Searching \ for '[EE]Flow Sensor possibilities' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/io/sensors.htm?key=sensor
Search entire site for: 'Flow Sensor possibilities'.

Exact match. Not showing close matches.
PICList Thread
'[EE]Flow Sensor possibilities'
2006\10\31@021929 by cdb

flavicon
face
Some years ago I designed a water usage meter project for N&V
magazine that used a ready made turbine flow sensor.  

The drawback with this component was that it was expensive from the
DIY makers point of view, exacerbated by the fact it was only
available from RS Components.

I'm thinking of revamping the sensor to make it easier to 'roll your
own' and to this end I've got four ideas, which I'd like thoughts on.

Option 1. Heated coil around water pipe / downstream thermister
detection technique - measure temperature at coil and at thermister -
work out rate of cooling = flow speed in Ltrs/minute.

Option 2. Use either a differential or maybe two pressure sensors
inserted into pipe, difference in air pressure = rate of flow - not
too sure on this.

Option 3. Use a Piezo Film sensor like the Parallax tab device - to
measure pressure or vibration or maybe cable like the road speed
sensor units use. Perhaps this might work wrapped around a pipe or
easilt inserted into a section of pipe

Option 4. Similar to #3 use the flexiforce or muscle wire sensors.

What I'm looking for is something that would be easy for a magazine
reader to construct, is relatively easy to make reasonably accurate
and best of all slightly novel, hence options 3 and 4.

Naturally it would be Pic'arised at some point or maybe Zilogged

Thanks for your thoughts.

Colin
--
spam_OUTcolinTakeThisOuTspambtech-online.co.uk





--
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.427 / Virus Database: 268.13.18/506 - Release Date: 10/30/2006 6:15 PM


2006\10\31@024212 by Vasile Surducan

face picon face
On 10/31/06, cdb <.....colinKILLspamspam@spam@btech-online.co.uk> wrote:

> Option 1. Heated coil around water pipe / downstream thermister
> detection technique - measure temperature at coil and at thermister -
> work out rate of cooling = flow speed in Ltrs/minute.

water's temperature does not count ?

>
> Option 2. Use either a differential or maybe two pressure sensors
> inserted into pipe, difference in air pressure = rate of flow - not
> too sure on this.

depends what is flowing, a liquid is not compressible so the  pressure
inside a pipe  is quite constat  (assuming the pressure sensors are
one near the other and there is no pipe node or opened vane) Once a
vane is opening the preasure will decrease semnificatively only if
there isn't enough initial debit.

>
> Option 3. Use a Piezo Film sensor like the Parallax tab device - to
> measure pressure or vibration or maybe cable like the road speed
> sensor units use. Perhaps this might work wrapped around a pipe or
> easilt inserted into a section of pipe

any non-linearities on a pipe (ie a semiclosed vane) could cause vibrations
Some of them so big than could heat the water up to boiling. And this
is not a joke.

>
> Option 4. Similar to #3 use the flexiforce or muscle wire sensors.

flow means movement, right ?
A clean fluid flow could be detected optical ( or magnetical?).

Vasile

2006\10\31@025034 by Vasile Surducan

face picon face
On 10/31/06, Vasile Surducan <piclist9spamKILLspamgmail.com> wrote:
{Quote hidden}

like this: adsabs.harvard.edu/abs/2000SPIE.3990..264Y

2006\10\31@043821 by Brent Brown

picon face
cdb wrote:
> Option 1. Heated coil around water pipe / downstream thermister
> detection technique - measure temperature at coil and at thermister -
> work out rate of cooling = flow speed in Ltrs/minute.

Good start.. how about throw away the coil and just use the thermistor? Measure
the water temperature then run some current through the thermistor to try and heat
it, measure temperature again and/or temperature slope to compute flow rate.
Could be advantageous to pass a constant current through the thermistor and
measuring voltage drop. Few parts, I like it.

Some cars use a similar scheme to in mass air flow sensors in the intake manifold
(hot wire sensors)... so the idea is solid.

--
Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: 027 433 4069
eMail:  EraseMEbrent.brownspam_OUTspamTakeThisOuTclear.net.nz


2006\10\31@045438 by Gerhard Fiedler

picon face
Vasile Surducan wrote:

> On 10/31/06, cdb <colinspamspam_OUTbtech-online.co.uk> wrote:
>
>> Option 1. Heated coil around water pipe / downstream thermister
>> detection technique - measure temperature at coil and at thermister -
>> work out rate of cooling = flow speed in Ltrs/minute.
>
>  water's temperature does not count ?

It does; you need differential measurement. I don't think thermistors are
sufficiently precise, since you need pretty good synchronization between
the two sensors, at least over the range of water temperature. Which might
not be that much, so maybe you can calibrate thermistors sufficiently well.

