Boats and such sometimes have fuel flow meters that you may be able to
use.
Though I would think setting up a gauge in your tank might be cheaper.
May not give you the instantaneous rate of flow but will let you know
how much oil is being used.
I want to get 'litres per day' on LCD in my living room. So - 'go check
the tank' is not an option ;)
Aadu
> Though I would think setting up a gauge in your tank might be
> cheaper. May not give you the instantaneous rate of flow but
> will let you know how much oil is being used.
>
> Shahid
On Mon, Nov 29, 2004 at 12:46:24PM +0200, Aadu Adok wrote:
>
> I want to know how much oil goes for my house heating.
> Oil flows via 10mm pipe from tank to burner (burner 'sucks' oil from
> tank).
>
> Searching the net, I found Swissflow's SF800, which converts fluid flow
> to TTL level signal.
>
> Maybe someone has experience in measuring fluid flow? Any comments on
> that topic? Any alternatives to SF800?
I have an oil stove as well and I've thought some about doing this. My idea
was to attach a timer to the thermostat and measure the time that the oil
burner is on and then multiply by the oil burn rate. The burner nozzle in my
stove has 0.65 stamped on it meaning that 0.65 gallons of oil will be burnt
in one hour of operation.
This would be simple to do, but I doubt the results would be very
accurate. Still, I would be able to get a fairly close figure as to what my
oil usage is. I would at least be able to track when I should call for
another oil delivery.
Good luck, Matthew.
--
Everyone was born right-handed. Only the greatest overcome it.
____________________________________________
piclist-bounces@mit.edu <> wrote:
> I want to get 'litres per day' on LCD in my living room. So - 'go
> check the tank' is not an option ;)
Knowing how tricky it is to accurately measure flow rates (we design
flowmeters for the oil industry amongst other things), I built myself an
ultrasonic gauge for a similar requirement. At present, I just display
litres remaining so I know when to fill up, converting that to 'litres per
day' would be very simple...
There are some extra features for intrinsic safety (to minimise the risk of
fire!) which I won't go into for liability reasons etc, but basically the
PIC generates a ping directly into the transducer then waits for the echo.
The tank is a cuboid, so it's easy to calculate remaining litres.
Apart from the protection components, the circuit is trivial (PIC + opamp
amplifier stage and comparator).
It would be easy to store total litres every hour, and display a running
total of 'litres per day' updated hourly if that's what you want.
Well, you could do that too but that is not what I had in mind. I was
thinking more on the lines of a gauge in the tank read by a
microcontroller which drives a display that you can set display option
on - like display liters/gallons remaining, gallons used today, amount
used since the beginning of month, amount used since last refuel, amount
of last refuel, $s spent on heating (today, this month, this year, this
season), warning when you are low on oil, etc etc.
IMO with a gauge in the tank you can get more useful readings than with
a flow meter. You'll just have to design a compensated gauge if you have
a cylindrical tank.
Shahid
> I want to get 'litres per day' on LCD in my living room. So
> - 'go check the tank' is not an option ;)
>> Though I would think setting up a gauge in your tank might be
>> cheaper. May not give you the instantaneous rate of flow but
>> will let you know how much oil is being used.
>>
>> Shahid
I have 1500 litre tank. On 'proper Estonian winter' (with temperatures
below -20 celsius), stove consumes about 1cm per day (which is about 10
litres). So - 1mm of oil level in tank corresponds to 1 litre.
What is the maximum precision you can squeeze out from ultrasonic ping
measurement? Should be well under 1mm?!
Aadu
> It would be easy to store total litres every hour, and
> display a running total of 'litres per day' updated hourly if
> that's what you want.
>
> Nigel
Any ideas how to implement _safe_ (no explosions) gauge thats sitting in
oil tank and outputs electrical signal which is processable by PIC?
We've been talking about ultrasonic ping. Any other ideas?
Aadu
> IMO with a gauge in the tank you can get more useful readings
> than with a flow meter. You'll just have to design a
> compensated gauge if you have a cylindrical tank.
