bin size below 1000k gives unexpected output

[rvijgen]

I am trying to make a wide scan from 24M to 1700M using HackRF One by using: rx_power -f 24M:1700M:1000k -g 50 data.csv

This gives me a nice output like this. 2019-01-10, 15:18:26, 24000000, 25000000, 1000000.00, 1, 38.42, 38.42

When I try to increase the resolution by decreasing the bin size like this: rx_power -f 24M:1700M:500k -g 50 data.csv

I get this result: 2019-01-10, 15:23:52, 24000000, 26797996, 349749.50, 2048, -64.47, -inf, -inf, -inf, -inf, -inf, -inf, -inf, -inf

Even if I just decrease bin size by 1k like this: rx_power -f 24M:1700M:999k -g 50 data.csv

I get: 2019-01-10, 15:24:55, 24000000, 26797996, 699499.00, 4096, -inf, -inf, -inf, -inf, -inf

It seems like anything below 1000k gives this behaviour.

If I go slightly higher to 1001k, output seems fine

rx_power -f 24M:1700M:1001k -g 50 data.csv gives 2019-01-10, 15:29:29, 24000000, 25001000, 1001000.00, 1, 39.29, 39.29

I would expect I would be able to decrease the bin size and get a higher resolution output. Instead I get a lot of -inf values. What could be the cause of this?

[dernasherbrezon]

Having almost the same issue. Values are changing from "-4.06" to "-30.42". If I put 1001k, then values are smooth.