- Posts: 4402
Single-Board Universal Module
- PhracturedBlue
-
Topic Author
- Offline
Less
More
09 May 2014 22:58 - 13 May 2014 03:05 #23090
by PhracturedBlue
Single-Board Universal Module was created by PhracturedBlue
Over the past few weeks, I've talked to a few folks about designing a single-board Universal module. This would differ from the 'MultiMod' design in that it would be a single board with 4 transceivers populated on it.
It would be much smaller than an assembled MultiMod, and require signifcantly less power. It would also only have a single antenna, and would likely be directly swappable with the stock module in the Devo6/8/10/12 (sorry 7e folks, you'd need to do some soldering)
It sounds great, right? Well there are some downsides:
a) It is not something that can be hand assembled (well maybe by someone crazy, but not me). The chips are miniscule, and use leadless packages
b) Because it must be sold assembled, it will need to go through FCC compliance testing which is many thousands of dollars (assuming it passes the 1st time through)
c) development is costly due to the high cost of protocotypes. Mistakes are expensive.
That said, I've talked to some folks who have experience in designing RF boards, and getting them through FCC. I DO NOT at this time have any commitments that someone with experience will actually help, but there has been some interest.
So the 1st step was to get a feel for the design. To do that, I put together a circuit diagram. The board has a similar concept to the MultiMod board. I use a uC to act as a smart-switch todirect SPI signals to one of 4 transceivers. But whereas the MultiMod has 4 different boards attached to it, for this design, the 4 transceiver chips are directly mounted on a single PCB. The output of each transceiver is directed through a 4:1 multiplexer, and then into a single PA/LNA chip, finally terminating in a single U.FL antenna plug.
Because of the cost of FCC testing, it makes sense to build a multipurpose board. So instead of the ATTiny used in the MultiMod, I've chosen an STM32F0 uC. When used in a Devo Tx, the programming would be very similar to the MultiMod, and this would be hugely overkill (though it adds very little additional cost). However, this design allows for attaching an LCD screen and rotary-encoder to enable using the same board as a stand-alone PPM-in design (like the MagicCube or Anylink) that would work with any transmitter. In this configuration, the board would be packaged in its own enclosure. The Deviation protocol stack would be loaded onto the uC, and you would select a protocol via a knob on the enclosure.
While I've done a little work on this, I want to stress that there is no guarantee this will ever see production. I don't have any experience in RF design, and unless I can find someone who has done it, this won't get off the ground. But I'm putting my work so far out there. I know a few of you have some RF experience, and could possibly give some feedback. I will start work on the circuit board next, and we'll see how that goes. Note that I decided to use Eagle for this design. It should make it easier for others to contribute should they desire.
It also has virtually no bearing on the MultiMod work. I'm going to continue to refine the software on that with the hope we can get all known issues with it resolved.
BOM
It would be much smaller than an assembled MultiMod, and require signifcantly less power. It would also only have a single antenna, and would likely be directly swappable with the stock module in the Devo6/8/10/12 (sorry 7e folks, you'd need to do some soldering)
It sounds great, right? Well there are some downsides:
a) It is not something that can be hand assembled (well maybe by someone crazy, but not me). The chips are miniscule, and use leadless packages
b) Because it must be sold assembled, it will need to go through FCC compliance testing which is many thousands of dollars (assuming it passes the 1st time through)
c) development is costly due to the high cost of protocotypes. Mistakes are expensive.
That said, I've talked to some folks who have experience in designing RF boards, and getting them through FCC. I DO NOT at this time have any commitments that someone with experience will actually help, but there has been some interest.
So the 1st step was to get a feel for the design. To do that, I put together a circuit diagram. The board has a similar concept to the MultiMod board. I use a uC to act as a smart-switch todirect SPI signals to one of 4 transceivers. But whereas the MultiMod has 4 different boards attached to it, for this design, the 4 transceiver chips are directly mounted on a single PCB. The output of each transceiver is directed through a 4:1 multiplexer, and then into a single PA/LNA chip, finally terminating in a single U.FL antenna plug.
Because of the cost of FCC testing, it makes sense to build a multipurpose board. So instead of the ATTiny used in the MultiMod, I've chosen an STM32F0 uC. When used in a Devo Tx, the programming would be very similar to the MultiMod, and this would be hugely overkill (though it adds very little additional cost). However, this design allows for attaching an LCD screen and rotary-encoder to enable using the same board as a stand-alone PPM-in design (like the MagicCube or Anylink) that would work with any transmitter. In this configuration, the board would be packaged in its own enclosure. The Deviation protocol stack would be loaded onto the uC, and you would select a protocol via a knob on the enclosure.