>
> Shahid
>We've been talking about ultrasonic ping. Any other ideas?
It will not give the accuracy that has been talked of in this thread, but
one form is to use an acrylic plastic rod which is moulded or machined to
have steps in it A light source then shines down it, and a suitable light
detector measures the reflected light. The liquid in the tank changes the
reflection co-efficient at the step, thereby changing the amount of light
reflected to the sensor. It may be possible to have a stepless sensor by
having a corkscrew groove around a tapered rod, thereby possibly giving the
resolution that others have been talking about.
> litres). So - 1mm of oil level in tank corresponds to 1 litre.
>
> What is the maximum precision you can squeeze out from ultrasonic ping
> measurement? Should be well under 1mm?!
Max resolution at 4MHz clock would be 1us, if all the rest of the circuit
is up to the job. In 1us, sound travelling at 330m/s would travel 0.33mm.
My display is at 1 litre resolution, but accuracy is probably much lower.
Each time I take a number of readings and display the average.
I haven't studied absolute accuracy etc. This was never meant to be a
commercial design, it was really a very quick project to replace a broken
mechanical gauge, and it worked 'good enough' straight off. At the moment,
it just streams out TTL serial for display on an LCD inside.
> hi Nigel,
>
> I have 1500 litre tank. On 'proper Estonian winter' (with temperatures
> below -20 celsius), stove consumes about 1cm per day (which is about 10
> litres). So - 1mm of oil level in tank corresponds to 1 litre.
Use a capacitive gauge, it will be accurate enough after calibration. But
1 liter/hour is pushing it a little imho. As to measuring oil flow speed,
that would be worse. You are trying to measure a flow of ~0.09 cm/sec in a
10mm pipe assuming 1 liter/hour consumption. That is hard. Maybe you can
use a dripper and count drops optically or by pressure impulse. I don't
think that something else will work very well at this speed.
> What is the maximum precision you can squeeze out from ultrasonic ping
> measurement? Should be well under 1mm?!
You can't probably. Express it as percentage of full scale. F.s. is 1500mm
and you want to read 1mm/1500mm = 0.00066 so you measure 0.00033. That's
0.033% . Very hard.
Peter
____________________________________________
Aadu Adok wrote:
> hello,
>
> I want to know how much oil goes for my house heating.
> Oil flows via 10mm pipe from tank to burner (burner 'sucks' oil from
> tank).
>
> Searching the net, I found Swissflow's SF800, which converts fluid flow
> to TTL level signal.
>
> Maybe someone has experience in measuring fluid flow? Any comments on
> that topic? Any alternatives to SF800?
>
> thanks,
> Aadu Adok
I seem to remember our furnace nozzles are rated in GPH (gallons per
hour) here , So run time should be an indicator?
D
____________________________________________
Hello,
Yup the nozzle flow is in gallons per hour. Run time = gallons used.
On some of the newer equipment they are running higher pump pressure
~ 140 psi opposed to 100psi for better atomization so .65 nozzle = .75.
There is some info at http://www.delavaninc.com/pdf/total_look.pdf for
pressures and flow.
You could sense run time using the thermostat between the R and W
terminal 24v = off 0v = on, Or at the unit might be neater.
Not sure how likely an explosion would be (the TWA800 center fuel tank
exploding was just a bunch of doggy poop.) After all the newer cars have
their fuel pumps completely submerged inside gasoline with plenty of
vapor trapped inside when the tanks no completely full.
Depending on the construction of the tank you may not be able to put
anything inside the tank. One option might be a capacitive sensor where
you can have a vertical column outside the tank that fills with oil to
the same level as the level in the tank. Have two vertical electrodes in
the column and measure the capacitance between the two electrodes which
should change with the level of oil.
Shahid
> Any ideas how to implement _safe_ (no explosions) gauge thats
> sitting in oil tank and outputs electrical signal which is
> processable by PIC?
>
> We've been talking about ultrasonic ping. Any other ideas?
here's an idea I had years ago for my motorbike fuel tank. Never tried or tested it, don't know even if it would work so it's was just thought at the time.
