Upgrading My Rocket Press (Would love help planning)

Off Topic
#1

Hey everyone,

I own a Rocket Press manufactured by Firesmith Tooling, here is a picture of the press:

image
image.jpg417×600 95.3 KB

This is an electronically controlled 14.4 ton hydraulic press and I would like to change how it displays pressure. Currently I’m relying on an analog pressure gauge that shows 0-3000 psi of the cylinder. I would like to change this to a digital gauge that shows the pressure applied in my rocket tooling. my tooling uses rammers that are .75 inch, 1 inch and 2 inch diameter and the hydrolic is a 3.5 inch diameter cylinder. So I’m often having to do non-trivial conversions of pressure to make sure I’m not over driving my tooling.

What I hope to have in the end is a hydraulic pressure sensor hooked into an arduino with a display. The arduino would have options for outputting the pressure for the different tooling im using with the pressure already converted for the tooling.

What do you all think, is this a minor project or a major one? what do you think I should set a side for budget? and suggestions on sensors that will work that cover 0 to 3000 psi?

Thanks,
Nick

#2

I dont know much about your hardware questions, but a similar problem exists at plants commonly. There is usually a laminated spreadsheet ziptied to the instruments with the conversions/readings at critical values and expected ranges. Nearly fail proof, cheap, and durable.

2 Likes
#3

Yeah,

I’ve done similarly, but run into the issue of constantly losing the sheet as I travel with my press. It never seems to fail that the sheet is torn off or missing when I need it. The second problem with that solution I am relying on the increments the gauge supports which are not fine enough for my smaller tooling as they tend to be used in the 300 to 400 psi range on a 3000 psi gauge. Which means I can be changing 1000s of psi on the tooling in an area of less than 1/8 of an inch on my gauge.

Thanks for the suggestion, it is a good idea, just doesn’t answer all the issues.

1 Like
#4

Interesting project!
Some suggestions/comments.
Keep the analog gauge. Plumb in a “T” fitting for the sensor.
Did a very quick look at Mouser, Digikey, Allied and found prices and availability all over the map.
Also looked at Omega. https://www.omega.com/
Configured one of their sensors w/ usb output. Cost ~$500.
I think sensors w/ digital output might be easier to interface to but will cost more.
Sensors w/ analog output may require some kind buffer/gain stage before the signal goes to the ADC.
Depends on sensor output range and the Input range on the ADC (on the embedded/micro brd)
Pressure sensor will probably be the most expensive part of you project.
Once the embedded system has the sensor reading, it can translate/convert to what ever units you need.

Some questions:
What kind of accuracy do you need?
Are you considering using the embedded controller to control hydraulic pressure to a preset value?

Hope this helps.

#5

Totally planning on keeping the gauge somewhere. That way I can make due if my electronics ever go kaput.
I am looking at analog sensors due to the drastic price difference. I’ve found this one in particular, https://www.transducersdirect.com/products/pressure-transducers/low-cost-pressure-transducers/tdl35-series-pressure-transducer-0-3000-psi/?gclid=EAIaIQobChMIh8eplYfr3AIVhbrACh0OQgZyEAYYASABEgI0gPD_BwE

It has an accuracy rating of 1%, which I understand to mean 1% of total range of 0 to 3000 psi. So effectively it reads in increments of 30 psi, if my understanding is correct. Not super happy with that accuracy as I think my 1lb tooling is a ration of 21 to 1, meaning one tick higher could be a shift of 630 psi on the tool, however this is already much finer than my current gauge, which is in increments of 62.5 psi till you get over 500 psi. You could try to get closer to the 30 psi by putting the needle between increments, but your really are guessing with a blunt tool at that point.

I’d love to find a .25% accuracy sensor over the range of 0-3000 psi, but haven’t found a cost effective option for that yet, but I am looking.

Here is an omega I worked up, but probably outside my price range. https://www.omega.com/pxconfig/cfg_final.html?pxapp=General&PRESSURETYPE=G&RANGE_UNITS=3.5K_P&UNITS=P&OUTPUT=V1&ACCURACY=4&FITTING=C&TERMINATION=6&TEMP_COMPEN=T3&THERM_ACC=A5&OPTIONS=CE

#6

1% of the indicated value is the typical meaning, otherwise would be ± 30 psi.

#7

I’ve read it to be accuracy of 1% of the Full Scale. I took that to mean, 0 to 3000 psi, so 1% would be 30psi. That said I’ve also read accuracy explained as a moving point based on temperature, so over the range of operating temperature the reading is within 1% of the full scale or 30 psi.

