Page 1 of 3 123 LastLast
Results 1 to 15 of 40
  1. #1
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default DC temperature/pressure meter

    This thread is about a recently home made Digital Dust Collector Pressure/Temperature meter.
    The first post will be a general description of meter. Subsequent posts will provide additional detail of some of the technical detail such as design, build, testing and calibration and use .

    The story is I’m still feeling too knackered to be standing at the bench in the shed for more than about half an hour before I need a rest, plus it's been a tad hot so I decided to retreat to the house and work on some electronics projects. This is a project I have been meaning to tackle for some time but it’s taken me a while to get all the pieces together.

    The temperature being measured is the temperature of the 3 phase, 4HP, DC impeller motor - DC located in an enclosure outside my shed. The motor runs on a 240V VFD and although the VFD has current limits set , it does draw quite a high current at higher speeds so I would like to know the temperature of the motor. Whatever it is, it should be considered in relation to the temperature of the air inside the DC enclosure so the meter provides that information as well.

    The Pressure reading output depends what the sensor hose is hooked up to.
    Usually it will be the pressure (relative to atmosphere)in between the impeller and the DC filters, as this can be used as a measure of when the filters need cleaning.

    IMG_3557.jpg

    What's inside the box.
    Guts.jpg
    The unit is based on an Ardunio “Zero” micro controller (MC). Labelled "A" in the photo above.

    The black rotary switch (#data points in photo) sets the number of readings averaged before displaying data. This can be 1, 2, 5, 10, 20, 40, 60, 80 or 99 readings.
    The DST data switch toggles the screen display between P/T data, and the raw digital data output (numbers would always between 0 and 4095) by the MC.
    Reset (R) - resets the MC.

    The temperature sensors are K type thermocouples (TCs) are located in the DC enclosure and require a small operation amplifier (OA in photo) to boost the small microvolt signals into the sensing range of the MC.

    The pressure sensor (P orange arrow) is located inside the MC box as it is sensitive (see subsequent posts about this) and can measure pressure differences of ~0.01 inches (0.25mm) of water column (or 0.0025 kPa).

    A small daughter board is used for the OA and P and for additional ground, 3.3V & 5V supply points . P is underneath the daughter board.
    PressSensor.jpg

    Referring to 2 photos back, brass flange (F) connect the pressure sensor to an outlet (P in photo below) outside the box suited to 4 mm diameter black HSPE trickle irrigation tubing which is used between the MC box to hook up and measure up any required pressure measuring points/holes in/on the DC system.

    Inputs.jpg

    Plug and socket (T) connects via an ~ 3 m long cable to the temperature sensors in the enclosure .
    The connections are housed in the small black plastic boxes (M & E) shown below.
    The thermocouples (temperature sensors) are the fine brown and yellow insulated wires coming out of the plastic boxes - these will be tucked away under something thermally conductive like a small Al block.

    Heads.jpg
    The Motor Temp thermocouple (Motor T) is embedded inside an Aluminium block attached to the cast iron motor housing inside the motor power junction box.
    The other (Enclosure T sensor) just hangs from the ceiling of the DC enclosure

    Display.
    The P/T data display page is pretty self explanatory
    Motor T is the motor temperature
    Enclosure T is the enclosure temperature
    Diff is the the between the two measured temperatures.
    DCP is the pressure relative to atmosphere in kPa of whatever the pressure sensor is connected to .

    The number "9" on the bottom right of the screen is a count down number and shows the current data point being displayed.
    This also tells the operator where it is in the data collection cycle.
    Data is accumulated from every sensor at a rate of about 100 points per minute.
    Disp2.jpg

    The 12 Bit ADC Raw data display is a diagnostic display screen and is the output of the analog sensor ports (S1 =enclosure, and S2= motor) these should always be between Amax (a constant 4095) and Amin (should be zero but there is always an offset)
    AP is the digital output of the analog port the pressure sensor is connected to.
    The 12 Bit ADC Raw data is also used to calibrate the sensors.
    Disp1.jpg

    Clear as mud?
    Further details are provided in subsequent posts.

  2. # ADS
    Google Adsense Advertisement
    Join Date
    Always
    Location
    Advertising world
    Posts
    Many





     
  3. #2
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    Now here's a ramble about the background to this project (YAWN! - yeah I know)

    Rationale
    A pressure meter comes in very handy when installing a DC where a typical use is to measure the basic “sucking” performance (static Pressure) of a DC, and how well machinery is breathing when connected to ducting.