Gerhard

2006\10\31@060142 by Genome

picon face
Good .. How does mass flow rate relate to temperature integral/slope/change
etc.... anybody can post the equation...

"Brent Brown" <@spam@brent.brownKILLspamspamclear.net.nz> wrote in message
news:KILLspam4547CFD4.10961.55795B36KILLspamspambrent.brown.clear.net.nz...> {Quote hidden}

> --

2006\10\31@061211 by cdb

flavicon
face
I hadn't thought about just using just a thermistor. I'll think I'll
delve into that a bit more.

Thanks

Colin

--
cdb, spamBeGonecolinspamBeGonespambtech-online.co.uk on 31/10/2006

Web presence: http://www.btech-online.co.uk  

Hosted by:  http://www.1and1.co.uk/?k_id=7988359

Do No Harm.

We must never do evil that good may come of it: William Penn 1693

.



--
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.427 / Virus Database: 268.13.18/506 - Release Date: 10/30/2006 6:15 PM


2006\10\31@062013 by cdb

flavicon
face
Magnetic might be interesting, optical will need holes drilled and
might be fiddlely to set up - the original flow sensor used was
optical - a hampster type turbine interrupted an IR diode - the setup
is simple, but making your own would be difficult especially as far as
inserting it into the flow.

Hence the reason for wanting something that could either be external,
or, perhaps just one or two easily drilled holes in a pipe.



:::: A clean fluid flow could be detected optical ( or magnetical?).
::::
:: like this: http://adsabs.harvard.edu/abs/2000SPIE.3990..264Y

Colin
--
cdb, TakeThisOuTcolinEraseMEspamspam_OUTbtech-online.co.uk on 31/10/2006

Web presence: http://www.btech-online.co.uk  

Hosted by:  http://www.1and1.co.uk/?k_id=7988359

Do No Harm.

We must never do evil that good may come of it: William Penn 1693

.



--
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.427 / Virus Database: 268.13.18/506 - Release Date: 10/30/2006 6:15 PM


2006\10\31@065305 by Alan B. Pearce

face picon face
>Good .. How does mass flow rate relate to temperature integral/slope/change
>etc.... anybody can post the equation...

I believe it is something you need to determine empirically. My memories of
folk here attempting to calibrate a similar system for determining helium
flow rates quickly realised the data given by the manufacturer was
simplistic.It depended on gas pressure (affects the density being heated,
probably not a problem for a liquid), the flow rate, and the initial
temperature. From these they ended up with a multidimensional curve to
determine flow rate.

2006\10\31@070522 by cdb

flavicon
face


:: Could be advantageous to pass a constant current through the
:: thermistor and measuring voltage drop.

That's a shame that technique is subject to  US patent 5493100  1996.  
:(

Though I suppose, taking the idea and making something of its kind
wouldn't be breaking the patent seeing as how it is analogue in design
and meant for blood flow measurement.

Colin

--
cdb, RemoveMEcolinspamTakeThisOuTbtech-online.co.uk on 31/10/2006

Web presence: http://www.btech-online.co.uk  

Hosted by:  http://www.1and1.co.uk/?k_id=7988359

Do No Harm.

We must never do evil that good may come of it: William Penn 1693

.



--
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.427 / Virus Database: 268.13.18/506 - Release Date: 10/30/2006 6:15 PM


2006\10\31@085657 by Genome

picon face
exactly how does this work...
Does this work by comparing the temperature reached when the thermistor is
heated after a constant time interval when the flowrate is zero(basis) to
that when flowrate is greater than zero.. the change would then be
proportional to flowrate assuming all other variables are constant like
presure and density and viscosity.. but there is also a problem of ambient
temperature when the (basis) temperature is taken versus the ambient
temperature of the current measurement..

anyone comment on this.. or am I wrong..


"Alan B. Pearce" <A.B.PearceEraseMEspam.....rl.ac.uk> wrote in message
news:01b301c6fce3$1f2dd170$EraseME1979f682spamspace.rl.ac.uk...
{Quote hidden}

> --

2006\10\31@093513 by Cristóvão Dalla Costa

picon face
Another idea is to use ultrasonic transducers. Send a ping down the pipe and
record the time it takes for the ping to travel to the second transducer at
a known distace. Calculate speed, subtract speed of sound in the fluid,
done.

Calibration might be easy: using a still fluid you can work out the speed of
sound in that fluid or the precise distance between the transducers.

Some industrial plants use that method.

On 10/31/06, cdb <RemoveMEcolinEraseMEspamEraseMEbtech-online.co.uk> wrote:
{Quote hidden}

> -

2006\10\31@102126 by Peter P.

picon face

Passing a constant current through a thermisor does not make so much sense from
the flow measurement point of view, but it is interesting that someone managed
to patent a standard ohmmeter in 1996 (never mind it is used to measure a
thermistor's resistance). I think that you can guess by now how your standard
DVM measures ohms.