I’m honestly foggy on the meaning of accuracy as it pertains to the sensors, here is the data sheet on the $49 sensor I’m most intrigued by currently. http://www.transducersdirect.com/wp-content/uploads/2018/03/TDL35-1.17.pdf

other options I’ve found,
https://www.aliexpress.com/store/product/0-0-2-60Mpa-Silicon-Pressure-Transmitter-Pressure-Transducer-G1-4-4-20mA-output/100271_32797583298.html?spm=2114.search0104.3.1.1fde47d814NfIf&ws_ab_test=searchweb0_0,searchweb201602_5_10152_10151_10065_10344_10130_10068_10324_10342_10547_10325_10343_10546_10340_10548_10341_315_10545_5012515_10696_5012615_10084_531_10083_10618_10307_10059_100031_10103_10624_10623_10622_10621_10620,searchweb201603_56,ppcSwitch_4&algo_expid=c8cac08a-5b5c-4443-b74b-6513c5500811-0&algo_pvid=c8cac08a-5b5c-4443-b74b-6513c5500811&transAbTest=ae803_1&priceBeautifyAB=0

https://www.aliexpress.com/store/product/0-25Mpa-Waterproof-Lead-Wire-Straight-Line-Pressure-Sensor-Transmitter-4-20-mA-M20-1-5/345733_32763134931.html?spm=2114.search0104.3.1.636c43965B0Bop&ws_ab_test=searchweb0_0,searchweb201602_5_10152_10151_10065_10344_10130_10068_10324_10342_10547_10325_10343_10546_10340_10548_10341_315_10545_5012515_10696_5012615_10084_531_10083_10618_10307_10059_100031_10103_10624_10623_10622_10621_10620,searchweb201603_56,ppcSwitch_4&algo_expid=e4f873a2-22cc-4c10-86cc-4813e9b56c3e-0&algo_pvid=e4f873a2-22cc-4c10-86cc-4813e9b56c3e&transAbTest=ae803_1&priceBeautifyAB=0

#8

When tools come back from calibration they marked with tolerance limits - this would indicate their “accuracy” or the ability to indicate the true value.

If it reads 1% it is referring to the indicated value. If it were1% of the total value it would just say ±30 psi in your case. At 500psi it would be ± 5psi.

What Is percent accuracy?
_Accuracy is the agreement between an experimental value, or the average of several. determinations of the value, with an accepted or theoretical (“true”) value for a quantity.

Accuracy is usually expressed as a percent difference: % difference = (experimental – true) x 100%_

#9

Thanks for further explanation, but you are not mentioning Full Scale in your explanation. This is what I read to understand accuracy compared to full scale,

FS = FULL SCALE = maximum reading.

It means that the accuracy is such that the reading is probably within + or - 0.5% of the FULL SCALE reading.

This is a very important and easily overlooked qualification of the result.
If I have a reading of 1 Volt and the accuracy is +/- 0.5% it means that the actual result should lie in the range 1 - 0.5% x 1 to 1 + 0.5% of 1
= 0.995V to 1.005 V

However - if I measure the result on the 10V range then 0.5% of 10V = 0.5% of Full Scale
= 0.05V. So 1V +/- 0.5% of FS
= 0.95V to 1.05V.

On the 100V range, 1V +/- 0.5% FS lies in the range
0.5V to 1.50 V. !!!

The reason for specifying results in this manner is that the error experienced on a given range tends to largely be constant regardless of the actual reading. So, as the input gets smaller the error becomes increasingly large in proportion.

So eg on a 100V range a reading of 0.5V +/- 0.5% FS lies in the range
0 to 1V!

Here is the source, https://engineering.stackexchange.com/questions/1718/what-does-pm-0-5-f-s-mean

Is this explanation incorrect?

#10

This is somewhat outside my expertise, but IIRC analog signals use a 4-20mA output. You measure the current over the 4-20 range, 4mA being 0 psi 20 mA being 3000psi. It’s a narrow range, so small variations may not affect a detectable change in current. We use digital controls in critical process measurements for this reason.

I guess what I’m saying is make sure you have good enough resolution with small enough error before you buy. The sensor itselfay not be the limiting factor.

1 Like
#11

If something is within 1% accuracy (at any given point) then true value lay within 1% of that.

