Author Topic: SpaceX Falcon 9 v1.1 - SES-8 - DISCUSSION THREAD (Read 611299 times)

Okie_Steve · « **Reply #160 on:** 10/03/2013 05:15 pm »

Even automotive cam/crank position sensors are not what you would call large, and as for reliability automotive component rating is not much below aerospace in terms of harsh environment. So, I think "very small" and "very reliable" fills the bill.

kevin-rf · « **Reply #161 on:** 10/03/2013 05:24 pm »

Why are we debating if the telemetry captured the RPM of the pump? One would assume that is a critical parameter that they always measure and is available to them. Just because they said the engine tripped on pressure does not mean they do not have the pump RPM data.

I can not see how you would man rate the engine without some way to detect turbo pump overspeed and shut down and gracefully abort before the pump gets into a RUD situation.

ugordan · « **Reply #162 on:** 10/03/2013 05:26 pm »

Quote from: kevin-rf on 10/03/2013 05:24 pm

Why are we debating if the telemetry captured the RPM of the pump?

The Forum works in mysterious ways...

LouScheffer · « **Reply #163 on:** 10/03/2013 05:31 pm »

Quote from: Lar on 10/03/2013 04:42 pm

Wouldn't pump RPM require some sort of tachometer on the shaft? With concomitant weight, reliability, and power consumption (shaft power) issues compared to a pressure sensor? How small and how reliable can you make a tach? I understand tachs are done using magnets bonded to the shaft (extension) and stator loops with current sensors, but I may be confused, so there's no actual physical contact but still...

Here's a case where a pump shaft sensor, installed as an instrumentation upgrade to diagnose why the pump was catching on fire, itself caused the pump to catch on fire. On the brighter side, it was possible to diagnose this problem from the extra instrumentation, presumably including the sensor that was at fault.

http://www.enginehistory.org/SSME/SSME5_2.pdf

Here's a paper on 14 ways to make this measurement. It explicitly says the SSME reads and controls this parameter.
http://archive.org/details/nasa_techdoc_19850018597

Antares · « **Reply #164 on:** 10/03/2013 08:59 pm »

Hall effect sensors are unreliable for health checks. You could make them redundant, but there are sensor failure modes (where an engine is otherwise healthy) that could take out all redundant HE sensors. Pressure is a better choice.

To me extended spin start sounds like the spin-up helium flowed too long or a check valve (which is necessary for a flow path transitioning between sources) didn't check and GG gas went in the wrong direction.

baldusi · « **Reply #165 on:** 10/04/2013 01:52 pm »

I think that you don't use pressure transducers, but temperature sensors and make your calculations backwards from there.

Antares · « **Reply #166 on:** 10/04/2013 03:25 pm »

Depends on what you're looking to measure and check. Rapid, large magnitude temperature changes are suitable for health checks (ambient to cryo, ambient to combustion), but it's not going to work well for other things. Heat transfer takes much longer than the application or removal of pressure.

savuporo · « **Reply #167 on:** 10/04/2013 04:27 pm »

Quote from: Antares on 10/03/2013 08:59 pm

Hall effect sensors are unreliable for health checks.

Huh ? Cite ?
There are probably hundreds of millions if not billions of HE sensors out in critical applications working every moment. In some cases they are redundant but not due to inherent fault with the sensing mechanism, and in a lot of apps they are complemented by other sensors for readings, but this is the first i hear anyone make a statement of HE sensors being "unreliable".

Antares · « **Reply #168 on:** 10/04/2013 08:22 pm »

I spoke too generically. I wouldn't use them in a rocket turbopump application and certain other rotating machinery, where axial shifts and some radial modes of the shaft could affect the precision of the measurement.

Robotbeat · « **Reply #169 on:** 10/04/2013 08:27 pm »

Quote from: Antares on 10/04/2013 08:22 pm

I spoke too generically. I wouldn't use them in a rocket turbopump application and certain other rotating machinery, where axial shifts and some radial modes of the shaft could affect the precision of the measurement.

What difference would that make? You're just looking at the frequency of the Hall Effect sensor output, you don't actually care about the amplitude (other than simply having enough amplitude to detect a frequency).

mlindner · « **Reply #170 on:** 10/04/2013 08:28 pm »

Quote from: Antares on 10/04/2013 08:22 pm

I spoke too generically. I wouldn't use them in a rocket turbopump application and certain other rotating machinery, where axial shifts and some radial modes of the shaft could affect the precision of the measurement.

I'm not sure how axial shift could cause measurement precision issues here unless we're talking about _a_lot_ of shift, enough to get out of range.

savuporo · « **Reply #171 on:** 10/05/2013 04:33 am »

Quote from: Antares on 10/04/2013 08:22 pm

I spoke too generically. I wouldn't use them in a rocket turbopump application and certain other rotating machinery, where axial shifts and some radial modes of the shaft could affect the precision of the measurement.

I'd like to see even a single reference here, especially if we are talking about simple RPM measurement.

You are of course aware of things like these ( just an example, these are COTS solutions from myriad of vendors, sub $2 range ) :
http://www.renishaw.com/en/magnetic-rotary-encoders--9801

These things normally give you 10+bit ABSOLUTE shaft position and are impressively immune to magnet placement, non-aligment and gap size.
RPM measurement is super trivial compared to this.

laika_fr · « **Reply #172 on:** 10/05/2013 08:48 am »

this thread will have to be cleaned up again

Enlightening paper about the ongoing showdown between SES and insurers :

http://www.spacenews.com/article/launch-report/37547ses-approves-satellite-shipment-for-falcon-9-launch-despite-questions

cambrianera · « **Reply #173 on:** 10/05/2013 11:09 am »

Quote from: laika_fr on 10/05/2013 08:48 am

this thread will have to be cleaned up again

Agree completely

Quote from: laika_fr on 10/05/2013 08:48 am

Enlightening paper about the ongoing showdown between SES and insurers :

http://www.spacenews.com/article/launch-report/37547ses-approves-satellite-shipment-for-falcon-9-launch-despite-questions

Enlightening indeed, thanks!

kevin-rf · « **Reply #174 on:** 10/05/2013 11:13 am »

Personally if I was a GTO customer I would now demand SpaceX demonstrate an in orbit restart before risking my bird. The easiest way would be to shuffle the schedule moving CRS-3 up, add a restart demo to the mission, then fly SES-8 and Thiacom 6.

Be interesting to see what SES does, If they keep the schedule, I would bet they will amend the policy and pay a higher premium.

I wonder if to satisfy the insurers restart analysis will cause an additional slip...

ugordan · « **Reply #175 on:** 10/05/2013 11:17 am »

Quote from: kevin-rf on 10/05/2013 11:13 am

The easiest way would be to shuffle the schedule moving CRS-3 up

It can't be moved up because Dragon and its upgrades won't be ready until the Jan/Feb timeframe.

Jakusb · « **Reply #176 on:** 10/05/2013 11:18 am »

Quote from: ugordan on 10/05/2013 11:17 am

Quote from: kevin-rf on 10/05/2013 11:13 am
The easiest way would be to shuffle the schedule moving CRS-3 up

It can't be moved up because Dragon and its upgrades won't be ready until the Jan/Feb timeframe.

And CSR-3 demands 3 successful flights.

Jim · « **Reply #177 on:** 10/05/2013 11:19 am »

Quote from: Jakusb on 10/05/2013 11:18 am

And CSR-3 demands 3 successful flights.

Where is that documented?

Ben the Space Brit · « **Reply #178 on:** 10/05/2013 01:13 pm »

Quote from: Jakusb on 10/05/2013 11:18 am

And CSR-3 demands 3 successful flights.

I'm not sure it 'demands' anything but Gerst said he wants two or three successful flights before NASA approves v.1.1 for use.

Regarding the restart issue, the fact that SES have shipped the bird suggests to me that their technical people are satisfied with SpaceX's diagnostic and proposed fix for the issue. They wouldn't risk their payload otherwise. But, yeah, the insurers are probably not going to be happy and will likely demand higher premiums or, possibly, an Excess (where they only cover a claim, minus a penalty that must be paid by the claimant).

rds100 · « **Reply #179 on:** 10/05/2013 01:52 pm »

So isn't there any cheaper payload that is ready to be sent and can be sent first?