Air pressure.........tape a balloon of suitable material at the bottom of an empty tank. pump air into it and then pour measured amounts of fuel into tank taking readings as it full up.
the problem would be to maintain a constant air pressure in the balloon and still measure a drop in air pressure due fuel loss.
In large pump stations we use two transducers and the SCADA software
checks both are within tolerance - alarms if one goes out of tolerance
an these switches control to the one still within tolerance i.e. within
4-20ma range.
They do fail - we replace about 5 a year in a total of 200 wet wells.
Usually they either fail completely or they drift in there 4ma area.
But still recommended in that they are a passive way of sensing a liquid
level.
Pay my air fare and I will come and set one up for you :-)))
_______________________________________
Roy
Tauranga
New Zealand
_______________________________________
> Any ideas how to implement _safe_ (no explosions) gauge thats
> sitting in oil tank and outputs electrical signal which is
> processable by PIC?
>
> We've been talking about ultrasonic ping. Any other ideas?
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.802 / Virus Database: 545 - Release Date: 26/11/2004
> here's an idea I had years ago for my motorbike fuel tank. Never tried or
> tested it, don't know even if it would work so it's was just thought at
> the time.
Air pressure.........tape a balloon of suitable material at the bottom of an
empty tank. pump air into it and then pour measured amounts of fuel into
tank taking readings as it full up.
/>
Washing machines (amongst other things) use air pressure liquid level
sensors. Pressure at the sealed top of a vertical tube with an open end
increases as liquid had increases. OK as long as you don't lose any air or
whatever gas used out of the tube.
Washing machines (amongst other things) use air pressure liquid level
sensors. Pressure at the sealed top of a vertical tube with an open end
increases as liquid had increases. OK as long as you don't lose any air or
whatever gas used out of the tube.
> an even better idea as the sensors r already made up. just need to
> interface them.
> wonder if this will work in my swimming pool:-))
The general method will definitely work for a swimming pool.
Washing machines tend to have a finite number of steps and often use
mechanical pressure sensors. You can in some cases hear them trip as the
water level rises and falls (a lot easier to hear if you play with the
sensor out of the machine ;-) ). If all you want is a control of fill level
then a washing machine unit should work fine.
There are numerous commercial level sensors available that use electronic
pressure sensors to do this job. You could build one yourself with nothing
more than a pressure sensor of approriate range and a piece of tubing.
Dave Smith wrote:
> here's an idea I had years ago for my motorbike fuel tank. Never tried
> or tested it, don't know even if it would work so it's was just thought
> at the time.
>
> Air pressure.........tape a balloon of suitable material at the bottom
> of an empty tank. pump air into it and then pour measured amounts of
> fuel into tank taking readings as it full up.
>
> the problem would be to maintain a constant air pressure in the balloon
> and still measure a drop in air pressure due fuel loss.
>
>
> any comments on the idea??
Even a tiny air leak would mess it up, and it would be highly temperature
dependent.
*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
____________________________________________
Washing machines do not have to hold the water for extended periods of time,
if they did the pressure would cause some of the O2 to be dissolved into the water resulting in false readings.
Each time the machine empties the tubing gets drained.
Oil might not absorb air (I do not know) so it might work using a pressure sensor and tubing as opposed to a pressure switch.
A better idea might be to fill the tube with mineral oil and read the vacuum created as the level drops.
But I think someone else mentioned runtime times nozzle flow (which is a constant for heaters) so the simplest and most reliable would be to measure run-time and calculate.
Might also want something to tell the µp when the tank has been filled for calculating fuel remaining.
KF4HAZ - Lonnie
----- From: "Dave Smith" <dave@
I shall keep this in mind.
I thought about sensing the level using 2 electrodes, but being a salt water pool I don't think the electrodes will last long.
Nobody has mentioned putting a heating wire in the flow, then a
temperature sensor downstream a little bit. The slower the fluid
flow, the hotter it will get as it moves past the heater. You'll need
another sensor before the heater, and take the temperature difference.