I suggest using a pair of transducers (piezo disks) on two adjacent branches of
a T about 10cm long and 1cm diameter. This should give good signal probably (the
fluid needs to travel through the T, using a second T probably. The Doppler
shift for 1cm/sec flow (0.78 cm^3/sec or about 50ml/min) in that tube and
c=1500m/sec at 2kHz is about 3Hz which is measurable with a microprocessor.
Temperature compensation will be necessary. By choosing the carrier cleverly
(different with temperature f.ex.) most painful calculations can be removed from
the output side. For example it should be possible to shift the carrier trying
to keep the output frequency constant.

Peter P.


2006\10\31@103202 by Martin K

flavicon
face
Doesn't the resistance of a nichrome wire change with temperature?
If you put a fixed current through the nichrome wire (enough to heat it
a little) and measure the voltage difference you could figure out how
much it was being cooled by the flow.
--
Martin K

cdb wrote:
{Quote hidden}

2006\10\31@111139 by Alan B. Pearce

face picon face
>Passing a constant current through a thermisor does not make so much sense
>from
>the flow measurement point of view,

It does make sense, as what you measure is the cooling of the internal
thermistor heating by the liquid flow. As others have pointed out here it is
used to measure air flow in automotive intake manifolds, and it is sort of
used in the safety cut-outs in electrical heaters.

2006\10\31@121000 by Peter P.

picon face

> Doesn't the resistance of a nichrome wire change with temperature?
> If you put a fixed current through the nichrome wire (enough to heat it
> a little) and measure the voltage difference you could figure out how
> much it was being cooled by the flow.

Actually the resistance of Nichrome wire varies very little with temperature, it
is specially formulated for that purpose.

Using another material works much better. This is known as hot wire anemometry
when used in air or gas.

When using a thermistor the problem is its very high non-linearity in this mode.
The blood application probably got away with this because it measures a very
narrow range of flows over a very narrow range of media temeperatures.

The thermistor-heater-thermistor method has the problem of heat capacity. Water
as you know has a very high heat capacity so to register 1 deg C between the
thermometers (assuming no losses and so on) for 1cm^3/second flow one has to put
in about 4Watts into the heater. Obviously this gets out of hand very fast.
Heating flow meter that consumes more power than the heater proper ? For the
same reason the hot wire type of system is not usually used with liquids. Think
about how much power an on-demand water heater draws to heat a few liters per
minute of water by only 90 degrees C at most.

The same considerations apply for an integrating heater/thermometer flow meter.
To get a meaningful reading on the thermometer significant power must be input
during the heating cycle.

Peter P.


2006\10\31@121056 by Robert A LaBudde

flavicon
face
At 10:29 AM 10/31/2006, Martin wrote:
>Doesn't the resistance of a nichrome wire change with temperature?
>If you put a fixed current through the nichrome wire (enough to heat it
>a little) and measure the voltage difference you could figure out how
>much it was being cooled by the flow.

Called a "hot-wire anemometer". Used in all commercial aircraft to measure
windspeed.
================================================================
Robert A. LaBudde, PhD, PAS, Dpl. ACAFS  e-mail: RemoveMEralTakeThisOuTspamspamlcfltd.com
Least Cost Formulations, Ltd.            URL: http://lcfltd.com/
824 Timberlake Drive                     Tel: 757-467-0954
Virginia Beach, VA 23464-3239            Fax: 757-467-2947

"Vere scire est per causas scire"
================================================================

2006\10\31@122511 by Peter P.

picon face

Alan B. Pearce <A.B.Pearce <at> rl.ac.uk> writes:

> >Passing a constant current through a thermisor does not make so much sense
> >from the flow measurement point of view,
>
> It does make sense, as what you measure is the cooling of the internal
> thermistor heating by the liquid flow. As others have pointed out here it is
> used to measure air flow in automotive intake manifolds, and it is sort of
> used in the safety cut-outs in electrical heaters.

Yes, however both applications aim for relatively constant mass flow, i.e. the
setpoint and not the linearity is interesting. Additionally an auto ECU has
enough computing power and storage to linearize almost anything and the mass
flow on an engine varies less than 8:1 over all afaik. That is hardly a
'measuring instrument' imho. Knowing that hot thermistor anemometry is the poor
cousin of hot wire anemometry one can probably say that someone saved $20 on a
$10,000 car using the thermistor (but this verly likely does not reflect in the
price of the spare part when it needs to be bought retail). Come to think of it,
I think that a pressure differential type of mass flow metering on a car would
be more accurate and also affordable now (dual pressure gauge on chip, one for
static one for inlet manifold vacuum at constant area and shape of the manifold
should do it).

Peter P.