    Using a pressure meter has some problems and can be counter intuitive, one example being the flow is highest when the pressure drop is lowest.
    Pressure readings are also sensitive to the placement/orientation of the sensing tube eg placing a sensing tube at right angles to the flow will produce lower pressures compared to head onto the flow.

    Leaving those matters aside for the moment because multiple pressure readings are usually made before and after making specific changes to a DC system with the hope of making improvements, so as long as the pressure readings are done in the much the same way the results may sill beusable.

    U-tube Manometers.
    The cheapest possible pressure measuring instrument for DC use is simpleU-Tube manometer. This can be made from an ~1m length of transparent PVC tubing bent in half to form a U shape about 500mm long on each leg. Here's one I made about 10 years ago.
    IMG_3108.jpg
    The tube is fixed vertically to a stand/backboard and partially filled with water plus a bit of food colouring and a ruler is located vertical in between the arms of the manometer to measure the differences in height of the water columns on either side, Moving the two sides of the U close to the sides of a roller makes it easier to read.

    BTW If you decide too make one and try to measure the ma pressure difference produce by a vacuum cleaner like I was trying above WATCH OUT - vacs have as much a 30” (750 mm) of vacuum so you will need a manometer with at least meter long arms or you could end up lwith all the water in your vac! Fortunately on mine the vac was a piece of junk and only about 1/4 of the water went in!

    The typical maximum height difference observed between the water heights for conventional DCs is around 10” mm (250mm) of water column which is equivalent to ~2.5 kPa or 0.36 psi.

    The uncertainty of an "absolute pressure measurement" performed in this way depends on many factors like, the diameter of the tube, the viscosity of the water and its temperature, and things like parallax errors in reading the ruler and water positions.

    However, if we just want to make a comparison between two situations (eg with and without a BMH) we can using what is called a relative reading system where we assume we can use the same setup and measure everything the same way so most of these things can be ignore.

    To make a single position reading the tolerance in reading a ruler and water meniscus is at best (eg with 20/20 use, ruler hard up against the U-tube arms, good lighting etc) +/-1/2 mm, Then there is the other measurement on the other arm, so +/-1/2mm for that.
    So a single pressure reading with have +/-1 mm
    Then after making the required change or modification a second reading has to be made, +/- 1mm for that so +/-2 mm all up
    Now in reality the tolerance is more like double that so +/-4 mm which over say a 200 mm measurement is +/- 2%.
    However most pressure changes that result in improvements to DC systems are small, some a very small ie about 10 mm, so now you have 10mm +/- 4 mm or +/-40%!
    One you get below this you can end up chasing your tail and making things worse rather than better because very small changes (ie just bumping) a sensor placement can make a much big difference in readings.
    Besides taking all these readings and calculating even this simple difference becomes a PITA.

    Digital meters and pressure sensors
    So a few years ago I went a bought a cheap pressure meter like this.
    Manometer.jpg
    Here you can see the section of end caps and llttlemicro irrigation taps I use to measure pressures long the ducting as I built modified my system.
    Its really nice to use and measures a single pressure reading to +/-1mm, however once again all measurements have to be zeroed so that adds a +/-1mm, and there is is the second reading so we are back to +/- 4mm ie the basic U-Tube performance.

    So then I started to investigate pressure sensors. Most sensors are used for higher pressures, like that generated by a conventional compressor, which means we’re talking in the atmospheres, tens of psi or hundreds of kPa ranges.
    One psi is equivalent to 28” of water column so a 1 atmosphere meter will measure ~400" of water, 14.5 psi, or 100 kPa. Even if it can measure the pressure to +/-0.1% that translates to +/- 0.4” of WC, 0.015 PSI or 0.1 kPa which is again similar to the meters mentioned above.

    What would be needed would be something with a max pressure reading of say 20” of WC, 0.75 PSI or 5kPa so I looked around and there are a few but a) they are quite expensive and b) few have a good measurement tolerance.