Obviously, you don't want to heat flammable liquids too much!
> hello,
>
> I want to know how much oil goes for my house heating.
> Oil flows via 10mm pipe from tank to burner (burner 'sucks' oil from
> tank).
>
> Searching the net, I found Swissflow's SF800, which converts fluid flow
> to TTL level signal.
>
> Maybe someone has experience in measuring fluid flow? Any comments on
> that topic? Any alternatives to SF800?
>
> thanks,
> Aadu Adok
>
If NiChrome wire is fed from a constant current source the voltage drop across it will be proportional to the temp.
So the heating wire could serve as the 2nd temp sensor as well.
No oil flow = max voltage drop across the heating wire, max oil flow = min voltage across the wire.
Voltage rise above this point would indicate out of oil, or burned out wire.
Compensating for oil temp with the 1st temp sensor this could work.
Only one problem, if much heating is generated, especially with little or no oil-flow the oil will probably begin to "coke up" on the wire.
KF4HAZ - Lonnie
----- From: "Mark Rages" <markrages@
> Nobody has mentioned putting a heating wire in the flow, then a
> temperature sensor downstream a little bit. The slower the fluid
> flow, the hotter it will get as it moves past the heater. You'll need
> another sensor before the heater, and take the temperature difference.
> Obviously, you don't want to heat flammable liquids too much!
>
> Regards,
> Mark
> markrages@gmail
Washing machines (amongst other things) use air pressure liquid level
sensors. Pressure at the sealed top of a vertical tube with an open end
increases as liquid had increases. OK as long as you don't lose any air or
whatever gas used out of the tube.
> an even better idea as the sensors r already made up. just need to
> interface them.
> wonder if this will work in my swimming pool:-))
The general method will definitely work for a swimming pool.
Washing machines tend to have a finite number of steps and often use
mechanical pressure sensors. You can in some cases hear them trip as the
water level rises and falls (a lot easier to hear if you play with the
sensor out of the machine ;-) ). If all you want is a control of fill level
then a washing machine unit should work fine.
There are numerous commercial level sensors available that use electronic
pressure sensors to do this job. You could build one yourself with nothing
more than a pressure sensor of approriate range and a piece of tubing.
John carmack (of Doom fame) is using such a sensor to monitor fuel level in
his rocket ship.
At 06:00 AM 11/30/2004, Dave Smith wrote:
>I shall keep this in mind.
>
>I thought about sensing the level using 2 electrodes, but being a salt
>water pool I don't think the electrodes will last long.
They can last a long time if you use AC drive and stainless steel
electrodes. We build automatic level controls for greenhouses that grow
cucumbers hydroponically. (no, they are NOT growing that other green stuff
<grin>). The nutrient solution is extremely active, containing various
acids and other nasty stuff (information that they choose not to share with
us) - we have not yet had to replace electrodes after several years of use.
dwayne
PS - controller is dirt cheap: 24 Vac transformer, a few resistors and
capacitors, simple half-wave power supply feeding a zener regulator, 4093
quad schmitt trigger NAND, 3011 opto, triac, single-sided PCB, case - it
feeds a 24 Vac solenoid but was also designed to be used with line-voltage
solenoids. It maintains nutrient level between upper and lower electrodes
suspended in the metal channel. I also did a version based upon a 12c508
but haven't yet built many (the 4093 version just works too well and I
don't yet need the special timing features that the PIC allowed me to
incorporate).
Celebrating 20 years of Engineering Innovation (1984 - 2004)
.-. .-. .-. .-. .-. .-. .-. .-. .-. .-
`-' `-' `-' `-' `-' `-' `-' `-' `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.
This method (measuring voltage drop across a nichrome wire) would seem to
depend on external environment too much to be a great solution. I'm
assuming the tank is outside, exposed to the weather. In the winter, the
incoming temperature would be colder than the incoming temperature in
summer, which would screw up the reading (?).