2006\10\31@122901 by Gerhard Fiedler

picon face
Peter P. wrote:

> The thermistor-heater-thermistor method has the problem of heat capacity.
> Water as you know has a very high heat capacity so to register 1 deg C
> between the thermometers (assuming no losses and so on) for 1cm^3/second
> flow one has to put in about 4Watts into the heater. Obviously this gets
> out of hand very fast.

This method works and is being used in industrial flow sensors. Your
mistake is that you are trying to heat the water. You don't heat the water,
you heat the heater/sensor element. The water cools the heater/sensor
element. The thermal resistance between the heater/sensor element and the
water is a function of flow speed. Therefore eg. the power to keep the
heater/sensor element at a constant temperature difference to the water
temperature is a function of flow speed. (There are other ways to measure
the same effect.)

The non-trivial bit here (depending on the precision requirements and the
temperature range of the water) is to get the two temperature sensors (the
heater/sensor and the base water temperature sensor) to run in sync. Don't
forget that you are measuring a difference, and all the precision problems
related to measuring differences. (Say you want to measure a difference of
1% full range. If you have a precision of each sensor of 1% full range, the
result is meaningless.)

Gerhard

2006\10\31@132040 by Peter P.

picon face
Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

> Peter P. wrote:
>
> > The thermistor-heater-thermistor method has the problem of heat capacity.
> > Water as you know has a very high heat capacity so to register 1 deg C
> > between the thermometers (assuming no losses and so on) for 1cm^3/second
> > flow one has to put in about 4Watts into the heater. Obviously this gets
> > out of hand very fast.

> The non-trivial bit here (depending on the precision requirements and the
> temperature range of the water) is to get the two temperature sensors (the
> heater/sensor and the base water temperature sensor) to run in sync. Don't
> forget that you are measuring a difference, and all the precision problems
> related to measuring differences. (Say you want to measure a difference of
> 1% full range. If you have a precision of each sensor of 1% full range, the
> result is meaningless.)

I know that this method is in use. But still, I do not understand how the high
heat capacity problem is circumvented. Heat flow through an exchanger is
proportional to the mass of liquid that goes past it at constant temperature
f.ex. (in laminar flow). So to detect a range of 10:1 in flow the thermal
resistance to the sensor should be of the order of magnitude of the equivalent
thermal resistance at high flow, or there will be huge errors. So imho assuming
the temperature difference is only 1C and knowing the static cooling capacity of
water which is around 0.2K/cm^2*W, the heater requires at least 5W per cm^2 of
exposed  heated element at rest! I don't think that there are low power
versions, unless they are micromachined or such.  

Here is an example of device:

http://www.bronkhorst.com/en/products/liquid_flow_meters_&_controllers/liqui-flow_l30_series,_digital_style/

Note temp. rise min. 1C ! At 1kg/hour this is at least 300mW in the heater, but
1kg/h is a very low flow.

Omega has a listing of methods and important advice imho:

http://www.omega.com/techref/flowcontrol.html

Peter P.


2006\10\31@134221 by Orin Eman

picon face
On 10/31/06, Peter P. <EraseMEplpeter2006spamspamspamBeGoneyahoo.com> wrote:
>
> Alan B. Pearce <A.B.Pearce <at> rl.ac.uk> writes:
>
> > >Passing a constant current through a thermisor does not make so much sense
> > >from the flow measurement point of view,
> >
> > It does make sense, as what you measure is the cooling of the internal
> > thermistor heating by the liquid flow. As others have pointed out here it is
> > used to measure air flow in automotive intake manifolds, and it is sort of
> > used in the safety cut-outs in electrical heaters.
>
> Yes, however both applications aim for relatively constant mass flow, i.e. the
> setpoint and not the linearity is interesting. Additionally an auto ECU has
> enough computing power and storage to linearize almost anything and the mass
> flow on an engine varies less than 8:1 over all afaik. That is hardly a
> 'measuring instrument' imho. Knowing that hot thermistor anemometry is the poor
> cousin of hot wire anemometry one can probably say that someone saved $20 on a
> $10,000 car using the thermistor (but this verly likely does not reflect in the
> price of the spare part when it needs to be bought retail).

The car has a feedback system (oxygen sensor) so it doesn't need to be
too accurate... but I doubt a thermister in a car intake system is
measuring air flow, just temperature.

Come to think of it,
> I think that a pressure differential type of mass flow metering on a car would
> be more accurate and also affordable now (dual pressure gauge on chip, one for
> static one for inlet manifold vacuum at constant area and shape of the manifold
> should do it).

Two main ways of EFI - a hot 'wire' mass air flow sensor or "speed
density" which uses manifold pressure, temperature, RPM and an engine
specific volumetric efficiency table.  Plenty of fun reading on
diy-efi.org on these subjects.

