Da ist offenbar jemand auf ein Problem mit einem Matek Rx24D-Empfänger gestossen und einen offenen Brief an Matek auf Discord gepostet:
- Open letter to @Matek_Sampson :
- Hi Matek, My name is Chris Seto (I'm Death By Hot Glue on the ExpressLRS discord). A few weeks ago, while flying my RX24D, I had an abrupt signal loss which caused a crash of my airplane. Doing some research, I found that the RX24D Xtal is far from the 52Mhz standard; in fact it's so bad that of the +/-200KHz width ELRS has to correct TX to RX mismatches, all of the silver button RX24Ds I tested were ~150Khz too slow (frequency correction ticks -82, -92, and even worse, for example). Frequency correction ticks can be reported over the "RSNR" telemetry field in the TX by flashing ELRS v3.0.0-RC1 with the DEBUG_FREQ_CORRECTION define to both TX and RX. This number ranges from -127 to +127, and each tick is 1.57Khz of RF frequency difference between TX and RX. The larger the number, the farther apart the TX and RX are. Past 127 in either direction, the link will be lost. Note that temperature causes this number to change, of course. If the RX is already using -97 ticks of this number, a slight temperature change is all it takes to cause the link to fail. This was the case with the crash I experienced. I calibrated my TX against my SDR, so I know the Xtal in my TX is operating at the proper reference frequency. The new gold button RX24Ds are better, but still not perfect. (I saw frequency correction ticks of around -30 on fresh gold button RX24Ds out of the bag). While this is far more in margin and much less likely to result in a sudden link loss, it still indicates matek has work to do to ensure their hardware is as compliant as possible with the established ELRS standards.
I believe Matek to be a premium manufacturer of FPV electronics, but I am very disappointed such premium hardware has this critical issue. At best, this means that any silver button RX may not bind out of the bag (I have seen this too with some RX24Ds I tested). At worst, it is a very real timebomb which could suddenly lose connection, as in my case. I believe at a minimum, Matek needs to re-test the current revision of the RX24D and make changes to reduce the Xtal error. ExpressLRS, as a free open source project, is only able to thrive if manufacturers commit themselves to providing high quality hardware to support it. I would like to buy Matek hardware in the future, but will be unable to purchase hardware myself, or encourage others to do so, if the community cannot be confident that this hardware is as compatible as possible with the software.
Meine Frage, ob gleiches Problem auch bei den anderen Matek-Empfängern erwartet werden kann, ist noch unbeantwortet.
Update: Es sollen offenbar nur ältere Matek-Empfänger betroffen sein, die neueren mit "goldenem" Bindeknopf sollen wohl okay sein.