- Posts: 4402
Single-Board Universal Module
- PhracturedBlue
-
Topic Author
- Offline
Less
More
25 Mar 2015 03:44 - 25 Mar 2015 03:45 #30222
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I've got the UniversalTx board working in my Devo radio now, and it seems to be very reliable. It fits exactly as I'd expected (pull out the stock module and plug in the UTx, no soldering needed), it detects properly and can communicate with all modules fine. Power output is good for all modules except the CC2500 which is about 15dBm down from the other 3. I verified the PA is working properly, so I am not sure what is going on. and I have no idea how to debug it. I verified with the MultiMod that power output is good, so it doesn't seem to be a software issue.Maybe there is a manufacturing defect, I haven't rechecked all solder joints yet. Or maybe there is something wrong with the components or layout.
Last edit: 25 Mar 2015 03:45 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- JustenCase
-
- Offline
Less
More
- Posts: 13
25 Mar 2015 05:46 #30227
by JustenCase
Replied by JustenCase on topic Single-Board Universal Module
Simply amazing! I didn't realize it was going to be a "pull out CC2500, insert UTx". That's such a great concept. The work you and the others have done is super. I may quit troubleshoot my 3 multi-mods (each one gives me a different error), and wait for this awesome baby!
Please Log in or Create an account to join the conversation.
- Fernandez
-
- Offline
Less
More
- Posts: 983
25 Mar 2015 08:23 #30232
by Fernandez
Replied by Fernandez on topic Single-Board Universal Module
Congratulations, this is a major breaktrue!!!
What I am also wondering, do we need to put a "smd 2.4ghz bandpass filter".
I see some designs use it some don't. Like yeti and here another example with bandpass Flysky Gt3, photo I found on the web.
What I am also wondering, do we need to put a "smd 2.4ghz bandpass filter".
I see some designs use it some don't. Like yeti and here another example with bandpass Flysky Gt3, photo I found on the web.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 13:12 #30237
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I use an LC filter instead. a proper rf filter would likely be a bit better, but it won't explain such a big power loss (it would also go on the output, so it would affect all 4 modules evenly
Please Log in or Create an account to join the conversation.
- phantom8
-
- Offline
Less
More
- Posts: 109
25 Mar 2015 13:58 #30240
by phantom8
Replied by phantom8 on topic Single-Board Universal Module
Excellent work, PB! And thank you for all the hard work that you keep pumping out. I suppose plug n play is only for Devo 10. Does it use the stock Devo antenna? I currently have a Devo 7e, but I will buy a Devo 10 just for the UniversalTX when it's available.
Please Log in or Create an account to join the conversation.
- Fernandez
-
- Offline
Less
More
- Posts: 983
25 Mar 2015 13:59 #30241
by Fernandez
Replied by Fernandez on topic Single-Board Universal Module
Are you sure that you use sufficient decoupling capacitors directly to chip and gnd. Is the ground around chip a real ground, or act like an inductor? Use plenty of VIA and make solid ground at decoupling c's.
Did you measure that Vin voltage is indeed correct level at chip pin?
Did you not miss any aditional pin, so all the pins required are connected GNd and VCC, according datasheet?
RF lanes, are no longer than few mm and that those lanes cannot form a capacitive link with ground?
Could it be due a software mistake?
Or else a defect chip end stage or board with faulty soldering.
At high freq like 2.4ghz RF can do very strange things, so even if there is a short, or open, RF could still come out......
Did you measure that Vin voltage is indeed correct level at chip pin?
Did you not miss any aditional pin, so all the pins required are connected GNd and VCC, according datasheet?
RF lanes, are no longer than few mm and that those lanes cannot form a capacitive link with ground?
Could it be due a software mistake?
Or else a defect chip end stage or board with faulty soldering.
At high freq like 2.4ghz RF can do very strange things, so even if there is a short, or open, RF could still come out......
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 17:42 #30249
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
I think the design is solid in general. I followed the recommended circuit for the CC2500. I just noticed that on pins 4&5 i'm missing47pF cpas, so there may be a high-frequency ripple that may have an effect. If I was good at soldering, I could stack one on the existing 0.1uF cap and see if it helps.
Voltage is correct and nice and stable on the board.
It is very unlikely to be a software mistake, since the Devo is driving the CC2500 chip, and the same code works fine with an external module. I measured a 20dB gain across the PA, so I don't think there is a misconfiguration on the switch or the PA. I actually see the 15dBm difference at the input to the switch.
It could absolutely be a soldering defect. I'll go over it again with my microscope. I also have the 1st board, which has the same circuit, but with a defective switch). I can still measure the power at the switch input though.
There are a few other folks here who build the 0.9 boards, and maybe one of them can verify my findings too.
Voltage is correct and nice and stable on the board.
It is very unlikely to be a software mistake, since the Devo is driving the CC2500 chip, and the same code works fine with an external module. I measured a 20dB gain across the PA, so I don't think there is a misconfiguration on the switch or the PA. I actually see the 15dBm difference at the input to the switch.
