Hacker News Reader

Comment by robertk | Hacker News Reader

I don't think the comparison on their site is fair. The Ettus B200 series uses an Analog Devices chip that has a 640MSPS sampling rate, the Limemicro chip can sample at 61.4MSPS at most. Even if the RF bandwidth is the same, the Limemicro chip will have poorer performance by perhaps an order of magnitude. So I'm not quite sure how it qualifies as next generation, even the 4 year old HackRF has better frequency specs.

View on HN

Are you sure you mean 640MSPS re. the Ettus B200, that sounds very high for an entry level SDR.

Looking at the Ettus spec page - https://www.ettus.com/product/details/UB200-KIT they quote 61.44MS/s.

Also the HackRf isn't full duplex and the ADCs sample rate is a lot lower however it does have a greater frequency range.

Throwaway to avoid giving too much away but I have used the AD9364 in a few point to point links at my current employer. Its ADC sampling rate is 640MSPS [1]. This oversampling is important, if not outright necessary for noise shaping, and appropriate baseband filtering. I am not quite sure how the LMS7002 chip can achieve a usable 56MHz channel bandwidth with only 61MSPS.

[1] https://wiki.analog.com/resources/eval/user-guides/ad-fmcomm...

That's very interesting I didn't realise that. So it's oversampling ~10x?

The datasheet for the LMS7002 chip seems to indicate its sample rate is 160 MHz? Which you can grab from:

http://www.limemicro.com/products/field-programmable-rf-ics-...

They are not using industry standard terms in the datasheet for some reason. I believe the 160MHz is the analog bandwidth of the device. There is no real-time bandwidth (RTBW/IBW) specification, which is the number the defines the widest channel bandwidth of a device. It is likely that the 60MSPS bandwidth can be shifted within that 160MHz analog bandwidth using the on-chip NCO to hop around.

Could you elaborate on why oversampling is important? With Nyquist I was generally of the impression that 2 x bandwidth should be just fine, but in practice oscilloscopes (just one example) seem to have five times or more sampling rate than bandwidth.

The Nyquist rate is the theoretical minimum that allows you to distinguish one frequency from another. Sampling a higher frequency will cause an alias component to appear. This means that in practice, there must be analog band-limiting before the signal is sampled to reject any out of band signals before they are aliased to an in-band frequency and cannot be removed. The closer the band limiting cutoff comes to the nyquist frequency the more extreme the roll-off requirements for this band-limiting. Oversampling relaxes the band-limiting requirements, and can also increase the bit depth at lower sample rates. See PDM microphones for an example where this is taken to the extreme.

As a happy owner of a HackRF, I can say that this is better than the HackRF in basically every way other than the 5GHz band coverage. They have the advantage of producing their own custom chip, unlike the HackRF. To offer this at the same price point* seems pretty "next-gen" to me.

*Kickstarted HackRF One was $275 with case, Lime is $300 without a case and $500 with so it's still a bit more expensive.

I'm pretty glad to have jumped on one the 199$ early birds... though I'm not entirely sure what I intend to use this for.

Reply on news.ycombinator.com