Not the Mercdog again

[larrycavan]

Some testing you can try as well..

Testing with orifice on the bench.

Dial up 10" of test pressure & note stability of screen readings. Notate variance.
Dial up 28" and repeat above process.
Dial up 36" and repeat process.

Analyze results

Mount head to bench and repeat above test pressure series tests while going through lift range of .100 - .500 inches. You're looking for two things here. Comparison of variance in stability of readings through full test range AND discovering if the head has turbulence issues at specific lifts. It may not have turbulence issues at 10" at specific lifts but have issues at 28 or 36.........All reasonable variables that can show you things about the bench and the head itself if you look at the results carefully.... Pain in the ass....Oh yea

[Old Grey]

Some testing you can try as well..

Testing with orifice on the bench.

Dial up 10" of test pressure & note stability of screen readings. Notate variance.
Dial up 28" and repeat above process.
Dial up 36" and repeat process.

I ran a quick orifice test(200cfm PAP), and it seams that the bench has a natural 1 cfm bounce with sample duration set at #2.
Set at #20 it seams to be less than 0.5, maybe even lower.

I will run more tests latter

The other option is use slow #20 and the fluid manometer that I'm making.

[Old Grey]

Using the 200cfm PAP with sample duration at #20
number order : Dep : ACFM : SCFM
10" variance :- 0.0 : 0.0 : 0.0
20" variance :- 0.2 : 0.4 : 0.6
30" variance :- 0.1 : 0.7 : 0.7

Using the 200cfm PAP with sample duration at #2
number order : Dep : ACFM : SCFM
10" variance :- 0.0 : 0.4 : 0.6
20" variance :- 0.1 : 1.2 : 1.1
30" variance :- 0.1 : 1.3 : 1.1

Occasionally there is 1 erroneous number that jumps 0.5cfm, but that only happens once in maybe 50 samples, and only on 3 of the tests.

Just looking at the numbers they don't look too bad, maybe the head is a little turbulent.

The PTS guys have spent enough years testing to know what works best, so after testing their advice I would say that their high sample rate and software averaging looks like a winner. I'm a bit behind the game with this old equipment, so I'm just trying to simulate something similar to make something better than it originally was. It would still be nice to have the averaging option in Excel, even for curiosity sake, so if I have time I will look into it again.

[larrycavan]

https://www.facebook.com/photo.php?v=657059517679751 draw your own conclusions.... Looks to me like both dogs pretty much wag the same tail

From the beginning, when guys have called and asked me which box to get I've told them all.....pick the best price, they both do the same thing equally well. That's my story and I'm sticking to it ...

Watching the video with mouse pointer on the pause button....tracking numbers at the .100 check lift portion of the video and carefully grabbing numbers where the depression hasn't varied [locked on 28"]. A variance of 1.4 CFM.

28 - 35.9
28 - 36.1
28 - 35.7
28 - 35.9
28 - 36.1
28 - 35.8
28 - 35.3
28 - 34.7

That's not meant to slam the PTS box. It is what it is. The numbers are what they are.

Visit one of your friends with a blue bench with a Flowcom unit and get some numbers. I'm curious to see how it stacks up against the PTS and FP1 for displayed number variance...

Also wondering what a chunk of very loose cell foam would yield if it was places over the end of a pressure sensor.... would it dampen things out physically without falsifying the numbers....

[Brucepts]

Watching the video with mouse pointer on the pause button....tracking numbers at the .100 check lift portion of the video and carefully grabbing numbers where the depression hasn't varied [locked on 28"]. A variance of 1.4 CFM.

28 - 35.9
28 - 36.1
28 - 35.7
28 - 35.9
28 - 36.1
28 - 35.8
28 - 35.3
28 - 34.7

That's not meant to slam the PTS box. It is what it is. The numbers are what they are.

