placement of md0550 and stevenson screen

Questions about Modern Device and JeeLabs Sensors
Post Reply
rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

placement of md0550 and stevenson screen

Post by rynait » Wed Jul 01, 2020 11:24 pm

I searched the internet for any information recommending good case to place the md0550. and outside.

did anyone do experiment finding optimal placement for this device inside the stevenson screen??

paul
Site Admin
Posts: 764
Joined: Mon May 12, 2008 4:19 pm

Re: placement of md0550 and stevenson screen

Post by paul » Thu Jul 02, 2020 4:57 pm

rynait,

I don't have any special knowledge about this. But some facts about the sensor that may or may not be useful.

 • The sensor is pretty easy to spit the sensing head off from the electronics just cut it with tin snips between the headers and then join back together with three wires. Distances between up to 1 ft should be no problem. You should probably use shielded cable for anything longer than about a foot with the shield grounded only at the electronics end.

• The sensor is directionally sensitive and readings from the backside are about 50% of readings with wind from parts side.

rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

Re: placement of md0550 and stevenson screen

Post by rynait » Fri Jul 03, 2020 11:57 am

thank you for the information about

the j2 and j3 found on schematic. i did puzzled over that. now understand device design approach.

and the alert regarding back side wind reading.

I studied your script and realized you left out a lot of information. I am not willing to use your script, based on several facts that you have not indicated nor provided the solver output data (just add excel file showing the data please).

and how did you arrive at the equation Vraw = V0 + b * WindSpeed ^ c. I know in your blog you derived this from an formula done by university of E2 – A2 = B U0.5. can i have the name or reference to the university white paper that you read about.

rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

Re: placement of md0550 and stevenson screen

Post by rynait » Fri Jul 03, 2020 12:39 pm

reading the schematic...

i dont understand the design. you had U1A reading bridge voltage, but output of U1A(1) is feeding to the 2n222 transistor thus controlling the voltage going back in the bridge. Vraw is voltage across r5 and r6. not the bridge voltage.

is the schematic showing some errors?

paul
Site Admin
Posts: 764
Joined: Mon May 12, 2008 4:19 pm

Re: placement of md0550 and stevenson screen

Post by paul » Sat Jul 04, 2020 3:25 pm

rynait

r5 & r6 are setting a reference voltage to be compared the active arm of the circuit. When the voltage changes the heated sensor changes resistance, balancing the circuit. So voltage is proportional to the square of wind speed, approximately.

R5 representing a small value in proportion to other voltage is added in. The error is taken care of in the regression, and from my reading this is common practice in hot wire anemometers. That said it would be easy to add another op amp section and pick the voltage off only the active sensor. This is what we did in Rev P I believe.

Another way of thinking about Kings law is in terms of power
Power = heat so the added heat is represented by P = V^2 / R
Square root of the voltage = wind speed

paul
Site Admin
Posts: 764
Joined: Mon May 12, 2008 4:19 pm

Re: placement of md0550 and stevenson screen

Post by paul » Sat Jul 04, 2020 3:30 pm

and how did you arrive at the equation Vraw = V0 + b * WindSpeed ^ c. I know in your blog you derived this from an formula done by university of E2 – A2 = B U0.5. can i have the name or reference to the university white paper that you read about.
Caimbridge University in UK has lots of great information on hot wire anemometers and King's Law.
http://www-g.eng.cam.ac.uk/whittle/curr ... -wire.html

We just looked at the data and found a regression that fit the data fairly well. Since it was all empirical there were not a lot of things to argue with (but plenty could go wrong with future changes I guess). These were the first regressions I ever did with excel so my chops have improved. Our test equipment has too, so I am hoping to finally get better curves and regressions this summer.

rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

Re: placement of md0550 and stevenson screen

Post by rynait » Mon Jul 06, 2020 1:39 am

paul wrote:
Sat Jul 04, 2020 3:25 pm
r5 & r6 are setting a reference voltage to be compared the active arm of the circuit. When the voltage changes the heated sensor changes resistance, balancing the circuit. So voltage is proportional to the square of wind speed, approximately.

R5 representing a small value in proportion to other voltage is added in. The error is taken care of in the regression, and from my reading this is common practice in hot wire anemometers. That said it would be easy to add another op amp section and pick the voltage off only the active sensor. This is what we did in Rev P I believe.
studying your schematic (since you are not saying it is not in error)

realizing that the cambridge paper is referencing to the "upper arm voltage". unfortunately that online whitepaper did not present concept schematic establishing their bridge definitions. Common wheatstone bridge schematic design is used, translating cambridge's statement into bridge voltage.

other white paper i could find on the king's law. indicates voltage E is sensor voltage, not the bridge voltage. since you use CTA bridge method, that Vraw is bridge voltage.

If Vraw reading is plugged into Cambridge formula, then there is need to address the voltage across r3, because resistor is not contributing to flow heat loss reading.

From Schematic, is possible to get voltage reading across the r4, utlizing the trace pads on j2-2 and j2-3. thus establishing sensor voltage E. should help calculate the constant correction in this vraw = Vo + b u0.5 equation.

Roy
Last edited by rynait on Mon Jul 06, 2020 3:44 am, edited 6 times in total.

rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

Re: placement of md0550 and stevenson screen

Post by rynait » Mon Jul 06, 2020 3:02 am

looks like there is CTA design error on this device.

according to CTA wheatstone design, r2/r1 = rx/r3. r2 is adjustable and underlined, r1 & r3 is fixed, rx is sensor

your schematic looks like it is saying r6/r5 = r4/r3. problem that I see r5 is adjustable [underlined] and comparing the formulas, on wrong side of wheatstone.

Roy

paul
Site Admin
Posts: 764
Joined: Mon May 12, 2008 4:19 pm

Re: placement of md0550 and stevenson screen

Post by paul » Mon Jul 06, 2020 9:28 am

Roy,

The bridge is R3/R4 R5/R6
As the sensor heats up the bridge is balanced.

Paul

rynait
Posts: 6
Joined: Wed Jul 01, 2020 11:21 pm

Re: placement of md0550 and stevenson screen

Post by rynait » Tue Jul 07, 2020 1:56 pm

after looking at schematic and analyzing the board.

realized my eyesight is so bad!. broke out magnifying glass... that pointeometer is not r5. it is r9. !!

r5 is fixed value of 75k, not 2k. r3 code is 34Y, = 2.21 ohms. r6 is 2k, not 165k. found there is r13 [10m ohms], but not on the schematic.

i had trouble reading r1 with my magnifying glass. r8 code could be read as 00E. since 00x is jumper, I see from rev.1 schematic this supposed be 10k, reading from other direction would make r8 code 300 = 30 ohms?

i noted throughout your blog(s), that bunches revisions was considered based on .1 % tolerance resistor and as well different resistor selections.

since this is a revison C4. is there a file or statements declaring what changes was done? and your test charts in blog(s) made 2014, was that based on rev c4 sensor?

roy

paul
Site Admin
Posts: 764
Joined: Mon May 12, 2008 4:19 pm

Re: placement of md0550 and stevenson screen

Post by paul » Tue Jul 07, 2020 5:41 pm

Roy,

Yes - maybe email support at moderndevice.com and I'll answer some of the technical questions about where we're at with testing and new regressions. As you mention the listed code is old.

Paul Badger

Post Reply