It could absolutely be a soldering defect. I'll go over it again with my microscope. I also have the 1st board, which has the same circuit, but with a defective switch). I can still measure the power at the switch input though.
There are a few other folks here who build the 0.9 boards, and maybe one of them can verify my findings too.
Please Log in or Create an account to join the conversation.
- moss
-
- Offline
Less
More
- Posts: 40
25 Mar 2015 18:57 #30251
by moss
The schematic has 0.1uF caps on those pins, not very effective at 2.5GHz. These lines
are digital, and may not be too sensitive for that. Pick a good bypass caps with low impedance at the frequency of interest. The lower Z the better. A 47pF Murata an impedance of 7 Ohms, while a 8pF has 0.4Ohm at 2.5GHz. Other frequencies also need bypassing, but the lower ones are assisted by multiple bypass caps on the PCB.
LC131 (45°) and LC122 may have mutual coupling, how much and to what effect I don't
know.
The CC2500 eval board layout has a lot more bypass caps but is a 2-layer (65mil) board, necessitating better bypass caps including 2 added ones those on the digital supply lines.
Adding 8pF should bring down the 2.4GHz signal better than 0.1u or 47pF. If you cant piggyback that on the 0.1uF I would replace them with 8pF.
Replied by moss on topic Single-Board Universal Module
PhracturedBlue wrote: I think the design is solid in general. I followed the recommended circuit for the CC2500. I just noticed that on pins 4&5 I'm missing 47pF caps, so there may be a high-frequency ripple that may have an effect. If I was good at soldering, I could stack one on the existing 0.1uF cap and see if it helps.
Voltage is correct and nice and stable on the board.
It is very unlikely to be a software mistake, since the Devo is driving the CC2500 chip, and the same code works fine with an external module. I measured a 20dB gain across the PA, so I don't think there is a misconfiguration on the switch or the PA. I actually see the 15dBm difference at the input to the switch.
It could absolutely be a soldering defect. I'll go over it again with my microscope. I also have the 1st board, which has the same circuit, but with a defective switch). I can still measure the power at the switch input though.
There are a few other folks here who build the 0.9 boards, and maybe one of them can verify my findings too.
The schematic has 0.1uF caps on those pins, not very effective at 2.5GHz. These lines
are digital, and may not be too sensitive for that. Pick a good bypass caps with low impedance at the frequency of interest. The lower Z the better. A 47pF Murata an impedance of 7 Ohms, while a 8pF has 0.4Ohm at 2.5GHz. Other frequencies also need bypassing, but the lower ones are assisted by multiple bypass caps on the PCB.
LC131 (45°) and LC122 may have mutual coupling, how much and to what effect I don't
know.
The CC2500 eval board layout has a lot more bypass caps but is a 2-layer (65mil) board, necessitating better bypass caps including 2 added ones those on the digital supply lines.
Adding 8pF should bring down the 2.4GHz signal better than 0.1u or 47pF. If you cant piggyback that on the 0.1uF I would replace them with 8pF.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 20:05 #30253
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
i am only going by the cc2500em schematic which has the same number of bypass caps and recommends the 0.1uF and 47pF caps in parallel.
Please Log in or Create an account to join the conversation.
- Fernandez
-
- Offline
Less
More
- Posts: 983
25 Mar 2015 20:06 #30254
by Fernandez
Replied by Fernandez on topic Single-Board Universal Module
Decent 100p, as close as possible should do the trick to short the RF and caps are pretty good nowdays, but I think not critical from 47p to 1n should all do it.
Personally I would like the caps direct to gnd layer on component side. The grounds on both sides of the board to be "perforated" with loads of via's to link the grounds from the both layers, and form a solid induction free gnd.
But would also not surprice me if PB is right here and a small soldering fault is present.
Personally I would like the caps direct to gnd layer on component side. The grounds on both sides of the board to be "perforated" with loads of via's to link the grounds from the both layers, and form a solid induction free gnd.
But would also not surprice me if PB is right here and a small soldering fault is present.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 20:20 #30257
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
FYI, I was referring to the 2500EM RevA schematic
I found the latest spec here:
www.ti.com/tool/cc2500em_refdes
My scheamtic matches this design (only 0.1uF caps on the 4/5 pins) so it is likely that is what I used as the reference.
So while there may be an issue with the layout, the schematic itself should be ok.
I found the latest spec here:
www.ti.com/tool/cc2500em_refdes
My scheamtic matches this design (only 0.1uF caps on the 4/5 pins) so it is likely that is what I used as the reference.
So while there may be an issue with the layout, the schematic itself should be ok.
Please Log in or Create an account to join the conversation.
- Fernandez
-
- Offline
Less
More
- Posts: 983
25 Mar 2015 21:00 #30261
by Fernandez
Replied by Fernandez on topic Single-Board Universal Module
Electrically it is the same, however a DC ground is not the same as a true RF ground. On the reference pcb the Capacitors are connected to GND top layer, than bottom and top GND layers with many VIA connected.
But i think it is not your problem here but for final revision could be small point of improvement from RF perspective.
But i think it is not your problem here but for final revision could be small point of improvement from RF perspective.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 21:19 #30263
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
what version of the board are you looking at? The latest versions of my board have a full top pour and connect all cap grounds to the top layer. I want to make sure we're talking about the same thing.
Please Log in or Create an account to join the conversation.
- Fernandez
-
- Offline
Less
More
- Posts: 983
25 Mar 2015 21:30 #30265
by Fernandez
Replied by Fernandez on topic Single-Board Universal Module
Hmm ooops yes forget my comments.
It looks like the board as on page 15 update of 3 weeks ago is adressing my points toward gnd all look fine. I was indeed refering at the older pcb revision back in topic.
It looks like the board as on page 15 update of 3 weeks ago is adressing my points toward gnd all look fine. I was indeed refering at the older pcb revision back in topic.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
25 Mar 2015 21:52 - 25 Mar 2015 21:55 #30268
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
Well, I looked at the board and found one marginal solder joint on the RF path. The part didn't tombstone, but there didn't look to be any solder rising up on the side of the cap. I added a dab of solder, and my gain went up 10dBm!. So while it still roughly 2dBm down from the CC2500 module I have, it is now much closer to the other transceiver chips. I am not sure how to optimize the design further to maximize gain, but the performance is pretty reasonable now.
Edit: I have no idea how some of you guys do it. All I needed to do was add a little solder to one component. I ended up shorting 2 adjacent components to that one and then had to spend a lot longer unshorting them. I really don't like hand-working 0402 components.
Edit: I have no idea how some of you guys do it. All I needed to do was add a little solder to one component. I ended up shorting 2 adjacent components to that one and then had to spend a lot longer unshorting them. I really don't like hand-working 0402 components.
Last edit: 25 Mar 2015 21:55 by PhracturedBlue.
Please Log in or Create an account to join the conversation.
- octagon
-
- Offline
Less
More
- Posts: 58
26 Mar 2015 02:20 - 26 Mar 2015 02:21 #30279
by octagon
Replied by octagon on topic Single-Board Universal Module
Are you using chem-wick or some good quality desoldering braid of suitable width? 63/37 or SAC alloy? (I hope not the latter)
Working on 0402 parts is a bit of a pain.
Do you have a small heat gun or some hot air resoldering station?
Working on 0402 parts is a bit of a pain.
Do you have a small heat gun or some hot air resoldering station?
Last edit: 26 Mar 2015 02:21 by octagon.
Please Log in or Create an account to join the conversation.
- PhracturedBlue
-
- Offline
Less
More
- Posts: 4402
27 Mar 2015 13:19 #30326
by PhracturedBlue
Replied by PhracturedBlue on topic Single-Board Universal Module
Since the board seems to be working quite well in my Devo radio, I've moved on to PPM support. I have PPMIn working now, as well as bluetooth communication. So I'm working on the software to interface with my phone App. Once that is done I should be able to control any supported model from any transmitter.
I still need to spend some time understanding why the board doesn't boot reliably with the USB jumper installed. After that, I'll either spin a new board (if fixes are needed) or build a couple more from the pcbs I have. Then I need to start looking into manufacturing options.
I still need to spend some time understanding why the board doesn't boot reliably with the USB jumper installed. After that, I'll either spin a new board (if fixes are needed) or build a couple more from the pcbs I have. Then I need to start looking into manufacturing options.
Please Log in or Create an account to join the conversation.
- SadSack
-
- Offline
Less
More
- Posts: 317
27 Mar 2015 14:11 - 27 Mar 2015 14:12 #30328
by SadSack
Replied by SadSack on topic Single-Board Universal Module
Excellent work and saw android app...the HID option would that mean Pc transmitter?
Saw this
vpapanik.blogspot.gr/2012/11/low-budget-...-pick-place.html?m=1
Add witness board could be very handy for short run of pcb?
So want to build one...
Nice work and thanks.
Saw this
vpapanik.blogspot.gr/2012/11/low-budget-...-pick-place.html?m=1
Add witness board could be very handy for short run of pcb?
So want to build one...
Nice work and thanks.
Last edit: 27 Mar 2015 14:12 by SadSack.
Please Log in or Create an account to join the conversation.
- SadSack
-
- Offline
Less
More
- Posts: 317
27 Mar 2015 14:20 #30329
by SadSack
Replied by SadSack on topic Single-Board Universal Module
And Sky switch could be used or happy with other?
Please Log in or Create an account to join the conversation.
- moss
-
- Offline
Less
More
- Posts: 40
27 Mar 2015 14:43 #30330
by moss
Replied by moss on topic Single-Board Universal Module
If you are interested in producing prototypes a bit easier, or low volume production, you could search eBay for "Semi-Auto SMT Pick and Place Machine".
Please Log in or Create an account to join the conversation.
Time to create page: 0.115 seconds
-
Home
-
Forum
-
Development
-
Development
- Single-Board Universal Module