Ken,
I can totally understand how the discussion gradually shifted from the
context of a low-end solution to measure this test to a more generic data
acquisition discussion especially when you have contributors like myself saying
things that aren’t quite 100% spot on 😊.
Contributions from experts in the field is ultimately a good thing for the
knowledge base.
Troy
From: arocket-bounce@xxxxxxxxxxxxx <arocket-bounce@xxxxxxxxxxxxx> On Behalf Of
roxanna Mason
Sent: Friday, February 24, 2023 1:17 PM
To: arocket@xxxxxxxxxxxxx
Subject: [AR] Re: Subject: Re: S-26.500 KNSB Sugar Shot test tomorrow
There are a few exceptions for the need of high sampling rates would be to
analyze a flawed grain;
a sample rate in the 100's to 1KHz would help.
But, as it was stated, this wasn't a motor development program, just a cast and
burn.
Ken
On Thu, Feb 23, 2023 at 2:59 AM Ben Brockert <wikkit@xxxxxxxxx
<mailto:wikkit@xxxxxxxxx> > wrote:
The tiniest of caveat for people thinking of doing a DAQ for the first time:
loadcell amplifiers and pressure transducers have bandwidth specifications, and
they're often much lower than what you can get the A/D to do. So while you can
capture the signal at 20khz (40ksps/2), there's a good chance that there's only
100 Hz of actual data in the loadcell and a few khz in the pressure transducer.
(And 5Hz would be incredibly fast for a thermocouple amplifier.)
That can be used to your advantage for reducing noise, in that oversampling and
averaging is an effective way to increase the precision of the measurement.
Basically trading resolution on the X axis for resolution on the Y.
If you ever need 10khz of real data, you have to have that spec in mind through
the process, and it leads to strange pressure transducers and tiny
accelerometers.
This part is currently a stretch if you're not a computer person, but loading
the data into a time series database like Timescale or Influx and then doing
your data analysis in Grafana can make that process fairly quick and
repeatable, as you can configure the averaging in an interactive plot, and also
do things like integrating vs. time to get total impulse. An Arduino or analog
shield feeding data into a Pi running that software is fantastically powerful.
You could use that idea as the data visualization side to launch an orbital
rocket from practically another country, for example.
On Wed, Feb 22, 2023 at 8:55 PM Terry McCreary <prfesser@xxxxxxxxxxx
<mailto:prfesser@xxxxxxxxxxx> > wrote:
Cheaper, maybe. Easier...I don't think so. Dataq's starter units require no
programming. None. The free WinDaq software is quite versatile and has worked
well for me for 25 years. I use a DI-1100 right now ($80 but out of stock; damn
chip shortage). It's powered from the computer's USB, has 12-bit precision with
four differential inputs. Using just one input permits up to 40ksps---literally
almost 100 times faster than I use for routine thrust measurement. No auxiliary
parts needed, no wiring other than connecting to the load cell amp or other
transducer---and they supply both the USB cable and a screwdriver. Don't even
need an enclosure.
I guess it depends on how you want to spend your money and your time. And on
the value you place on your time. The time I've saved by using a turnkey
product that cost as little as the Dataq units was definitely worthwhile---for
me. However, for those who enjoy working with an Arduino, the time spent
setting up the unit may not be objectionable.
Terry
_____
From: arocket-bounce@xxxxxxxxxxxxx <mailto:arocket-bounce@xxxxxxxxxxxxx>
<arocket-bounce@xxxxxxxxxxxxx <mailto:arocket-bounce@xxxxxxxxxxxxx> > on behalf
of Ramizcan Umut Seckin <Rocketry101@xxxxxxxxxxx
<mailto:Rocketry101@xxxxxxxxxxx> >
Sent: Wednesday, February 22, 2023 10:39 AM
To: arocket@xxxxxxxxxxxxx <mailto:arocket@xxxxxxxxxxxxx>
<arocket@xxxxxxxxxxxxx <mailto:arocket@xxxxxxxxxxxxx> >
Subject: [AR] Re: Subject: Re: S-26.500 KNSB Sugar Shot test tomorrow
Troy,
Isn’t it easier using an Arduino laying around (also cheaper) instead of the
DataQ module?
I never tried transducers before, do you need to make a hole in the bulk head
or can it be done without (used load cells till now)?
Last question is it possible to just use a transducer in order to make a
impulse graph?