Not taken as a "slam" but would like to clarify; On my flowbench I keep my averaging set so I have some number roll on screen, that is what you are seeing. I prefer to do a "mental average" for my testing. I could jack the averaging up and have rock steady numbers on the video and will probably do that on future videos. So the variation that shows in the video is how I have my personal averaging set. I've always stated that the PTS DM can be set for the end user's preference, we built that into the design, you can set your averaging from 1 to 1000+ if you would like. Each flowbench and computer is going to be different on the settings, no two benches are going to be set the same in configuration.

Looking back we actually have made it to end user adjustable for some people, it can be quite confusing for those who do not want to fiddle with the settings and to many setting to fiddle with for the tinkerers to add confusion.

[larrycavan]

That I knew [adjustment feature]. But it still remains the same scenario. Setting the box to lock on vs roll is more or less just a screen buffer feature. It has nothing at all to do with accuracy. THAT has been my point all along through this discussion. I have never played favorites to either box. I have used both. They both do the job required equally well for most people.

What grey is after is a screen buffer that isn't part of the FP1 features. Rick went an extra measure to provide that & it was a good thought on his part. Also, this matter never came up prior to this particular thread. Had it been a complaint, I would have addressed it in FPexcel a long time ago.

For what grey is trying to accomplish, averaging an array is overkill & doing it after the fact isn't affecting anything but displayed numbers.......

I believe there's a better way to go about building in a displayed number buffer using a the delay feature in VBA along with a comparison place holder cell and some code. It may need to be a worksheet change event to be handled properly....anyway I think it's doable.

I'm up to my ears in something new in Visual Basic 2010 or I'd delve into it in Excel for him. I have confidence he can sort out a solution.

FWIW, for the time being, I still use a PTS box and it's still hooked to an MSD bench which is actually what that Mercdog is a copy of. With the mods I did to it the only time I get unstable numbers is if the port has turbulence in it....with the FP1 it was the same

[Old Grey]

Just looking at the PTS video it looks like the FP1 can match it at sample duration #20 - at least where I'm happy with it -. The problem is that because the FP1 does everything before the display, everything coming out of it is slow. The PTS has the advantage in that they can make the display fast, like in the video, and average it after when it gets pasted to the table - this is what I'm trying to do artificially with the FP1 -.

It's still a good project to work on, I just have to play around some more.

I still have to test Larry's screen freeze code, and also 1960FL changing Private Sub to some sort of Global Sub, so I've got some clues to work on.

This is the code that I last left it on

Code: Select all

Private Sub Calc()

Beep 2000, 30
Range("T11:W20").ClearContents
Range("G28").ClearContents

Dim i As Integer
For i = 1 To 10
Range("G28").Value = i
Sleep 500

Range("A1:D1").Select
Selection.Copy
Range("T" & i + 10).PasteSpecial Paste:=xlPasteValues
Application.CutCopyMode = False

Next

Beep 3000, 300
Sleep 1000
Range("G28").ClearContents

End Sub

[larrycavan]

Good Morning ....just so you know, sometimes you run into problems between 32 and 64 bit versions of OS and/or Office. This will give you a quick heads up on that..

http://www.exceltrick.com/formulas_macr ... functions/

Also, sleep is an API, not an Excel internal function. That can tie into the first link I posted. Sleep will freeze the app interface during processing. You may find this useful http://www.cpearson.com/excel/WaitFunctions.aspx

I know what it's like to work all day on adding a feature via code and run smack into a wall after hours of code attempts. Just giving you some hopefully helpful info that will alleviate that situation.... Good Luck

[Old Grey]

This is the ball bearing version.
When you run a rubber seal on the bowl it needs to spin independently of the screw. You can just loosen the nuts, but this set-up makes it more rigid.
The only problem is that the screw is unsupported on this end, and coupled with a slight bend it orbits, making slight dead spots where nothing happens when you turn the knob. A longer screw and a support will probably fix that.



Over designed control valves to follow.

[larrycavan]

I found some youtube videos on the flowcom. Pretty hard to see the numbers in the videos. https://www.youtube.com/watch?v=nGov3CHc4SU