Mar/2014: Hier eine Anleitung von Marcus und Kollegen, EDGE 120HV und Microbeast, Futaba-Telemetrie. Das Problem war, dass Castle Link Live nicht funktionieren wollte, der EDGE den invertierten Gasimpuls nicht verstand, wenn der Impuls direkt aus dem Futaba-Empfänger (R7008) stammte. Daher verband man Empfänger und Beast per S.Bus und zweigte den Gasimpuls aus dem Beast ab, um ihn JLog zwecks Invertierung zuzuführen. Meine Vermutung ist, dass die Gasimpulsfrequenz aus dem 7008 zu hoch ist. Für Castle Link Live sollte sie Standard sein, 50Hz, maximal jedoch 100Hz. (German)
Important is that you understand how it works. Yes *rolleyes*, I’m sorry, you have to.
You need the appropriate firmware for your ESC, 4.02 or higher!Castle Link Live must be enabled in the setup of the ESC. You need to have the Castle Link USB interface for setup! Now the ESC expects the throttle pulse inverted, not reversed as you know it from your transmitter, long pulse -> short pulse, short pulse -> long pulse, than electrically inverted, LOW->HIGH, HIGH->LOW. No currently known (to me) receiver supporting that. – JLog is doing that inversion.
But there is a 2nd reason why the throttle pulse has to go via JLog, FBL system/receiver -> JLog -> ESC: 500 microseconds after the end of an inverted throttle pulse (rising edge) the ESC answers by a data pulse, that is a short negative spike. The time lag between the end of the throttle pulse and the falling edge of the data pulse, minus 500 microseconds, describes the value of the data item. Well, interpretation of that is different for each different type of data item. There’re 12 types. whereas the 12th data pulse is omitted, for synchronization purposes.
OK.. A throttle pulse can be theoretically up to 3,000 microseconds long. The maximum length of a data pulse is 5,000 microsends, plus 500 microseconds before the time lag counts. That is in summary 8,500 microseconds. Now let’s add some safe distance, so let’s say 10,000 microseconds. – Standard R/C pulse frequency is 50Hz, every 20 milliseconds a pulse, here throttle pulse. OK, that fits to the above. Summarized does that mean: Try to use your throttle source (receiver or FBL system) at the default pulse frequency of 50Hz (or 20ms pulse rate). At least do not exceed 100Hz, 10ms pulse rate!
OK.. but:
Castle says in their protocol description: rpm (motor) 100,000 at max. This is a 2-pole-normalized rpm! The maximum effective telemetry pulse distance is 5,000us, this corresponds to numerically (calculation factor) 10,208 == 102,080 rpm (2-pole) == 14,583 rpm at max for a 14-pole motor! Uhh.., no Sir! JLog accepts a pulse distance of up to 8572us which is 17,500 == 175,000 rpm (2-pole) and so 25,000 on a 14-pole. If we assume that a throttle pulse worst case can be 3,000us long, plus 500us security distance, plus 8,572us (see above) == 12,072us. That way we have a distance of ~8,000us from the telemetry pulse to the next throttle pulse. But only if the pulse frequency is standard, 50Hz == every 20,000us a pulse! – The conclusion is: Do not fall below a pulse rate of 13ms == do not exceed 77 Hz!
Concerning the red wire in the DIY “Y” or “Y+” cable: It connects (+) of the ESC to (+) of the receiver or FBL system. On the other hand this is the voltage supply of JLog, by the BEC or the receiver/FBL. If the ESC has a BEC then it may supply the remaining setup via that wire. Make sure to use wires of sufficient width for (+) AND Ground! If the ESC has no BEC and no opto coupler then its electronics will be supplied that way by the receiver/FBL. If the ESC is a HV version without BEC and with an opto coupler in the throttle line then connecting it to (+) has no function.
<– normal throttle pulse
inverted throttle + data pulses –>
(Figures from the public protocol description by Castle Creations.)
Well.., from the viewpoint of noise immunity this procedure is not unproblematic compared to “digitally coded” methods.. Especially ground loops may cause problems in your specific setup (-> wrong/zombie pulse reading).
.
Element of the data protocol but not yet fed, at least not by an “ICE”: Ubec and Ibec. (Possibly an “Edge” will send now Ubec into Link Live.)
Throttle is the pulse length in microseconds, not very informative for our purpose. So we show up with throttle twice in the log: Throttle pulse length and throttle relative (percent 0..100%) whereas the reference point is a free choice of mine (1100 microseconds==0%, so 1940==100%).
The ESC temperature comes also twice in log, in Celsius and Fahrenheit. There can two different types of temperature sensors be used in a CC, a linear or a non-linear one (NTC). The file “CCinfo.txt” on the SD tells us which type is existent in our ESC. I saw huge accuracy variations across different exemplars of ESCs (ICE).
mAh is calculated by JLog on base of Imot from the CC. There’re limits of accuracy: Calibration, temperature compensation? No data on Ibec.
All-in-all: Not that complete data set as with a JIVE for example but nice to have, especially for use by telemetry (currently 7 types of telemetry supported).
