PTS Flowbench Forum

David,

I here what you are saying and YES in a orifice i a TUBE (large Beta) You would need to concern yourself with what you talk about STREAMS! but if you have followed the thread on Wet Flow and or any of the discussions on Swirl or Tumble you must admit to yourself it is a far stretch that the Orifice in a PTS style bench would ever see a STREAM of air flow. The air leaving the valve (Intake Mode) is almost chaotic swirl, tumble spin and that is just at the chamber. Then it must try to find some order in a Pipe say 4" ID and become a high powered stream I think not.

I concur that when one develops a bench using Pito or OIT and uses motor boxes settling chambers and lengths of PVC to connect them they are focusing this air to a stream and inducing more potential flow problems, thus the beauty of the orifice between two box's design . Remember the term stream inherently refers to laminar and in almost any flow bench of the orifice betreen box design i have seen the only laminar flow is that in a good port on the top and that going through the orifice.

You seem to have passion for this subject build one and test it, test it against a standard of some type and that standard for sure is not Big Blue.

Well... this is interesting. It seems others have questions about "Orifice in a Tube".

From the beginning and until the end of this forum the discussion of flowbench design will be debated tested and challenged. The fact is if you are using the tool to compare to your own work and don't care about others results it really does not matter what you build as long as you always calibrate to your same standard. I have often threatened Bruce that i was going to build a flowbench from two cardboard computer boxes with a 45RPM record as the internal orifice just to prove it really does not matter it the attention to detail in your workmanship and the accuracy of your math and measuring devices.

Rick

I think I may have been misunderstood. When I was talking about variable baffle sizes etc ,I was talking about having them in the typical settling chamber box environment and what effect the size , thickness , shape, location, etc, would have depending on the chamber size

Bakerlite,

I do not think any of us really put any scientific data to work on this piece it was more an addition to protect the orifice from rampant air currents within the chamber that were stirred by the air coming from the test piece or from the orifice to the test piece. I will say there is a calculation i think that is a relationship between the orifice and the Vena Contracta I/E the center line of the orifice plane to the smallest diameter steam on the outlet side of the plate. As for the baffle shape itself I believe at lease 1/2" thick with full radius edges thus disrupting the air as little as possible, i have often thought about doing one in a round shape instead of a square just to see if it has any affect on linearity data.

Rick

1960FL Wrote:

i have often thought about doing one in a round shape instead of a square just to see if it has any affect on linearity data.

I will be doing that in my build my round baffle will be 11" 7/8" around. I hope to be posting my PTS build soon I have taken many pictures Ask Bruce he will be receiving 10 more pictures of the work Gary & I did today!!!!!!

As I am one of the "orifice in the tube" guys who have seen issues and countered some of them - like using bigger than 2" orifices in 4" pipes - using very good flow software designed to take the guesswork out of wall proximity and such, I would like to know: How big a chamber is big enough for a good Beta ratio according to you guys? Flow orifice is < 3", would a 2ft x 2ft x 2ft box with the orifice centered in one wall, leading into a similarly sized chamber do? That size would make Bruce's 625cfm orifice have about a 0.133 BR, a smaller orifice should be even better.

I'll add a large, circular baffle 2.5" off the orifice either side if that would be needed, as well as a baffle plate under the head fixture. Just doing it to see if the numbers are really that far off between designs. That should end that discussion with real facts. After all, what else is the time off during Christmas for? Any input on the sizes needed to make a good result is welcome.

Hi,

Place a tube of at least 3x longer than the the orifice diam and at least 2,5 x the orifice on both sides of the orifice.
calculate the beta and then adjust the cf value so that you get the same flownumbers for different size of testplates.

what you are actually doing is installing a measuring element in the chamber design that will convert it to a orifice in a tube design.


you can also switch to a nozzle or a pitot by just repostioning the the measuring pipes to the mano and use a little bit different math.

/kimi

Kimi wrote:Hi,

Place a tube of at least 3x longer than the the orifice diam and at least 2,5 x the orifice on both sides of the orifice.

I did. I had 8D before it and 5D after it, and even adjusting for Beta, which my software does automatically, I would still need a higher rise (11.64") on the inclined than what the orifices are calculated to, namely 10" rise, for the inclined to show the correct percentage of flow. That's when I started wondering if I should say my orifices are calibrated to 10" or 11.64"... a 2" orifice has a higher cfm number when calculated at 11.64 than 10... so to compensate for that increase in cfm, I would have to increase the rise of the inclined a bit more.. and a bit more.... and a bit more... never-ending issue.

calculate the beta and then adjust the cf value so that you get the same flownumbers for different size of testplates.

what you are actually doing is installing a measuring element in the chamber design that will convert it to a orifice in a tube design.


you can also switch to a nozzle or a pitot by just repostioning the the measuring pipes to the mano and use a little bit different math.

/kimi

Hi,
Adjust the cf value . Use that to calibrate.

Decide from where you measure the 10" dp and stick to it. its good to measure it close to the orifice.
Measure the flow for a known orifice that is in the upper range you use most and adjust the cf to show the correct flow.

An orifice has a calculated flow based on the diameter and a cf value based on the shape of the orifice edge( contour and thickness).

If you have 10 inch at the orifice 50cm from the testpiece and flowing 100cfm you propably have only 9 below the testpiece. With a different flow the dp will change below the testpiece and is not linear to the flow.

The orifice flow is calculated from the pressure difference from the both side of the orifice and the magnitude from the dp.
The absolute correct way should be to measure the dp as close to the testpiece as possible , but the swirl and tumble makes this very difficult.

I also recognised a difference in calibration when i switched from soft silicone to hard pvc tubes for the sensors. The soft silicone was shrinking/sqweezed from the depression and the real dp never reached the sensor. This fixed some of the unlinearity but it disturbed my mind.
It took a week for me to find the reason. My digital mano takes about 8900 samples per sensor per second and i could measure a 5ms delay before the dp reached the sensor. I placed a sensor close to the orifice with a t-piece to measure the difference.

lycka till!
/kimi

Kimi wrote:The absolute correct way should be to measure the dp as close to the testpiece as possible , but the swirl and tumble makes this very difficult.

I also recognised a difference in calibration when i switched from soft silicone to hard pvc tubes for the sensors. The soft silicone was shrinking/sqweezed from the depression and the real dp never reached the sensor. This fixed some of the unlinearity but it disturbed my mind.
It took a week for me to find the reason. My digital mano takes about 8900 samples per sensor per second and i could measure a 5ms delay before the dp reached the sensor. I placed a sensor close to the orifice with a t-piece to measure the difference.

I have observed the same effects with my bench. I believe Bruce's product would be improved a good deal if the sensors were removed from the board and instead were potted into, say, a billet aluminium housing that was directly bolted to the side of the cabinet. Leads would then connect to the board/box. Added advantage would be that if a sensor went down it could be unplugged and replaced.

Also with long tubes you get resonance effects which also adds an uncertaincy value to the readings.

I'm making my tubes for pressure sensing from brakeline copper tubing now.. and when I go digital I'll place the sensors as close as I can to the pickup locations. I see no reason not to be as accurate as you can..