toggle quoted messageShow quoted text
Thanks for the update Youssef. Did you have any chance to evaluate or gather first-hand information about the performance of R820T2 siblings( (R828D2, R836, R840
)mentioned by you?
On Thu, Mar 8, 2018 at 4:59 PM, prog <info@...>
On Thu, Mar 8, 2018 at 02:17 am, Siegfried Jackstien wrote:
and will be the only solution if chip will be obsolete
Not quite true. The problem with these tuners is that they are designed to be just good enough for 8 bit dongles, but when you plug them in front of a 12bit ADC you start seeing all sorts of artifacts. In that case the silicon variability becomes a bigger concern. An R820T2 can pass the QA tests in a 8bit dongle, but could still fail the tests in an Airspy.
ps i have some dongles with t2 tuner and also some with old fcd tuner ... so i have some spare
In our QA platform when a unit fails the spur, NF, or PLL lock tests, we systematically replace the tuner once or twice and it fixes the problem. This is more work for the production, but that's how we ensure everyone gets a good quality radio. People tend to overlook these details. "Transplanting" a tuner is an oversimplification of a more complex reality.
Now recently, one of the trusted Rafael distributors shipped us a reel of "last chance" tuners after a long lead time, we discovered that only 3 out of 10 tuners were good for a 12 bit SDR. Though, I believe they can still be OK for 8bit dongles due to the limited dynamic range and higher quantization noise of the RTL chips.
Now after returning the whole reel, Rafael Micro are pushing hard so we adopt one of the more recent siblings (R828D2, R836, R840).
I think that refreshing the design might be a good idea, after all.