    However I did stumble across a unit made by NXP called a "MPXV4006DP - differential Pressure Sensor”
    https://au.element14.com/nxp/mpxv400...a%2Fw%2Fsearch

    Screen Shot 2019-02-08 at 4.52.32 pm.png
    It's a 0 - 6kPa (24” WC), with a 766mV/kPA output from a ~5V input which means it will be nicely Arduino compatible.

    If you can reliably measure 1mV from this sensor then the theoretical uncertainty for a single read across full scale is 1:5000 or 0.02%!!!
    Across the 4 reads to make a complete relative measurement that s 0.08%
    Across full scale 6kPa that is equivalent of +/-0.0048 kPA or +/- 0.02”, while for a pressure difference of 10 mm of WC that will be +/- 0.5mm or +/- 5% which is an 8 fold improvement over the U-tube or cheap pressure meter.


    Of course provided the input and output are stable enough and if digitising systems have sufficient “ADC bit resolution” then it may be possible to get even more out of these sensors - see next post. Then if the data collection system is handled by a micro controller (MC) and the pressure being measured is stable enough then the MC could repeatedly measure the pressure and we could explore if the tolerance could be reduced with some statistical trickery??


    When I bought it the cost for just the sensor was $32r. The price unit now is $28 and if you buy >1000 you can get them for $16. I have since found out that similar sensors are used in water saving wash machines to measure the depth of water in the machine although I doubt that most will have a 0.5mm depth measurement capability. I actually bought the sensor a year or so but another project that never got started so I had it on hand and had to just collect up all the other parts.


    So the ultimate aim was to build a low pressure sensor with as low a tolerance or as high a sensitivity as possible. Of course cost was a factor but I was confident that the sensor would be by far the most expensive component of this project.


    In the next post I will describe the MC in a bit more detail.

  4. #3
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    As promised here is some more stuff about this project, this time some details about the Arduino micro controller (MC), and calibration
    Micro Controller
    I am far from even a half bottle on this topic - I just know enough to use basic Arduino MCs but I’m not even a half bottle on these either.
    Once you get used to programming one type of MC its just easier to stick with those until the disadvantages outweigh the laziness factor of sticking with what you know.

    Most basic/budget/entry level MCs are what is known as "10 bit boards". This means input analog signals like voltages can be sliced up or sampled into 2^10 (or 1024) steps or bits. The theoretical resolution or ultimate tolerance is then 1 bit in 1024 bits or ~0.1% at full range - in practice a zero point measurement is needed so a single read is more like +/-0.2%. Better boards are 12 or 14 bit (1:4096 bits or 1: 16k bits) but expect to pay more above that while basic sound sampling boards are 16 bit (1:64k).

    You need to make sure you you have enough input ports for all the sensors and as the Zero has 6 analog input ports (A0 to A5) see orange line) that is more than enough for the 2 temperature inputs and the pressure sensor. The 13 digital I/O ports are just blow the blue line near the top of teh photos

    Arduino-pm.jpg

    The first generation of Arduino boards were all known as "10 bit" boards but I decided I would go for a "12 bit" board like the “Arduino Zero”. The board is now officially outdated by Arduino but clones are still readily available and sell on eBay for about $10. The digital and analog input ports on the Zero are all 3.3V max boards so the input analog or digital signals cannot exceed 3.3V or the ports or even board may be damaged. In practice they can handle just over 3.3V for a short time but not much higher and not for long. This means any sensors like the pressure sensor I’m using that uses 5V and has a max analog output signal of close to 5V has to be used with care.

    The 5V max output for the pressure sensor is 5V for its full range of 6kPa (or 24" or 600mm of WC) but as my DC will not exceed more than 14” of WC (3.5 kPa) of vacuum this means the sensor will not output more than ~3V anyway so the Arduino ADC input port will not be damaged.

    The ~3V int signal from the pressure sensor will be spilt up across the 10^12 Bits or 4096 slices so each bit or slice represents 0.7 mV or ~1Pa of pressure/bit. This means at the full range of around 3500 Pa the singe "bit" or "slice" of 1Pa is 0.03% of the full scale - using the fact that 4 bits of resolution are needed for a relative measurement this makes the uncertainty around 0.12%.

    When measuring much smaller differences like the 10mm of WC (or 100Pa) pressure difference referred to in the first post, the 4 bits now represent 4/100 or +/- 4% which is 10 time better than than the U-tube method.