Raum
On 21 Feb 2023, at 22:47, Troy Prideaux <troy@xxxxxxxxxxxxxxxxxxxxx
<mailto:troy@xxxxxxxxxxxxxxxxxxxxx> > wrote:
John,
Seriously, you can pick up a cheap Chinese transducer up for not very much at
all these days eg:
https://www.aliexpress.com/item/4000704177712.html?spm=a2g0o.productlist.main.37.36306a5c49iiuO&algo_pvid=8e31b2f2-0268-4b88-aa52-48ab4ca50125&aem_p4p_detail=202302211321284979106786310800001783306&algo_exp_id=8e31b2f2-0268-4b88-aa52-48ab4ca50125-18&pdp_ext_f=%7B%22sku_id%22%3A%2210000006206659883%22%7D&pdp_npi=3%40dis%21AUD%2124.09%2122.89%21%21%21%21%21%40211be59e16770144882715524d06f7%2110000006206659883%21sea%21AU%21100784320&curPageLogUid=jwiSi6bY02xM&ad_pvid=202302211321284979106786310800001783306_19&ad_pvid=202302211321284979106786310800001783306_19
and a basic DataQ module for $50 which you shouldn’t lose in a mishap
https://www.dataq.com/products/di-188/
Not much investment provided you have a PC to run it with or even some flight
electronics have the capability to log additional data channels. Craig
Strudwicke’s old flight computer was one example of that. In fact, my
multimeter does data logging.
I mean, for *candy* motors, you generally don’t worry, but for something this
size – as you would know, there’s a fair bit of work in producing something on
this scale irrespective.
Troy
From: arocket-bounce@xxxxxxxxxxxxx <mailto:arocket-bounce@xxxxxxxxxxxxx>
<arocket-bounce@xxxxxxxxxxxxx <mailto:arocket-bounce@xxxxxxxxxxxxx> > On Behalf
Of John DeMar
Sent: Wednesday, February 22, 2023 6:29 AM
To: arocket@xxxxxxxxxxxxx <mailto:arocket@xxxxxxxxxxxxx>
Subject: [AR] Re: Subject: Re: S-26.500 KNSB Sugar Shot test tomorrow
Ken,
The one thing it may tell us with some certainty is that the cause wasn't
the hardware. Even with an aluminum case, it should have held 1500 psi. But, it
would be helpful to see the retaining ring, to see if it's intact or sheared or
rolled.
Regarding the instrumentation, burn time plus success is good enough for the
relatively low cost of this motor. The loss of any sensors and related
equipment could exceed the investment in the test article.
-John DeMar
Las Cruces, NM
https://www.linkedin.com/in/jsdemar/
On Tue, Feb 21, 2023 at 12:02 PM roxanna Mason <rocketmaster.ken@xxxxxxxxx
<mailto:rocketmaster.ken@xxxxxxxxx> > wrote:
Other than the test video, without any instrumentation, post test hardware
inspection (and speculation) will be all you can do to conduct your analysis.
One channel minimum of Pc, chamber pressure, trace would be helpful. Even an
analog pressure gauge on video would give
at least a low frequency reading and maybe give the failure pressure...
Posting some pics of the post test hardware would give more eyes on the subject
and more chances of zeroing in on the overpressure mechanism.
Ken.
On Tue, Feb 21, 2023 at 8:25 AM John DeMar <jsdemar@xxxxxxxxx
<mailto:jsdemar@xxxxxxxxx> > wrote:
Rick,
From those specs on the case and screws, I get about 2800 PSI to shear the
screws (derate grade 8 tensile strength by 50%). Assuming 1/4" wall
standard-grade oil pipeline steel, the case should have deformed the holes at
2000 PSI and sheared at 2400 PSI. High grade pipeline steel would shear at 3000
PSI. I didn't derate for temperature based on the short time to heat soak.
So, plenty of margin on the hardware for a normal operating pressure in the
1000 to 1500 PSI range.
-John DeMar
Las Cruces, NM
https://www.linkedin.com/in/jsdemar/
On Mon, Feb 20, 2023 at 7:18 PM Rick Maschek <dmarc-noreply@xxxxxxxxxxxxx
<mailto:dmarc-noreply@xxxxxxxxxxxxx> > wrote:
The case was approximately 12 inch by 9 feet long, quarter inch wall oil
pipeline steel pipe.
24 grade 8 half inch bolts in two rows holding the quarter inch wall steel
retaining ring.
The forward bulkhead was a steel plate welded in the motor case.
5 grains of KNSB propellant 20 inch long with the cores varying from 2.5 inches
at the bulkhead to 4 inches at the nozzle
We used aluminum motor cases as our mandrels covered in silicon sheet and
greased for release.
The nozzle grain weighed in at 124 pounds with the others slightly heavier
because of smaller cores.
The nozzle had a 94% ideal density, normal density for me is 96-98%
I was up all night Thursday and much of Friday night helping student teams so I
left soon after our 12 inch test.
Not sure what caused the anomaly, lots of speculation and guessing. I'm still
investigating the cause.
We had two previous successful 12 inch KNSB sugar tests, one with 2 grains and
one with 3 grains.
This test had five grains and was designed to be progressive to test our three
piece safety nozzle.
Design pressure at start up was higher than the two previous 12 inch tests.
Flew a 4 inch KNSB motor just a couple hours before the 12" test
<https://www.youtube.com/watch?v=jWspWbTCXbM> UCLA Rocket Project Avionics &
Recovery test
<~WRD0000.jpg>
UCLA Rocket Project Avionics & Recovery test
<https://www.youtube.com/watch?v=jWspWbTCXbM>
https://www.youtube.com/watch?v=jWspWbTCXbM ;
Rick