Coming out of lurk mode… 
I’m looking at the best way of implementing sensors for a ringing
simulator for church bells. The bells are swung via a rope that runs
over a wheel (think cart wheel) attached to the bell itself. Current
sensor systems use serial interfaces, which of of course virtually no
PCs have any more and which therefore require an intermediate
usb-to-serial converter. The obvious thing is just to replace that
with a USB-capable AVR with the sensors connected directly to the AVR.
That bit is relatively straightforward.
The bit I’m looking for suggestions for is the sensors themselves.
There will be 8 bells, and there are 3 main sorts of sensors that are
used:
-
Reed switches. Generally not a good solution as they need to be
close to the magnet to work which is difficult with 100+ year old bell
wheels which are often warped.
-
Optical sensors. Generally OK, but may be subject to interference
from reflections/stray light.
http://www.abelsim.co.uk/doc/dbagley.htm
http://www.abelsim.co.uk/doc/tbellex1.htm
-
Tuned circuit, similar to metal detectors. Not clear if the
components can still be sourced, and seem a bit fiddly to make.
http://www.abelsim.co.uk/doc/tbellex2.htm
The other thing I’ve looked at is a TTL-compatible IR sensor that has
an amplifier and schmitt trigger all in one package:
http://uk.farnell.com/optek-technology/opl550a/photo-sensor-lateral-tm-btp/dp/1678676
That looks like it would be a mimimal-component solution and that
would be easy to debounce etc.
Anyone have any thoughts, or better suggestions?
Thanks,
Alan Burlison
Accelerometers ? Then you can get speed / amplitude as well as position
detection.
Or if you only want simple presence detection, industrial inductive
sensors. Less fussy than reed switches, no problems with dirt. Essentially
the tuned circuit but packaged and off-the-shelf.On Fri, Jun 3, 2016 at 11:14 PM, Alan Burlison alan.burlison@gmail.com wrote:
Coming out of lurk mode…
I’m looking at the best way of implementing sensors for a ringing
simulator for church bells. The bells are swung via a rope that runs
over a wheel (think cart wheel) attached to the bell itself. Current
sensor systems use serial interfaces, which of of course virtually no
PCs have any more and which therefore require an intermediate
usb-to-serial converter. The obvious thing is just to replace that
with a USB-capable AVR with the sensors connected directly to the AVR.
That bit is relatively straightforward.
The bit I’m looking for suggestions for is the sensors themselves.
There will be 8 bells, and there are 3 main sorts of sensors that are
used:
-
Reed switches. Generally not a good solution as they need to be
close to the magnet to work which is difficult with 100+ year old bell
wheels which are often warped.
-
Optical sensors. Generally OK, but may be subject to interference
from reflections/stray light.
http://www.abelsim.co.uk/doc/dbagley.htm
http://www.abelsim.co.uk/doc/tbellex1.htm
-
Tuned circuit, similar to metal detectors. Not clear if the
components can still be sourced, and seem a bit fiddly to make.
http://www.abelsim.co.uk/doc/tbellex2.htm
The other thing I’ve looked at is a TTL-compatible IR sensor that has
an amplifier and schmitt trigger all in one package:
http://uk.farnell.com/optek-technology/opl550a/photo-sensor-lateral-tm-btp/dp/1678676
That looks like it would be a mimimal-component solution and that
would be easy to debounce etc.
Anyone have any thoughts, or better suggestions?
Thanks,
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
Accelerometers ? Then you can get speed / amplitude as well as position
detection.
I did think of that initially, but there were two main issues: firstly,
getting power/data to the sensor, because it would have to be mounted on
the wheel, which of course rotates. Secondly, the protocol used by the
simulator software just needs a single trigger notification as the bell
passes the vertical position, so the extra data you’d get from an
accelerometer would have to be discarded anyway.
Or if you only want simple presence detection, industrial inductive
sensors. Less fussy than reed switches, no problems with dirt. Essentially
the tuned circuit but packaged and off-the-shelf.
Interesting idea. I looked on Farnell and ones that work at a reasonable
distance (2cm or more) are around £50 each, which unfortunately means
they are too expensive. I also looked at magnetorestrictive sensors, but
they too suffer from a sensing range that’s too small.
Alan Burlison
Could you superglue plastic reflectors to the sides of the bells or wheels
at the vertical point?
You could then use retroreflective sensors with infrared beams similar to
those used on automatic gates
You have probably been past them at work or a multistory car park before.
You have an infrared beam emmitor / sensor on one side of the road and
something similar to a single direction version of a bicycle reflector.
Modern turnstile systems also use them
Accelerometers ? Then you can get speed / amplitude as well as position
detection.
I did think of that initially, but there were two main issues: firstly,
getting power/data to the sensor, because it would have to be mounted on
the wheel, which of course rotates. Secondly, the protocol used by the
simulator software just needs a single trigger notification as the bell
passes the vertical position, so the extra data you’d get from an
accelerometer would have to be discarded anyway.
Or if you only want simple presence detection, industrial inductive
sensors. Less fussy than reed switches, no problems with dirt. Essentially
the tuned circuit but packaged and off-the-shelf.
Interesting idea. I looked on Farnell and ones that work at a reasonable
distance (2cm or more) are around £50 each, which unfortunately means they
are too expensive. I also looked at magnetorestrictive sensors, but they
too suffer from a sensing range that’s too small.
Alan Burlison
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
Good quality sensors that I’d use on something like a machine guard are
indeed expensive. But as with most such devices, much cheaper ones are
available from China, eg
http://www.ebay.co.uk/itm/LJ18A3-8-Z-AX-18mm-Thread-8mm-Detect-Inductive-Sensor-Proximity-Switch-DC-6-36V-/291537749606On Sat, Jun 4, 2016 at 7:58 AM, Alan Burlison alan.burlison@gmail.com wrote:
On 04/06/2016 00:32, Adrian Godwin wrote:
Accelerometers ? Then you can get speed / amplitude as well as position
detection.
I did think of that initially, but there were two main issues: firstly,
getting power/data to the sensor, because it would have to be mounted on
the wheel, which of course rotates. Secondly, the protocol used by the
simulator software just needs a single trigger notification as the bell
passes the vertical position, so the extra data you’d get from an
accelerometer would have to be discarded anyway.
Or if you only want simple presence detection, industrial inductive
sensors. Less fussy than reed switches, no problems with dirt. Essentially
the tuned circuit but packaged and off-the-shelf.
Interesting idea. I looked on Farnell and ones that work at a reasonable
distance (2cm or more) are around £50 each, which unfortunately means they
are too expensive. I also looked at magnetorestrictive sensors, but they
too suffer from a sensing range that’s too small.
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
You should check sensing distance on these. Retroreflective sensors
certainly have a much greater range but at the cost of more sensitivity to
cobwebs, dirt etc.On Sat, Jun 4, 2016 at 9:06 AM, Adrian Godwin artgodwin@gmail.com wrote:
Good quality sensors that I’d use on something like a machine guard are
indeed expensive. But as with most such devices, much cheaper ones are
available from China, eg
http://www.ebay.co.uk/itm/LJ18A3-8-Z-AX-18mm-Thread-8mm-Detect-Inductive-Sensor-Proximity-Switch-DC-6-36V-/291537749606
On Sat, Jun 4, 2016 at 7:58 AM, Alan Burlison alan.burlison@gmail.com wrote:
On 04/06/2016 00:32, Adrian Godwin wrote:
Accelerometers ? Then you can get speed / amplitude as well as position
detection.
I did think of that initially, but there were two main issues: firstly,
getting power/data to the sensor, because it would have to be mounted on
the wheel, which of course rotates. Secondly, the protocol used by the
simulator software just needs a single trigger notification as the bell
passes the vertical position, so the extra data you’d get from an
accelerometer would have to be discarded anyway.
Or if you only want simple presence detection, industrial inductive
sensors. Less fussy than reed switches, no problems with dirt.
Essentially
the tuned circuit but packaged and off-the-shelf.
Interesting idea. I looked on Farnell and ones that work at a reasonable
distance (2cm or more) are around £50 each, which unfortunately means they
are too expensive. I also looked at magnetorestrictive sensors, but they
too suffer from a sensing range that’s too small.
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
Could you superglue plastic reflectors to the sides of the bells or wheels
at the vertical point?
Yes, that’s the method that seems to be commonly used. Either that or
aluminium foil, or reflective tape, e.g.
You could then use retroreflective sensors with infrared beams similar to
those used on automatic gates
You have probably been past them at work or a multistory car park before.
You have an infrared beam emmitor / sensor on one side of the road and
something similar to a single direction version of a bicycle reflector.
Yes, I think that’s what is commonly used. Stray light might be a
problem but bell rooms are generally dark (and full of pigeon sh*t) and
if the sensor & LED were set well back inside a project box with holes
in it I think they’d probably work fine.
Alan Burlison
Good quality sensors that I’d use on something like a machine guard are
indeed expensive. But as with most such devices, much cheaper ones are
available from China, eg
http://www.ebay.co.uk/itm/LJ18A3-8-Z-AX-18mm-Thread-8mm-Detect-Inductive-Sensor-Proximity-Switch-DC-6-36V-/291537749606
Interesting, thanks. I wonder what the sensing range is? Also, whilst
many bell frames are made of wood (oak), the more modern ones (i.e.
post-1890s) are often steel, so mounting them might be tricky.
Alan Burlison
You should check sensing distance on these. Retroreflective sensors
certainly have a much greater range but at the cost of more sensitivity to
cobwebs, dirt etc.
Yeah, dirt/cobwebs are indeed an issue, bell rooms are not exactly clean
as they are open to the outside. As you have to go into the bell room to
silence the bells anyway, cleaning the sensors occasionally should be
acceptable.
Examples:
http://www.erringdenbenefice.org.uk/content/pages/uploaded_images/56.jpg
http://www.stgite.org.uk/library/bells02.jpg
http://www.kcacr.org.uk/lewidist/lpics/wilmington.jpg
http://www.pershoreabbey.org.uk/wpimages/wp0b051be3_05_06.jpg
Alan Burlison
I’ve been looking at the OPL550 and they come in 3 flavours, totem
pole, open collector, 10k pullup and each of those has active high and
active low variants. Any recommendations for which variant would be
best for directly connecting to an AVR? I’m thinking of using a
Minimus as I have some left over from the group buy we did ages ago.
The ringing room is 2 floors down from the belfry so I need to get the
two connected. I could either put the AVR in the ringing room and run
the 5V power/signals over multicore (I’d need 18 way) down from the
belfry to there or alternatively put the AVR in the belfry and then
use an active USB extender cable to connect it to the PC in the
ringing room. I’m leaning towards the active USB extender option as it
probably better shielded and cheaper to do it that way, but I’m
wondering if anyone has had any experience with them?
Alan Burlison
There’s been a bit of a trend of them being used in recent bots such as 3d
printers and cnc’s recently
Inductive sensors only detect metal
Capactive sensors detect metal and other types of material
Normally it’s a 3 core wire, 2 of them being a 12V supply, the 3rd being
the detect line which returns a similar voltage
(so you need to use something like an opto isolator to get it down to 5V /
3.3V)
They usually activate within a couple of mm of the surface of the sensor
and are ether on or off
They also come in Normally Open or Normally Closed configurationsOn Saturday, June 4, 2016 at 9:31:54 AM UTC+1, Alan Burlison wrote:
On 04/06/2016 09:06, Adrian Godwin wrote:
Good quality sensors that I’d use on something like a machine guard are
indeed expensive. But as with most such devices, much cheaper ones are
available from China, eg
http://www.ebay.co.uk/itm/LJ18A3-8-Z-AX-18mm-Thread-8mm-Detect-Inductive-Sensor-Proximity-Switch-DC-6-36V-/291537749606
Interesting, thanks. I wonder what the sensing range is? Also, whilst
many bell frames are made of wood (oak), the more modern ones (i.e.
post-1890s) are often steel, so mounting them might be tricky.
–
Alan Burlison
Normally it’s a 3 core wire, 2 of them being a 12V supply, the 3rd
being the detect line which returns a similar voltage (so you need to
use something like an opto isolator to get it down to 5V / 3.3V) They
usually activate within a couple of mm of the surface of the sensor
and are ether on or off They also come in Normally Open or Normally
Closed configurations.
Yeah, I don’t think that will work as it needs to be further away than a
couple of mm. I’ve been in touch with someone who has built some using
hall effect sensors and rare earth magnets - I’ve had a look on farnell
and something like this might do the trick:
http://uk.farnell.com/hamlin/55100-3h-02-a/sensor-hall-effect-flange-24vdc/dp/5052567
They operate down to 3.8V and have a sense distance of 18mm, which I’m
guessing can be increased if you use a couple of magnets rather than
just one.
Alan Burlison
I think the hall effect sensor you chose would work. It has quite good
sensitivity at 59 Gauss from a distance of 18 mm.
These magnets would probably work for the sensing part:
http://e-magnetsuk.com/magnet_products/neodymium_magnets/ring_magnets.aspx
100 Gauss = 0.01 Tesla - so any magnet which had a magnetic field strength
greater than this would probably work. I would get stronger to be on the
safe side. A magnet with a decent pulling force will probably work.
Be aware you may need to buffer the switch output if you are going to place
the sensors a long way from the microcontroller.
Best of luck
AlexOn 5 June 2016 at 18:51, Alan Burlison alan.burlison@gmail.com wrote:
On 05/06/2016 18:02, ‘garlicbread (Richard)’ via Hackspace Manchester wrote:
Normally it’s a 3 core wire, 2 of them being a 12V supply, the 3rd
being the detect line which returns a similar voltage (so you need to
use something like an opto isolator to get it down to 5V / 3.3V) They
usually activate within a couple of mm of the surface of the sensor
and are ether on or off They also come in Normally Open or Normally
Closed configurations.
Yeah, I don’t think that will work as it needs to be further away than a
couple of mm. I’ve been in touch with someone who has built some using hall
effect sensors and rare earth magnets - I’ve had a look on farnell and
something like this might do the trick:
http://uk.farnell.com/hamlin/55100-3h-02-a/sensor-hall-effect-flange-24vdc/dp/5052567
They operate down to 3.8V and have a sense distance of 18mm, which I’m
guessing can be increased if you use a couple of magnets rather than just
one.
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
I think the hall effect sensor you chose would work. It has quite good
sensitivity at 59 Gauss from a distance of 18 mm.
Thanks Alex, I had no real idea what was good or bad in terms of
sensitivity.
These magnets would probably work for the sensing part:
http://e-magnetsuk.com/magnet_products/neodymium_magnets/ring_magnets.aspx
100 Gauss = 0.01 Tesla - so any magnet which had a magnetic field strength
greater than this would probably work. I would get stronger to be on the
safe side. A magnet with a decent pulling force will probably work.
I got some details from someone who had made a homebrew sensor, they
used these:
http://uk.farnell.com/comus-assemtech/m1219-5/magnet/dp/7230382, 3500
gauss at the surface but I’m not sure how that compares with the ones
you linked to, which seem to use kgs of pulling power.
Be aware you may need to buffer the switch output if you are going to place
the sensors a long way from the microcontroller.
I was thinking of putting the AVR somewhere in the middle of the belfry
and running a USB extender from there to the PC 2 floors down. That
would mean that the transducer-AVR distances would be no more than 3m at
most. It should be a benign environment from an electrical noise
perspective - unless they get a lightning strike of course…
Alan Burlison
Other option for communication is Bluetooth LE or Zigbee - arduino
daughterboard models for both are available from ebay relatively cheaply
and you can do serial comms over them quite easily.
Not sure what the latency will be like though
(Either that or you have one minuture bluetooth LE / accelerometer / avr /
battery package per bell (mounted on the bell) and change the batteries
every few weeks / hope they dont change the sound the bell makes)On 5 June 2016 at 23:12, Alan Burlison alan.burlison@gmail.com wrote:
On 05/06/2016 19:51, Alexander Lang wrote:
I think the hall effect sensor you chose would work. It has quite good
sensitivity at 59 Gauss from a distance of 18 mm.
Thanks Alex, I had no real idea what was good or bad in terms of
sensitivity.
These magnets would probably work for the sensing part:
http://e-magnetsuk.com/magnet_products/neodymium_magnets/ring_magnets.aspx
100 Gauss = 0.01 Tesla - so any magnet which had a magnetic field strength
greater than this would probably work. I would get stronger to be on the
safe side. A magnet with a decent pulling force will probably work.
I got some details from someone who had made a homebrew sensor, they used
these: http://uk.farnell.com/comus-assemtech/m1219-5/magnet/dp/7230382,
3500 gauss at the surface but I’m not sure how that compares with the ones
you linked to, which seem to use kgs of pulling power.
Be aware you may need to buffer the switch output if you are going to place
the sensors a long way from the microcontroller.
I was thinking of putting the AVR somewhere in the middle of the belfry
and running a USB extender from there to the PC 2 floors down. That would
mean that the transducer-AVR distances would be no more than 3m at most. It
should be a benign environment from an electrical noise perspective -
unless they get a lightning strike of course…
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
The BBC microbit boards are about £10 and have accelerometer, bluetooth LE
and a reasonable processor. I would think you could get power requirements
pretty low if you had it go to sleep when the bell isn’t ringing - but I
haven’t proved this.On Mon, Jun 6, 2016 at 1:18 PM, Stu Walker stuartw1@gmail.com wrote:
Other option for communication is Bluetooth LE or Zigbee - arduino
daughterboard models for both are available from ebay relatively cheaply
and you can do serial comms over them quite easily.
Not sure what the latency will be like though
(Either that or you have one minuture bluetooth LE / accelerometer / avr
/ battery package per bell (mounted on the bell) and change the batteries
every few weeks / hope they dont change the sound the bell makes)
On 5 June 2016 at 23:12, Alan Burlison alan.burlison@gmail.com wrote:
On 05/06/2016 19:51, Alexander Lang wrote:
I think the hall effect sensor you chose would work. It has quite good
sensitivity at 59 Gauss from a distance of 18 mm.
Thanks Alex, I had no real idea what was good or bad in terms of
sensitivity.
These magnets would probably work for the sensing part:
http://e-magnetsuk.com/magnet_products/neodymium_magnets/ring_magnets.aspx
100 Gauss = 0.01 Tesla - so any magnet which had a magnetic field
strength
greater than this would probably work. I would get stronger to be on the
safe side. A magnet with a decent pulling force will probably work.
I got some details from someone who had made a homebrew sensor, they used
these: http://uk.farnell.com/comus-assemtech/m1219-5/magnet/dp/7230382,
3500 gauss at the surface but I’m not sure how that compares with the ones
you linked to, which seem to use kgs of pulling power.
Be aware you may need to buffer the switch output if you are going to
place
the sensors a long way from the microcontroller.
I was thinking of putting the AVR somewhere in the middle of the belfry
and running a USB extender from there to the PC 2 floors down. That would
mean that the transducer-AVR distances would be no more than 3m at most. It
should be a benign environment from an electrical noise perspective -
unless they get a lightning strike of course…
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
The BBC microbit boards are about £10 and have accelerometer, bluetooth LE
and a reasonable processor. I would think you could get power requirements
pretty low if you had it go to sleep when the bell isn’t ringing - but I
haven’t proved this.
Other option for communication is Bluetooth LE or Zigbee - arduino
daughterboard models for both are available from ebay relatively cheaply and
you can do serial comms over them quite easily.
Not sure what the latency will be like though
(Either that or you have one minuture bluetooth LE / accelerometer / avr
/ battery package per bell (mounted on the bell) and change the batteries
every few weeks / hope they dont change the sound the bell makes)
I need to keep this as simple as possible - bell rooms access is
generally tricky, they are full of heavy lumps of rapidly rotating
metal weighing up to a tonne and are also pretty hostile to
electronics as they are open to the elements - full of blown stone
dust, cobwebs, pigeon sh*t, leaks/condensation and blown rain. Plus I
don’t need to know acceleration, just a single pulse when the bell
passes the vertical position.
I managed to find someone who has done these sorts of installations in
the past (around 30 so far, he tells me) and he uses these
magnetoresistive sensors from Farnell:
http://uk.farnell.com/honeywell/sm451r/magnetoresistive-sensor-7g-to/dp/2507582
They only need around 7 gauss to trigger, work at 5V and only need
~100mA for an 8 bell setup, so can be USB-powered. He also said he
just runs multicore from the belfry to the ringing room and puts the
sensitive electronics in there. I can improve on what he uses for the
sensor/PC interface by using an AVR, but the sensor config he uses
seems difficult to improve on.
Thanks,
Alan Burlison
I’d worry slightly about having a long run of cable on the AVR pins, but a
long run of USB isn’t good either (even with active repeaters). I’d also
want to avoid supplying 5V over a long cable, but since those work off
3-24V they should be OK, and if your contact has used them succesfully
they’re presumably happy driving the long cable.
My preferred system would have the AVR in the bell chamber with RS422 to
the PC. But I accept you can get away with less, and the only thing I’d add
to what you’re proposing is to protect the AVR inputs with 1K resistors and
4v7 zener diodes (or 1n914 to ground and 5v). Ideally, use one of these :
http://www.rugged-circuits.com/ruggeduino/
and some sort of daughter board so you can connect the bell sensors to
terminal block instead of solder pins. Fuse the supply going to the
sensors and check that it’s still a healthy >3V at the sensors : if it’s
not, send 12V and fit an attenuation / clipping network at the AVR (the 1K
- zener will do it, as long as the 1K is between zener and sensor).On Mon, Jun 6, 2016 at 3:49 PM, Alan Burlison alan.burlison@gmail.com wrote:
On 6 June 2016 at 15:27, Adrian Godwin artgodwin@gmail.com wrote:
The BBC microbit boards are about £10 and have accelerometer, bluetooth
LE
and a reasonable processor. I would think you could get power
requirements
pretty low if you had it go to sleep when the bell isn’t ringing - but I
haven’t proved this.
On Mon, Jun 6, 2016 at 1:18 PM, Stu Walker stuartw1@gmail.com wrote:
Other option for communication is Bluetooth LE or Zigbee - arduino
daughterboard models for both are available from ebay relatively
cheaply and
you can do serial comms over them quite easily.
Not sure what the latency will be like though
(Either that or you have one minuture bluetooth LE / accelerometer /
avr
/ battery package per bell (mounted on the bell) and change the
batteries
every few weeks / hope they dont change the sound the bell makes)
I need to keep this as simple as possible - bell rooms access is
generally tricky, they are full of heavy lumps of rapidly rotating
metal weighing up to a tonne and are also pretty hostile to
electronics as they are open to the elements - full of blown stone
dust, cobwebs, pigeon sh*t, leaks/condensation and blown rain. Plus I
don’t need to know acceleration, just a single pulse when the bell
passes the vertical position.
I managed to find someone who has done these sorts of installations in
the past (around 30 so far, he tells me) and he uses these
magnetoresistive sensors from Farnell:
http://uk.farnell.com/honeywell/sm451r/magnetoresistive-sensor-7g-to/dp/2507582
They only need around 7 gauss to trigger, work at 5V and only need
~100mA for an 8 bell setup, so can be USB-powered. He also said he
just runs multicore from the belfry to the ringing room and puts the
sensitive electronics in there. I can improve on what he uses for the
sensor/PC interface by using an AVR, but the sensor config he uses
seems difficult to improve on.
Thanks,
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
I’d worry slightly about having a long run of cable on the AVR pins, but a
long run of USB isn’t good either (even with active repeaters). I’d also
want to avoid supplying 5V over a long cable, but since those work off 3-24V
they should be OK, and if your contact has used them succesfully they’re
presumably happy driving the long cable.
The cable run should be around 40ft, he told me he hasn’t even found
shielded cable is necessary. The circuit he uses has a 74HC14 (schmitt
trigger) and a 74HC74 (edge triggered flip-flop) between the sensors
and the Serial/USB converter, I’m not sure what the 74HC74 is for. The
USB-Serial converter he uses is based on the Prolific 2303 so I don’t
think it is anything special.
My preferred system would have the AVR in the bell chamber with RS422 to the
PC. But I accept you can get away with less, and the only thing I’d add to
what you’re proposing is to protect the AVR inputs with 1K resistors and 4v7
zener diodes (or 1n914 to ground and 5v). Ideally, use one of these :
There’s a heating blower in the void between the belfry and the bell
room but no other electrical stuff so I hope electrical noise won’t be
a problem. The zener/resistor sounds like a good idea though, if just
to protect from shorts & reversed polarity problems during the initial
installation.
http://www.rugged-circuits.com/ruggeduino/
and some sort of daughter board so you can connect the bell sensors to
terminal block instead of solder pins. Fuse the supply going to the sensors
and check that it’s still a healthy >3V at the sensors : if it’s not, send
12V and fit an attenuation / clipping network at the AVR (the 1K + zener
will do it, as long as the 1K is between zener and sensor).
Yes, I was going to use a daughter board with the Minimus. The systems
he’s built draw ~100mA so are just powered directly from the USB port
they are plugged in to. There’s no power available in the belfry
itself so being able to power it from the ringing room is a big plus.
It should be easy enough to measure the voltage drop along the cable
before installation and adapt the design if necessary. In fact getting
the whole thing set up before going anywhere near a tower sounds
sensible anyway.
When the system is not in use I expect it to be unplugged. The most
likely electrical mishap in a bell tower is probably a lightnigh
strike, in which case no amount of zener diodes is going to help
anyway 
Thanks for the hints and tips,
Alan Burlison
Yes, I don’t think there’s much you can do about a lighting strike without
going over the top electrically. And the tower might fall down anyway
That blower is probably OK when running if it’s an induction motor. If it’s
a universal motor (brushes and commutator) it might make some noise but the
worst is when it’s turned on or off - if the cables are near each other you
might get a spike induced that’s well outside the AVR’s specs. The zener
will probably stop this from destroying it. You might also get a strong
50Hz signal induced in it - this might give you false triggering. A load at
the AVR (1k to earth as well as 1k in series) might fix that. If you use
open collector drivers the pullup will probably help too.On Mon, Jun 6, 2016 at 5:09 PM, Alan Burlison alan.burlison@gmail.com wrote:
I’d worry slightly about having a long run of cable on the AVR pins, but
a
long run of USB isn’t good either (even with active repeaters). I’d also
want to avoid supplying 5V over a long cable, but since those work off
3-24V
they should be OK, and if your contact has used them succesfully they’re
presumably happy driving the long cable.
The cable run should be around 40ft, he told me he hasn’t even found
shielded cable is necessary. The circuit he uses has a 74HC14 (schmitt
trigger) and a 74HC74 (edge triggered flip-flop) between the sensors
and the Serial/USB converter, I’m not sure what the 74HC74 is for. The
USB-Serial converter he uses is based on the Prolific 2303 so I don’t
think it is anything special.
My preferred system would have the AVR in the bell chamber with RS422 to
the
PC. But I accept you can get away with less, and the only thing I’d add
to
what you’re proposing is to protect the AVR inputs with 1K resistors and
4v7
zener diodes (or 1n914 to ground and 5v). Ideally, use one of these :
There’s a heating blower in the void between the belfry and the bell
room but no other electrical stuff so I hope electrical noise won’t be
a problem. The zener/resistor sounds like a good idea though, if just
to protect from shorts & reversed polarity problems during the initial
installation.
http://www.rugged-circuits.com/ruggeduino/
and some sort of daughter board so you can connect the bell sensors to
terminal block instead of solder pins. Fuse the supply going to the
sensors
and check that it’s still a healthy >3V at the sensors : if it’s not,
send
12V and fit an attenuation / clipping network at the AVR (the 1K + zener
will do it, as long as the 1K is between zener and sensor).
Yes, I was going to use a daughter board with the Minimus. The systems
he’s built draw ~100mA so are just powered directly from the USB port
they are plugged in to. There’s no power available in the belfry
itself so being able to power it from the ringing room is a big plus.
It should be easy enough to measure the voltage drop along the cable
before installation and adapt the design if necessary. In fact getting
the whole thing set up before going anywhere near a tower sounds
sensible anyway.
When the system is not in use I expect it to be unplugged. The most
likely electrical mishap in a bell tower is probably a lightnigh
strike, in which case no amount of zener diodes is going to help
anyway
Thanks for the hints and tips,
–
Alan Burlison
–
You received this message because you are subscribed to the Google Groups
"Hackspace Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to hacman+unsubscribe@googlegroups.com.
To post to this group, send an email to hacman@googlegroups.com.
Visit this group at https://groups.google.com/group/hacman.
For more options, visit https://groups.google.com/d/optout.