Orin.

2006\10\31@145711 by David P Harris

picon face
cdb wrote:

>Option 2. Use either a differential or maybe two pressure sensors
>inserted into pipe, difference in air pressure = rate of flow - not
>too sure on this.
>  
>
Seems to me this might work using the Venturi and *Bernoulli's
principles*.  Depending on the flow rate, this could generate a good
pressure.  Wikipedia claims there are two different equations for
compressible and non-compressible fluids.

See: <http://en.wikipedia.org/wiki/Venturi> and
<http://en.wikipedia.org/wiki/Bernoulli%27s_principle>

Cheers,
David


2006\10\31@160710 by Gerhard Fiedler

picon face
Peter P. wrote:

> I know that this method is in use. But still, I do not understand how
> the high heat capacity problem is circumvented. Heat flow through an
> exchanger is proportional to the mass of liquid that goes past it at
> constant temperature f.ex. (in laminar flow).

It's the medium speed rather than the mass flow, unless you're talking
about pipes that are very thin relative to the sensors. As long as you're
using small sensor areas, it's not really the mass flow. The sensor gives
the same output for a small pipe and a large pipe as long as the medium
speed is the same. Mass flow is of course higher in the large diameter
pipe.

> So to detect a range of 10:1 in flow the thermal resistance to the sensor
> should be of the order of magnitude of the equivalent thermal resistance
> at high flow, or there will be huge errors. So imho assuming the
> temperature difference is only 1C and knowing the static cooling
> capacity of water which is around 0.2K/cm^2*W, the heater requires at
> least 5W per cm^2 of exposed  heated element at rest! I don't think that
> there are low power versions, unless they are micromachined or such.

1 cm² is huge, though. The ones I know expose a few mm² (eg. a part of a
thick film substrate).

> Here is an example of device:
>
> http://www.bronkhorst.com/en/products/liquid_flow_meters_&_controllers/liqui-flow_l30_series,_digital_style/

Here are some more:
http://www.captor.com/index.php?option=com_content&task=view&id=39&Itemid=69

Typ. power consumption is around 5 W, for flow speeds up to 2 m/s.

Gerhard

2006\10\31@173231 by Peter P.

picon face

Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

> Peter P. wrote:
> > So to detect a range of 10:1 in flow the thermal resistance to the sensor
> > should be of the order of magnitude of the equivalent thermal resistance
> > at high flow, or there will be huge errors. So imho assuming the
> > temperature difference is only 1C and knowing the static cooling
> > capacity of water which is around 0.2K/cm^2*W, the heater requires at
> > least 5W per cm^2 of exposed  heated element at rest! I don't think that
> > there are low power versions, unless they are micromachined or such.
>
> 1 cm² is huge, though. The ones I know expose a few mm² (eg. a part of a
> thick film substrate).

Ok, but thick film hybrids are out for homebrew. And a homebrew unit made with
say  thin PCB and SMD transistors used as heaters would be limited to about 10W
and expose at least 100mm^2. So imho this technology is not suitable for
homebrew.

Peter

2006\10\31@184812 by Gerhard Fiedler

picon face
Peter P. wrote:

>> 1 cm² is huge, though. The ones I know expose a few mm² (eg. a part of a
>> thick film substrate).
>
> Ok, but thick film hybrids are out for homebrew. And a homebrew unit
> made with say  thin PCB and SMD transistors used as heaters would be
> limited to about 10W and expose at least 100mm^2. So imho this
> technology is not suitable for homebrew.

I wouldn't use an external heater. Using the temperature sensor itself as
heater (which is usually possible with resistive sensors), you can make a
suitable sensor yourself and you also don't have really a problem with the
required power. Unless you want to make a battery-operated device :)

Gerhard

2006\10\31@190400 by Peter P.

picon face
Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

> I wouldn't use an external heater. Using the temperature sensor itself as
> heater (which is usually possible with resistive sensors), you can make a
> suitable sensor yourself and you also don't have really a problem with the
> required power. Unless you want to make a battery-operated device :)

I looked it up some more and the method is limited to low flows. When high flows
are measured then the measuring branch is configured as a side branch that short
circuits a metering orifice or restriction. Also most instruments are limited to
a  flow range of 20:1 or so.

I think that it makes sense to consider a flow meter a specialised instrument
and use one that suits tha application. I.e. in the context of the o.p. there is
no such thing as a universal flowmeter. And the water wheel type may be the best
for his type of application.

Peter


2006\10\31@193604 by Genome

picon face
Is it really not very accurate ... then why do they use hotwire anemometry
in investigation of boundary layers on windtunnels, aircraft tubomachinerry,
propellers.. etc.. they say it could measure flow well beneath 1% accuracy
and very reactive too as they use it also on turbulence investigation too..


"Peter P." <RemoveMEplpeter2006KILLspamspamyahoo.com> wrote in message
news:loom.20061031T180956-162STOPspamspamspam_OUTpost.gmane.org...> {Quote hidden}

> --

2006\10\31@194544 by Genome

picon face
and that measurement using ultasonic by measuring doppler is great too but
speed of sound in water is affected by many things including presure density
viscosity and not to mention salinity too.. I think they use this ultrasonic
to measure salinity on seawater... but that measures only the speed of
sound in water and correlating it to salinity with all other variables
constant maybe take a look at sutherlands formula.. however extracting speed
of flow from doppler effect is going to be tricky...

"Peter P." <spamBeGoneplpeter2006STOPspamspamEraseMEyahoo.com> wrote in message
news:KILLspamloom.20061031T180956-162spamBeGonespampost.gmane.org...> {Quote hidden}

> --

2006\10\31@195152 by cdb

flavicon
face
Well I don't need the accuracy to be to industrial standards this is
for home construction after all.

So boring one or two holes for a thermistor or two (perhaps
thermistor/thermocouple combination) would be easy from a construction
point of view.

The air pressure one, I'm sure would work, but obviously some sort of
membrane would have to be used to stop water getting into the pressure
sensor.

For some reason doing something 'weird' with piezo stuff appeals to
me. Is there a formula that might correlate force/compression into
flow? I'm thinking along the lines of the faster or slower a liquid
flows, pressure is expanded against the sides of the vessel (in this
case a pipe) which a thin film piezo device might be able to pick up,
two would be needed one above a constriction and the other below for
measuring time difference - or have I just confused myself?

Still the thermistor or possibly ultrasonic methods might be the way
to go. I looked for some capacitive sensors that might do the trick,
but drew a blank.

Colin

--
cdb, EraseMEcolinspamEraseMEbtech-online.co.uk on 1/11/2006

Web presence: http://www.btech-online.co.uk  

Hosted by:  http://www.1and1.co.uk/?k_id=7988359

Do No Harm.

We must never do evil that good may come of it: William Penn 1693

.



--
No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.427 / Virus Database: 268.13.21/509 - Release Date: 10/31/2006 4:10 PM


2006\10\31@201010 by Genome

picon face
I dont think so for gas flow venturi is good but for water its not good as
water is treated as incompressible at low velocities unless you are
measuring speed of water at xtremely high velocities 300m/s or so.. if you
measure the preasure difference in a venturi at low speed ther would be
almost no difference and probably wahat ever difference you measure is due
to measurement error or simply due to turbulence.. we use to measure long
ago preasure gradient over and bellow a wing section on a wind tunnel using
arrays of utube manometers.. and we joke sometimes about our eyes going up
and down trying to read the scales heheheeh....

"David P Harris" <@spam@dpharris@spam@spamspam_OUTtelus.net> wrote in message
news:spamBeGone4547A6F5.1020502spamKILLspamtelus.net...
{Quote hidden}

> --

2006\10\31@203734 by Genome

picon face
omg wat am I thinking its the density that is constant.. hehehe.. but as its
flowrate were after.. I dont think it would work..

"Genome" <.....jtroxasspam_OUTspamgmail.com> wrote in message
news:ei8s52$fhp$TakeThisOuT1.....spamTakeThisOuTsea.gmane.org...
{Quote hidden}

>> --

2006\10\31@213840 by Genome

picon face
Maybe you can use 1 LM35 temp sensor then tie a heating coil around it and
cover it with maybe some epoxy. Measure the initial temperature of the
water.. then heat the lm35 through the coil a few degrees above the initial
reading and then wait for the time it took for the LM35 to cool back.. the
time elapsed would be a good indication of flow speed change.. you just have
to calibrate it with some empirical measurement... this way its cheap though
maybe not be too accurate..


"cdb" <RemoveMEcolinspamspamBeGonebtech-online.co.uk> wrote in message
news:20061031172040.343508@btech-882bd9da7...
{Quote hidden}

> --

2006\10\31@215138 by David P Harris

picon face
Then how do the venturi apparatus in chemistry labs produce meaningful
suction connected to a water tap?

David

Genome wrote:

{Quote hidden}

>>--


'[EE]Flow Sensor possibilities'
2006\11\01@091103 by Gerhard Fiedler
picon face
Genome wrote:

> Maybe you can use 1 LM35 temp sensor then tie a heating coil around it
> and cover it with maybe some epoxy. Measure the initial temperature of
> the water.. then heat the lm35 through the coil a few degrees above the
> initial reading and then wait for the time it took for the LM35 to cool
> back.. the time elapsed would be a good indication of flow speed
> change.. you just have to calibrate it with some empirical
> measurement... this way its cheap though maybe not be too accurate..

Something like this can work, but you'll have a real hard time to measure
the cooling time. It's an inverse exp function, and it slowly smoothes down
to 0. It's not easy to measure the exact point in time when it goes below a
certain threshold. The error (in time) can easily become bigger than
desired.

Gerhard

2006\11\01@093907 by M. Adam Davis

face picon face
The pressure sensor method is usually performed as follows:

Assume you have a pipe with two sections, one which is 1" in diameter,
the other is 3/4" in diameter.  If the liquid is not moving, then the
pressure is the same in both sections of pipe.

If the liquid is moving at one gallon per minute, then the flow
through the larger pipe is slower than the flow through the narrower
pipe.  This translates to a pressure difference in the two pipes,
which is related to the speed of the flow.  A single sensor can be
used for this - connect one side to one pipe, the other side of the
sensor to the other pipe, and the sensor only outputs the difference.

Not great for a wide range of flows - a setup that can handle and
measure hundreds of gallons per minute won't be able to easily detect
a 1 gallon per minute flow.  Conversely, a setup that can measure
differences of mL per second isn't going to be able to handle gallons
per second.  So as long as you have a reasonably narrow range where
you need accurate results you should be able to design a cheap sensor
of this type.

-Adam

On 10/31/06, cdb <RemoveMEcolinEraseMEspamspam_OUTbtech-online.co.uk> wrote:
{Quote hidden}

> -

2006\11\01@115755 by Vasile Surducan

face picon face
On 11/1/06, cdb <EraseMEcolinspam@spam@btech-online.co.uk> wrote:
> Well I don't need the accuracy to be to industrial standards this is
> for home construction after all.

http://en.wikipedia.org/wiki/Flow_measurement

read carefully the links.

Vasile

2006\11\02@015948 by Peter P.

picon face
Genome <jtroxas <at> gmail.com> writes:

> Is it really not very accurate ... then why do they use hotwire anemometry
> in investigation of boundary layers on windtunnels, aircraft tubomachinerry,
> propellers.. etc.. they say it could measure flow well beneath 1% accuracy
> and very reactive too as they use it also on turbulence investigation too..

They are using it in air (not liquid), as precision laboratory research
equipment. Not exactly what you'd solder up on a saturday evening from mail
order parts. 1% accuracy does not say anything about the span.

Peter P.


2006\11\02@022052 by Peter P.

picon face

Genome <jtroxas <at> gmail.com> writes:

> and that measurement using ultasonic by measuring doppler is great too but
> speed of sound in water is affected by many things including presure density
> viscosity and not to mention salinity too.. I think they use this ultrasonic
> to measure salinity on seawater... but that measures only the speed of
> sound in water and correlating it to salinity with all other variables
> constant maybe take a look at sutherlands formula.. however extracting speed
> of flow from doppler effect is going to be tricky...

The speed of sound in air is affected by even more things than the speed of
sound in water. Ultrasound speed measurement in liquids is a standard method in
f.ex. medicine (blood speed among others). This despite the liquid being
'contaminated' with live cells and other things (and contained in a living,
pulsing container, as in heart for example). Imaging doppler ultrasound can be
used to image turbulence in liquid in real time. But not homebrew ;-)

Extracting the doppler shifted return is not so hard over the distances I
suggested. At 2kHz and 10cm piezo separation in water without bubbles I'd expect
less than 20dB loss even if the sound turns a 90degree corner in the middle of
the T (or cross) where the flow passes. That would be like 500mV pk-pk out from
5Vpk-pk drive. A bigger problem would be getting the air out of the closed 'T'
or 'X' branches. Also at constant frequency (or wavelength <g>) it is possible
to use a trick to cancel the 'carrier' using standing waves at the receiver
(wavelength is about 75cm at 2kHz in water so a ~foot long sensor tube could use
the standing wave effect already).

Peter P.


2006\11\02@112301 by Genome

picon face
I've seen something called called clamp on ultrasonic flow meters on the net
that rely on doppler effect and some on measuring transit flight for clear
fluids and they dont use large distances but only about the diameter size of
a pipe.. from what i read it function by receiving the refflected signals
from microscopic particles embedded in the fluid and measuring doppler.. in
laymans terms like the way police measure speeding cars on a roadway.. they
use high freaquencies with wavelength less than the diameter of the
particles it will reflect to.. if i remember correctly from college days
waves only reflect on objects with diameters greater than their wavelength..
so they might be using ultrasound frequencies around 1mhz to have it reflect
on particles microns in size.. Im interested too on building a flowsensor so
if the guy maybe succeeds on building a homemade lowcost flowmeter.. I would
love to read that issue...