1% at 3000 is means the true value is ±30 from the indicated value. It can also be stated as a tolerance band of 3970 ~3030. This is typically seen when the engineer is more interested that it is within this range than at a set nominal.

Percent values of indicated are typically found where actual deviation from the expected is not a uniform value, that the value varies along the scale in terms of absolute variance, e.g. 5psi @ 500 and 30psi @3000. But the deviation is proportionate along the scale.

Temperature and pressure typically have a percent of scale, especially over wider ranges. I believe this is typical for electrical measuring devices. When i had to witness electrical testing the specification is usually a tolerance band at a given range of values and then the measuring device will have it own limits. As the values start to approach the limits of the tolerance bands you may have to run a statistical analysis of the measuring device and tighten/reduce the band so that you don’t accept a bad measurement as acceptable (It also means you are likely to reject a acceptable reading and bad). This of course is highly dependent or how critical the feature is.

Engineers have a tendency to over tolerance things, it is costly practice in terms manufacturing costs. But does increase the products “as designed” characteristics. There is very much a safe but sorry view … which manufacturing engineers want safe but affordable and realistic.

Measurement analysis in Six Sigma is a critical component - it’s really fun to calculate/measure system stack up through the process. If you have slightly bent personally it is fun and interesting. And a powerful tool when applied properly.

#12

Here’s a tech paper on error, various accuracy issues.


This may be the “Golden” answer - of course at a “Golden” price.
Get the sensor and the free sftwr - almost plug n play.

Analog is cheaper - in part because you have to design/deal w/ all of the analog and ADC issues.

1 Like
#13

Thanks for all the information about accuracy, I’m trying to understand the white paper to corroborate the the explanations I’ve now been informed on. Didn’t expect this level of detail right off the bat. :smile: But, had I known all this it wouldn’t be worth learning.

2 Likes
#14

Just had a chat with an engineer that works with sensors like we are talking about and he explained to me that full scale accuracy is as I read about before. Quote above.

#15

Also keep in mind the resolution may not be great if you dont have a higher bit ADC.

1 Like
#16

That is a good point @TBJK. The arduino comes stock with a 10bit ADC, meaning that it can see a range of 1024 readings. So if I can take advantage of that full range with my sensor, I would have a resolution of about 3 psi per increment with an accuracy of plus or minus 30psi. 3 psi is probably a good enough resolution, given the less than ideal accuracy. If I can find a cost effective sensor option with a higher accuracy, I might spring for a higher resolution ADC chip for the arduino.

That all said, I had an idea for increasing the accuracy of my lower PSI readings. I might be able to get a second pressure sensor that has a lower range and thus more accurate reading at the lower psi. Just have to make sure it can handle being over driven without breaking.

Wow, this is a very interesting problem. I’m starting to think I should look into a large face analog gauge that I can get custom scales printed for. Might be the best solution for simplicity as this is getting kind of complex give a limited budget.

Well scratch that idea, it won’t be any cheaper than buying a higher end digital pressure gauge and and would be less accurate than the digital pressure gauge. As the analog gauges tend to have an accuracy of 2% or higher full scale. So the cheap digital sensor will give me both higher accuracy and higher resolution than I currently have. The cheap $49 TDL35 sensor may be more than good enough given how I have been working with a much lower resolution and much lower accuracy solution this whole time.

#17

Adafruit has a 12 bit. That gives you 4x as much. I’m not sure on what they are doing to midigate noise, that becomes more apparent in the higher adc’s or so I’ve read. Keep in mind I’m not an electronics guru. Just telling you some of the issues I was running into when I was working on a differential pressure transducer, that was only 0-200 psi too.

1 Like
#18

Yeah,

12 bit would be plenty, as it would allow me a large buffer if my signal range isn’t ideal. But, noise may be an issue to consider given the higher detail. :smile: Every solution comes with it’s own new problem.

This has been a truly enlightening discussion so far. To shift the topic, any suggestions on a neat Arduino compatible display that would be daylight and low light visible. Thinking at least 2 inches and no bigger than 5 or 6 inches?

#19

What is the accuracy and resolution you actually need? This will drive you measurement needs. Not sure how much variance handling the material makes and what the difference is between too much pressure vs too little. The process may be fairly tolerant or not.

If your target value is say 2000 psi, how much variation can to accept?

#20

I’m going to ask how smart it is to start hooking electronics up to explode-y things, when it’s likely to need to operate in an environment rife with other explode-y things, the timing for which is crucial to the enjoyment of…