Perhaps a better method would be to pump a constant amount of power into the
oil flow, with a thermocouple both upstream and downstream. Then compute
the change in temperature and extrapolate from the heat constant of kerosene
what the flow must be. A lot of math for a PIC, but doable.
Perhaps measuring the pressure drop across an orifice would be easier? This
might also vary somewhat with temperature, but the calculations would be
easier after the initial calibration. If you want to look into this LMK, I
*love* fluid dynamics and would be thrilled to help out with the math.
--------------------------------------------
Robert B.
Same idea, but instead of 'balloon' say 'diaphram' or 'bellows'.
Put the pressure sensor on the bottom of the tank. Make sure it can
take the oil and pressure. If it can't handle the oil, put it on the
end of a wide pipe. Fill the pipe with some inert oil that is
compatible, then place a metal diaphram on the other end of the pipe,
the whole assembly then screws to the drain pipe on the bottom of the tank.
>here's an idea I had years ago for my motorbike fuel tank. Never tried or tested it, don't know even if it would work so it's was just thought at the time.
>
>Air pressure.........tape a balloon of suitable material at the bottom of an empty tank. pump air into it and then pour measured amounts of fuel into tank taking readings as it full up.
>
>the problem would be to maintain a constant air pressure in the balloon and still measure a drop in air pressure due fuel loss.
>
>
>any comments on the idea??
>
>
>Dave Smith
>Cape Town
>South Africa
>
Read more closely and you will see that the first temp sensor takes into account the initial oil temp, thereby compensating for it.
It is the difference in temp readings that tells the true flow rate.
> This method (measuring voltage drop across a nichrome wire) would seem to
> depend on external environment too much to be a great solution. I'm
> assuming the tank is outside, exposed to the weather. In the winter, the
> incoming temperature would be colder than the incoming temperature in
> summer, which would screw up the reading (?).
>
> Perhaps a better method would be to pump a constant amount of power into the
> oil flow, with a thermocouple both upstream and downstream. Then compute
> the change in temperature and extrapolate from the heat constant of kerosene
> what the flow must be. A lot of math for a PIC, but doable.
>
> Perhaps measuring the pressure drop across an orifice would be easier? This
> might also vary somewhat with temperature, but the calculations would be
> easier after the initial calibration. If you want to look into this LMK, I
> *love* fluid dynamics and would be thrilled to help out with the math.
>
> --------------------------------------------
> Robert B.
>
As you say....repeatability may not be any where near the technicaly
achievable resolution, But......the basic concept would be absolutly
perfect to monitor the water level in our house water tank.
What sort of transducer did you use ?
Would it be suitable for use in a damp environment ?
>
> Unrelatedly, but someone lately was looking for help with an air-flow
> sensor.
> The dandiest design I've seen uses a (ptc) thermistor in the flow
> (I *know*, I *know* those things are very non-linear...)
> Anyhows, this little matter was gotaround by configuring the part in a
> simple feedback cct. that maintained the ugly thermistor at a constant
> temperature.... >> ie constant PD.
>
> As flow changes, the current required to maintain the PD changes, and this
> is measurable & linear. Well, I suppose it is if you assume the heat
> transfer
> rate is proportional to flow rate, maybe only true in laminar conditions,
> or maybe only true in turbulence, or, ... what is this leading to...
>
best regards, John
email from the desk of John Sanderson.
JS Controls, PO Box 1887, Boksburg 1460, Rep. of S. Africa.
Tel/Fax 011 893 4154,
Cell 082 741 6275,
web http://www.jscontrols.co.za
Manufacturer & purveyor of laboratory force testing apparatus &
related products & services.
On Tue, 30 Nov 2004, Falcon Wireless Tech Support - KF4HAZ wrote:
> If NiChrome wire is fed from a constant current source the voltage drop
> across it will be proportional to the temp. So the heating wire could
> serve as the 2nd temp sensor as well. No oil flow = max voltage drop
> across the heating wire, max oil flow = min voltage across the wire.