    In practice a number of things like poor V control and noise can kick in to make things some what worse than this. One factor is the quality of the power supply being used. The tiny matchbox size USB wall adapters are handy but are also notorious for being poorly regulated and noisy so much so that the Arduino Board voltage regulator cannot control the Voltage all that well. If you start to get crazy results, start by checking out your power. Thos small adapters should at most be left to charging batteries.

    The "data scatter dragon" can often be beaten back with repeated measurements and a bit of stats and this is why I have implemented variable numbers of data collection points. More about this in a later post.

    The thermocouples are fairly straight forward to implement. The K type thermocouples (Nickel/Chromium - Nickel/Alumel) wire junction produce about 20 microvolts/ºC at around 20ºC so cannot be plugged directly into the Arduino board and must be amplified to the mV/ºC range. I used a basic dual input/output op amp and a factor 50 amplification. It's a tad fiddly but it works.

    Calibration:
    The Thermocouples have to be calibrated to get meaningful results out of them.

    There are calibration tables available on the web but I reckon it is important to calibrate not just the thermocouples but the measuring system as a whole, including the actual power supply, actual Arduino board, any flying leads connection and junction boxes. If you later swap things around don’t assume the calibration will stay as it is - this is why I included a calibration/diagnostic data display in the Arduino software as I can then calibrate the thermocouples in situ if needed. Just need two thermos (one of hot and one of cold water ) and I can do a two point calibration in about 5 minutes.

    A full temp calibration involves measurements from about 80ºC down to near 0ºC. I use a small thermos and fill it full of a water ice mix and while measuring the temperature with a Fluke digital Thermometer (0.1ºC tolerance) and the water cools slowly I record the ADC outputs every ~5ºC. Then the data is plotted and two functions generated relating Temperature to ADC values for each sensor. It’s standard calibration stuff that I used to do at work - I can’t believe I’m doing it for fun right now.


    I just fit a linear function to the measured temp V ADC data in Excel and got very linear fits, (R^2-=0.9996 and 0.9990 ) for the two thermocouples.
    CalibS1S2.jpg
    The two functions are then put back into the Arduino program and the actual temperatures it measures with the two sensors are compared to an independent temperature source (eg my iced or hot coffee ). Several repeated calibrations may be required to get everything lined up across the temperature range but high accuracy is not a requirement for this purpose. If you are thorough and careful you should be able to get the two thermocouples to agree to within 1ºC, if not it might be to 2ºC - but ether way it’s not a big deal.

    The pressure calibration can be handled in two ways.
    Because we are not interested in absolute accuracy, calibration against even a coarse resolution comparator and fitting a function to the data will work because most measurements will be done across small pressure ranges where we assume there will be a reasonably tight linear response.

    The other method is to accept the manufacturers spec of 7.511 mV/mm of WC, and measure mV and convert that to pressure.

    Using my 1mm resolution manometer as the calibration reference I still obtained a very linear output response (R^2 - 0.9998!) and calculated the mV/mm of W to be 7.452 mV/mm, which is a difference of <1% from the manufacturers spec, which I am pretty happy with.
    Sorry about the "Prewsure" typo on the Y-axis label
    PressVADC.jpg
    What matters for relative measurements over a small pressure range is the reproducibility of measurements - more about this in later posts.

    Sorry this has been a long technical post but I hope maybe someone can gain something useful from it.
    tOmorrow is another hot one here i Perth and Monday we are baby sitting so it looks like it will be tuesday before I can fully install this unit an try it in situ.

  5. #4
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

    Default

    Really nice work and project built Bob, and a great find on that differential pressure sensor. Looking forward to follow up info, I can see me ditching the old wall mounted manometer I use to check filter constriction and knocking up something like this instead.

    Mike.

  6. #5
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

    Default

    Stupid forum gave an error and said I had to try again, then promptly posted two messages the same - This was the second - now edited.

  7. #6
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    Now a bit about the importance of measuring the Zero pressure, and demonstrating the measurement of a very small pressure difference.

    Normally we just measure a zero pressure reading (ie have both sensor inlets tat the same atmospheric pressure) and then subtract the zero reading from the pressurised reading, but when measuring small differences it it is really important that the Zero be measured reliably.

    This graph shows the importance of repeated measurements of the zero for long enough and often enough to reduce the statistical variability to a minimum.
    It is more important to measure the tolerance or uncertainty of the Zero value because its the tolerance that more fully defines the Zero value - ie upper a lower limits..