"Peter P." <@spam@plpeter2006spam_OUTspam.....yahoo.com> wrote in message
news:spamBeGoneloom.20061102T080004-707EraseMEspampost.gmane.org...> {Quote hidden}

> --

2006\11\02@130043 by Aaron

picon face


Peter P. wrote:

{Quote hidden}

I formerly worked here:
http://www.vacuuminst.com/airleaktesters.htm

We had our own patented mass flow sensors that we primarily integrated
into industrial test systems.  Our spec claimed 1% (of full scale)
accuracy and 0.1% repeatablilty.  When operating below about 5% of the
full-scale calibration, my opinion was that this spec was a bit optimistic.

Our mass flow sensors used two thermisters suspended in the air stream.  
One was heated by a constant current and the other one was used for
temperature compensation.

Datasheet link to the basic flowmeter:  
http://www.vacuuminst.com/PDF/7000SERIES.pdf

Aaron

2006\11\02@152933 by Art

picon face

>
>
>We had our own patented mass flow sensors that we primarily integrated
>into industrial test systems.  Our spec claimed 1% (of full scale)
>accuracy and 0.1% repeatablilty.  When operating below about 5% of the
>full-scale calibration, my opinion was that this spec was a bit optimistic.
>
>Our mass flow sensors used two thermisters suspended in the air stream.
>One was heated by a constant current and the other one was used for
>temperature compensation.
>
>Datasheet link to the basic flowmeter:
>http://www.vacuuminst.com/PDF/7000SERIES.pdf

Thanks Aaron, at last a voice of reason!

When I was working in the semiconductor industry (20 years ago), we
use MFM's and MFR (mass flow regulators and mass flow meters). In a
typical application, we need to hold the volume of a gas flow at very
constant rate of SLM (Standard Liters per Minute).

The basic meter was a high flow tube (that carried the bulk of the
actual volume). In parallel with that was a much smaller tube, which
carried anywhere from 100 to 1000 times less volume. It was a simple
matter to heat the smaller tube at one end and then to detect the
temperature difference about 1 inch downstream. So, there was a
thermocouple just after the flow diverter to measure the inlet
temperate of the small tube and a second thermocouple 1 inch
downstream of the heater (which was located between the inlet
thermocouple and the downstream thermocouple). In those days,
thermocouples were the rule of the day, but these days simpler
temperature measuring systems are probably adequate.

A simpler system was developed just as I left the semiconductor
business...it used the same basic hardware, but measured the current
necessary to maintain the temperature sensor at a constant
temperature. It used insulation which was a little bulky, but only
one thermocouple was needed. I didn't actually work on these newer
units, so I don't have actual experience with them.

I think the company that made them was called MKS. I didn't google
them before writing this.

My guess is that the concept should work just as well for liquids as
it does for gasses.

Art


--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.409 / Virus Database: 268.13.23/513 - Release Date: 11/2/2006


2006\11\02@184242 by Marco Genovesi

flavicon
face

> Something like this can work, but you'll have a real hard
> time to measure the cooling time...


I'm thinking if it isn't better  measuring the "warming time" instead of "cooling time"..

For example, using a thermistor and a PIC with a comparator.
First step, measure the fluid temperature (same method as reading a potentiometer).
Second step, measuring the time the thermistor took to warm to a definite threshold.
If necessary (fluid flow is very variable) the internal voltage reference may be adjusted on the fly for an adapted  threshold.


regards
Marco




---------- Initial Header -----------

>From      : piclist-bouncesspamBeGonespammit.edu
To          : RemoveMEpiclist@spam@spamspamBeGonemit.edu
Cc          :
Date      : Wed, 1 Nov 2006 11:04:36 -0300
Subject : Re: [EE]Flow Sensor possibilities







{Quote hidden}

> --

2006\11\02@212350 by Richard Prosser

picon face
Using this technique I'd be tempted to look at increasing the
temperature 2 dgrees and timeing the drop back to one degree
difference. That should sort out some of the uncertainty. Of coarse,
if you measure too frequently then you're going to keep increasing the
temperature which could lead to additional errors!

RP

On 02/11/06, Marco Genovesi <.....marco.genovesi@spam@spamEraseMElibero.it> wrote:
{Quote hidden}

2006\11\03@040617 by Alan B. Pearce

face picon face
>The basic meter was a high flow tube (that carried the bulk of the
>actual volume). In parallel with that was a much smaller tube, which
>carried anywhere from 100 to 1000 times less volume. It was a simple
>matter to heat the smaller tube at one end and then to detect the
>temperature difference about 1 inch downstream.

This is the principle used in the Helium Gas flow meter I mentioned earlier
in the thread. The manner of heating the measurement tube was a little
different though, in that the smaller tube was copper, and fed from the
secondary of a step down drive transformer, so the tube acted as a short
circuit (or close to it) on the transformer, thereby getting heated from the
current through it. The copper tube also formed one of the metals in the two
thermocouples IIRC. It always seemed a real strange arrangement, but it
works.

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...