openXvario development

[rainer]

If you are testing like I saw in your video raising the hand up and down then you won't probably see something I see with my method. I got a well sealing plastic bottle and cut a hole in the lid for the wires and a small tube which are sealed in with hot glue. (BTW this would be a good way to connect to a TE probe as the bottle acts as a low pass filter) Or for a quick test you can grab the sensor in your closed fist and put side of fist to mouth. By drawing with my lungs I can simulate whatever climb rate including high rates. It seems when the climb rate exceeds the outer thresholds the audio goes wonky.

ok I changed my testing method
as i now have thepossibility in the vario to control a parameter from the transmitter, i added a debugging option to manually override the vertical speed using the transmitted ppm value.
so i can now define that my potentiometer on the transmitter directly set the vertical speed to any value inbetween 2 configurable limits. ( only for type=vario in open9x, as only that types can handle the direct transmitted vertical speed)
EDIT: i was wrong on that. the ALTI+ works the same way. it uses the direct transmission of vertival speed as well

i tried with -3,-0.1,+0.1,+3. it seems to work fine. inside of the limits i get a sound relative to the vertical speed, exceeding the limits it stays on the min/max value of the tone.
i tried some other settings like -17,-1,-1,17 => no problem as well. at least for type vario it seems to work fine.
i'll change my source again in order to directly change the altitude now and test with alti and alti+ as well.

rainer

[RightRudder]

if somebody with some more electronic knowledge than me comes up with an easy circuit to connect a balance connector to the pins of the arduino, the code to use it as a 6S cell monitor is already in there as well. if you want to test it, just install some pulldown resistors (e.g. 30k to GND ) to the pins below.
#define PIN_VoltageCell1 0 // Pin for measuring Voltage of cell 1 ( Analog In Pin! )
#define PIN_VoltageCell2 1 // Pin for measuring Voltage of cell 2 ( Analog In Pin! )
#define PIN_VoltageCell3 2 // Pin for measuring Voltage of cell 3 ( Analog In Pin! )
#define PIN_VoltageCell4 3 // Pin for measuring Voltage of cell 4 ( Analog In Pin! )
#define PIN_VoltageCell5 6 // Pin for measuring Voltage of cell 5 ( Analog In Pin! )
#define PIN_VoltageCell6 7 // Pin for measuring Voltage of cell 6 ( Analog In Pin! )
then you can put 0..5V (relative to GND!!) to those pins and get it displayed in open9x.
we just can´t connect the balance connector directly to the arduino as only the 1st cell has it´s voltage relativ to GND.

You could set up a series of voltage divider networks between each pin of the balance plug and each pin of the arduino. Each arduino pin gets a pull down resistor. If you choose high value resistors then the current drain on the battery is minimal but the plug would still have to be disconnected when not in use. Then a scaled resistor goes between each balance pin and each arduino pin. So on the first (ground end) cell you choose a voltage divider such that the cell voltage is scaled to the input range of the arduino. Call this ratio X. It could be 1 for the first cell. The second cell has a divider ratio of 2X and the third cell 3x and so on. Now you never have more than the single cell voltage on any pin. The second pin voltage represents (V1+V2) through the second pin's divider so you need the software to read V1 first and subtract it from twice the voltage measured on pin 2. Likewise once you know V1 and V2 you can calculate V3 by subtracting V1 and V2 from 3 times the voltage on pin 3. Precision resistors should be used but this could be a small board which goes between the balance plug and arduino. Since the scaling resistor values get progressively higher there would be a progressive error due to the sampling current into the arduino, so what you do is add a small chip capacitor in parallel with each shunt resistor. This capacitor charges up to the measurement voltage in between samples. When the sample happens the current required for proper sampling is drawn from the shunt cap for the few cycles when the sample is happening and then it recharges before the next sample period.

Linear Technology corp sells some devices like the LTC6802 which can connect to the balance plug and give you all the data through serial but they are big 44 pin chips, rather overkill for what we are talking about.

[RightRudder]

i tried with -3,-0.1,+0.1,+3. it seems to work fine. inside of the limits i get a sound relative to the vertical speed, exceeding the limits it stays on the min/max value of the tone. rainer

Maybe it has to do with the way the Frsky hub is assembling packets that gives me the strange audio effect?? I can't wait to get my openXvario going!

My sensors have shipped but not the arduino yet. Somebody is still on vacation over there......

Joe

[Rob Thomson]

China is closed till then end of the week


Sent from my GT-I9300 using Tapatalk 2

[rainer]

You could set up a series of voltage divider networks between each pin of the balance plug and each pin of the arduino. Each arduino pin gets a pull down resistor. If you choose high value resistors then the current drain on the battery is minimal but the plug would still have to be disconnected when not in use. Then a scaled resistor goes between each balance pin and each arduino pin. So on the first (ground end) cell you choose a voltage divider such that the cell voltage is scaled to the input range of the arduino. Call this ratio X. It could be 1 for the first cell. The second cell has a divider ratio of 2X and the third cell 3x and so on. Now you never have more than the single cell voltage on any pin. The second pin voltage represents (V1+V2) through the second pin's divider so you need the software to read V1 first and subtract it from twice the voltage measured on pin 2. Likewise once you know V1 and V2 you can calculate V3 by subtracting V1 and V2 from 3 times the voltage on pin 3. Precision resistors should be used but this could be a small board which goes between the balance plug and arduino. Since the scaling resistor values get progressively higher there would be a progressive error due to the sampling current into the arduino, so what you do is add a small chip capacitor in parallel with each shunt resistor. This capacitor charges up to the measurement voltage in between samples. When the sample happens the current required for proper sampling is drawn from the shunt cap for the few cycles when the sample is happening and then it recharges before the next sample period.

Linear Technology corp sells some devices like the LTC6802 which can connect to the balance plug and give you all the data through serial but they are big 44 pin chips, rather overkill for what we are talking about.

Yes i thought about some resistors as well. but then again if we need to build something it would either be big and quite some work to build or it needs it own pcb and smd components, but let´s really try to keep this as simple as possible to build. if possible: use ready build cheap modules and put them together. let the software do the rest. this way almost everybody would be able to build the whole thing.
the solution that i have on my list is leraning to read the data IN port of the 2 wire smart port of the new frsky sensors. their lipo cell monitor is a nice 10-12€ multiplexer for cell voltages it´s not really worth spending the time and money to build something if a nice piece of tech is available... and being able to read that data would re-open the full world of the frsky sensors to our solution.
I succeeded to read the serial port of the frsky ampere meter ( standard RS232) , and that one has a data inport, so by being able to read that one we could allready read the cell monitor. but then again if we would just want the lipo monitor that wouldn´t work as of today. can´t be long till somebody finds a way of reading that protocoll and then we are going to be able get all other sensor data in.

[jhsa]

use ready build cheap modules and put them together. let the software do the rest. this way almost everybody would be able to build the whole thing.
their lipo cell monitor is a nice 10-12€ multiplexer for cell voltages it´s not really worth spending the time and money to build something if a nice piece of tech is available...

Would it be possible to have also the option available for those who would like to build their own voltage sensor? Schematic was posted somewhere in this forum

Would you still have the pins available as you had before on the arduino? Don't know if you also had some code for that??

Thanks,

João

[rainer]

use ready build cheap modules and put them together. let the software do the rest. this way almost everybody would be able to build the whole thing.
their lipo cell monitor is a nice 10-12€ multiplexer for cell voltages it´s not really worth spending the time and money to build something if a nice piece of tech is available...

Would it be possible to have also the option available for those who would like to build their own voltage sensor? Schematic was posted somewhere in this forum

Would you still have the pins available as you had before on the arduino? Don't know if you also had some code for that??

Thanks,

João

yes all the pins are still there. if you just uncomment the defines you can enable the required code to display the voltage on those analog pins on the transmitter. if you need a calculation that shouldn´t be a problem as well.

[RightRudder]

I guess I was assuming that once these prototypes are mature that we would be building our own board for the sensor and mpu and then the resistor networks could be on the board. Really though for 3 or 4 cells the circuit could be on a 1cm x1cm board for now inline between our prototype and the balance plug. I would just do it with a dremel tool to cut the copper. For an electric glider I don't need much more than a vario and battery monitor. In fact I was planning to just use A2 to monitor the total pack voltage and use some conservative limit. A cell monitor is only needed if you want to squeeze every drop safely out of the power source.

[rainer]

In the moment u i use the A1/2 to monitor total cell voltage as well. for my DLG i will try to use the internal voltage monitor of the openXvario.
For my electric glider i already have one of those small current monitor´s available. that can be connected to one of the analog input pins and won´t be a problem to implement. I just didn't get to it yet. if you want to give that resistor based 2-4s cell monitor a try using the simple resistor circuit , i can do the required code changes to do the calculations. we still have lots of flash left, so no problem to add options.
the layout files for the arduino pro mini are freely available online it shouldn't be a too big problem to add the ms5611 sensor , a 3.3v low drop regulator and some resistors. it´s just that most people do not have the tools / time / technique to build those small boards. look at the halcyon project. really nice stuff.. but just too few people build it as it requires too high skills to build.
Of course if there would be enough demand it could make sense to build a custom board and get it manufactured.

[RightRudder]

I made a post over on RCGroups. Dan has done exactly that but he still populated the board himself. Not a big deal really, so few parts. So I asked him if he would share his layout files or the board fab info. Generally they keep the info in case you want to re-order. We'll see what he says. I could do a layout but like you, time is the factor, I've got a lot going on already. I could certainly populate some boards though. Even for others if we can get the boards. They fit right inside the box for the frsky 6ch rx!

Joe

[jhsa]

Does the single cell monitor use only some resistors?? I saw a schematic that needed some op amps to isolate the different cells from each other..

Thanks

João

[RightRudder]

There are different ways this can be approached. A differential multiplexer is one solution, or FET switching of some sort etc but the resistor idea I proposed is a compromise biased in favour of simplicity and compactness. Not sure what you are referring to exactly with your question but a single cell because it is less than 5v can usually be measured without a divider. If you wan to measure more than that you need a divider so that you don't exceed the max permissible input voltage to your measurement device. If you use say a 10:1 divider you then have to 10x scale the result at the tx end of the telemetry to display the proper voltage. Not sure if that is the info you are looking for but I hope it helps.

[jhsa]

Here, this is what I meant.. I don't know hows we can isolate the individsual Lipo cells with resistors.

viewtopic.php?f=86&t=2780&p=37302&hilit=opamp#p37302

João

[rainer]

I received one more sensor today. its a GY-63 MS5611 5V module. It is the MS56110BA03 type ( the one with the stainless steel cap).
I'll build another prototype next and will take more photos, as this sensor's layout is the more common one. So, if you receive your modules, this will be like yours.

[MikeB]

Rainer: have you got the actual for this output documented? e.g. voltage to A1/A2 and the serial data id and values or do I need to go through the source code?

Thanks,

Mike.

[rainer]

Rainer: have you got the actual for this output documented? e.g. voltage to A1/A2 and the serial data id and values or do I need to go through the source code?

Thanks,

Mike.

Hi Mike,

here are the ID´s that i use:

The new ID for the vertical Speed:
#define FRSKY_USERDATA_VERT_SPEED 0x30 // open9x Vario Mode Only

A new way to write to the VFAS FIeld that bertrand implemented for us:
#define FRSKY_USERDATA_VFAS_NEW 0x39 // Sends internal read voltage to VFAS

Altitude in cm resolution
#define FRSKY_USERDATA_BARO_ALT_B 0x10
#define FRSKY_USERDATA_BARO_ALT_A 0x21

Temperature (normaly only t1, but i do send T2 for other configurable output)
#define FRSKY_USERDATA_TEMP1 0x02
#define FRSKY_USERDATA_TEMP2 0x05

RPM FIeld will be (mis-)used for other parameter output
#define FRSKY_USERDATA_RPM 0x03

The Analog In voltages can be written to the cell voltages using this ID
#define FRSKY_USERDATA_CELL_VOLT 0x06
#define FRSKY_USERDATA_CURRENT 0x28 (not yet)

There is another option to output an analog voltage representing the vertical speed which goes to either A1 or A2. This one is only used when there is no serial port for telemetry on the receiver. this pretty much behaves like the wssh vario then.

Oh, almost forgot:
the GPS Dist field: 0x3C I´m using this oone for various values as i am able tho write to it with relative high precision ( e.g. absolute height or debug parameter output )


I added another option to "force" open 9x to display the absolute altitude instead of the relative one by directly overwriting the internal offset. here is the code i use for that:
SendAlt(1); // send initial height
SendValue(0x00,int16_t(1)); //>> overwrite alt offset in open 9x in order to start with display of absolute altitude...
SendValue(0x30,(int16_t)(alt/100)+2); //>> overwrite min alt in open 9x
SendValue(0x31,(int16_t)(alt/100)-2); //>> overwrite max alt in open 9x
mySerial.write(0x5E); // End of Frame 1!

if you have any other question, just ask in this post or via PM

[RightRudder]

Here, this is what I meant.. I don't know hows we can isolate the individsual Lipo cells with resistors.

viewtopic.php?f=86&t=2780&p=37302&hilit=opamp#p37302

João

I made a quick drawing of what I am talking about. The plug on the right connects to the arduino AD inputs. I should have drawn the balance plug between the battery and divider chain because it needs to be disconnected when not in use. Here the battery is connected directly but you get the idea.

[rainer]

Hope it doesn´t get to confusing, but i am going to change the default pinouts (again) in this next prototype.

Pin 2 = PPM Signal to receiver ppm signal (in the servo plug)
Pin 3 = Analog vertical speed signal to receiver A1/A2
Pin 4 = RS232 to receiver RX

by changing this, we achieve that no cables have to cross each other inside of the shrink tube, and it look a bit tidier

to those that perhaps already built it using the other pin assignment, no problem, all the pins are configurable in the config section of the code.

i´ll update the code to use these new values as the default with the next submit.

Rainer

[Flaps 30]

Sounds logical to alter the pinout. Most of us are still waiting for parts to arrive from China so no problem.

[jhsa]

Yeah, I think that you might be the only one that built it till now

[Rob Thomson]

I have everything... But the barometers!

I am away next week so hopefully they will have arrived by the time I get back

Sent from my GT-I9300 using Tapatalk 2

[rainer]

puhh... something must be wrong with the new sensor.....
... its way more sensitive

[BBvideo 425,350]http://www.youtube.com/watch?v=5uz_x-nj ... e=youtu.be[/BBvideo]
don't just listen to the tone... it doesn´t start before reaching 10cm/s

edit: i compared again.. its not better. the other sensor is not different in the current version of the code.
i asked bertrand if it would be possible to change the audio tone to a finer stepping, The current +-10cm is to course for use
i changed the led on the oxv to light up when reaching a vertical speed of +5cm(!) and that seems to be working good now.

[rainer]

I have everything... But the barometers!

I am away next week so hopefully they will have arrived by the time I get back

Sent from my GT-I9300 using Tapatalk 2

fingers crossed.....
btw; there is at least one UK based seller on ebay. i got 2 sensors from him ( the CJMCU-5611) for 11.9 GBP they work fine, they are just not the "common type". shipping was really from UK.

[RightRudder]

I am looking at putting an Frsky Rx in my Spectre (HK version of Passer) electric glider. There is not much room and I don't think I can fit a D8R rx. I have a D4R-II but that means I would loose a channel (need 5) I could Y the v-tail servos and use it with Ail/ele only with the 4ch rx, that's one option. I've been looking all over to find out if the D6R-II has the rs232 input on the telemetry but can't find it even on the frsky manual for the Rx. I see that it has the A2 input but it doesn't show if the serial is there. Does anybode here know for certain? I prefer to use the serial data rather than connect openXvario to the A2 input.

Joe

[jhsa]

No, The D6FR doesn't have the serial port only internal A1 and external A2.
Some of us already asked on the rcg frsky forum if they could build these receivers with serial..
The problem is not frsky. some other guys start saying, "oh, you should really use the 8 channel, why should frsky do that? bla, bla.. bla bla bla.." you know the score.. But let's see what they do..

João

[Kilrah]

If you remove the case of a D8R and shrink-wrap it instead you'll save quite a lot of volume. It's not really much bigger than the D6FR.

[RightRudder]

Darn! Well thanks for the quick answer guys. I'll get the D8R and see if I can make it fit. My other option is always to build a 4m glider

[RightRudder]

edit: i compared again.. its not better. the other sensor is not different in the current version of the code.
i asked bertrand if it would be possible to change the audio tone to a finer stepping, The current +-10cm is to course for use
i changed the led on the oxv to light up when reaching a vertical speed of +5cm(!) and that seems to be working good now.

Tha Kalman filter is doing its majik! That's awesome Rainer. I can't wait till my parts arrive.

No response from the guy on RCGroups who made his own board. I guess I will do this myself then. I have one board design finished and another I have yet to start but when I get that done I should add this and get them all made at the same time. By that time we'll have a better idea of any extra options to put in the design like the resistor divider chain, serial in/out for the new Frsky protocol, any level shifters, inverters etc.

Joe

[rainer]

edit: i compared again.. its not better. the other sensor is not different in the current version of the code.
i asked bertrand if it would be possible to change the audio tone to a finer stepping, The current +-10cm is to course for use
i changed the led on the oxv to light up when reaching a vertical speed of +5cm(!) and that seems to be working good now.

Tha Kalman filter is doing its majik! That's awesome Rainer. I can't wait till my parts arrive.

No response from the guy on RCGroups who made his own board. I guess I will do this myself then. I have one board design finished and another I have yet to start but when I get that done I should add this and get them all made at the same time. By that time we'll have a better idea of any extra options to put in the design like the resistor divider chain, serial in/out for the new Frsky protocol, any level shifters, inverters etc.

Joe

Count me in for a couple of boards
I didn´t do a board design yet on my own, but i was already tempted to start with this project. We´ll just have to define all the options that we want. And we´ll have to make it in a way that it is compatible. SO that it will still work without doing your own PCB, but that won´t be a problem.
So if you need any help with the boards, i´m more than happy to help where i can.

I Just commited another version of the source. The latest changes are the full support of the "remote control" option for the vario sensitvity.
I decided to actually Enable it as the default. If you do not want to use a receiver channel for this, just comment out the following line in the config section:

#define PIN_PPM 2

There are a couple of other bits that i mainly added for the next step:

Tomorow i will send out one of these little toys to Ingo for a test in a pressure chamber... so hopefully we will get some nice test results in the near future. Thanks for that Ingo!

The compile options page in the wiki still has to be updated, so please look at the code comments in the top part of the code for now.

Rainer.

[Flaps 30]

I Just commited another version of the source. The latest changes are the full support of the "remote control" option for the vario sensitvity.
I decided to actually Enable it as the default. If you do not want to use a receiver channel for this, just comment out the following line in the config section:

#define PIN_PPM 2

There are a couple of other bits that i mainly added for the next step:.

I take it that this change applies to REV 97 of the code?

To disable remote option I take it that all we have to do is this - //#define PIN_PPM 2

Sorry for the silly questions. New to this aspect of things.