While I've done a little work on this, I want to stress that there is no guarantee this will ever see production. I don't have any experience in RF design, and unless I can find someone who has done it, this won't get off the ground. But I'm putting my work so far out there. I know a few of you have some RF experience, and could possibly give some feedback. I will start work on the circuit board next, and we'll see how that goes. Note that I decided to use Eagle for this design. It should make it easier for others to contribute should they desire.
It also has virtually no bearing on the MultiMod work. I'm going to continue to refine the software on that with the hope we can get all known issues with it resolved.
BOM
Part Value Device Package Description DIGIKEY# MANF MANF# MOUSER# PROD_ID VALUE
C1 33nF C_0402 C0402 NON-POLARIZED CAP 490-6333-6-ND Murata GRM155R71C333KA01J 81-GRM155R71C333KA1J
C2 10nF C_0402 C0402 NON-POLARIZED CAP 490-1313-6-ND Murata GRM155R71C103KA01D 81-GRM155R71C103KA1D
C3 1nF C_0402 C0402 NON-POLARIZED CAP 490-3244-6-ND Murata GRM1555C1H102JA01D 81-GRM1555C1H102JA1D
C4 22pF C_0402 C0402 NON-POLARIZED CAP 490-8173-6-ND Murata GRM1555C1E220GA01D 81-GRM1555C1E220GA1D
C5 22pF C_0402 C0402 NON-POLARIZED CAP 490-8173-6-ND Murata GRM1555C1E220GA01D 81-GRM1555C1E220GA1D
C6 2.2nF C_0402 C0402 NON-POLARIZED CAP 490-6358-6-ND Murata GRM155R71H222JA01D 81-GRM155R71H222JA1D
C7 4.7pF C_0402 C0402 NON-POLARIZED CAP 490-6244-6-ND Murata GRM1555C1H4R7BA01D 81-GRM1555C1H4R7BA1D
C8 1.5pF C_0402 C0402 NON-POLARIZED CAP 490-6212-6-ND Murata GRM1555C1H1R5BA01D 81-GRM1555C1H1R5BA1D
C9 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-1-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C10 680pF C_0402 C0402 NON-POLARIZED CAP 490-3240-6-ND Murata GRM1555C1H681JA01D 81-GRM1555C1H681JA1D
C11 10nF C_0402 C0402 NON-POLARIZED CAP 490-1313-6-ND Murata GRM155R71C103KA01D 81-GRM155R71C103KA1D
C12 0.1uF C_0402 C0402 NON-POLARIZED CAP 490-1318-6-ND Murata GRM155R61A104KA01D 81-GRM155R61A104KA1D
C13 27pF C_0402 C0402 NON-POLARIZED CAP 490-6224-6-ND Murata GRM1555C1H270GA01D 81-GRM1555C1H270GA1D
C14 33pF C_0402 C0402 NON-POLARIZED CAP 490-6232-6-ND Murata GRM1555C1H330GA01D 81-GRM1555C1H330GA1D
C15 2.2uF C_0402 C0402 NON-POLARIZED CAP 399-7375-6-ND Kemet C0402C225M8PACTU 80-C0402C225M8P
C16 2.2uF C_0402 C0402 NON-POLARIZED CAP 399-7375-6-ND Kemet C0402C225M8PACTU 80-C0402C225M8P
C17 2.2uF C_0402 C0402 NON-POLARIZED CAP 399-7375-6-ND Kemet C0402C225M8PACTU 80-C0402C225M8P
C18 120pF C_0402 C0402 NON-POLARIZED CAP 490-8183-6-ND Murata GRM1555C1H121GA01D 81-GRM1555C1H121GA1D
C19 1.5nF C_0402 C0402 NON-POLARIZED CAP 490-6354-6-ND Murata GRM155R71H152KA01J 81-GRM155R71H152KA1J
C20 2pF C_0402 C0402 NON-POLARIZED CAP 490-6226-6-ND Murata GRM1555C1H2R0BA01D 81-GRM1555C1H2R0BA1D
C21 47pF C_0402 C0402 NON-POLARIZED CAP 490-8206-6-ND Murata GRM1555C1H470GA01D 81-GRM1555C1H470GA1D
C22 0.1uF C_0402 C0402 NON-POLARIZED CAP 490-1318-6-ND Murata GRM155R61A104KA01D 81-GRM155R61A104KA1D
C23 47pF C_0402 C0402 NON-POLARIZED CAP 490-8206-6-ND Murata GRM1555C1H470GA01D 81-GRM1555C1H470GA1D
C24 0.1uF C_0402 C0402 NON-POLARIZED CAP 490-1318-6-ND Murata GRM155R61A104KA01D 81-GRM155R61A104KA1D
C25 220pF C_0402 C0402 NON-POLARIZED CAP 490-1293-6-ND Murata GRM1555C1H221JA01D 81-GRM1555C1H221JA1D
