IMPORTANT! THE BEST WAY OF POWERING THE UDB, SERVOS, RECEIVER,etc.
140 views
Skip to first unread message
nickarsow
Jun 14, 2012, 1:19:51 AM6/14/12
to uavde...@googlegroups.com
Hi guys,
I think it's time to place such discussions as the RC electronics become more and more complicated and the headaches become more and more.
I invoke all experienced pilots to place their oppinions and advices here.
Please do correct me and throw in in any point you think I'm wrong or I'm missing something! Thus the thread will become more detailed.
PART ONE
THE PURPOSE OF THIS THREAD IS TO DISCUSS AND OFFER:
1. The optimal power sourcing;
2. The optimal connections between devices;
3. Minimizing the EMI and current bounces;
4. Minimizing the cable count;
5. Achieving the best way of powering the different devices;
6. Achieving the seamless coexistence between different devices.
I'll start with my thoughts as follow:
A/ What hardware we have?
a/ base hardware - a battery, an ESC, a motor, a receiver, an UDB, servos, a GPS, a camera, an OSD, a video transmitter and a lot of cables;
b/ optional hardware - different modules - telemetry, dataloggers..etc.
B/ What problems we have?
a/ too many different boards by size, by function and by sensitivity;
b/ too many cables - high current, high EMI, low current, data, RF, etc.;
c/ too many sensitive electronic modules - receiver, UDB, IMU, other sensors, RF, etc.
C/ What power sources we have?
a/ batteries.
D/ What voltage regulators we have in common?
a/ 78x05 from the ESC;
b/ any type of BEC;
c/ in some cases a StepUp ( for 12VDC where needed ).
E/ What filtering we have?
a/ different types of ferrite toroids - both with proper and improper Al, ui and material;
b/ different types of LC filters - both proper calculated and improper calculated.
F/ What are the devices with high radiated EMI, RF and high current bounces?
a/ the motor;
b/ the OSD ( if MAX7456 is in use );
c/ servos;
d/ the onboard transmitters;
e/ DC-DC Buck, Boosts or any other SMPS DC-DC wich uses non shielded inductors or poor input/output filtering.
G/ What are the grounds with specific connection point?
a/ the ESC GND;
b/ the BEC GND;
c/ the transmitter GND;
d/ the OSD GND.
e/ servo GNDs.
H/ There are also many other questions concerning this matter.
SO, LET'S START THINKING!
I. Battery power connector and current sensor board.
I'll start with the nearest connection to battery. On my oppinion it should be the current sensor.
There are too many ICs for current sensing - INA138/139/168/169 ( TI ) , ZXCT10xx ( Diodes Inc. ), ACS7xx ( Allegro ), etc.
What we should have on current sensor board:
1. Input terminals/solder pads;
2. Output terminals/solder pads. I would advice two pairs + one optional - one for the ESC, one for the servo power board and the optional one ( but preferable for me ) for the BEC, powering the other electronics. I prefer this approach as each current flow contour is separate and currents from one device won't flow through other contours;
3. Voltage divider for voltage sensing;
4. Current sensor. I do not have a favorite. I've been using current sensors from Allegro Micro and from Diodes Inc. Allegro's do not need a current schunt, while the Diodes need a high current one.
5. Also ( optional ) a low voltage emergency sounder could be implemented.
The best solution for the current and voltage sensor outputs is to have them optoisolated, but this is more expensive.
II. ESC and motor
What we have here - cables between the current sensor board and the ESC input capacitor. These two cables ( + and - ) radiate a lot of EMI.
Some good readings say that:
1. Shorts cables = low EMI;
2. Good ESC decoupling capacitors ( low ESR ) = low EMI. Each phase current loop flows through the phase winding, the appropriate mosfet and the input cap.
3. Output shielding. There's a good approach with a shielding of all 3 phase cables with only one shielding point in the (-) of the input capacitor. The motor end of the shielding tube should NOT be grounded.
4. The ESC voltage regulator. Commonly it is 5V 1A to 3A. This voltage could be used with a good low ESR cap, placed as close as possible to the powered device power connector. Thus the inserted EMI is highly redused. On my opinion this voltage regulator should be used in a small models/trainers with max 4 small servos.
END OF PART ONE
Please do comment this part.
I think it's time to place such discussions as the RC electronics become more and more complicated and the headaches become more and more.
I invoke all experienced pilots to place their oppinions and advices here.
Please do correct me and throw in in any point you think I'm wrong or I'm missing something! Thus the thread will become more detailed.
PART ONE
THE PURPOSE OF THIS THREAD IS TO DISCUSS AND OFFER:
1. The optimal power sourcing;
2. The optimal connections between devices;
3. Minimizing the EMI and current bounces;
4. Minimizing the cable count;
5. Achieving the best way of powering the different devices;
6. Achieving the seamless coexistence between different devices.
I'll start with my thoughts as follow:
A/ What hardware we have?
a/ base hardware - a battery, an ESC, a motor, a receiver, an UDB, servos, a GPS, a camera, an OSD, a video transmitter and a lot of cables;
b/ optional hardware - different modules - telemetry, dataloggers..etc.
B/ What problems we have?
a/ too many different boards by size, by function and by sensitivity;
b/ too many cables - high current, high EMI, low current, data, RF, etc.;
c/ too many sensitive electronic modules - receiver, UDB, IMU, other sensors, RF, etc.
C/ What power sources we have?
a/ batteries.
D/ What voltage regulators we have in common?
a/ 78x05 from the ESC;
b/ any type of BEC;
c/ in some cases a StepUp ( for 12VDC where needed ).
E/ What filtering we have?
a/ different types of ferrite toroids - both with proper and improper Al, ui and material;
b/ different types of LC filters - both proper calculated and improper calculated.
F/ What are the devices with high radiated EMI, RF and high current bounces?
a/ the motor;
b/ the OSD ( if MAX7456 is in use );
c/ servos;
d/ the onboard transmitters;
e/ DC-DC Buck, Boosts or any other SMPS DC-DC wich uses non shielded inductors or poor input/output filtering.
G/ What are the grounds with specific connection point?
a/ the ESC GND;
b/ the BEC GND;
c/ the transmitter GND;
d/ the OSD GND.
e/ servo GNDs.
H/ There are also many other questions concerning this matter.
SO, LET'S START THINKING!
I. Battery power connector and current sensor board.
I'll start with the nearest connection to battery. On my oppinion it should be the current sensor.
There are too many ICs for current sensing - INA138/139/168/169 ( TI ) , ZXCT10xx ( Diodes Inc. ), ACS7xx ( Allegro ), etc.
What we should have on current sensor board:
1. Input terminals/solder pads;
2. Output terminals/solder pads. I would advice two pairs + one optional - one for the ESC, one for the servo power board and the optional one ( but preferable for me ) for the BEC, powering the other electronics. I prefer this approach as each current flow contour is separate and currents from one device won't flow through other contours;
3. Voltage divider for voltage sensing;
4. Current sensor. I do not have a favorite. I've been using current sensors from Allegro Micro and from Diodes Inc. Allegro's do not need a current schunt, while the Diodes need a high current one.
5. Also ( optional ) a low voltage emergency sounder could be implemented.
The best solution for the current and voltage sensor outputs is to have them optoisolated, but this is more expensive.
II. ESC and motor
What we have here - cables between the current sensor board and the ESC input capacitor. These two cables ( + and - ) radiate a lot of EMI.
Some good readings say that:
1. Shorts cables = low EMI;
2. Good ESC decoupling capacitors ( low ESR ) = low EMI. Each phase current loop flows through the phase winding, the appropriate mosfet and the input cap.
3. Output shielding. There's a good approach with a shielding of all 3 phase cables with only one shielding point in the (-) of the input capacitor. The motor end of the shielding tube should NOT be grounded.
4. The ESC voltage regulator. Commonly it is 5V 1A to 3A. This voltage could be used with a good low ESR cap, placed as close as possible to the powered device power connector. Thus the inserted EMI is highly redused. On my opinion this voltage regulator should be used in a small models/trainers with max 4 small servos.
END OF PART ONE
Please do comment this part.
nickarsow
Jun 14, 2012, 1:21:17 AM6/14/12
to uavde...@googlegroups.com
PART TWO
SERVOS AND SERVO POWER
There are several concepts for driving and powering the servos:
1. Direct connection and powering from the receiver/UDB. This approach is applicable when using a couple of small servos. The main problem is that all servos' current path overlaps the current power path of the receiver. That's not good.
2. A separate servo power with common ground with the receiver/UDB. This approach is commonly used with more powerfull servos. The problem here is almost the same as in p.1 - common ground current power path with the receiver/UDB.
3. Optoisolated servo board with separate power. A good approach with some remarks:
a/ a good decoupling capacitor lith loe ESR should be placed as close as possible to the input power connector;
b/ the GND and the Vcc should be captured directly from the current sensor board via a separate connector.
4. Optoisolated PPM->PWM servo board. THIS IS THE PERFECT SOLUTION. What we have here - totally optoisolated current path, power source - current sensor board. So we have the shortest current path possible - current sensor output, current sensor board capacitor, servo board capacitor, servos' power. Perfect!...just do it... ;-)
Some schematics attached.
END OF PART TWO
SERVOS AND SERVO POWER
There are several concepts for driving and powering the servos:
1. Direct connection and powering from the receiver/UDB. This approach is applicable when using a couple of small servos. The main problem is that all servos' current path overlaps the current power path of the receiver. That's not good.
2. A separate servo power with common ground with the receiver/UDB. This approach is commonly used with more powerfull servos. The problem here is almost the same as in p.1 - common ground current power path with the receiver/UDB.
3. Optoisolated servo board with separate power. A good approach with some remarks:
a/ a good decoupling capacitor lith loe ESR should be placed as close as possible to the input power connector;
b/ the GND and the Vcc should be captured directly from the current sensor board via a separate connector.
4. Optoisolated PPM->PWM servo board. THIS IS THE PERFECT SOLUTION. What we have here - totally optoisolated current path, power source - current sensor board. So we have the shortest current path possible - current sensor output, current sensor board capacitor, servo board capacitor, servos' power. Perfect!...just do it... ;-)
Some schematics attached.
END OF PART TWO
nickarsow
Jun 14, 2012, 1:39:10 AM6/14/12
to uavde...@googlegroups.com
PART THREE
THE VOLTAGE SENSOR
In this part I'll comment some spot on voltage sensor.
There are two concepts on this - optoisolated and not isolated. The not isolated one takes the voltage directly at the battery terminals and uses a simple R-R dividor to maintain the voltage within the Vcc range ( see the previous schematics applied ). The common problem here is the grounding. This schematics uses the GND of the battery (-) terminal and connects it to the sensitive analog input of the MPU. Thus a sensitive loop is overlapped with a high current loop and commonly causes noise problems.
Second approach uses optoisolated measuring ( see the attached schematics ) and thus there's no GND connection between the primary high current loop and the sensitive analog input. This approach should use an optoisolated amplifier like ISO122/124 ( http://www.ti.com/lit/ds/symlink/iso122.pdf ) , but due to it's high price it could be replaced by a less precise and cheap linear opto as PS8821 ( http://www.cel.com/pdf/datasheets/ps8821.pdf ).
END OF PART THREE
THE VOLTAGE SENSOR
In this part I'll comment some spot on voltage sensor.
There are two concepts on this - optoisolated and not isolated. The not isolated one takes the voltage directly at the battery terminals and uses a simple R-R dividor to maintain the voltage within the Vcc range ( see the previous schematics applied ). The common problem here is the grounding. This schematics uses the GND of the battery (-) terminal and connects it to the sensitive analog input of the MPU. Thus a sensitive loop is overlapped with a high current loop and commonly causes noise problems.
Second approach uses optoisolated measuring ( see the attached schematics ) and thus there's no GND connection between the primary high current loop and the sensitive analog input. This approach should use an optoisolated amplifier like ISO122/124 ( http://www.ti.com/lit/ds/symlink/iso122.pdf ) , but due to it's high price it could be replaced by a less precise and cheap linear opto as PS8821 ( http://www.cel.com/pdf/datasheets/ps8821.pdf ).
END OF PART THREE
nickarsow
Jun 14, 2012, 3:36:58 AM6/14/12
to uavde...@googlegroups.com
PART FOUR
THE ESC
We have two types of ESC - not optoisolated and optoisolated. Not isolated include voltage regulator, commonly 5V from 1A to 3A. Also it uses common grounding and thus the low current voltage regulator ( commonly a few 78M05 ) ground overlaps the high current ground of the ESC. As the ESC uses three Half Bridges for driving the motor, the switching mode causes a lot of noise. More, we have long wires betwenn the ESC and the motor. They in turn emit a lot of EMI. A good optional approach is to shield the wires between the ESC and motor with a PVC coated shielding tube. Grounding of the shielding tube must be performed in just one point - as close to the ESC and soldered to the ESC (-) power point ( see the attached schematics ).
The ESC voltage regulator is commonly used for powering different devices - the receiver, the UDB, etc. This kind of ESC is commonly used for powering a low current devices/servos as they can't supply current exceeding 3A.
What we have - a ground path which is used for powering the sensitive devices, overlapped with a high current ground path - ESC<->Battery.
How noise and EMI could be reduced in this case?
1. Good low ESR capacitors directly connected to the power cables of the ESC;
2. Good low ESR capacitor connected at the output of the voltage regulator and as close as possible to the ESC signal cable solderings ( between +5V and GND );
3. A ferrite toroid on the signal ESC cable ( this not helps too much );
4. Keeping the noisy high current cables away from sensitive devices;
5. Do power the ESC as close to the current sensor as possible.
When using this type of ESC, one should have in mind the following:
1. Do power the UDB. The receiver is powered on it's turm from the UDB. This approach is better because the receiver is a low current, sensitive device. If receiver is powered first, all current drawn from servos and other peripheral devices will flow through the signal cables in a random way. When powering the UDB, the receiver will distribute a very low current for it's powering through several cables thus avoiding high current loops.
2. NEVER use two power sources as ESC and additional BEC for powering the devices. Just choose one of them. I highly recommend BEC for powering the UDB, receiver, etc. When using BEC for powering the devices and ESC with volage regulator, DO NOT FORGOT to CUT the red wire of the ESC !
3. When using only a PPM from the receiver, do use the same cable for powering it....no problem, the receiver is a low power device. This cable takes the receiver power from the UDB power line, but as the receiver current is so small, no disturbances are expected.
4. When using BEC or other voltage regulator for powering the UDB and servos, make sure that the current it could supply is at least 10% more power than the worst case. Thus you'll have a cooler power supply without danger of tripping the current protection. The current protection tripping can left your devices without power.
OPTO DECOUPLING ( optoisolated ESC )
Optoisolated ECSs use optoisolator between the receiver/UDB and the ESC. They do not offer voltage regulators for powering the devices. Optoisolated ESCs use the same technics for lowering EMI as described in p.1, p.2, p.4 and p.5 . I do prefer these ESCs for less headaches I have with noise.
When using optoisolated ESCs, a few things should be taken in mind:
1. An additional power supply have to be used for powering the devices;
2. Points 3 and 4 from the writing a bit above should be also taken in mind;
3. Shielding and displacement of the cables between the ESC and the motor is the same as discussed above.
That's for now... ;-)
END OF PART FOUR
THE ESC
We have two types of ESC - not optoisolated and optoisolated. Not isolated include voltage regulator, commonly 5V from 1A to 3A. Also it uses common grounding and thus the low current voltage regulator ( commonly a few 78M05 ) ground overlaps the high current ground of the ESC. As the ESC uses three Half Bridges for driving the motor, the switching mode causes a lot of noise. More, we have long wires betwenn the ESC and the motor. They in turn emit a lot of EMI. A good optional approach is to shield the wires between the ESC and motor with a PVC coated shielding tube. Grounding of the shielding tube must be performed in just one point - as close to the ESC and soldered to the ESC (-) power point ( see the attached schematics ).
The ESC voltage regulator is commonly used for powering different devices - the receiver, the UDB, etc. This kind of ESC is commonly used for powering a low current devices/servos as they can't supply current exceeding 3A.
What we have - a ground path which is used for powering the sensitive devices, overlapped with a high current ground path - ESC<->Battery.
How noise and EMI could be reduced in this case?
1. Good low ESR capacitors directly connected to the power cables of the ESC;
2. Good low ESR capacitor connected at the output of the voltage regulator and as close as possible to the ESC signal cable solderings ( between +5V and GND );
3. A ferrite toroid on the signal ESC cable ( this not helps too much );
4. Keeping the noisy high current cables away from sensitive devices;
5. Do power the ESC as close to the current sensor as possible.
When using this type of ESC, one should have in mind the following:
1. Do power the UDB. The receiver is powered on it's turm from the UDB. This approach is better because the receiver is a low current, sensitive device. If receiver is powered first, all current drawn from servos and other peripheral devices will flow through the signal cables in a random way. When powering the UDB, the receiver will distribute a very low current for it's powering through several cables thus avoiding high current loops.
2. NEVER use two power sources as ESC and additional BEC for powering the devices. Just choose one of them. I highly recommend BEC for powering the UDB, receiver, etc. When using BEC for powering the devices and ESC with volage regulator, DO NOT FORGOT to CUT the red wire of the ESC !
3. When using only a PPM from the receiver, do use the same cable for powering it....no problem, the receiver is a low power device. This cable takes the receiver power from the UDB power line, but as the receiver current is so small, no disturbances are expected.
4. When using BEC or other voltage regulator for powering the UDB and servos, make sure that the current it could supply is at least 10% more power than the worst case. Thus you'll have a cooler power supply without danger of tripping the current protection. The current protection tripping can left your devices without power.
OPTO DECOUPLING ( optoisolated ESC )
Optoisolated ECSs use optoisolator between the receiver/UDB and the ESC. They do not offer voltage regulators for powering the devices. Optoisolated ESCs use the same technics for lowering EMI as described in p.1, p.2, p.4 and p.5 . I do prefer these ESCs for less headaches I have with noise.
When using optoisolated ESCs, a few things should be taken in mind:
1. An additional power supply have to be used for powering the devices;
2. Points 3 and 4 from the writing a bit above should be also taken in mind;
3. Shielding and displacement of the cables between the ESC and the motor is the same as discussed above.
That's for now... ;-)
END OF PART FOUR
nickarsow
Jun 15, 2012, 3:13:50 AM6/15/12
to uavde...@googlegroups.com
PART FIVE
CONNECTING PERIPHERAL DEVICES TO THE UDB
There are a lot of peripheral devices one uses to connect to the UDB - analog sensors, digital devices, I2C devices, SPI devices, OSD, etc.
In this chapter I'll explain how this devices should be connected to the system, but the OSD. As the OSD is connected to the high powe consuming transmitter, I'll explain it in the next chapter of this reading.
Commonly there are no headaches with devices like analog sensors, digital IN/OUT devices, I2C and SPI devices. Simply one uses twisted cabling and that is enough in common, BUT.....
Analog devices ase too sensitive to common noise. They should be shielded.
SHIELDING OF ANALOG DEVICE LINES
1. Do not ground both ends of the shielding;
2. Common grounding technics is to connect the shielding to the most "quiet" point - the nearest to the battery large capacitor (-)...usually it should be on the current sensor board. As this approach is difficult to realise, the good technics is to ground the shielding to the same sonnector GND - as shown on the applied schematics.
3. A good manner when designing analog peripherals is to place a tantalum cap 10-100uF + 100nF ceramic X7R capacitor as close to the power connector of the analog board as possible.
4. Do always check the current consumption of the analog device and make sure it draws less current than the UDB could supply.
SHIELDING ( OPTIONAL ) OF DIGITAL LINES
1. As the digital levels have enough voltage margin and are not so sensitive, the digital lines shielding is optional and usually the twisted cables works just fine.
2. A good technic is to use a ferrite toroid and pierce the twisted cabe set till the ferrite inner hole is fullfilled. A good ferrite materials for the purpose are N30, N47 and N87.
3. A good manner when designing digital peripherals is to place a tantalum cap 10-100uF + 100nF ceramic X7R capacitor as close to the power connector of the board as possible.
4. Do always check the current consumption of the digital device and make sure it draws less current than the UDB could supply.
POWERING HIGH POWER DIGITAL DEVICES / HIGH POWER LEDs
When powering high power LEDs or other high power devices through the digital lines, do always use an optoisolation with an optoisolator and power the high power device directly from the current sensor ( see the attached schematics ).
END OF PART FIVE
CONNECTING PERIPHERAL DEVICES TO THE UDB
There are a lot of peripheral devices one uses to connect to the UDB - analog sensors, digital devices, I2C devices, SPI devices, OSD, etc.
In this chapter I'll explain how this devices should be connected to the system, but the OSD. As the OSD is connected to the high powe consuming transmitter, I'll explain it in the next chapter of this reading.
Commonly there are no headaches with devices like analog sensors, digital IN/OUT devices, I2C and SPI devices. Simply one uses twisted cabling and that is enough in common, BUT.....
Analog devices ase too sensitive to common noise. They should be shielded.
SHIELDING OF ANALOG DEVICE LINES
1. Do not ground both ends of the shielding;
2. Common grounding technics is to connect the shielding to the most "quiet" point - the nearest to the battery large capacitor (-)...usually it should be on the current sensor board. As this approach is difficult to realise, the good technics is to ground the shielding to the same sonnector GND - as shown on the applied schematics.
3. A good manner when designing analog peripherals is to place a tantalum cap 10-100uF + 100nF ceramic X7R capacitor as close to the power connector of the analog board as possible.
4. Do always check the current consumption of the analog device and make sure it draws less current than the UDB could supply.
SHIELDING ( OPTIONAL ) OF DIGITAL LINES
1. As the digital levels have enough voltage margin and are not so sensitive, the digital lines shielding is optional and usually the twisted cables works just fine.
2. A good technic is to use a ferrite toroid and pierce the twisted cabe set till the ferrite inner hole is fullfilled. A good ferrite materials for the purpose are N30, N47 and N87.
3. A good manner when designing digital peripherals is to place a tantalum cap 10-100uF + 100nF ceramic X7R capacitor as close to the power connector of the board as possible.
4. Do always check the current consumption of the digital device and make sure it draws less current than the UDB could supply.
POWERING HIGH POWER DIGITAL DEVICES / HIGH POWER LEDs
When powering high power LEDs or other high power devices through the digital lines, do always use an optoisolation with an optoisolator and power the high power device directly from the current sensor ( see the attached schematics ).
END OF PART FIVE
nickarsow
Jun 15, 2012, 7:15:22 AM6/15/12
to uavde...@googlegroups.com
PART SIX
POWERING THE UDB, REDUCING THE CABLE COUNT BETWEEN THE UDB AND
RECEIVER/SERVOS
The main but not the most difficult part of this reading is the UDB powering
and connections to the receiver and servos.
What we have here?
There are two main approaches of connecting the UDB to the receiver and servos
- parallel and serial. Under the name of serial I mean PPM, not a kind of
serial bus although the concept is almost the same. Also we could have a mixed
connection - PPM input and parallel output and vise versa.
PARALLEL CONNECTION
This technics is commonly used when driving a few small power servos. What we
have connected - UDB to the receiver via a 3 wire cables and all servos
connected to the UDB via their 3 wire cables. The particularity here is the ESC
( see the "PoweringUDB_common.pdf" file attached ).
Whe using a few low power servos, one do not need a separate BEC/VR ( voltage
regulator ) for powering the UDB. The ESC onboard BEC could be enough for the
purpose. The only additional
good thing one could supply is a ferrite toroid over the ESC signal cable.
When using a lot of servos, you need a separate power supply - could be linear
or switched mode. In this situation there are two options:
1. Just cut the red ( +5V )
and the brown (GND) wires of the ESC cable when one uses not optoisolated
ESC. As the ESC is directly
connected to the current sensor, it gets it’s GND from there. Thus we have no
overlapped GND contours. ( if we connect the ESC signal GND to the UDB we have
overlapped grounding/crossed ground contours which will cause a lot of
disturbances).
2. When using optoisolated ESC, just use it as is. As it has no BEC/VR, one has
to power the UDB with a separate power supply - BEC/VR.
PPM CONNECTION
This approach is used for reducing cable count when a lot of channels and
servos are used. Using a lot
of servos requires a lot of power. The applied schematics ( Powering
UDB_PPM.pdf file ) describes the situation. What we have ?
1. Receiver with a PPM putput or a PPM encoder board between the receiver and
the UDB;
2. PPM to PWM decoder board with or without optoisolation;
3. A mixed situation.
According the type of input of the PPM->PWM servo decoder - optoisolated or
not we have two subsections:
1. The PPM->PWM servo decoder is not optically isolated ( I don’t recommend this ). In this
case we must either cut the RED (+5V) and BROWN ( GND ) wires of the ESC signal cable or use an
optoisolated ESC. The UDB is powered from it's own BEC/VR.
2. The PPM->PWM servo decoder is optically isolated ( the best case ). In this case we have again two subsections:
a/ The ESC is not optoisolated. We have to cut the RED ( +5V ) and BROWN ( GND ) wires of the ESC signal cable;
b/ The ESC is optoisolated – just use it as is…..this is the best situation ever!
In all cases when the RED (+5V) and BROWN (GND) ESC signal wires are cut, it’s a good practice to shield the ORANGE ( signal ) and grounding the shield JUST IN ONE POINT as close to the current sense output capacitor as possible.
I'd highly apreciate some discussions on cutting the brown(GND) wire with blog members as there are some controversial readings in the web, but I think this is the right approach.
CONCLUSIONS FROM THIS CHAPTER
1.Use an optoisolated input PPM -> PWM decoder . Thus you’ll have less headaches.
2.Use an optoisolated ESCs.
3. Do always avoid overlapped contours. Do never use one and the same ground wire for different contours.
4. If one decide to reduce the UDB wire count, the only way is to use PPM if applicable.
END OF PART SIX
Message has been deleted
nickarsow
Jun 18, 2012, 5:59:58 AM6/18/12
to uavde...@googlegroups.com
PART SEVEN
POWERING AND WIRING OSDs, CAMARAS AND VIDEO TRANSMITTERS
This is the "dirtiest" part of my writing as it is most difficult to realise noiseless wiring.
I'll explain some attached example schematics showing some proven technics.
What hardware we have:
1. A video transmitter, commonly with it's own linear or switched mode power supply. They are powered in common from 7 to 15VDC. The power dissipation from the onboard voltage regulator is proportional to the differential voltage difference;
2. A camera. We have two types according to power voltage ... 5VDC powered and wide voltage powered...for example 7-13VDC;
3. An OSD. There are so many OSD designs....really many. We have OSDs with onboard voltage regulators and with no onboard voltage regulators. Second ones are 5VDC driven;
4. A microphone. Also there are too many variants here...5VDC powered, 12VDC powered and wide voltage powered;
5. Filters and voltage regulators. I'll discuss this on the go.
I'll discuss here some examples which use the simplest OSD possible with SPI interface as it is the most used one. There are a lot of OSDs with a serial TxRx interfaces ( such as REMZIBI's ), but as they are a subcase I won't discuss them.
SPI OSD has 3 digital lines to the UDB - SDI, CLK and CS, 2 analog lines - VIDEOout and VIDEOin and 2 power lines - 5VDC and GND.
In common OSDs are using MAX7456, which is known as noisy. Of course, the good PCB design reduces the radiated noise to minimum.
I'll describe 5 refferense designs which cover almost all cases:
A. OSD powered from the UDB ( +5V ), separate battery for powering the Transmitter and camera ( I'll award this design with 3 stars *** )
The simplest and less noise schematics is shown on "Powering_OSD_Camera_VTx_BAT.pdf". In this case we have a separate battery for powering the OSD/VTx/Camera. This is the less noisiest schematics as there is no common power ground beteen the UDB and the OSD but just a signal ground. Although it is a good practice to power the transmitter via a LC filter to avoid polution of the other devices in this circuit. There is an optional low value OHm resistor along the GND between the OSD and the battery (-). This low OHm resistor ( bould be also ferrite bead ) lowers the common noise. As described above, the MAX7456 is a bit noisy IC, it is a good practice to place a ferrite toroid over the OSD power lines as shown in the schematics.
B. OSD powered from a separate 5V voltage regulator, separate battery for powering the Transmitter and camera ( I'll award this design with 4 stars **** )
This ( Powering_OSD_Camera_VTx_Bat_2.pdf ) is almost the same as previous one but the OSD is powered from it's own 5V regulator from the second battery which also powers the transmitter and camera. Towards the UDB we have only the 3 digital lines and a GND. This ground could optionally use a low value OHm resistor or a ferrite bead for additional noise reduction. This schematics offer an additional advantage for powering 5V cameras as shown on the schematics with a dashed lines.
C. OSD powered from a separate 5V voltage regulator, the transmitter and camera are powered from the BATT_OUT and the GND of the current sensor ( I'll award this design with 2 stars ** )
This schematics ( Powering_OSD_Camera_VTx.pdf ) uses the main power for powering the transmitter and the camera and additional 5V voltage regulator for powering the OSD. If the main battery is higher than 3S, an additional 12V regulator has to be used. If the main battery is 3S, the regulator has to be avoided. The OSD is again separately powered from a separate 5V regulator, thus allowing 5V cameras to be used. An optional low value resistor/ferrite bead brings an additional noise filtering.
D. OSD powered from the UDB, transmitter and camera powered from the main supply ( I'll award this design with 1 star * )
This schematics ( Powering_OSD_Camera_VTx_2.pdf ) uses the main power for powering the transmitter and the camera and additional 5V voltage regulator for powering the OSD. If the main battery is higher than 3S, an additional 12V regulator has to be used. If the main battery is 3S, the regulator has to be avoided. The OSD is powered from the UDB and again a good practice is to use a ferrite toroid for the two power lines ( +5V and the GND ). The low OHm resistor/ferrite bead in this case is mandatory. The BEST case is if this resistor is implemented over the OSD board between the UDB and the VIDEO sides !
E. Optoisolated OSD ( Powering_OSD_Camera_VTx_OPTO.pdf )......5 stars *****
In this schematics the datalines between the OSD and the UDB are totally optoisolated. Unidirectional optoisolators are enough as the communication in common is unidirectional. A good optoisolators are IL711/712 ( file attached ). In this case powering of the OSD is from a separate 5V regulator, thus alolowing 5V cameras to be used. Powering of the transmitter and the camera is performed as already described - either from 3S battery via the current sensor or via a 12V regulator if the main power is > 3S.
LC FILTERS USED
As the video transmitter is a noisy device, good LC filters should be used. Here are some good filter examples:
1. http://shop.righthere.nu/fpv-cameras-accessories/15a-25v-dc-power-line-filter-for-fpv-video-tx-rc-models/prod_1299.html
2. http://www.fpvmanuals.com/2011/06/17/new-lc-power-filter/
3. http://www.fpvflying.com/products/PSF-LC-power-filter.html
My favorite is number 1. This filter is based on Murata's BNX02 EMIFIL filter series ( a pdf file attached ).
DEAR FRIENDS, AS THIS PART IS THE MOST DIFFICULT PART, PLEASE PLACE HEREAFTER ALL YOUR COMMENTS AND REMARKS. THUS ALL WE COULD DISCOVER THE BEST DESIGN WE ASPIRE TO !
END OF PART SEVEN
POWERING AND WIRING OSDs, CAMARAS AND VIDEO TRANSMITTERS
This is the "dirtiest" part of my writing as it is most difficult to realise noiseless wiring.
I'll explain some attached example schematics showing some proven technics.
What hardware we have:
1. A video transmitter, commonly with it's own linear or switched mode power supply. They are powered in common from 7 to 15VDC. The power dissipation from the onboard voltage regulator is proportional to the differential voltage difference;
2. A camera. We have two types according to power voltage ... 5VDC powered and wide voltage powered...for example 7-13VDC;
3. An OSD. There are so many OSD designs....really many. We have OSDs with onboard voltage regulators and with no onboard voltage regulators. Second ones are 5VDC driven;
4. A microphone. Also there are too many variants here...5VDC powered, 12VDC powered and wide voltage powered;
5. Filters and voltage regulators. I'll discuss this on the go.
I'll discuss here some examples which use the simplest OSD possible with SPI interface as it is the most used one. There are a lot of OSDs with a serial TxRx interfaces ( such as REMZIBI's ), but as they are a subcase I won't discuss them.
SPI OSD has 3 digital lines to the UDB - SDI, CLK and CS, 2 analog lines - VIDEOout and VIDEOin and 2 power lines - 5VDC and GND.
In common OSDs are using MAX7456, which is known as noisy. Of course, the good PCB design reduces the radiated noise to minimum.
I'll describe 5 refferense designs which cover almost all cases:
A. OSD powered from the UDB ( +5V ), separate battery for powering the Transmitter and camera ( I'll award this design with 3 stars *** )
The simplest and less noise schematics is shown on "Powering_OSD_Camera_VTx_BAT.pdf". In this case we have a separate battery for powering the OSD/VTx/Camera. This is the less noisiest schematics as there is no common power ground beteen the UDB and the OSD but just a signal ground. Although it is a good practice to power the transmitter via a LC filter to avoid polution of the other devices in this circuit. There is an optional low value OHm resistor along the GND between the OSD and the battery (-). This low OHm resistor ( bould be also ferrite bead ) lowers the common noise. As described above, the MAX7456 is a bit noisy IC, it is a good practice to place a ferrite toroid over the OSD power lines as shown in the schematics.
B. OSD powered from a separate 5V voltage regulator, separate battery for powering the Transmitter and camera ( I'll award this design with 4 stars **** )
This ( Powering_OSD_Camera_VTx_Bat_2.pdf ) is almost the same as previous one but the OSD is powered from it's own 5V regulator from the second battery which also powers the transmitter and camera. Towards the UDB we have only the 3 digital lines and a GND. This ground could optionally use a low value OHm resistor or a ferrite bead for additional noise reduction. This schematics offer an additional advantage for powering 5V cameras as shown on the schematics with a dashed lines.
C. OSD powered from a separate 5V voltage regulator, the transmitter and camera are powered from the BATT_OUT and the GND of the current sensor ( I'll award this design with 2 stars ** )
This schematics ( Powering_OSD_Camera_VTx.pdf ) uses the main power for powering the transmitter and the camera and additional 5V voltage regulator for powering the OSD. If the main battery is higher than 3S, an additional 12V regulator has to be used. If the main battery is 3S, the regulator has to be avoided. The OSD is again separately powered from a separate 5V regulator, thus allowing 5V cameras to be used. An optional low value resistor/ferrite bead brings an additional noise filtering.
D. OSD powered from the UDB, transmitter and camera powered from the main supply ( I'll award this design with 1 star * )
This schematics ( Powering_OSD_Camera_VTx_2.pdf ) uses the main power for powering the transmitter and the camera and additional 5V voltage regulator for powering the OSD. If the main battery is higher than 3S, an additional 12V regulator has to be used. If the main battery is 3S, the regulator has to be avoided. The OSD is powered from the UDB and again a good practice is to use a ferrite toroid for the two power lines ( +5V and the GND ). The low OHm resistor/ferrite bead in this case is mandatory. The BEST case is if this resistor is implemented over the OSD board between the UDB and the VIDEO sides !
E. Optoisolated OSD ( Powering_OSD_Camera_VTx_OPTO.pdf )......5 stars *****
In this schematics the datalines between the OSD and the UDB are totally optoisolated. Unidirectional optoisolators are enough as the communication in common is unidirectional. A good optoisolators are IL711/712 ( file attached ). In this case powering of the OSD is from a separate 5V regulator, thus alolowing 5V cameras to be used. Powering of the transmitter and the camera is performed as already described - either from 3S battery via the current sensor or via a 12V regulator if the main power is > 3S.
LC FILTERS USED
As the video transmitter is a noisy device, good LC filters should be used. Here are some good filter examples:
1. http://shop.righthere.nu/fpv-cameras-accessories/15a-25v-dc-power-line-filter-for-fpv-video-tx-rc-models/prod_1299.html
2. http://www.fpvmanuals.com/2011/06/17/new-lc-power-filter/
3. http://www.fpvflying.com/products/PSF-LC-power-filter.html
My favorite is number 1. This filter is based on Murata's BNX02 EMIFIL filter series ( a pdf file attached ).
DEAR FRIENDS, AS THIS PART IS THE MOST DIFFICULT PART, PLEASE PLACE HEREAFTER ALL YOUR COMMENTS AND REMARKS. THUS ALL WE COULD DISCOVER THE BEST DESIGN WE ASPIRE TO !
END OF PART SEVEN
nickarsow
Jun 19, 2012, 3:09:58 AM6/19/12
to uavde...@googlegroups.com
PART EIGHT
REDUCING WIRINGS
Hi guys,
This is the final part of my thread, but not less important.
The RC wiring is so complex as it contains a mix of many different devices - low current sensitive devices and high current ones.
The ideal solution all we wish is just a few wires.....so sorry to dissapoint you.....no way to make it. Hrereafter I'll describe how to reduse most of the wires. The main problem remains with the power wires as the right way is connecting them as near the current sensor as possible.
Please note I'll not discuss here the serial bussess as SBUS, LIN, etc.
A/ Receiver
Just use the PPM as it uses the standard 3 wire #AWG26 wires. This cable is enough for powering the receiver and the data. If the receiver doesn't have a PPM output, just use a reduced count harness - NxSignal+GND+PWR. Example - if you have 8 channels, the wire harness could contain 10 wires. This approach is applicable as the receiver draws very low current.
B/ Analog devices
Analog devices draw are low power devices and 3 wires are enough in most cases. Also a shielding is mandatory.
C/ Servos, servo power boards...everything related to servos
The best solution described is the optoisolated PPM->PWM servo board with onboard DC-DC BUCK onboard. Thus we have 2 wires for PPM signal and independent power supply. Thus we'll not disturb any devices.
D/ OSD
....very interesting.....What I think the OSD should look like - optoisolated towards the UDB and with it's own voltage regulator - either linear or DC-DC. The best design should have also an additional low power 5V LDO for 5V powered cameras.
E/ ESC
I think this part is clear enough - the best are OPTO ESCs.
F/ Power output devices
The best solution here is if they are optoisolated. Thus all you need are 2 wires #AWG26 between the device and the UDB and 2 with appropriate #AWG according the current drawn between the device and the current sensor.
G/ Digital input/output devices
Optoisolation here is easy up to 100MBd. If the devices are low current, one can power them directly from the UDB. In this case no optoisolation is needed. If they are high power devices, one should power them via a BEC/VR if there is no such onboard.
Attached pdf file describes most of the devices explained above.
END OF PART EIGHT
I invite all members to discuss this thread and to offer their views on different circuit solutions.
REDUCING WIRINGS
Hi guys,
This is the final part of my thread, but not less important.
The RC wiring is so complex as it contains a mix of many different devices - low current sensitive devices and high current ones.
The ideal solution all we wish is just a few wires.....so sorry to dissapoint you.....no way to make it. Hrereafter I'll describe how to reduse most of the wires. The main problem remains with the power wires as the right way is connecting them as near the current sensor as possible.
Please note I'll not discuss here the serial bussess as SBUS, LIN, etc.
A/ Receiver
Just use the PPM as it uses the standard 3 wire #AWG26 wires. This cable is enough for powering the receiver and the data. If the receiver doesn't have a PPM output, just use a reduced count harness - NxSignal+GND+PWR. Example - if you have 8 channels, the wire harness could contain 10 wires. This approach is applicable as the receiver draws very low current.
B/ Analog devices
Analog devices draw are low power devices and 3 wires are enough in most cases. Also a shielding is mandatory.
C/ Servos, servo power boards...everything related to servos
The best solution described is the optoisolated PPM->PWM servo board with onboard DC-DC BUCK onboard. Thus we have 2 wires for PPM signal and independent power supply. Thus we'll not disturb any devices.
D/ OSD
....very interesting.....What I think the OSD should look like - optoisolated towards the UDB and with it's own voltage regulator - either linear or DC-DC. The best design should have also an additional low power 5V LDO for 5V powered cameras.
E/ ESC
I think this part is clear enough - the best are OPTO ESCs.
F/ Power output devices
The best solution here is if they are optoisolated. Thus all you need are 2 wires #AWG26 between the device and the UDB and 2 with appropriate #AWG according the current drawn between the device and the current sensor.
G/ Digital input/output devices
Optoisolation here is easy up to 100MBd. If the devices are low current, one can power them directly from the UDB. In this case no optoisolation is needed. If they are high power devices, one should power them via a BEC/VR if there is no such onboard.
Attached pdf file describes most of the devices explained above.
END OF PART EIGHT
I invite all members to discuss this thread and to offer their views on different circuit solutions.
0 new messages