> Voltage rise above this point would indicate out of oil, or burned out
> wire. Compensating for oil temp with the 1st temp sensor this could
> work. Only one problem, if much heating is generated, especially with
> little or no oil-flow the oil will probably begin to "coke up" on the
> wire.
Have you tried this ? Nichrome has a very low temperature coefficient
afaik. Also oil has relatively low thermal conductivity so a low power
device would work well. For example the transistor wind meter that is on
the web.
Peter
____________________________________________
> This method (measuring voltage drop across a nichrome wire) would seem to
> depend on external environment too much to be a great solution. I'm
> assuming the tank is outside, exposed to the weather. In the winter, the
> incoming temperature would be colder than the incoming temperature in
> summer, which would screw up the reading (?).
Oil tanks are never outside in cold areas. Oil gel is hard to pump ;-) It
would be at least buried and insulated if outside.
Peter
____________________________________________
> > assuming the tank is outside, exposed to the weather. In the winter,
the
> > incoming temperature would be colder than the incoming temperature in
> > summer, which would screw up the reading (?).
>
> Oil tanks are never outside in cold areas. Oil gel is hard to pump ;-) It
> would be at least buried and insulated if outside.
>
> Peter
I guess that depends on what you consider a "cold area". Here in Michigan
most people think it gets pleanty cold (about 28 F last night, expect -20F
before winter's over) and there are many uninsulates oil tanks outside
peoples homes.
At 12:46 PM 29/11/2004 +0200, you wrote:
>hello,
>
>I want to know how much oil goes for my house heating.
>Oil flows via 10mm pipe from tank to burner (burner 'sucks' oil from
>tank).
You could make a very accurate measurement using a small pump, two extra
small tanks, and a few opto sensors. If you feel like the heath robinson
approach, it's also cheap.
I presume your feed is a siphon out of the main tank as shown below. Move
your feed point to small tank B. Oil is pumped to tank A until full,
whereupon it drains into tank B. Tank A is calibrated, maybe 300mL. You
keep tank B full by delivering measured tots all day.
Use a stepper motor to drive the pump
The pipe between A and B should be ~5mm and the volume can easily be
accurately measured.
The method is to pump until fluid is sensed at B. Now reverse the pump
direction and count steps until sense at A. You now have steps/mL. Now pump
foward until the gravity feed tank is full. The calibration can be re-done
every half hour to cater for viscosity changes, pump wear.
While waiting for the gravity tank to drain, you could pump to B, add ten
steps, then hold. If opto B goes clear, then it's sucking back, and the
pump or tubes are leaking, and need repair. The drip point at the end
should be a U bend pointing upwards to prevent drain-out of the tube.
Actually oil has a high thermal conductivity.
And while I have not tried this, it is the same technique used in automotive Air-Flow sensors.
Everyone seems to be missing the fact that there would be 2 temp readings, one of the incoming temp, and the other that is altered by the dissipation of the flow.
Reading the incoming temp and comparing it to the voltage required to heat the wire gives a differential that is 1 relative to the flow, and 2 independent of the incoming temp.
Think about it, the air coming into your cars air filter can be as hot as 180f in summer just after a hot-restart, or as cold as -40 (in worst case locations)in winter on a cold-start.
If it works for Air-Flow which has a lower thermal conductivity than oil, it should work for oil-flow, provided the oil does not "coke-up" on the wire.
>
> On Tue, 30 Nov 2004, Falcon Wireless Tech Support - KF4HAZ wrote:
>
> > If NiChrome wire is fed from a constant current source the voltage drop
> > across it will be proportional to the temp. So the heating wire could
> > serve as the 2nd temp sensor as well. No oil flow = max voltage drop
> > across the heating wire, max oil flow = min voltage across the wire.
> > Voltage rise above this point would indicate out of oil, or burned out
> > wire. Compensating for oil temp with the 1st temp sensor this could
> > work. Only one problem, if much heating is generated, especially with
> > little or no oil-flow the oil will probably begin to "coke up" on the
> > wire.
