not clear to me that you'd actually get the response to expect.
That would certainly kill it if true. I'll have to get some ln2 and try it.
However, the current through a diode is also temperature dependent, which points the way to a more elegant implementation. Perhaps you could just have a line of plain diodes and sense the change in current through each one as the liquid level passes it. Thoughts?
If it's a large enough change then you could have a bunch of them in serial and have basically the same analog sensor as centaur but with a current change rather than capacitance. Easier to turn into something a uc reads.
-p
On Mon, Sep 14, 2015 at 2:14 PM, Ben Brockert <wikkit@xxxxxxxxx> wrote:One of the few disadvantages liquids have over solids is propellant depletion. A solid burns nearly all its propellant, though some of it tends to be at lower than nominal pressures during the tail off. For a liquid stage to get maximum impulse it is imperative that it deplete both propellants simultaneously. The last kilogram of propellant is the most important it. A real world example of this is that Stiga flew to ~90km but had 30lb of fuel remaining after it burned out the LOX; with equal depletion it would have reached roughly 120km.
Centaur accomplishes this with capacitive sensors in the propellant tanks. A rod is inside and electrically insulated from a tube that is open at bottom and top. This forms a capacitor. As the level of the propellant goes down, more of the capacitor contains gas than liquid and its capacitance changes. This is used as an input to the mixture ratio controller, which drives a harmonic drive on a throttle valve to slightly tweak the mixture. With this control a stage carrying thousands of kilograms of propellant is able to deplete down to tens of kg, helping give the stage it's incredible (and consistent) performance.
One downside of this system is that it is analog and it requires some extract structures and mass to hold up the tube and wire, and keep them electrically isolated.
Cryo and electronics geeks know that some types of light emitting diodes (LEDs) change their emission wavelength when they are cooled to cryogenic temperatures.
An optocoupler is a solid state semiconductor device essentially composed of an IR LED mechanically coupled to an IR transistor. The two components aren't electrically coupled, so it's a way to send information between two isolated DC systems without them needing to share a ground. I used optocouplers to run some of the solenoids on Xombie and Xoie; it was a convenient way to amplify a TTL signal to coil current while isolating the noise associated with quickly turning on and off inductive loads.
Background complete. You could make a chain of optocouplers inside a cryogenic tank, with them turned on all the time. It would consume just a few mA per sensor. If the wavelength alteration trick works on IR LEDs, the optocouplers would only pass a signal when they're above the level of the cryogen.
By watching the timing as the sensors toggle, you can then measure the depletion rate of the propellant in the tank and adjust mixture ratio accordingly. With SMD optocouplers, the whole thing could be quite light.
There is a similar concept with small heaters coupled to temperature sensors, where sensors that get warmer are above the liquid level. But I think my concept would put less heat into the tank and has digital outputs, rather than analog.
So that's my idea of the afternoon.
Ben
