A Look at Two Different Low-Power RF Switches.
Omron G5V-1 (left) and HMC349 (right).
Small signal relays, like the G5V-1 shown above, are frequently used to change RF paths in home-built amateur radio equipment. You could use two of them to switch an IF strip (with embedded filter, perhaps) from receive to transmit. (I used these in the IF strip of my 6m transceiver.) These devices are cheap (~$2.00 from the usual sources), switch relatively quickly (~5 ms), and aren't too power-hungry. Coil power consumption is rated at 150 mw. Of course, you have to supply your own board and connectors.
I've used 3253-type bus switches for similar low-power RF switching tasks with acceptable results, but wondered if dedicated RF switches (like the HMC349 shown above) could work acceptably well for ham radio-type tasks. I found these breakout boards widely available on the usual on-line auction and e-commerce sites for prices ranging from $10 - $20+. Not too salty and you don't have to make a board and supply SMA connectors. Switching levels are standard TTL at around 3 ma. and these devices do switch quickly! Total supply current at 5v VDD is 1ma to 3.5 ma. The big question is how these puppies will perform at HF as far as isolation and return loss, since the data sheet doesn't list values for signals below 100 MHz.
So the overall goal, here, is to see how these two technologies compare in the real world as far as (on-port) insertion loss and (off-port) isolation. I two-port calibrated my NanoVNA (oh no!) for a 1 MHz to 100 MHz sweep; the two small SMA jumpers I needed were included in the calibration. NanoVNA-saver was used to control and record all measurements. Return loss and insertion loss were measured for each switch type in both the "open" and "closed" states. Without further ado, results are depicted in the following graphs:
(G5V-1 closed)
(G5V-1 closed)
(G5V-1 open)
(G5V-1 open)
(HMC349 closed)
(HMC349 closed) 
(HMC349 open)
(HMC349 open)
The G5V-1 performed as expected, with very low insertion loss and isolation > 40 db under 50 MHz. The HMC349 exhibited isolation >60 db from 1 MHz to 100 MHz, but insertion loss really climbed below 35 MHz. The HMC349 did not behave as the non-reflective switch it's supposed to be.
So what's up with the 349? Could be a counterfeit chip (considering the source) or maybe something else. The insertion loss curve for the 349 in its "closed" state suggested that perhaps the blocking caps on the board might be too small (maybe in the 10 pF to 50 pF range), introducing a lot of unwanted reactance in the signal path. To test that, the existing caps were bridged with small 0.1 uF disc caps with very short leads (as short as possible) and the 349 re-evaluated as before.
Results follow:
(Modified HMC349 closed)
(Modified HMC349 closed)
(Modified HMC349 open)
(Modified HMC349 open)
Things are looking a bit better, now. The 349 still doesn't behave like a non-reflective switch, but return loss (for the "closed" state) was better (a more negative number means a better 50 ohm match) and isolation still ~60 db from 1 MHz to 100 MHz. Of course, the performance at > 100 MHz probably has been degraded significantly. I don't care - I originally bought two boards and modified one. So I have one for HF and one for VHF and above. Sometime a little experimentation is worthwhile and can make a possibly sketchy module work as desired. It just depends on your desire, skill, and willingness put cheap electronics in harms way.
Comments
Post a Comment