    Zero-readings.jpg
    There are 4 lines shown on teh graph
    The black line (labelled T) is a theoretical line that shows a plus and minus zigzag of the theoretical 1 bit (which fortunately also happens to be 1Pa or 0,001kPa) zigzagging back and forth from a nominal zero value of 0.017 kPa.

    The blue line/dots (labelled on graph as "60") are real measured data linking some 40 x 60 integrations over time. Each 100 integrations are averaged and each blue dot represents one of these averaged measurements which sit mostly inside the span of the black “zig zag” with occasional wanderings outs of ether black span.
    On average the blue line statistically achieves < +/- 1 bit of spread ie it's better than theory!

    The red line/dots (labelled as "5" on graph) is for 40 measurements each of 5 integrations - it's still pretty good but it deviates a further from the black "zig zag" than the blue line. Statistically it achieves a spread of +/- 2 bits of reproducibility - this will be fine for most pressure differences.

    The green line/dots (labelled "1") are for 40 single integrations (ie no averages) It’s 60 times faster to collect than the blue measurement but it is also statistically, on average, +/- 5 bits of reproducibility. ie not good for small differences.

    This takes me back to my days in the lab some 30 years ago - we spent 18 hours at a time perform some very small Isotope Ratio difference measurements (1 part in 36 million) and about a third of those hours was spent repeated measuring the zero or baseline.

    Here is an example of a very small (supposedly 20 Pa) measurement using the pressure sensor.

    What I did was attach a small PVC chamber made from a couple of white PVC irrigation pipe and brass pipe fittings to the sensor.
    DC temperature/pressure meter-chamber-jpg

    The chamber has a micro irrigation tap attached at one end so that the chamber can be opened to, or sealed off from, the atmosphere. Then I place the sensor and chamber on the floor and open the tap and set the sensor acquiring data (100 readings that are then averaged) for a few minutes (each dot on the graph represents one of these 100 averaged readings). Then I lift the Sensor and chamber up to a shelf ~1.8m above the floor and leave it there to collect data and then alternately move it from floor to shelf) to see what happens. I did this with the sensor connected to the laptop so I did not have to write down anything as the Arduino programming/debugging software allows for some rudimentary data collection

    The resulting graph shows the sensor can easily measure this change in vertical height. The difference it registers is - 10Pa whereas theory says it should be -20Pa, but I am not concerned about this discrepancy as there are many mitigating factors like; the fact that the room I'm doing this in is reasonably well sealed, while the chamber itself is not perfectly sealed, then there are temperature differences to consider etc . We are after all interested in just measuring small relative pressure changes.

    DC temperature/pressure meter-screen-shot-2019-02-10-3-33-32-pm-png

    Both the zero line and small pressure difference are being measured with a +/- of ~ 1Pa each so the difference is being measure to +/- 2Pa.

    A +/-2Pa uncertainty is equivalent to +/- 0.2mm, which is better by a factor of about 20 to a conventional U-tube measurement.

    I'm pretty happy with this.

    BTW to perform this experiment accurately the chamber, connection tubes and all fittings should be made of something like clean stainless steel, as PVC will expand and contract with changes in temperature and also absorb and release small amounts of gasses under changes in pressure.
    Attached Images Attached Images

  8. #7
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    The previous calibration turned out to be producing results that are about 3% too high and I put that down to the calibration rig being a bit leaky so I decided to make a more air tight calibration rig for the pressure sensor.
    This allows me to set any pressure over the range and it more or less stays at that pressure at least long enough for me to take calibration readings more reliably
    All stainless except for the screw piston knob and piston cap which are brass but are outside the pressure zone.

    Calibsetup1.jpg

    The screw piston is made from a SS hex socket screw, The head OD has been turned down to ~9mm and an O-ring groove cut into the piston head.
    The extra O-rings behind the piston are for extra security even though the piston is good to >40 PS (~100 over requirements)

    Piston.jpg

    Piston2.jpg

    The first thing this does is provide "not quite exquisite" but very fine input pressure adjustment - however, unfortunately it still leaked - ever tried to find a leak at low pressure - small bubbles take ages to develop under water, in fact at really low pressure they don't even develop! Initially I suspected the Poly pressure fittings as the tubing is only drip irrigation poly but I added an extra PVC olive (3mm long stub of 5mm ID clear PVC tubing) and this works fine. Then I suspected my screw valve so thats when I added the extra O-rings behind the piston

    After phaffing about a bit and getting totally bamboozled I gave up testing at low pressure and removed all the SS parts and tested them under water at 40 PSI! air pressure. I figured if I could seal it at 40 PSI it would stay sealed at 0.3PSI!. I did indeed find 2 leaks (see orange arrows two photos back). I was really pleased my home made piston did not leak even at those pressures.

    So I can now generate any pressure between 0 and ~3.5kPa and while it drifts slowly up and down over time it will stay steady for long enough at that pressure for me to get reliable readings of the sensor and digital manometer provided the temperature does not change. If anything the temperature effects are more obvious, Just handling the valves and taps for too long will raise the pressure up. Being all metal fittings thing equilibrate faster than plastic. There's also long term drift when I going in and out of the room, or the AC kicking in! Fortunately the software records everything so I can dial up a pressure and go away and come back in a few minutes and take a fair reading.

    Here is a graph of what's

    Presscurve.jpg

    A is the max value the sensor can read/output in bits (4096)
    B (purple line) is the sensor curve in bits (0 - 4095). At around "B" you can see the slight rise over about 2 minutes - that was me just entering the room!
    I then opened up T2 and let it show the sensors baseline, then at "O" I reapplied the pressure one turn at every step and left it stabilise about 20 seconds at each step - then let it sit for a while near max pressure.
    AT E2 I touched one of the SS Tee pieces with 2 fingers for about 3 seconds which sent it over the max limit. so I reduced the pressure and then did the 2 finger heat application again at E2. you can see how long it takes to stabilise. So "No touchy"when working - I might have to wear cotton gloves!

  9. #8
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    I finally got around to installing the sensors and micro controller for this system in the shed. Everything was working hunky-dory on the bench but I had no illusions that it would be a trouble free install. I was expecting a few problems, it's all part of the fun and frustration of trying new things out, and boy did I get it - in spades.

    The electrical noise being picked up by the two thermocouples (one attached to the motor and one hanging from an internal enclosure wall) used to measure the temperature was so erratic that their output was meaningless. When I stopped and thought about it the microvolt output of these sensors never stood a chance against the EM radiation by the 4HP motor.

    I thought about amplifying the thermocouple signal closes to the source instead of 4m back at the micro controller but instead decided to see what the common Arduino DHT22 temperature/humidity could do in these circumstances.

    I'm using these sensors on the compressed air humidity monitoring system so I have one of these sensors inside the compressor enclosure quite close to the 4HP motor and it seems to be providing reasonably accurate values back to the micro controller. However, both the sensors on the motor and the other one up against the inside wall of the enclosure failed miserably inside the DC enclosure.

    One major difference is that even though both the DC and compressor use 4HP motors the DC's motor is driven by a VFD and these produce a lot more of Electromagnetic Interference (EMI). The enclosure is clad in Colorbond sheet metal so would also constrain much of the EMI inside the enclosure making things worse. This effect is demonstrated by opening the enclosure doors and watching the sensor outputs. While the sensor up against the motor continues to fail the one up against the inside wall of the enclosure now works sometimes - maybe once every 4 or 5 seconds (these things have a read cycle of about 1s). With the doors open and taking the sensor off the motor and dangling it inside the enclosure alongside the other one up against the enclosure wall the output from the motor sensor is still garbage while the other still occasionally outputs a really number. I suspect this difference is because the motor sensor has a longer connecting lead which acts like and aerial and picks up more EMI. This is further reinforced by hanging up both sensors outside the closed door and now the wall sensor works continuously just fine and the ex-motor sensor with the longer lead works occasionally.

    So now I am off looking at ways to shield the sensors from EMI. So shielded cables and shielded connector boxes. This is OK for the wall sensor but what to do about the motor temperature as the sensor needs to be in failure close contact with the motor to have a chance of measuring something meaningful.

  10. #9
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

    Default

    Hi Bob, just had a quick read of the last post, I use two K type thermocouples in the vehicle, just the standard ones normally supplied with some DVM. K type with combined lead about a meter long. They run in the wiring looms and close to the 6 ignition coils, both sensor tips sit directly in the transmission fluid - I get zero interference:

    However I'm using special small interface boards made specifically for K-thermocouples, the thermocouple and lead (part of the thermocouple) must be soldered almost directly to the input pins of the interface board for temperature compensation and correct temperature readings to be outputted, the pickup boards are also mounted in the engine bay in an unshielded housing.
    Wiring from the boards and transmission pressure sensors run around the engine bay before entering the cabin, and straight into the pins of a micro, zero interference.

    Sounds like the simple Op-AMP might be the problem?

    Mike.

  11. #10
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

    Default

    BTW as you likely know, those sensor heads are conductive and have to be isolated from the measurement point if conductive, the special pickups use differential inputs and go nuts if any part of the pickup touches anything - got caught the first time

  12. #11
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    Quote Originally Posted by MandJ View Post
    BTW as you likely know, those sensor heads are conductive and have to be isolated from the measurement point if conductive, the special pickups use differential inputs and go nuts if any part of the pickup touches anything - got caught the first time
    Thanks for the info. I have only ever mounted them on or inside plastic boxes which probably explains why I have had no problems with them before. However they are still going nuts hanging mid air inside the enclosure so the EMI is still a problem.

  13. #12
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

  14. #13
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    Thanks for the links.

    I have been thinking of using shielded ethernet cable as a connector and an Al box with metal glands and ground all these (cable shield, box and glands) to the shed frame which is galv steel. The resistance between the shed frame and the shed electrical earth is 0.1Ω which is the limit of reading on my multimeter.

    The question then is, what to do with the other halves of the twisted pairs inside the cable, Nothing? or also wire those to the ground?

    I could of course try both and see what works.

  15. #14
    Join Date
    Nov 2013
    Location
    Caboolture QLD AU
    Posts
    781

    Default

    One question, if you plug the Sensor into your DVM, is there any sign of noise in the readings when it's sitting right next to the motor?

  16. #15
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,756

    Default

    Quote Originally Posted by MandJ View Post
    One question, if you plug the Sensor into your DVM, is there any sign of noise in the readings when it's sitting right next to the motor?
    Good question - I will investigate.

    AND ~ 1 hour later.

    Connected a sensor up to a V meter.
    Sensor has 3 leads Vcc, Grnd, and Data, just looking at the Grnd-Data V.

    First away from the motor.
    When sensor is unpowered voltages float around, DC <30mV, AC <1mv
    When powered (by microcontroller) the voltages fluctuate anyway because the data lines send data, , DC is 4.94 - 4.98V , AC 0 - 1.3V

    Now with sensor inside enclosure near the running motor
    Unpowered sensor (no leads) voltage floats around, DC is 0-300mV, AC is 0-400mV definitely more than away from the motor - especially teh AC.
    Powered (from microcontroller), DC output is same as away from the motor possibly wider range maybe 4.92 to 4.99V,
    AC is all over the place - up to 7V! this will only be the low frequency stuff.
    The AC depends on the length of the leads used, shorter leads generate ~3V AC.
    I'm surprise it has not damaged the ports on the micro controller.

    Same as before (ie near running motor) but now powered by battery (ie no microcontroller) - just a basic DC PS and no leads between PS and sensor
    DC <15mV. AC up to 300mV

    Might get my Oscilloscope down there and take a look

Page 1 of 3 123 LastLast

Similar Threads

  1. Rising temperature
    By Allan at Wallan in forum WOODIES JOKES
    Replies: 1
    Last Post: 15th January 2013, 06:35 PM
  2. dripping temperature/pressure relief valve
    By Reno RSS Feed in forum PLUMBING, ELECTRICAL, HEATING, COOLING, etc
    Replies: 0
    Last Post: 18th May 2009, 12:40 PM
  3. Problems with temperature
    By Allan at Wallan in forum WOODIES JOKES
    Replies: 2
    Last Post: 5th November 2008, 10:36 PM
  4. Temperature Switch
    By echnidna in forum PLUMBING, ELECTRICAL, HEATING, COOLING, etc
    Replies: 2
    Last Post: 23rd July 2006, 11:31 AM
  5. Pressure Pump - pressure switch setting?
    By Woody Allan in forum PLUMBING, ELECTRICAL, HEATING, COOLING, etc
    Replies: 4
    Last Post: 4th January 2006, 04:12 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •