Cold gas thrusters fire only after a short delay due to solenoid activation dynamics. The delay depends on the thrusters but I've seen numbers ~ 5 ms.

What I'm wondering: Is there a similar delay when deactivating the thrusters? It seems reasonable that a solenoid delay would affect both the activation and the deactivation of a thruster, but I don't know for sure---and I really don't want to take a blind guess.


  • 1
    $\begingroup$ I think by that time you're getting way out into the weeds of transfer functions and step response that isn't neccessarily easily described with a simple "delay". Controls theory -- here be complex-number nonlinear dragons! $\endgroup$
    – ikrase
    Commented May 16, 2021 at 1:35
  • $\begingroup$ Sure! I'm only trying to understand this qualitatively for now. There is an activation lag when firing a thruster. Is there also a deactivation lag when ending that firing? In terms of modeling, yeah, I would love to have those thruster transfer functions. Do you know where I can find them or at least read up on how to determine/estimate them? $\endgroup$
    – user39728
    Commented May 16, 2021 at 4:11
  • $\begingroup$ I think it might be better modelled as a minimum thruster pulse width / minimum thruster impulse in many cases. I am afraid I don't really know. $\endgroup$
    – ikrase
    Commented May 17, 2021 at 5:43
  • $\begingroup$ Yeah, I've seen it treated that way in papers I've come across. I was trying to model the firings with pulse-width modulation, and figured I could add in the activation lag plus rise time to the model... but that slowed simulation a lot... so now I'm taking a simpler approach: modeling the thrust force as a linear ramp with minimum corresponding to that minimum impulse bit and max at 100% duty cycle. Although it still would be nice to capture the activation delay and rise time there... $\endgroup$
    – user39728
    Commented May 17, 2021 at 5:50
  • $\begingroup$ And using transfer functions to capture the transient dynamics of the thrust signals would affect both the rise and "fall" time of those signals. I can work around this, but I guess my question was "should I?" $\endgroup$
    – user39728
    Commented May 17, 2021 at 5:53

1 Answer 1


This question is a yes/no type. The answer (TLDR: "probably not a similar delay") is going to depend on the nature of the circuit and the valve. From an attitude control perspective, it is likely that you will need to model the centroid of the delivered thrust pulse as being later than the middle of the electrical pulse.

Lets imagine that the thruster is operated with a flow control valve that is monostable, i.e. an electrical pulse has to open the valve against a spring and at the end of the pulse the electrical command pulse is removed and the valve closes under the spring action. Lets say that the logical command is a top-hat function with near vertical sides.

  1. The rise and fall of the electrical pulse that is actually transmitted to the thruster valve could be quite different. Imagine that there is a logical command pulse within an onboard computer and a drive circuit that can deliver current to open the valve. The latter electrical pulse isn't going to be as clean a top hat function at all and may not have completely vertical sides. However this is within the control of the circuit designer and probably any deviation from the logical pulse will be minimal compared to what comes next.
  2. The flow control valve has to physically move, i.e. accelerate, the valve poppet/other feature so that it moves to fully open. There are different designs around some of which may have less inertia than others.
  3. The propellant flow will rise to maximum pretty quickly following the movement of the valve seat for a small thruster. Just as an aside, for a thruster involving combustion there will be another rise time associated with the combustion building up to maximum temperature, particularly for a catalytic thruster.
  4. At the end of the logical pulse the current collapses from the valve opening coil and the poppet accelerates back to the valve seat.

If we assume items 1 and 3 are too small to be of interest in a cold gas thruster then I suggest that item 2 is not going to be the same as item 4 because of the different physical processes involved.

I'd welcome contributions from those with some figures to contribute. If you look up web pages for common thruster valve manufacterers, e.g. Moog, Vacco then maybe they will have some valve performance figures.

  • 3
    $\begingroup$ I'm not a valve engineer, but I wonder if the opening occurs faster, as the pressure is initially holding the valve stationary, and as it opens this friction decreases and the pressure will help push it open once it starts opening. However during closing both these forces work against the closing action. So I guess closing would take longer? $\endgroup$
    – Innovine
    Commented May 16, 2021 at 11:40
  • $\begingroup$ @Innovine This sounds quite plausible to me though, likewise, I'm not a specialist in that topic. $\endgroup$
    – Puffin
    Commented May 16, 2021 at 12:14
  • $\begingroup$ @Innovine I just stumbled upon this and thought of a bit more of an explanation as to why I'm hesitant. This could apply if the valve poppet is a barrel guided by a cylinder. There are other designs where the moving poppet is supported instead by bellows or on a lateral arm, in these cases the valve designer can change the shape of the moving part so as to influence closing or opening forces. $\endgroup$
    – Puffin
    Commented Feb 13, 2022 at 11:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.