TCXO on Chinese MMDVM boards
Posted: Mon Feb 13, 2023 3:53 pm
I recently bought a cheap Chinese MMDVM board and struggled to get it up and running.
By a process of elimination, I was able to determine that the board was significantly off frequency, so far off that it was not possible to transmit or receive with it on the indicated channel. The method to find the actual frequency was to change reflectors/unlink them on D-Star and observe the transmission on the spectrum display of my Icom IC-9700.
The frequency seemed to jump around, adding some soldering flux and re-flowing the solder under the TCXO "chip" seems to have cured the jumping in frequency, however it was still around 5 KHz off frequency. The TX and RX offset entries in Pi-Star would normally be around +/-200 or 500, in my case the offset needed was around 5000.
Having determined the offset and checked that it really did transmit on the indicated frequency courtesy of the Icom spectrum display, the board seemed to work reliably on D-Star. However it was intermittent on DMR, sometimes it worked perfectly, sometimes I could receive but not transmit, then not at all. By adjusting the final 70cms frequency up/down in 1 KHz increments, it appeared that the board was again off frequency by around 1 KHz. Back to tweaking the offsets... and once again the board works on DMR. On observing the transmitted frequency it was noticed that the oscillator is drifting up to 1 KHz from the indicated frequency as it warms up.
The bottom line is that a 22 GBP MMDVM board has a 10 pence TCXO that is nowhere near good enough to hold a frequency stable on 70 cms. Whether some are better than others is a matter of guesswork. Maybe the solution is to replace the TCXO with one of a higher grade, although if the board has an OLED display it means removing that to gain access.
Anyone else noticed similar problems with these boards?
By a process of elimination, I was able to determine that the board was significantly off frequency, so far off that it was not possible to transmit or receive with it on the indicated channel. The method to find the actual frequency was to change reflectors/unlink them on D-Star and observe the transmission on the spectrum display of my Icom IC-9700.
The frequency seemed to jump around, adding some soldering flux and re-flowing the solder under the TCXO "chip" seems to have cured the jumping in frequency, however it was still around 5 KHz off frequency. The TX and RX offset entries in Pi-Star would normally be around +/-200 or 500, in my case the offset needed was around 5000.
Having determined the offset and checked that it really did transmit on the indicated frequency courtesy of the Icom spectrum display, the board seemed to work reliably on D-Star. However it was intermittent on DMR, sometimes it worked perfectly, sometimes I could receive but not transmit, then not at all. By adjusting the final 70cms frequency up/down in 1 KHz increments, it appeared that the board was again off frequency by around 1 KHz. Back to tweaking the offsets... and once again the board works on DMR. On observing the transmitted frequency it was noticed that the oscillator is drifting up to 1 KHz from the indicated frequency as it warms up.
The bottom line is that a 22 GBP MMDVM board has a 10 pence TCXO that is nowhere near good enough to hold a frequency stable on 70 cms. Whether some are better than others is a matter of guesswork. Maybe the solution is to replace the TCXO with one of a higher grade, although if the board has an OLED display it means removing that to gain access.
Anyone else noticed similar problems with these boards?