C26 0.1uF C_0402 C0402 NON-POLARIZED CAP 490-1318-6-ND Murata GRM155R61A104KA01D 81-GRM155R61A104KA1D
C27 0.1uF C_0402 C0402 NON-POLARIZED CAP 490-1318-6-ND Murata GRM155R61A104KA01D 81-GRM155R61A104KA1D
C28 220pF C_0402 C0402 NON-POLARIZED CAP 490-1293-6-ND Murata GRM1555C1H221JA01D 81-GRM1555C1H221JA1D
C29 15pF C_0402 C0402 NON-POLARIZED CAP 490-6200-6-ND Murata GRM1555C1H150FA01D 81-GRM1555C1H150FA1D
C30 15pF C_0402 C0402 NON-POLARIZED CAP 490-6200-6-ND Murata GRM1555C1H150FA01D 81-GRM1555C1H150FA1D
C31 220pF C_0402 C0402 NON-POLARIZED CAP 490-1293-6-ND Murata GRM1555C1H221JA01D 81-GRM1555C1H221JA1D
C32 0.47uF C_0402 C0402 NON-POLARIZED CAP 490-3266-6-ND Murata GRM155R60J474KE19D 81-GRM155R60J474KE9D
C33 22pF C_0402 C0402 NON-POLARIZED CAP 490-8173-6-ND Murata GRM1555C1E220GA01D 81-GRM1555C1E220GA1D
C34 2.2uF C_0402 C0402 NON-POLARIZED CAP 399-7375-6-ND Kemet C0402C225M8PACTU 80-C0402C225M8P
C35 10pF C_0402 C0402 NON-POLARIZED CAP 490-6186-6-ND Murata GRM1555C1H100FA01D 81-GRM1555C1H100FA1D
C36 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-6-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C37 100pF C_0402 C0402 NON-POLARIZED CAP 490-6188-6-ND Murata GRM1555C1H101GA01D 81-GRM1555C1H101GA1D
C38 NC C_0402 C0402 NON-POLARIZED CAP
C39 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-6-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C40 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-6-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C41 100pF C_0402 C0402 NON-POLARIZED CAP 490-6188-6-ND Murata GRM1555C1H101GA01D 81-GRM1555C1H101GA1D
C42 100pF C_0402 C0402 NON-POLARIZED CAP 490-6188-6-ND Murata GRM1555C1H101GA01D 81-GRM1555C1H101GA1D
C43 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-6-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C44 1.0pF C_0402 C0402 NON-POLARIZED CAP 490-6206-6-ND Murata GRM1555C1H1R0BA01D 81-GRM1555C1H1R0BA1D
C45 1.8pF C_0402 C0402 NON-POLARIZED CAP 490-6215-6-ND Murata GRM1555C1H1R8BA01D 81-GRM1555C1H1R8BA1D
C46 1.5pF C_0402 C0402 NON-POLARIZED CAP 490-6212-6-ND Murata GRM1555C1H1R5BA01D 81-GRM1555C1H1R5BA1D
C47 15pF C_0402 C0402 NON-POLARIZED CAP 490-6200-6-ND Murata GRM1555C1H150FA01D 81-GRM1555C1H150FA1D
C48 2pF C_0402 C0402 NON-POLARIZED CAP 490-6226-6-ND Murata GRM1555C1H2R0BA01D 81-GRM1555C1H2R0BA1D
C49 1.5pF C_0402 C0402 NON-POLARIZED CAP 490-6212-6-ND Murata GRM1555C1H1R5BA01D 81-GRM1555C1H1R5BA1D
C50 2pF C_0402 C0402 NON-POLARIZED CAP 490-6226-6-ND Murata GRM1555C1H2R0BA01D 81-GRM1555C1H2R0BA1D
C51 1uF C_0805 C0805 NON-POLARIZED CAP 490-1695-6-ND Murata GRM21BR71A105KA01L 81-GRM40X105K10L
C52 10uF C_0805 C0805 NON-POLARIZED CAP 490-1709-6-ND Murata GRM21BR61A106KE19L 81-GRM21BR61A106KE19
C53 10uF C_0805 C0805 NON-POLARIZED CAP 490-1709-6-ND Murata GRM21BR61A106KE19L 81-GRM21BR61A106KE19
D1 LSM115J LSM115J DO214BA LSM115JE3/TR13DKR-ND Microsemi LSM115JE3/TR13
D2 LSM115J LSM115J DO214BA LSM115JE3/TR13DKR-ND Microsemi LSM115JE3/TR13
IC1 STM32-M-48 STM32-M-48 LQFP48 STM32F103 Medium Density 48pins 497-14645-ND STMicroelectronics STM32F072CBT6
IC2 3.3V LD1117SOT223 SOT223 Voltage Regulator LM1117 497-1241-1-ND STMicroelectronics LD1117S33CTR 511-LD1117S33C
J1 U.FL U.FL U.FL SMD antenna connector- WRL-09144 WM5587DKR-ND Molex 0734120110 CONN-09193 U.FL
