Catweasel, if you avoid a LiPo battery because of the cost I would consider using a Propeller. I built a BS2 project which cost a couple hundred dollars. It had a serial lcd, a unipolar stepper and a keypad.
A Propeller is also I use. The Stamp will be used to control some servos and read some sensors, while the Propeller is the "main" processor,
The The Propeller would just send an instruction to the Stamp like "Take Sample from A". And the Stamp would do all the PCM to move the servos and take the sample, store the value and send a "done" back to the Propeller. The Propeller is then spared from all the PCM and can also use its I/O pins for other things.
Depending on how the design evolves, I might even use a couple of Stamps that are controlled by the Propeller, possibly degrading the Stamp to a mere controller IC.
And making two processors directly "talk" to each other; from I/O pin to I/O pin so to speak; is part of what I want to explore.
The cost came from the extra PCBs needed for the hardware.
Are you referring to the costs when you order a custom made PCB from a shop, or simply the raw modules?
I will make my own PCBs. I like to combine the work with electronics and some pure crafting and building in my activities.
Slightly OT, just to give you an idea about the process:
1. Take an appropriately sized sheet of Peek. Available, amongst other places, at: http://www.sdplastics.com/peek.html
2. Shop for some good quality copper foil. For example here: http://allfoils.thomasnet.com/item/copper/copper-sheet/pn-1005?
3. Bond the coppersheet to the Peek with epoxy
4. Use your favorite method for mask transfer (toner / lithography)
5. Edge away!
6. Drill holes where needed or use as is for SMD soldering.
I made some experiments with some simple patterns and so far the results are promissing.
It still needs fiddeling and finetuning. For example: I edge the PCB inside an ultrasonic bath (these for cleaning silver cuttlery in).
Too little power, insufficient effect, too much power - the copper tends to come loose.
Maybe its cumbersome, but I like to do as much as possible myself.
A 9V battery would probably not receive any charge from a 12V solar cell using the '317. Better to charge the batteries directly from the solar cell. A small solar cell will not put out enough current to damage the batteries.
Hmmm. Good point. Didn't think it would be so tricky to build a charge/supply cuircuit.
Over the years I've designed many solar charger circuits. There are many complexities - eg maximum power point tracking, under and overvoltage protection, lowering the charge rate if the battery has gone undervoltage [...]
Still, I am not giving up. Maybe the 317 is not such a good choice afterall?
How would you do it?
I wonder how such solar/battery load driving systems work in satellites??
The The Propeller would just send an instruction to the Stamp like "Take Sample from A". And the Stamp would do all the PCM to move the servos and take the sample, store the value and send a "done" back to the Propeller. The Propeller is then spared from all the PCM and can also use its I/O pins for other things.
The Propeller can drive servos in its sleep (well almost). You don't have to make sure and refresh the servos every 20ms since the Prop does it for you.
If you're running out of pins and you have extra Stamps, then I suppose it could be useful. I personally think there are many better options (and much less expensive) for increasing the IO capability of a Prop than with a Stamp (including using a second Prop).
The Propeller can drive servos in its sleep (well almost). You don't have to make sure and refresh the servos every 20ms since the Prop does it for you.
If you're running out of pins and you have extra Stamps, then I suppose it could be useful. I personally think there are many better options (and much less expensive) for increasing the IO capability of a Prop than with a Stamp (including using a second Prop).
The I/O pins are indeed a bottleneck. At the moment the system consists of:
8 "Sampling units" (SU)
A SU consists of:
2 servo motors
2 photosensors
2 LEDs
The LEDs and the photosensors are used to verify the servo motions (so mechanical failure is noticed)
That makes 6 I/O lines per SU, or 48 for all SUs
Then it has 5 "Enviromental Control Units" (ECU)
An ECU consists of:
1 temperature sensor
1 humidity/water sensor
1 heating element (transistor controlled)
That requires another 15 input lines.
Last, but not least, the "Recovery Unit" (RU)
The RU consists of a GPS and a GSM module.
The design of this one is till in its concept phase. So I don't know how many I/O lines would be needed.
The idea is that if the probe descends at the end of the mission (verified by the altimeter), the GPS position shall be transmitted in a SMS via the GSM module.
It is assumed the proble lands within GSM network coverage.
Then, there is the "main unit" which is connected to
- a gyroscope
- an altimeter
- accelerometer
My plan was/is to use one Stamp per SU, one Stamp for the 5 ECU units and a Propeller for the main unit.
The SU-Stamps would then, if commanded, move and monitor the servos.
The ECU-Stamp could easily handle the enviromental control job, e.g. turn on the heating when needed, signal a water hazard, etc.
The propeller would then run the main mission program, including the sampling schedule, deciding not to take a sample if the vehicle is in an undesired orientation, etc.
The main "flight program" would be done by the Propeller. So I wanted to make the SUs and the other subcomponents as autonomous as possible.
Sure, with clever programming and some I/O expanders, everything could be one with a single Propeller. But if it fails, the probe is useless.
If, for example, a Stamp fails, only one SU is out of the game.
It also allows for more flexibility. If testing reveals more ECUs or SUs are needed, they can be added blockwise.
Although it is just a simple baloon probe, I try to build it as it it would have to survice on the moon :-)
Hmmm. Good point. Didn't think it would be so tricky to build a charge/supply cuircuit.
Then again, I have been warned.
Still, I am not giving up. Maybe the 317 is not such a good choice afterall?
How would you do it?
I wonder how such solar/battery load driving systems work in satellites??
Cheers,
Catweasel
In satellites the solar cells are always facing the sun and there are no clouds or atmosphere to block sunlight. This means they can be connected to provide the optimum voltage for the system to use and they will provide a steady voltage and current except when in earth's shadow. Battery charging and regulating power to the electronics is done by switching regulators for maximum efficiency.
In your case it would depend on how much power the solar cell can produce and how much it varies. If expense was not a consideration I would probably connect the solar cells for 24V output, use a switching charge regulator for charging a 12V battery, a switching regulator to provide +5V from the battery, a linear regulator to go from 5V to +3.3V if the current draw is low, or a switching regulator from the battery if the 3.3V power requirement is high.
BTW you could use an LDO (low drop out) regulator for the battery charging circuit. That would increase the voltage available to the battery by about 1.4 - 1.6V. If the solar cells provide no more than 200mA in bright sunshine you could connect it to the battery through a transistor controlled by the prop and control the charging through software.
In If expense was not a consideration I would probably connect the solar cells for 24V output, use a switching charge regulator for charging a 12V battery, a switching regulator to provide +5V from the battery, a linear regulator to go from 5V to +3.3V if the current draw is low, or a switching regulator from the battery if the 3.3V power requirement is high.
If you're going to be using GPS or other radio devices, be aware switching regulators can be very noisy.
When I used a switching regulator with a GPS logger I was making, the GPS module could no longer lock on the the satellites.
I've also had radio interference when using a switching regulator with a 2.4GHz RC transmitter.
Apparently there are ways to sheild the regulators, but I haven't tried this (successully) yet.
Duane has a good point about switching noise. If you don't have the space or weight allowance to provide adequate shielding they could cause problems with noise sensitive equipment like GPS receivers. In that case it may be better to trade off higher power use for lower noise by sticking to linear regulators and careful selection of components.
In satellites the solar cells are always facing the sun and there are no clouds or atmosphere to block sunlight. This means they can be connected to provide the optimum voltage for the system to use and they will provide a steady voltage and current except when in earth's shadow. Battery charging and regulating power to the electronics is done by switching regulators for maximum efficiency.
Was't that how NASA "discovered" th memory effect in rechargable batteries. They magically ajusted their capacity to the day/night charge/drain cycle.?
In your case it would depend on how much power the solar cell can produce and how much it varies. If expense was not a consideration I would probably connect the solar cells for 24V output, use a switching charge regulator for charging a 12V battery, a switching regulator to provide +5V from the battery, a linear regulator to go from 5V to +3.3V if the current draw is low, or a switching regulator from the battery if the 3.3V power requirement is high.
That would be four cells in parallel. A 12V battery should be posible. (Still cheaper than LiPo). A single C-Block should suffice. The 3.3V is only used to drive the Propeller and its main sensors. The Stamps are powered by the 5V rail. Also the Servos and other "high-load" components are powered directly by the source (not via the Stmp I/O pins).
BTW you could use an LDO (low drop out) regulator for the battery charging circuit. That would increase the voltage available to the battery by about 1.4 - 1.6V. If the solar cells provide no more than 200mA in bright sunshine you could connect it to the battery through a transistor controlled by the prop and control the charging through software.
That is going into the direction od an "active power control". I thought about that.
Hooking a Stamp with through an RC circuit to the battery and solar cells, scanning for over/under voltage and activel switching the charger/battery on/off.
I just thought this could be realized by rather "dumb" feedback circuits. The same way cou can make a blinking LED without any microcontroler. Just with a battery, capacitor, resistor an LED and one transistor. Instead of programming a high-low sequence, the value of C determines the frquency, the vakue of R controls the brightness.
What type of LDO regulator would you recommend?
Duane has a good point about switching noise. If you don't have the space or weight allowance to provide adequate shielding they could cause problems with noise sensitive equipment like GPS receivers. In that case it may be better to trade off higher power use for lower noise by sticking to linear regulators and careful selection of components.
Hmmm. Never thought about noise.
Would it be sufficient if the GPS module had ome spacing between itself and the rest of the system (for example having it tethered some meters below the probe)?
Hi,
I don't believe the noise being referred to is acoustical but rather rf and power. Just moving the gps away from the switching regulator will not solve the power noise issue.
Jim
Hi,
I don't believe the noise being referred to is acoustical but rather rf and power. Just moving the gps away from the switching regulator will not solve the power noise issue.
Jim
So not decoupled HF radiation but ripples in the DC lines?
I thought he meant that the regulators would work like HF senders....
Why would distance not lessen the effect of these?
In my experience moving the GPS unit away from the switching regulator greatly reduces the interference.
A change in GPS/regulator distance of a few inches can be the difference between the GPS unit not getting a lock and it getting a lock. Of course a few feet of separation would reduce the interference much more than a separation of a few inches.
In my experience moving the GPS unit away from the switching regulator greatly reduces the interference.
A change in GPS/regulator distance of a few inches can be the difference between the GPS unit not getting a lock and it getting a lock. Of course a few feet of separation would reduce the interference much more than a separation of a few inches.
Thats what I thought. So a 5 meter tether should be plenty.
Thats what I thought. So a 5 meter tether should be plenty.
A 5 meter tether might introduce new problems. I don't think you'd need that much distance.
I used this switching regulator with my GPS logger hoping to have the batteries' charge last longer. I was using a plastic Sucrets box as a case so everything was very close together. I had previously logged some data as I walked using a linear regulator. The log file after using the switching regulator was full of error messages stating a lock hadn't been made with the satellites.
After mentioning the problem of interference from a switching regulator, Mike Green suggested there are ways of shielding the offending regulator. I think he mentioned using some sort of metal cage and the correct filtering capacitors (this isn't a direct quote).
I tried wrapping the regulator with 5mm copper tape. I don't know if the copper tape helped or not, but the regulator still interfered with the GPS unit.
Many GPS units will give the signal strength from each satellite. It should be possible to quantify the interference of a regulator and monitor and change based on shielding attempts. I watched the signal strength numbers as I was using the LCD and back light. When the LCD was next to the GPS unit, all signals were lost. As I moved the LCD away from the GPS unit the signal strength would increase. Just a few inches made the difference between no signal and very poor signal strength. I think 12 inches of distance between LCD and GPS was enough to return the signals to almost full strength. I don't recall the exact distances and strength readings from these tests. It should be very difficult to repeat these tests with whatever regulator you wish to use.
I've recently purchased a roll of 2" copper tape. I'm hoping the wider tape will make a better Faraday cage than the thinner 5mm tape. I haven't tried making a new cage yet.
I also found the back light voltage booster on Rayman's 4.3" LCD breakout board also causes interference problems with my GPS unit.
I haven't tested the back light voltage booster with a RC transmitter, but I have tested the small DE regulator with a 2.5GHz Spektrum transmitter and I found the regulator severely interfered with the RC transmission. It's ironic that DE links to a tutorial of how to replaces a linear regulator with their switching regulator inside a Spektrum transmitter.
I've also been warned against using a switching regular in cell phone projects.
I used this switching regulator with my GPS logger hoping to have the batteries' charge last longer. I was using a plastic Sucrets box as a case so everything was very close together. I had previously logged some data as I walked using a linear regulator. The log file after using the switching regulator was full of error messages stating a lock hadn't been made with the satellites.
After mentioning the problem of interference from a switching regulator, Mike Green suggested there are ways of shielding the offending regulator. I think he mentioned using some sort of metal cage and the correct filtering capacitors (this isn't a direct quote).
I tried wrapping the regulator with 5mm copper tape. I don't know if the copper tape helped or not, but the regulator still interfered with the GPS unit.
I used to work for TomTom. When we opened defective devices (faulty battery, etc), I could never see any special RF protection and the internal batteries, which run at 3,7V. So I guess the trick is not to use any regulators and tailor the power source to the desired level. Of course those devices only need one voltage level.
Many GPS units will give the signal strength from each satellite. It should be possible to quantify the interference of a regulator and monitor and change based on shielding attempts. I watched the signal strength numbers as I was using the LCD and back light. When the LCD was next to the GPS unit, all signals were lost. As I moved the LCD away from the GPS unit the signal strength would increase. Just a few inches made the difference between no signal and very poor signal strength. I think 12 inches of distance between LCD and GPS was enough to return the signals to almost full strength. I don't recall the exact distances and strength reads from these tests. It should be very difficult to repeat these tests with whatever regulator you wish to use.
I've recently purchased a roll of 2" copper tape. I'm hoping the wider tape will make a better Faraday cage than the thinner 5mm tape. I haven't tried making a new cage yet.
I also found the back light voltage booster on Rayman's 4.3" LCD breakout board also causes interference problems with my GPS unit.
I haven't tested the back light voltage booster with a RC transmitter, but I have tested the small DE regulator with a 2.5GHz Spektrum transmitter and I found the regulator severely interfered with the RC transmission. It's ironic that DE links to a tutorial of how to replaces a linear regulator with their switching regulator inside a Spektrum transmitter.
Would the regulators from the Propeller eduction board (LM2940 ad LM2937) be problematic?
I've also been warned against using a switching regular in cell phone projects.
It is reasonable to assume, if HF interference is a problem, that this also affects GSM connections.
Depending on the GSM module I'm gonne be using (haven't decided on one yet), I could hook it up with a very small battery. If it can send the SMS more or less autinomous, I could power the probe down before sending. The GSM module could only be used after landing anyway (No GSM coverage at 8000 meters). So periodical location updates are not possible anyway.
I also don't know how much static electricity will have an effect on the system. (Iimagine the probe flying through a highly charged cloud).
At the moment I consider for the structural parts polystyrol that is reinforced with fiberglass gaze and epoxy. The shape is cylindical.(Diameter 60cm; height 125cm). The maximum weight I can afford is 10Kg, or I have to go with a different baloon (which I on't really want to do).
If I wrap the polystyrol in aluminium foil first, would that help in terms of shielding?
After a little more consideration, I get more and more convinced that I should not cut costs on the battery. But if I am going to spend a bit more money on a battery, it should be the optimal battery for the job. If you had to shop for a battery here (I get some discount there), whichone would you pick? http://www.conrad.com/Rechargeable-Batteries.htm?websale7=conrad-int&ci=SHOP_AREA_14717
From a quality perspective, I always had good expieriences with the "Emmerich" batteries....
I made some tests with the servos and they can handle the torque, so I will definitely not need a 12V rail. Just a strong 5V for the stamps and the servos and the 3.3V for the Propeller.
Not knowing the capacity of your battery pack or the output of your solar cells I would not know which one to recommend. There are lots of choices available. You need to get one that can handle the current, power dissipation, and a low enough drop out voltage to charge the batteries.
I just thought this could be realized by rather "dumb" feedback circuits. The same way you can make a blinking LED without any microcontroler.
This can be done with analog circuits but it would simpler to have one of the micros handle it. I would suggest doing two tests to determine how simple the charging circuit can be.
The first test would be to plot a current/voltage curve for the solar cells in full sunshine.
The second test would be to see if there is any current flow from the battery through the solar cell when it is in the dark.
The other test to do is to determine the current draw of the electronics and the servos/motors.
Good that you can drop down to a lower voltage. That makes everything simpler.
4 NiMh cells will provide 5.2V and the solar cells can be connected to provide 6V for charging.
The servos can connect directly to the batteries.
An LDO regulator for the 3.3V and 5V (although you may need to drop the 5V to 4.8V).
Not knowing the capacity of your battery pack or the output of your solar cells I would not know which one to recommend. There are lots of choices available. You need to get one that can handle the current, power dissipation, and a low enough drop out voltage to charge the batteries.
Hmm. Sounds I should make some power consumption tests first.
This can be done with analog circuits but it would simpler to have one of the micros handle it. I would suggest doing two tests to determine how simple the charging circuit can be.
After reading into the "maximum power point tracking" concept a bit further, I think a pure analog solution it a bit over my head. Especially since I don't have that much expierience in analog circuits, its propably not the best one to start with. I will draw up a schematic for a Stamp controled power distributions. Lets see how that turns out :-)
The first test would be to plot a current/voltage curve for the solar cells in full sunshine.
I planned to set up a performance test for the solar cells next week. I will test sunlight at various angles as well es occluding the cells with colored foils to simulate dimmed lighting conditions.
The second test would be to see if there is any current flow from the battery through the solar cell when it is in the dark.
A tough one to test as I haven't decidd on a battery yet.
The choice of battery is dependent of the design of the power supply which is depending on the choice of battery.
So many options...
The other test to do is to determine the current draw of the electronics and the servos/motors.
That will depend mainly on the mission profile. If I take samples on a lower frequency, the servos would have to act less often.
The sampling sequence uses several servos, but only one at a time. (One servo moves the sample container in front of the sample port. Then another servo opens the port hatch.
My approach was to see what power allowance I have and then design the mission accordingly.
But I could indeed make some power consumption profiles for each sampling process.
4 NiMh cells will provide 5.2V and the solar cells can be connected to provide 6V for charging.
The servos can connect directly to the batteries.
An LDO regulator for the 3.3V and 5V (although you may need to drop the 5V to 4.8V).
The LM2940 and LM2937 should do the job.
The attached schematic does not have any voltage regulation yet. I just tried to set up RC loops for measuring the solar and battery output.
A on/off switch for the powered circuits. For example P14 controls the RC charge period, then P15 reads the discharge time.
Would that work?
I thought you were going to go with 5-6V for the servos. You now have the solar cells connected in series which will give you 24V at low current. For 6V operation they could be connected in parallel to provide higher current.
You are putting the battery/solar cell voltage on the prop pin P15 through a 220 ohm resistor. This will fry the pin and possibly the whole chip.
None of the transistors will do what you expect them to do. They are all NPN transistors. Some are connected backwards and the others can never be turned off as shown.
I don't have time right now but I will take a look later tonight and make some suggestions unless someone else pipes up before then.
In the mean time don't connect it up as shown or you will most likely let out some smoke.
Comments
A Propeller is also I use. The Stamp will be used to control some servos and read some sensors, while the Propeller is the "main" processor,
The The Propeller would just send an instruction to the Stamp like "Take Sample from A". And the Stamp would do all the PCM to move the servos and take the sample, store the value and send a "done" back to the Propeller. The Propeller is then spared from all the PCM and can also use its I/O pins for other things.
Depending on how the design evolves, I might even use a couple of Stamps that are controlled by the Propeller, possibly degrading the Stamp to a mere controller IC.
And making two processors directly "talk" to each other; from I/O pin to I/O pin so to speak; is part of what I want to explore.
Are you referring to the costs when you order a custom made PCB from a shop, or simply the raw modules?
I will make my own PCBs. I like to combine the work with electronics and some pure crafting and building in my activities.
Slightly OT, just to give you an idea about the process:
1. Take an appropriately sized sheet of Peek. Available, amongst other places, at:
http://www.sdplastics.com/peek.html
2. Shop for some good quality copper foil. For example here:
http://allfoils.thomasnet.com/item/copper/copper-sheet/pn-1005?
3. Bond the coppersheet to the Peek with epoxy
4. Use your favorite method for mask transfer (toner / lithography)
5. Edge away!
6. Drill holes where needed or use as is for SMD soldering.
I made some experiments with some simple patterns and so far the results are promissing.
It still needs fiddeling and finetuning. For example: I edge the PCB inside an ultrasonic bath (these for cleaning silver cuttlery in).
Too little power, insufficient effect, too much power - the copper tends to come loose.
Maybe its cumbersome, but I like to do as much as possible myself.
Cheers,
Catweasel
Then again, I have been warned. Still, I am not giving up. Maybe the 317 is not such a good choice afterall?
How would you do it?
I wonder how such solar/battery load driving systems work in satellites??
Cheers,
Catweasel
The Propeller can drive servos in its sleep (well almost). You don't have to make sure and refresh the servos every 20ms since the Prop does it for you.
If you're running out of pins and you have extra Stamps, then I suppose it could be useful. I personally think there are many better options (and much less expensive) for increasing the IO capability of a Prop than with a Stamp (including using a second Prop).
While surfing around, I found this one:
http://www.youtube.com/watch?v=EJcbxrdErkY
The I/O pins are indeed a bottleneck. At the moment the system consists of:
8 "Sampling units" (SU)
A SU consists of:
2 servo motors
2 photosensors
2 LEDs
The LEDs and the photosensors are used to verify the servo motions (so mechanical failure is noticed)
That makes 6 I/O lines per SU, or 48 for all SUs
Then it has 5 "Enviromental Control Units" (ECU)
An ECU consists of:
1 temperature sensor
1 humidity/water sensor
1 heating element (transistor controlled)
That requires another 15 input lines.
Last, but not least, the "Recovery Unit" (RU)
The RU consists of a GPS and a GSM module.
The design of this one is till in its concept phase. So I don't know how many I/O lines would be needed.
The idea is that if the probe descends at the end of the mission (verified by the altimeter), the GPS position shall be transmitted in a SMS via the GSM module.
It is assumed the proble lands within GSM network coverage.
Then, there is the "main unit" which is connected to
- a gyroscope
- an altimeter
- accelerometer
My plan was/is to use one Stamp per SU, one Stamp for the 5 ECU units and a Propeller for the main unit.
The SU-Stamps would then, if commanded, move and monitor the servos.
The ECU-Stamp could easily handle the enviromental control job, e.g. turn on the heating when needed, signal a water hazard, etc.
The propeller would then run the main mission program, including the sampling schedule, deciding not to take a sample if the vehicle is in an undesired orientation, etc.
The main "flight program" would be done by the Propeller. So I wanted to make the SUs and the other subcomponents as autonomous as possible.
Sure, with clever programming and some I/O expanders, everything could be one with a single Propeller. But if it fails, the probe is useless.
If, for example, a Stamp fails, only one SU is out of the game.
It also allows for more flexibility. If testing reveals more ECUs or SUs are needed, they can be added blockwise.
Although it is just a simple baloon probe, I try to build it as it it would have to survice on the moon :-)
Cheers,
Catweasel
In satellites the solar cells are always facing the sun and there are no clouds or atmosphere to block sunlight. This means they can be connected to provide the optimum voltage for the system to use and they will provide a steady voltage and current except when in earth's shadow. Battery charging and regulating power to the electronics is done by switching regulators for maximum efficiency.
In your case it would depend on how much power the solar cell can produce and how much it varies. If expense was not a consideration I would probably connect the solar cells for 24V output, use a switching charge regulator for charging a 12V battery, a switching regulator to provide +5V from the battery, a linear regulator to go from 5V to +3.3V if the current draw is low, or a switching regulator from the battery if the 3.3V power requirement is high.
BTW you could use an LDO (low drop out) regulator for the battery charging circuit. That would increase the voltage available to the battery by about 1.4 - 1.6V. If the solar cells provide no more than 200mA in bright sunshine you could connect it to the battery through a transistor controlled by the prop and control the charging through software.
If you're going to be using GPS or other radio devices, be aware switching regulators can be very noisy.
When I used a switching regulator with a GPS logger I was making, the GPS module could no longer lock on the the satellites.
I've also had radio interference when using a switching regulator with a 2.4GHz RC transmitter.
Apparently there are ways to sheild the regulators, but I haven't tried this (successully) yet.
That would be four cells in parallel. A 12V battery should be posible. (Still cheaper than LiPo). A single C-Block should suffice. The 3.3V is only used to drive the Propeller and its main sensors. The Stamps are powered by the 5V rail. Also the Servos and other "high-load" components are powered directly by the source (not via the Stmp I/O pins).
That is going into the direction od an "active power control". I thought about that.
Hooking a Stamp with through an RC circuit to the battery and solar cells, scanning for over/under voltage and activel switching the charger/battery on/off.
I just thought this could be realized by rather "dumb" feedback circuits. The same way cou can make a blinking LED without any microcontroler. Just with a battery, capacitor, resistor an LED and one transistor. Instead of programming a high-low sequence, the value of C determines the frquency, the vakue of R controls the brightness.
What type of LDO regulator would you recommend?
Cheers,
Catweasel
Would it be sufficient if the GPS module had ome spacing between itself and the rest of the system (for example having it tethered some meters below the probe)?
Cheers,
Catweasel
I don't believe the noise being referred to is acoustical but rather rf and power. Just moving the gps away from the switching regulator will not solve the power noise issue.
Jim
I thought he meant that the regulators would work like HF senders....
Why would distance not lessen the effect of these?
Cheers,
Catweasel
A change in GPS/regulator distance of a few inches can be the difference between the GPS unit not getting a lock and it getting a lock. Of course a few feet of separation would reduce the interference much more than a separation of a few inches.
A 5 meter tether might introduce new problems. I don't think you'd need that much distance.
I used this switching regulator with my GPS logger hoping to have the batteries' charge last longer. I was using a plastic Sucrets box as a case so everything was very close together. I had previously logged some data as I walked using a linear regulator. The log file after using the switching regulator was full of error messages stating a lock hadn't been made with the satellites.
After mentioning the problem of interference from a switching regulator, Mike Green suggested there are ways of shielding the offending regulator. I think he mentioned using some sort of metal cage and the correct filtering capacitors (this isn't a direct quote).
I tried wrapping the regulator with 5mm copper tape. I don't know if the copper tape helped or not, but the regulator still interfered with the GPS unit.
Many GPS units will give the signal strength from each satellite. It should be possible to quantify the interference of a regulator and monitor and change based on shielding attempts. I watched the signal strength numbers as I was using the LCD and back light. When the LCD was next to the GPS unit, all signals were lost. As I moved the LCD away from the GPS unit the signal strength would increase. Just a few inches made the difference between no signal and very poor signal strength. I think 12 inches of distance between LCD and GPS was enough to return the signals to almost full strength. I don't recall the exact distances and strength readings from these tests. It should be very difficult to repeat these tests with whatever regulator you wish to use.
I've recently purchased a roll of 2" copper tape. I'm hoping the wider tape will make a better Faraday cage than the thinner 5mm tape. I haven't tried making a new cage yet.
I also found the back light voltage booster on Rayman's 4.3" LCD breakout board also causes interference problems with my GPS unit.
I haven't tested the back light voltage booster with a RC transmitter, but I have tested the small DE regulator with a 2.5GHz Spektrum transmitter and I found the regulator severely interfered with the RC transmission. It's ironic that DE links to a tutorial of how to replaces a linear regulator with their switching regulator inside a Spektrum transmitter.
I've also been warned against using a switching regular in cell phone projects.
I used to work for TomTom. When we opened defective devices (faulty battery, etc), I could never see any special RF protection and the internal batteries, which run at 3,7V. So I guess the trick is not to use any regulators and tailor the power source to the desired level. Of course those devices only need one voltage level.
Would the regulators from the Propeller eduction board (LM2940 ad LM2937) be problematic?
It is reasonable to assume, if HF interference is a problem, that this also affects GSM connections.
Depending on the GSM module I'm gonne be using (haven't decided on one yet), I could hook it up with a very small battery. If it can send the SMS more or less autinomous, I could power the probe down before sending. The GSM module could only be used after landing anyway (No GSM coverage at 8000 meters). So periodical location updates are not possible anyway.
I also don't know how much static electricity will have an effect on the system. (Iimagine the probe flying through a highly charged cloud).
At the moment I consider for the structural parts polystyrol that is reinforced with fiberglass gaze and epoxy. The shape is cylindical.(Diameter 60cm; height 125cm). The maximum weight I can afford is 10Kg, or I have to go with a different baloon (which I on't really want to do).
If I wrap the polystyrol in aluminium foil first, would that help in terms of shielding?
Cheers,
Catweasel
After a little more consideration, I get more and more convinced that I should not cut costs on the battery. But if I am going to spend a bit more money on a battery, it should be the optimal battery for the job. If you had to shop for a battery here (I get some discount there), whichone would you pick?
http://www.conrad.com/Rechargeable-Batteries.htm?websale7=conrad-int&ci=SHOP_AREA_14717
From a quality perspective, I always had good expieriences with the "Emmerich" batteries....
I made some tests with the servos and they can handle the torque, so I will definitely not need a 12V rail. Just a strong 5V for the stamps and the servos and the 3.3V for the Propeller.
Cheers,
Catweasel
Not knowing the capacity of your battery pack or the output of your solar cells I would not know which one to recommend. There are lots of choices available. You need to get one that can handle the current, power dissipation, and a low enough drop out voltage to charge the batteries.
This can be done with analog circuits but it would simpler to have one of the micros handle it. I would suggest doing two tests to determine how simple the charging circuit can be.
The first test would be to plot a current/voltage curve for the solar cells in full sunshine.
The second test would be to see if there is any current flow from the battery through the solar cell when it is in the dark.
The other test to do is to determine the current draw of the electronics and the servos/motors.
Good that you can drop down to a lower voltage. That makes everything simpler.
4 NiMh cells will provide 5.2V and the solar cells can be connected to provide 6V for charging.
The servos can connect directly to the batteries.
An LDO regulator for the 3.3V and 5V (although you may need to drop the 5V to 4.8V).
Hmm. Sounds I should make some power consumption tests first.
After reading into the "maximum power point tracking" concept a bit further, I think a pure analog solution it a bit over my head. Especially since I don't have that much expierience in analog circuits, its propably not the best one to start with. I will draw up a schematic for a Stamp controled power distributions. Lets see how that turns out :-)
I planned to set up a performance test for the solar cells next week. I will test sunlight at various angles as well es occluding the cells with colored foils to simulate dimmed lighting conditions.
A tough one to test as I haven't decidd on a battery yet.
The choice of battery is dependent of the design of the power supply which is depending on the choice of battery.
So many options...
That will depend mainly on the mission profile. If I take samples on a lower frequency, the servos would have to act less often.
The sampling sequence uses several servos, but only one at a time. (One servo moves the sample container in front of the sample port. Then another servo opens the port hatch.
My approach was to see what power allowance I have and then design the mission accordingly.
But I could indeed make some power consumption profiles for each sampling process.
Yes. I underestimated the actual strength of those 6V motors.
The LM2940 and LM2937 should do the job.
The attached schematic does not have any voltage regulation yet. I just tried to set up RC loops for measuring the solar and battery output.
A on/off switch for the powered circuits. For example P14 controls the RC charge period, then P15 reads the discharge time.
Would that work?
Cheers,
Catweasel
You are putting the battery/solar cell voltage on the prop pin P15 through a 220 ohm resistor. This will fry the pin and possibly the whole chip.
None of the transistors will do what you expect them to do. They are all NPN transistors. Some are connected backwards and the others can never be turned off as shown.
I don't have time right now but I will take a look later tonight and make some suggestions unless someone else pipes up before then.
In the mean time don't connect it up as shown or you will most likely let out some smoke.
Is that what you want?
I have sent you a PM and have some more suggestions for you.