>
> Have you tried this ? Nichrome has a very low temperature coefficient
> afaik. Also oil has relatively low thermal conductivity so a low power
> device would work well. For example the transistor wind meter that is on
> the web.
>
> Peter
> And while I have not tried this, it is the same technique used in
> automotive Air-Flow sensors. Everyone seems to be missing the fact that
> there would be 2 temp readings, one of the incoming temp, and the other
> that is altered by the dissipation of the flow. Reading the incoming
> temp and comparing it to the voltage required to heat the wire gives a
> differential that is 1 relative to the flow, and 2 independent of the
> incoming temp.
The problem with this and all differential measurements is precision. If
you run potential circuits through a precision analysis, you'll see what I
mean...
The full range (flow) corresponds to only a few degrees (temperature)
difference. You need to use a good circuit to get that difference with a
few percent precision. And it's not linear, so you need to figure out the
actual flow through experimenting (with a reference sensor, of course).
A practical way to get a reasonably precise consumption measurement is
probably to have a small reservoir between the tank and the burner, with a
lower and a higher threshold, and counting the number of times this
reservoir gets filled.
Gerhard
____________________________________________
>
>A practical way to get a reasonably precise consumption measurement is
>probably to have a small reservoir between the tank and the burner, with a
>lower and a higher threshold, and counting the number of times this
>reservoir gets filled.
Like rain, drip it through a tipping bucket sensor.
At least it will be well lubricated.
> And while I have not tried this, it is the same technique used in
> automotive Air-Flow sensors. Everyone seems to be missing the fact that
> there would be 2 temp readings, one of the incoming temp, and the other
> that is altered by the dissipation of the flow. Reading the incoming
> temp and comparing it to the voltage required to heat the wire gives a
> differential that is 1 relative to the flow, and 2 independent of the
> incoming temp. Think about it, the air coming into your cars air filter
> can be as hot as 180f in summer just after a hot-restart, or as cold as
> -40 (in worst case locations)in winter on a cold-start. If it works for
> Air-Flow which has a lower thermal conductivity than oil, it should work
> for oil-flow, provided the oil does not "coke-up" on the wire.
At the speed of the oil in the pipe at 1l/hour (op given figures) and 10mm
pipe diameter, diffusion, convection, and other effects (like swirling and
vibration induced externally as well as friction on the pipe walls, giving
uneven speed distribution in the pipe) will likely prevail over the actual
motion of the oil column so the heat based approach will work but it will
not necessarily measure oil flow. Again 1l/hour in a 10mm i.d. pipe is a
liquid column about 1.28 meters long (~4 feet), and the flow per second is
0.35 mm/sec. This is hard to measure. Things like a dripper with counter
or a capillary tube with bubbing would work here. Reducing the pipe
diameter to get more speed may not work if it is gravity fed (very likely
at this kind of consumption rate), due to head loss by friction in the
small metering tube. Even with a 2mm metering hole the speed in the hole
will be only about 9 mm/sec. Maybe enough for the thermal approach if the
tube is built well (insulated etc). Using a day tank as someone else
suggested would work better.
Peter
____________________________________________
>>> assuming the tank is outside, exposed to the weather. In the winter,
> the
>>> incoming temperature would be colder than the incoming temperature in
>>> summer, which would screw up the reading (?).
>>
>> Oil tanks are never outside in cold areas. Oil gel is hard to pump ;-) It
>> would be at least buried and insulated if outside.
>>
>> Peter
>
> I guess that depends on what you consider a "cold area". Here in Michigan
> most people think it gets pleanty cold (about 28 F last night, expect -20F
> before winter's over) and there are many uninsulates oil tanks outside
> peoples homes.
No, it depends on the 'poring temperature' of the oil ;-) US domestic
heating oil is apparently of good quality. Other places you need to heat
the oil if it's less than 10C.
Peter
____________________________________________
>
> >
> >A practical way to get a reasonably precise consumption measurement
> is
> >probably to have a small reservoir between the tank and the burner,
> with a
> >lower and a higher threshold, and counting the number of times this
> >reservoir gets filled.
