Just noticed Rob tested the SunSDR2 DX, with very poor results, way down the list at 68 dB or so. I am hoping something was wrong with his test methodology in this particular case. If the guys at EE were not involved with the testing, they should take a look.

Doug

Dear Doug!
Sherwood's reputation is questionable!
On alternative site you can check SUN products compared with some popular rigs.
Read this here: dc4ku.darc.de and dc4ku.darc.de/Dynamic-range_of_direct-sampling_receivers_DC4KU.pdf
Me personaly I sold K3, because it can not compete with SUN in CW.
In the tests for SDRs with main aplication on PC (which is doing basicly all the work with filtering and decoding), you have to know, that it is important if you are using I5, I7 or AMD Ryzen 9 or Ryzen 7 CPU and amount of RAM. Also GPU is important. And also ESDR2 or ESDR3 are not the same... There is no info what filter taps, AGC and ATT or preamps were used.
Better use your energy to have good antennas...Take Sherwood with big reserve and try to listen in some contest (this weekend is CQ WW 160 m CW contest) with SUN and some other RIG on switch. Comparing RIGs is not a rocket science.
But that is only my humble oppinion and my way of thinking... I will keep my SUN SDR2DX
SRI for my bad English; It is not my native language.

Doug, I think it may be useful to try a 'gut feel' test here...recreating a situation whereby you have 2 close stations, one of which is very strong, is not difficult. Even using a simple text like this makes me highly sceptical of Sherwood's -68dB measurement.
I've used my SunSDR2DX (along with V3 software since its early days) in all sorts of band conditions, and it feels to me like it is more likely in the Top 20. I really would take Sherwood's figures with a pinch of salt...
Regards
Paul M0CVX

S58WW wrote:
Pibo nailed it.

And I mentioned while back Sherwood doesn't have any concept in how to really setup a true SDR rig for performance.

One thing I think has happened here and I have said this a few times to Roman- the default filter taps are way to low also the TX taps as well.

My gut feeling- Rob- used out of the box default settings- on a so -so PC with most likely onboard graphics which will raise the CPU rate and can/cause other issues in overall performance.

While not so pleased with state of missing promised functions, I seriously doubt Rob ratings are close to accurate.

Having had a Flex series- I would gladly take the Sun 2DX over the 6400, the rig is much more capable of noise filtering, smother cw- better close in signal filtering, and better ImD, from firsthand experience!

again, my thoughts alongside Pibo

Poor setup of PC running the software- and lack of taking the time to choose anything outside the default parameters in the setup!

Yes, I understand all of the comments and generally agree (I am an EE who specializes in RF and microwave design). But:

1. Many people will read the comparisons and make purchasing decisions.
2. Other direct sampling receivers such as the Anan/Flex/IC7300/Yaesu FTDX10 perform much better in his testing. Why? Is there some reason the other direct sampling RXs perform better? Some test conditions that are wrong in this case? This type of measurement should not be dependent on the software version, v2 or v3.

I am not getting rid of my Sun either. Nor my Anan ... I like them both. And there are much more important things to consider when evaluating receivers. But, such a low dynamic range result should be explained per item 2 above.

The point is, EE should take his results seriously, and try to reproduce and understand them, and improve the results if possible either by test condition changes or design changes. Otherwise over time they will loose sales that they might otherwise enjoy.

Doug

Very interesting discussion; thanks for the referal to DC4KU's work, I learned something!

Quick question for CASDR re: default ESDR filter taps (incl. TX taps) being way too low - which values would you recommend?

73,
JS

W3UB wrote: Doug, it is not because "other" perform any better. It is because "Sherwood says" they are better"
"Sherwood's" approach looks suspicious and that is what Pibo (+1 to him of my support) means mentioning "Sherwood's reputation".

W3UB wrote: Sherwoord's whole concept (mixing analogue and ADC-based TX/RX into one testing) is wrong. Apples should be to apples.

W3UB wrote: Really? Same as to measure your car performace with 2L and 4L engine and then say "nothing has changed" ... in the rear doors.
Actually, to get proper measurements with v2 and v3, the reseacher - first and foremost - has to completely understand the difference between v2 and v3, otherwise the "measurements" will result in waste of time..

W3UB wrote: Are you proposing EE to educate "Sherwood" about ADC/DDC basics? Would be bloody waste of time (IMHO), but ... well, "Sherwood" may sit (as a student) in the classes, which EE are performing (as the professor) in the local Universities ...

Doug, most of EE customers prefer EE to continue working on the software and future hardware, rather than merry chase improvment of the "place" in self-proclamated "expert's" ratings.

And as for the measurements, we have enough educated enthusiasts in the community to perform the measuring.
V2 already completed, Pibo already have given you links.
V3 will be measured sooner or later, once final release of V3 will be out.

JS. wrote:
for me I have them maxed out, what you should look for is artifacts or glitching audio, dropouts, or breaks, and abnormalities that shouldn't be there.

I generally will MAX out a setting of taps or buffers, then slowly bring them down till I do not see or hear any artifact.

granted this will require a good computer system overall, not only in CPU and GPU but networking as well. as well as lots of patients and a good ear!
remember the computer is doing the filtering - it controls how smooth or sharp the filter skirts are and you need everything to be setup and working well.

having a fast CPU but so - so graphics isn't going to help (ESP onboard in the intel series), you need both alongside good networking.
Perhaps Sherwood used a crappy Wi-Fi connection, he has yet, and I suspect will never list the operating settings in the application he used along with his complete computer specs and network setup.

Perhaps he used a Fisher Price 2GB child's touchpad with a 300 Bps WiFi connection and some mickey mouse graphic processor. one will never know!