JP1 2X5 2X5 2X5
JP2 JST-PH7 JST-PH7
JP3 JST-PH6 JST-PH6
JP4 JST-PH4 JST-PH4
JP5 JST-PH4 JST-PH4
JP6 USBSMD USBSMD USB-MINIB USB Connectors 670-1190-6-ND JAE DX2R005HN2E700 CONN-08193
L1 8.2nH L_0402 L0402 INDUCTOR PCD1909DKR-ND Panasonic ELJ-RF8N2ZFB 667-ELJ-RF8N2ZFB
L2 3.9nH L_0402 L0402 INDUCTOR PCD1905DKR-ND Panasonic ELJ-RF3N9ZFB 667-ELJ-RF3N9ZFB
L3 3.9nH L_0402 L0402 INDUCTOR PCD1905DKR-ND Panasonic ELJ-RF3N9ZFB 667-ELJ-RF3N9ZFB
L4 1.2nH L_0402 L0402 INDUCTOR PCD1899DKR-ND Panasonic ELJ-RF1N2ZFB 667-ELJ-RF1N2ZFB
L5 4.7nH L_0402 L0402 INDUCTOR PCD1906DKR-ND Panasonic ELJ-RF4N7ZFB 667-ELJ-RF4N7ZFB
L6 22nH L_0402 L0402 INDUCTOR PCD1914DKR-ND Panasonic ELJ-RF22NGFB 667-ELJ-RF22NGFB
L7 1.2nH L_0402 L0402 INDUCTOR PCD1899DKR-ND Panasonic ELJ-RF1N2ZFB 667-ELJ-RF1N2ZFB
L8 1.2nH L_0402 L0402 INDUCTOR PCD1899DKR-ND Panasonic ELJ-RF1N2ZFB 667-ELJ-RF1N2ZFB
L9 1.2nH L_0402 L0402 INDUCTOR PCD1899DKR-ND Panasonic ELJ-RF1N2ZFB 667-ELJ-RF1N2ZFB
L10 1.8nH L_0402 L0402 INDUCTOR PCD1901DKR-ND Panasonic ELJ-RF1N8ZFB 667-ELJ-RF1N8ZFB
L12 1.5nH L_0402 L0402 INDUCTOR PCD1900DKR-ND Panasonic ELJ-RF1N5ZFB 667-ELJ-RF1N5ZFB
L13 2.7nH L_0402 L0402 INDUCTOR PCD1903DKR-ND Panasonic ELJ-RF2N7ZFB 667-ELJ-RF2N7ZFB
R1 1M R_0402 R0402 RESISTOR P1.0MJDKR-ND Panasonic ERJ-2GEJ105X 667-ERJ-2GEJ105X
R2 22k R_0402 R0402 RESISTOR P22KJDKR-ND Panasonic ERJ-2GEJ223X 667-ERJ-2GEJ223X
R3 200 R_0402 R0402 RESISTOR P200JDKR-ND Panasonic ERJ-2GEJ201X 667-ERJ-2GEJ201X
R4 56.2k R_0402 R0402 RESISTOR P56.2KLDKR-ND Panasonic ERJ-2RKF5622X 667-ERJ-2RKF5622X
R5 10k R_0805 R0805 RESISTOR P10KADKR-ND Panasonic ERJ-6GEYJ103V 667-ERJ-6GEYJ103V
R6 NC R_0402 R0402 RESISTOR
R7 NC R_0805 R0805 RESISTOR
U$1 A7105 A7105 QFN20 Amiccom
U$2 NRF24L01 NRF24L01 QFN20 nRF24L01 1490-1033-6-ND Nordic NRF24L01P-R7
U$3 CC2500 CC2500 QFN20 296-19586-2-ND TI CC2500RTKR
U$4 CYRF6936 CYRF6936 QFN40-6X6 428-2962-ND Cypress CYRF6936-40LTXC
U$5 RFX2401C RFX2401C QFN16 RF Axis RFX2401C
U$6 SKY13384-350LF SKY13384-350LF QFN16 863-1456-6-ND Skyworks SKY13384-350LF
Y1 26MHz CRYSTAL5X3 CRYSTAL-SMD-5X3 Various standard crystals. Proven footprints. Spark Fun Electronics SKU : COM-00534 887-1716-6-ND TXC AB-26.000MAGE-T 717-AB-26.000MAGE-T
Y2 12MHz CRYSTAL5X3 CRYSTAL-SMD-5X3 Various standard crystals. Proven footprints. Spark Fun Electronics SKU : COM-00534 887-1701-6-ND TXC AB-12.000MALE-T 717-AB-12.000MALE-T
Y3 16MHz CRYSTAL5X3 CRYSTAL-SMD-5X3 Various standard crystals. Proven footprints. Spark Fun Electronics SKU : COM-00534 887-1707-6-ND TXC AB-16.000MAGE-T 717-AB-16.000MAGE-T
Y4 16MHz CRYSTAL5X3 CRYSTAL-SMD-5X3 Various standard crystals. Proven footprints. Spark Fun Electronics SKU : COM-00534 887-1707-6-ND TXC AB-16.000MAGE-T 717-AB-16.000MAGE-T
Last edit: 13 May 2014 03:05 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- Freezywilly
-
- Offline
Less
More
- Posts: 51
10 May 2014 00:30 #23093
by Freezywilly
Replied by Freezywilly on topic Single-Board Universal Module
PB I would like your permission to bring this topic up in professional engineering forums. There many talented people that do not either know about the RC hobby or just have other interests. This is important enough that getting some real experts in RF
could be so big none of us know just where it may go.
could be so big none of us know just where it may go.
Please Log in or Create an account to join the conversation.
- cmpang
-
- Offline
Less
More
- Posts: 296
10 May 2014 01:43 #23097
by cmpang
Replied by cmpang on topic Single-Board Universal Module
sounds great. How about a module that can be plug into FrSky Transis?
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
10 May 2014 02:53 #23098
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
FreezyWilly,
The hardware design will be entirely open-source. For various reasons, it is extraordinarily unlikely I will ever sell them myself, or try to make any profit on this design. My goal is to provide a robust design, and work with a manufacturer to build and sell the boards. I am open to any and all assistance I can get to make that happen.
cmpang,
the multimod board will (soon) work in the Taranis, but unfortunately it requires some soldering in tight spaces.
The design of this board could make it possible to plug into the Taranis with no hardware modifications. I don't know what sort of interface will be necessary for that to happen. My priority will likely be to make a drop-in module for the Devo 1st.
But again, I expect this to be a long process, and there is no guarantee it will ever come to fruition.
The hardware design will be entirely open-source. For various reasons, it is extraordinarily unlikely I will ever sell them myself, or try to make any profit on this design. My goal is to provide a robust design, and work with a manufacturer to build and sell the boards. I am open to any and all assistance I can get to make that happen.
cmpang,
the multimod board will (soon) work in the Taranis, but unfortunately it requires some soldering in tight spaces.
The design of this board could make it possible to plug into the Taranis with no hardware modifications. I don't know what sort of interface will be necessary for that to happen. My priority will likely be to make a drop-in module for the Devo 1st.
But again, I expect this to be a long process, and there is no guarantee it will ever come to fruition.
Please Log in or Create an account to join the conversation.
- Fixy
-
- Offline
Less
More
- Posts: 26
10 May 2014 04:10 #23099
by Fixy
Replied by Fixy on topic Single-Board Universal Module
This is a great idea many people would be interested. I think companies(and probably many users too) might be more interested in a JR type module rather than DIY one. Anyway a few comes to mind that might be interested.
Hobbyking/Turnigy/OrangeRX and the like are always interested in making easy money on open source projects.
FrSky is already in the open source hardware development so they might want to continue the adventure after the Taranis success.
Also HiSky seems to be very interested in the deviation project, they might want to get invested in this.
Good luck with this very nice project.
Hobbyking/Turnigy/OrangeRX and the like are always interested in making easy money on open source projects.
FrSky is already in the open source hardware development so they might want to continue the adventure after the Taranis success.
Also HiSky seems to be very interested in the deviation project, they might want to get invested in this.
Good luck with this very nice project.
Please Log in or Create an account to join the conversation.
- victzh
-
- Offline
Less
More
- Posts: 1386
10 May 2014 05:35 #23102
by victzh
Replied by victzh on topic Single-Board Universal Module
The existing design is slightly problematic from the point of view of nRF24L01 and several protocols running on it. The problem is in PA switch from transmit to receive. In this design it is controlled from the MCU, while there is an auto-acknowledgement mode in nRF which is used by some protocols, when the radio switches to receive to get an incoming packet automatically. To catch this exact moment and switch the PA fast enough from the MCU seems unfeasible. I am not aware of the situation with other chips, I don't know them enough, but if there is an auto-acknowledgement mode or its analogue there it will be problematic also.
Other than that I think your schematic is a very good start.
Other than that I think your schematic is a very good start.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
10 May 2014 12:47 - 10 May 2014 13:27 #23112
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
victzh I actually thought of that. It is why the VDD_PA signal is sent back to the MCU from the NRF24L01. My plan is to use that pin as an interrupt to switch the tx/rx enables. The ARM core will run at 48MHz, and has ~1instruction per cycle. It should be plenty fast enough to react to the Rx/Tx change. The other option was to use some logic gates to achieve the same behavior. I like this better, because it is overall simpler. We can also be smarter. All nrf24l01 designs I've seen always have either the PA or LNA enabled. with this setup we have the option to disable both when the chip is idle.
One thing I need to check is that 1.8V (the value on VDD_PA) is sufficient to trigger a logic change on the stm32. Otherwise I'll run it through an nfet to get a full-rail transition
Edit: I took a look, and I'm skeptical that 1.8V will flip the Shmitt trigger input, so I've rerouted VDD_PA to PA1 which has a comparator function with an adjustable internal reference
One thing I need to check is that 1.8V (the value on VDD_PA) is sufficient to trigger a logic change on the stm32. Otherwise I'll run it through an nfet to get a full-rail transition
Edit: I took a look, and I'm skeptical that 1.8V will flip the Shmitt trigger input, so I've rerouted VDD_PA to PA1 which has a comparator function with an adjustable internal reference
Last edit: 10 May 2014 13:27 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
10 May 2014 12:56 #23113
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
Fixy,
The JR modules are just PWM devices in a different package. It isn't really any different than an Anylink module which I discussed above. I agree that there are many more users who'd like such a thing, but as I've said in the past, it isn't really something I'm interested in. The problem is that many of the protocols have options that provide control of them, and it will require some sort of interface to provide access to those. It adds complication and expense.
So I'm focusing on radios that use Deviation.
The Taranis is a special case, because it is open source, we can (hopefully) use Deviation to run an I2S protocol over the PWM line and use the STM32 as a i2s to spi translator. Without understanding the technical side, it means that with some software changes, it should be possible to allow Deviation to use one of these boards plugged into the Taranis's module socket, and have full access to telemetry and all protocol controls.
The JR modules are just PWM devices in a different package. It isn't really any different than an Anylink module which I discussed above. I agree that there are many more users who'd like such a thing, but as I've said in the past, it isn't really something I'm interested in. The problem is that many of the protocols have options that provide control of them, and it will require some sort of interface to provide access to those. It adds complication and expense.
So I'm focusing on radios that use Deviation.
The Taranis is a special case, because it is open source, we can (hopefully) use Deviation to run an I2S protocol over the PWM line and use the STM32 as a i2s to spi translator. Without understanding the technical side, it means that with some software changes, it should be possible to allow Deviation to use one of these boards plugged into the Taranis's module socket, and have full access to telemetry and all protocol controls.
Please Log in or Create an account to join the conversation.
- Fixy
-
- Offline
Less
More
- Posts: 26
10 May 2014 17:56 #23117
by Fixy
Replied by Fixy on topic Single-Board Universal Module
Ok I think I understand. PWM is not enough to unlock the full potential of all protocols but on the taranis you can run I2S(the protocol taranis uses!?) through it which is then translated to SPI(the deviation protocol!?) 
it surely make it less interesting from a business point of view except for frsky.
So this means that the taranis won't run deviation but a tweaked opentx with a "deviation module" right?
Anyway taranis/devo users will love this, I wonder if frsky next gen tx will be compatible cause I'm not really interested in a taranis but the horus might be tempting if it ever gets released.
Out of curiosity is there any other open source radio this might be compatible with like the 9X/9XR series?
it surely make it less interesting from a business point of view except for frsky.So this means that the taranis won't run deviation but a tweaked opentx with a "deviation module" right?
Anyway taranis/devo users will love this, I wonder if frsky next gen tx will be compatible cause I'm not really interested in a taranis but the horus might be tempting if it ever gets released.
Out of curiosity is there any other open source radio this might be compatible with like the 9X/9XR series?
Please Log in or Create an account to join the conversation.
- victzh
-
- Offline
Less
More
- Posts: 1386
10 May 2014 19:07 #23118
by victzh
Replied by victzh on topic Single-Board Universal Module
PhracturedBlue, I noticed that circuit, but I am not sure when the VDD_PA becomes active - if at the end of transmission then you have 130 microseconds for PLL lock to switch the PA, but if after PLL lock then it may be very fast for interrupt handling. Unfortunately, datasheet is not clear about this. What I've seen in PA-enabled nRF implementations is that it is connected through low pass filter to TXEN of (typically) RFAXIS RFX2401c PA. May be it's possible to figure out it's time constant, otherwise I don't know how to reliably measure it.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
10 May 2014 20:41 - 10 May 2014 20:42 #23119
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I have an nrf24L01 with RFX2401C that uses some sort of switch to set the PA_EN/LNA_EN signals. It is labeled as 'V4' and in SOT23-6, but I have no idea how to figure out what it actually is. I don't know how fast it reacts. I'd be really surprised if they don't leave enough time to transition the amplifier, but figuring that out would be challenging. I really don't want to add the discrete components to bypass the uC control unless it is necessary.
One possibility would be to test it on an existing board by cutting the enable traces to the PA, and routing them (and VDDPA) through a STM32 board, and see if the ACK can still be detected.
That shouldn't be too hard to do, and is probably a good mitigation plan.
One possibility would be to test it on an existing board by cutting the enable traces to the PA, and routing them (and VDDPA) through a STM32 board, and see if the ACK can still be detected.
That shouldn't be too hard to do, and is probably a good mitigation plan.
Last edit: 10 May 2014 20:42 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- victzh
-
- Offline
Less
More
- Posts: 1386
10 May 2014 22:23 #23123
by victzh
Replied by victzh on topic Single-Board Universal Module
This 'V4' chip is not always present - RFX2401C does not need it, it interprets RXEN and TXEN both active as TXEN, so RXEN is always set to Vdd. They used the chip as either a precaution, or as a diversity guarantee that they can use some other PA which does not have this behavior.
Cutting it and passing through the board is a good idea, though soldering is not very easy.
Cutting it and passing through the board is a good idea, though soldering is not very easy.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
10 May 2014 22:35 #23124
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I know it isn't needed, but I thought if I knew what it was, it could provide another data point on switching speed. I've been looking around, and while most PA/LNA are 16-pin QFN packages, they don't seem to use a common pinout. At least of the ones I've looked at so far, every one has unique I/O locations. It is a good point on the rfx2401c not needing PA_EN and LAN_EN to be mutex though. If we end up using some boolean logic to control the PA when using the nrf24L01, it would simplify things a bit.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
11 May 2014 03:38 #23125
by PhracturedBlue
For the Devo, it makes sense to Hook BOOT0 to the USB VBUS. I believe doing so would allow programming without any jumpers. However, it may make it harder to use in PWM mode where it may be necessary to be able to access a filesystem for model configuration. for that you'd need to 1st power on the module and THEN connect the USB cable (like the Taranis)
I was thinning about the 'PWM' variant, and am now thinking that a bluetooth module may be the best option. It could provide a full interface without the complication of a a screen interface. The downside is, of course, that you'd need to have a smartphone to configure the module (and probably switch protocols)
Replied by PhracturedBlue on topic Single-Board Universal Module
I think I'll just pull off the resistor (or the V4) on one of my boards. That will leave me with nice (but small) pads to attach to. I also ordered a F072 discovery kit (since they are pretty cheap) so I can play with programming it. I want to make sure I know how all pins need to be tied off to properly initialize.victzh wrote: This 'V4' chip is not always present - RFX2401C does not need it, it interprets RXEN and TXEN both active as TXEN, so RXEN is always set to Vdd. They used the chip as either a precaution, or as a diversity guarantee that they can use some other PA which does not have this behavior.
Cutting it and passing through the board is a good idea, though soldering is not very easy.
For the Devo, it makes sense to Hook BOOT0 to the USB VBUS. I believe doing so would allow programming without any jumpers. However, it may make it harder to use in PWM mode where it may be necessary to be able to access a filesystem for model configuration. for that you'd need to 1st power on the module and THEN connect the USB cable (like the Taranis)
I was thinning about the 'PWM' variant, and am now thinking that a bluetooth module may be the best option. It could provide a full interface without the complication of a a screen interface. The downside is, of course, that you'd need to have a smartphone to configure the module (and probably switch protocols)
Please Log in or Create an account to join the conversation.
- victzh
-
- Offline
Less
More
- Posts: 1386
11 May 2014 03:49 #23126
by victzh
Replied by victzh on topic Single-Board Universal Module
Choices, choices. Bluetooth is probably better than separate screen. Who doesn't have a smartphone these days? Did you program for Android or iPhone? I know that interface should not be that complex as for Deviation - you probably don't need any curves, all that fancy stuff, you have a host TX for this. But anyway, it should be coded 
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
11 May 2014 04:21 - 13 May 2014 03:07 #23128
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I've done a little work for Android. I have never even attempted iPhone (and don't have one to test with).
Ihave gone through the schematics again and fixed a few issues, as well as verifying all the packages are right (I still need to check all of the RF packages since apparently not all QFN20 packages are the same size).
The BOM is pretty long, but I don't think the parts costs will be too high.
Edit: See 1st post
Ihave gone through the schematics again and fixed a few issues, as well as verifying all the packages are right (I still need to check all of the RF packages since apparently not all QFN20 packages are the same size).
The BOM is pretty long, but I don't think the parts costs will be too high.
Edit: See 1st post
Last edit: 13 May 2014 03:07 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- victzh
-
- Offline
Less
More
- Posts: 1386
11 May 2014 05:56 #23131
by victzh
Replied by victzh on topic Single-Board Universal Module
Wow! 84 small parts! I would not like to put so many of them to the board manually. It will take ages! I put around 40 parts sometime ago, I cursed the design all the way.
I talked with MassiveOverkill sometime ago - he has friends, RF engineers at a factory, he was considering them for the RF design if the need arises. They can probably correct the RF part of the design, do the matching, prototype it. May be it make sense to talk to him about it.
I talked with MassiveOverkill sometime ago - he has friends, RF engineers at a factory, he was considering them for the RF design if the need arises. They can probably correct the RF part of the design, do the matching, prototype it. May be it make sense to talk to him about it.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
13 May 2014 03:38 #23169
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I've spent some more time cleaning up the schematic, and updated the 1st post. I've also detailed the BOM there.
Pricing of the BOM for a single unit is ~$40. That would go way down with even a small amount of volume, but still not cheap.
I've updated the schematic to support a Bluetooth adapter, and power over USB.
This should allow programming without setting any jumpers.
I'm about ready to start work on the circuit board now. We'll see how it goes.
Pricing of the BOM for a single unit is ~$40. That would go way down with even a small amount of volume, but still not cheap.
I've updated the schematic to support a Bluetooth adapter, and power over USB.
This should allow programming without setting any jumpers.
I'm about ready to start work on the circuit board now. We'll see how it goes.
Please Log in or Create an account to join the conversation.
- SeByDocKy
-
- Offline
Less
More
- Posts: 1016
13 May 2014 05:18 #23171
by SeByDocKy
Around 40 USD, it's still very interesting ... coz actually with MM I am not sure that BOM is much lower than 40 USD when you add extra pins headers, extra uFL cables, wires ... it should be close to 35 USD. By the way, If I can order one of this single board, I will do ...
Replied by SeByDocKy on topic Single-Board Universal Module
PhracturedBlue wrote: I've spent some more time cleaning up the schematic, and updated the 1st post. I've also detailed the BOM there.
Pricing of the BOM for a single unit is ~$40. That would go way down with even a small amount of volume, but still not cheap.
I've updated the schematic to support a Bluetooth adapter, and power over USB.
This should allow programming without setting any jumpers.
I'm about ready to start work on the circuit board now. We'll see how it goes.
Around 40 USD, it's still very interesting ... coz actually with MM I am not sure that BOM is much lower than 40 USD when you add extra pins headers, extra uFL cables, wires ... it should be close to 35 USD. By the way, If I can order one of this single board, I will do ...
Please Log in or Create an account to join the conversation.
- Mmon77
-
- Offline
Less
More
- Posts: 14
13 May 2014 07:16 #23177
by Mmon77
Yes, but to sell the assembled boards, you would also have to add the cost of assembly and FCC certification (at least for sale in the US). So that would change the price accordingly. Unless of course you have the means to assemble one yourself.
It is still an interesting proposition and hopefully it goes somewhere.
Replied by Mmon77 on topic Single-Board Universal Module
Around 40 USD, it's still very interesting ... coz actually with MM I am not sure that BOM is much lower than 40 USD when you add extra pins headers, extra uFL cables, wires ... it should be close to 35 USD. By the way, If I can order one of this single board, I will do ...
Yes, but to sell the assembled boards, you would also have to add the cost of assembly and FCC certification (at least for sale in the US). So that would change the price accordingly. Unless of course you have the means to assemble one yourself.
It is still an interesting proposition and hopefully it goes somewhere.
Please Log in or Create an account to join the conversation.
Time to create page: 0.096 seconds
-
Home
-
Forum
-
Development
-
Development
- Single-Board Universal Module