>
> Like rain, drip it through a tipping bucket sensor.
> At least it will be well lubricated.
I would go with a constant displacement pump and possibly a small
accumulaor. Use the level or pressure in the accumulator to cycle
demand, each pump cycle produces a certain metric, count the pump
cycles and do the math.
Parker or Enerpac has all the hydraulic stuff you need for this (state
side), not sure if I would mess around with home made or hobby stuff
for this project.
Regards, John.
__________________________________
Do you Yahoo!?
The all-new My Yahoo! - What will yours do? http://my.yahoo.com
____________________________________________
Compared with Air, the technique works with air over a temp range of -40 to over 180f
and oil is a much better thermal conductor than air.
Why do you think they use mineral oil in "dummy loads" a 100w resistor can carry 1000w when submerged in a gallon of mineral oil, ok not indefinitely but long enough, more oil or a flowing oil would lengthen this, but the point is if a resistor that can handle 100w max in free air can handle 1000w in a gallon of mineral oil without any flow to carry away the hot oil, oil is definitely a better conductor of thermal energy.
(if the mineral oil contains pcb's the figure jumps to 1500w that is why the power companies used to use pcb's, now they use mineral oil)
> On Fri, 3 Dec 2004, Falcon Wireless Tech Support - KF4HAZ wrote:
>
> > Actually oil has a high thermal conductivity.
>
> Compared with what? Water's is higher foer example. So are all currently
> used refrigerents when liquefied.
>
> > And while I have not tried this, it is the same technique used in
> > automotive Air-Flow sensors. Everyone seems to be missing the fact that
> > there would be 2 temp readings, one of the incoming temp, and the other
> > that is altered by the dissipation of the flow. Reading the incoming
> > temp and comparing it to the voltage required to heat the wire gives a
> > differential that is 1 relative to the flow, and 2 independent of the
> > incoming temp. Think about it, the air coming into your cars air filter
> > can be as hot as 180f in summer just after a hot-restart, or as cold as
> > -40 (in worst case locations)in winter on a cold-start. If it works for
> > Air-Flow which has a lower thermal conductivity than oil, it should work
> > for oil-flow, provided the oil does not "coke-up" on the wire.
>
> At the speed of the oil in the pipe at 1l/hour (op given figures) and 10mm
> pipe diameter, diffusion, convection, and other effects (like swirling and
> vibration induced externally as well as friction on the pipe walls, giving
> uneven speed distribution in the pipe) will likely prevail over the actual
> motion of the oil column so the heat based approach will work but it will
> not necessarily measure oil flow. Again 1l/hour in a 10mm i.d. pipe is a
> liquid column about 1.28 meters long (~4 feet), and the flow per second is
> 0.35 mm/sec. This is hard to measure. Things like a dripper with counter
> or a capillary tube with bubbing would work here. Reducing the pipe
> diameter to get more speed may not work if it is gravity fed (very likely
> at this kind of consumption rate), due to head loss by friction in the
> small metering tube. Even with a 2mm metering hole the speed in the hole
> will be only about 9 mm/sec. Maybe enough for the thermal approach if the
> tube is built well (insulated etc). Using a day tank as someone else
> suggested would work better.
>
> Peter
> Compared with Air, the technique works with air over a temp range of -40
> to over 180f
The first problem is not the temperature range, it is the flow speed. The
calorimetric measurement you suggested doesn't work well at all at very low
speeds. There are all kinds of effects that influence the measurement far
more than the speed of the liquid once you get down to mm/s or slower. I
have designed such sensors; it is not a simple thing.
The air speeds of interest in the intake of cars are probably in a
relatively small range, and that range is far enough away from 0. This
doesn't apply to the oil consumption measurement. Also, they probably don't
care in the car whether the absolute measurement result is off by 20%. But
a 20% error would render the consumption measurement pretty useless -- you
think you have still a 20% full tank, but in reality it's already empty. A
dip stick would be more precise (and safer!) in this case :)
Gerhard
____________________________________________
hansa.ee> wrote:
