Author Topic: Real time RSSI measurement broken RFM69CW?  (Read 27419 times)

emjay

  • Full Member
  • ***
  • Posts: 119
  • Country: nl
Re: Real time RSSI measurement broken RFM69CW?
« Reply #60 on: January 27, 2016, 04:58:43 PM »
@WhiteHare,

Quote
I'm not sure how to check whether or not they (AGC, AFC) are disabled, because I don't see an explicit enabled/disabled bit for either one.

I think you have found already - for completeness AGC is controlled by RegLna (0x18), AFC by RegAfcFei (0x1E)

Quote
Not to overcomplicate the discussion, but does DC cancellation figure into this at all?

In brief, no.  ZeroIF is when the local oscillator is approximately the same frequency as the nominal transmit carrier frequency, the one that is waggled +/- fdev for the frequency shift keying.  The local oscillator Flocal is fed with the (amplified) input signal Fc into the mixer section to produce the classic sum and difference results, k x Flocal +/- Fc. 
Now some local oscillator leaks into the area around the ANT pin, so that appears to be input signal (self reception) at exacty Flocal. Plug that into the mixer equation and the result is something at zero frequency.  Huh?  What does that mean?
Well that is just a D.C. term.  There is no useful information content and it is not healthy for the linearity of the following stages, so a sharp roll-off filter is introduced just after the mixer to cut off DC - Fcutoff.
The default setting for Fcutoff can be trusted.




perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Real time RSSI measurement broken RFM69CW?
« Reply #61 on: January 27, 2016, 05:47:15 PM »
Which is also why I think adding a small offset for AFC when using low beta signals where most of the enegy is concentrated around the carrier frequency gives better sensitivity, there is an effective  'notch' in the filter due to the DC cancelling roll-off.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #62 on: January 28, 2016, 02:06:24 AM »
This post confirms a finding that Perky posted just recently:

If I setup to do an RxRestart after every measurement of RSSI, and if I record the time interval between RSSI measurements, an interesting pattern emerges:

Code: [Select]
---------------------------------------------------------------------
Radio Configuration:
====================
Receiver-Mode(RX)
Packet Mode, FSK modulation scheme
[ActualBitrate] 4799.760bps = nominal bitrate of 4800bps
[RxBw] 10416Hz
  [RxBwExp] B101 (Bits 0-2 of regRXBW) = 5
  [theRxBwMant] B10 (Bits 3-4 of regRXBW) -> [RxBwMant] 24
---------------------------------------------------------------------
0.0. RSSI=-116
1.0. RSSI=-116, elapsed=108uSec
2.0. RSSI=-116, elapsed=104uSec
3.0. RSSI=-116, elapsed=116uSec
4.0. RSSI=-116, elapsed=104uSec
5.0. RSSI=-116, elapsed=108uSec
6.0. RSSI=-115, elapsed=1752uSec
7.0. RSSI=-115, elapsed=112uSec
8.0. RSSI=-115, elapsed=108uSec
9.0. RSSI=-115, elapsed=108uSec
10.0. RSSI=-115, elapsed=104uSec
11.0. RSSI=-115, elapsed=108uSec
12.0. RSSI=-116, elapsed=1760uSec
13.0. RSSI=-116, elapsed=104uSec
14.0. RSSI=-116, elapsed=108uSec
15.0. RSSI=-116, elapsed=108uSec
16.0. RSSI=-116, elapsed=104uSec
17.0. RSSI=-116, elapsed=108uSec
18.0. RSSI=-115, elapsed=1760uSec
19.0. RSSI=-115, elapsed=104uSec
20.0. RSSI=-115, elapsed=108uSec
21.0. RSSI=-115, elapsed=104uSec
22.0. RSSI=-115, elapsed=108uSec
23.0. RSSI=-115, elapsed=108uSec
24.0. RSSI=-113, elapsed=1756uSec
25.0. RSSI=-113, elapsed=108uSec
26.0. RSSI=-113, elapsed=108uSec
27.0. RSSI=-113, elapsed=112uSec
28.0. RSSI=-113, elapsed=108uSec
29.0. RSSI=-113, elapsed=104uSec
30.0. RSSI=-116, elapsed=1752uSec
31.0. RSSI=-116, elapsed=116uSec
32.0. RSSI=-116, elapsed=104uSec
33.0. RSSI=-116, elapsed=108uSec
34.0. RSSI=-116, elapsed=108uSec
35.0. RSSI=-116, elapsed=104uSec
36.0. RSSI=-119, elapsed=1736uSec
37.0. RSSI=-119, elapsed=104uSec
38.0. RSSI=-119, elapsed=108uSec
39.0. RSSI=-119, elapsed=104uSec
40.0. RSSI=-119, elapsed=108uSec
41.0. RSSI=-119, elapsed=108uSec
42.0. RSSI=-115, elapsed=1756uSec
43.0. RSSI=-115, elapsed=108uSec
44.0. RSSI=-115, elapsed=108uSec
45.0. RSSI=-115, elapsed=104uSec
46.0. RSSI=-115, elapsed=108uSec
47.0. RSSI=-115, elapsed=108uSec
48.0. RSSI=-114, elapsed=1756uSec
49.0. RSSI=-114, elapsed=108uSec
50.0. RSSI=-114, elapsed=108uSec
51.0. RSSI=-114, elapsed=104uSec
52.0. RSSI=-114, elapsed=112uSec
53.0. RSSI=-114, elapsed=108uSec
54.0. RSSI=-119, elapsed=1752uSec
55.0. RSSI=-119, elapsed=108uSec
56.0. RSSI=-119, elapsed=112uSec
57.0. RSSI=-119, elapsed=108uSec
58.0. RSSI=-119, elapsed=104uSec
59.0. RSSI=-119, elapsed=108uSec
60.0. RSSI=-117, elapsed=1732uSec
61.0. RSSI=-117, elapsed=108uSec
62.0. RSSI=-117, elapsed=108uSec
63.0. RSSI=-117, elapsed=104uSec
64.0. RSSI=-117, elapsed=108uSec
65.0. RSSI=-117, elapsed=108uSec
66.0. RSSI=-114, elapsed=1756uSec
67.0. RSSI=-114, elapsed=108uSec
68.0. RSSI=-114, elapsed=108uSec
69.0. RSSI=-114, elapsed=104uSec
70.0. RSSI=-114, elapsed=108uSec
71.0. RSSI=-114, elapsed=108uSec
72.0. RSSI=-115, elapsed=1756uSec
73.0. RSSI=-115, elapsed=108uSec
74.0. RSSI=-115, elapsed=108uSec
75.0. RSSI=-115, elapsed=104uSec
76.0. RSSI=-115, elapsed=108uSec
77.0. RSSI=-115, elapsed=112uSec
78.0. RSSI=-116, elapsed=1752uSec
79.0. RSSI=-116, elapsed=108uSec
80.0. RSSI=-116, elapsed=112uSec
81.0. RSSI=-116, elapsed=108uSec
82.0. RSSI=-116, elapsed=104uSec
83.0. RSSI=-116, elapsed=108uSec
84.0. RSSI=-115, elapsed=1760uSec
85.0. RSSI=-115, elapsed=104uSec
86.0. RSSI=-115, elapsed=108uSec
87.0. RSSI=-115, elapsed=108uSec
88.0. RSSI=-115, elapsed=104uSec
89.0. RSSI=-115, elapsed=108uSec
90.0. RSSI=-115, elapsed=1732uSec
91.0. RSSI=-115, elapsed=108uSec
92.0. RSSI=-115, elapsed=108uSec
93.0. RSSI=-115, elapsed=104uSec
94.0. RSSI=-115, elapsed=108uSec
95.0. RSSI=-115, elapsed=104uSec
96.0. RSSI=-116, elapsed=1760uSec
97.0. RSSI=-116, elapsed=108uSec
98.0. RSSI=-116, elapsed=104uSec
99.0. RSSI=-116, elapsed=108uSec
Very often when the RSSI does change, it coincides with a much longer time to take the RSSI measurement (an order of magnitude longer).

That is to say: at lower bitrates such as these, the period between fresh RSSI measurements is noticeably longer.
« Last Edit: January 28, 2016, 02:42:09 AM by WhiteHare »

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Real time RSSI measurement broken RFM69CW?
« Reply #63 on: January 28, 2016, 01:28:09 PM »
This is exactly what I saw, a significant delay introduced some time after the RxRestart was issued. It's strange because it wasn't immediate. I think that the MODEREADY bit also went inactive then active again at some point after issuing the restart as well, but again this wasn't immediate.
Mark.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #64 on: January 29, 2016, 04:45:47 PM »
I unsoldered Felix's wire antenna and then soldered on a 47 ohm resistor to short antenna to ground (see attached photo). ...

I can see how using an SMD resistor would have been ideal.  Unfortunately, I don't have any on-hand.

It seems I can get 47 ohm 1% SMD resistors from China for the cost of a postage stamp and waiting a month, so I'm game.  Which do you suppose would be easier to solder onto the RFM69 to short the antenna: a 1206 or 0805? 

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Real time RSSI measurement broken RFM69CW?
« Reply #65 on: January 29, 2016, 06:06:27 PM »
The ANT input has an impedance anyway (either 200R or 50R however it is programmed), it is effecftively already pulled low at the sx1231h. The only thing left is the small stub on the PCB to the actual pin which could act like a tiny antenna. Tacking the ANT pin to GND would I think be sufficient, terminating with 50R I don't think will make a great deal of difference. You'd normally match the impedances for maximum power transfer, but in this case you explicitly want no power transferred.


WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #66 on: January 29, 2016, 06:21:32 PM »
Excellent point.  No need to wait for 30 days either.   :) 

In summary, I plan to un-install the 47 ohm resistor sometime this weekend and then bridge between ANT and GND using a solder tack.    Afterward, I'll post the resulting RSSI measurements.

Thanks!

emjay

  • Full Member
  • ***
  • Posts: 119
  • Country: nl
Re: Real time RSSI measurement broken RFM69CW?
« Reply #67 on: January 30, 2016, 01:55:00 AM »
@WhiteHare - with your "short" in place, remember not to turn on Tx full blast !

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #68 on: January 30, 2016, 03:10:28 AM »
Yup, no plans to turn on the Tx at all on the shorted unit.

BTW, I completed uninstalling the 47 ohm resistor and adding the solder tack.  My initial impression, though, is that the solder tack seems to perform about the same when running the RSSI sketches.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #69 on: January 30, 2016, 01:10:14 PM »
Using the solder tacked antenna, if I stretch the time between RxRestart's, then the measured RSSI's become seemless at the transitions between just before the RxRestart and just after the RxRestart:

Code: [Select]
Radio Configuration:
====================
Receiver-Mode(RX)
Packet Mode, FSK modulation scheme
[ActualBitrate] 4799.760bps = nominal bitrate of 4800bps
[theDccFreq] B10 (equals Semtech default)
[RxBw] 10416Hz
  [RxBwMant] 24
  [RxBwExp] 5
[theDccFreqAfc] B100 (equals Semtech default)
[RxBwAfc] 50000Hz
  [RxBwMantAfc] 20
  [RxBwExpAfc] 3
[FDev] 5004Hz
[frequency-center] 915000000Hz
[regLNA] B1001
  [LnaZin] B0 -> Input Impedance = 50 ohm (Warning:  Semtech default is 200 ohm).
  [LnaCurrentGain] B1 -> G1
  [LnaGainSelect] B1 -> G1  (Note: Manually selected.  AGC disabled.)
                   ^Warning: does not match Semtech default.
[regAFCFEI] B10000
  [FeiDone] B0 -> FEI is on-going (i.e. not finished).  Matches Semtech default.
  [AfcDone] B1 -> AFC is finished (i.e. not on-going).  Matches Semtech default.
  [AfcAutoclearOn] B0 -> Invalid, because AfcAutoOn is not set.
  [AfcAutoOn] B0 -> AFC is performed each time AfcStart is set (i.e. not each time Rx mode is entered).
                    Matches Semtech default.
---------------------------------------------------------------------
[deleted beginning because of 10,000 character posting limit]
0.246. RSSI=-117
0.247. RSSI=-117
0.248. RSSI=-117
0.249. RSSI=-117
1.0. RSSI=-117
1.1. RSSI=-117
1.2. RSSI=-117
1.3. RSSI=-117
1.4. RSSI=-117
1.5. RSSI=-117
1.6. RSSI=-117
1.7. RSSI=-117
1.8. RSSI=-117
1.9. RSSI=-117
1.10. RSSI=-117
1.11. RSSI=-117
1.12. RSSI=-117
1.13. RSSI=-117
1.14. RSSI=-117
1.15. RSSI=-117
1.16. RSSI=-117
1.17. RSSI=-117
1.18. RSSI=-117
1.19. RSSI=-117
1.20. RSSI=-117
1.21. RSSI=-117
1.22. RSSI=-117
1.23. RSSI=-117
1.24. RSSI=-117
1.25. RSSI=-117
1.26. RSSI=-117
1.27. RSSI=-117
1.28. RSSI=-117
1.29. RSSI=-117
1.30. RSSI=-117
1.31. RSSI=-117
1.32. RSSI=-117
1.33. RSSI=-117
1.34. RSSI=-117
1.35. RSSI=-117
1.36. RSSI=-117
1.37. RSSI=-117
1.38. RSSI=-117
1.39. RSSI=-117
1.40. RSSI=-117
1.41. RSSI=-117
1.42. RSSI=-117
1.43. RSSI=-117
1.44. RSSI=-117
1.45. RSSI=-117
1.46. RSSI=-117
1.47. RSSI=-117
1.48. RSSI=-117
1.49. RSSI=-117
1.50. RSSI=-117
1.51. RSSI=-117
1.52. RSSI=-117
1.53. RSSI=-117
1.54. RSSI=-117
1.55. RSSI=-117
1.56. RSSI=-117
1.57. RSSI=-117
1.58. RSSI=-117
1.59. RSSI=-117
1.60. RSSI=-117
1.61. RSSI=-117
1.62. RSSI=-117
1.63. RSSI=-117
1.64. RSSI=-117
1.65. RSSI=-117
1.66. RSSI=-117
1.67. RSSI=-117
1.68. RSSI=-117
1.69. RSSI=-117
1.70. RSSI=-117
1.71. RSSI=-117
1.72. RSSI=-117
1.73. RSSI=-117
1.74. RSSI=-117
1.75. RSSI=-117
1.76. RSSI=-117
1.77. RSSI=-117
1.78. RSSI=-117
1.79. RSSI=-116
1.80. RSSI=-116
1.81. RSSI=-116
1.82. RSSI=-116
1.83. RSSI=-116
1.84. RSSI=-116
1.85. RSSI=-116
1.86. RSSI=-116
1.87. RSSI=-116
1.88. RSSI=-116
1.89. RSSI=-116
1.90. RSSI=-116
1.91. RSSI=-116
1.92. RSSI=-116
1.93. RSSI=-116
1.94. RSSI=-117
1.95. RSSI=-117
1.96. RSSI=-117
1.97. RSSI=-117
1.98. RSSI=-117
1.99. RSSI=-117
1.100. RSSI=-117
1.101. RSSI=-117
1.102. RSSI=-117
1.103. RSSI=-117
1.104. RSSI=-117
1.105. RSSI=-117
1.106. RSSI=-117
1.107. RSSI=-117
1.108. RSSI=-117
1.109. RSSI=-114
1.110. RSSI=-114
1.111. RSSI=-114
1.112. RSSI=-114
1.113. RSSI=-114
1.114. RSSI=-114
1.115. RSSI=-114
1.116. RSSI=-114
1.117. RSSI=-114
1.118. RSSI=-114
1.119. RSSI=-114
1.120. RSSI=-114
1.121. RSSI=-114
1.122. RSSI=-114
1.123. RSSI=-114
1.124. RSSI=-114
1.125. RSSI=-114
1.126. RSSI=-114
1.127. RSSI=-114
1.128. RSSI=-114
1.129. RSSI=-114
1.130. RSSI=-114
1.131. RSSI=-114
1.132. RSSI=-114
1.133. RSSI=-114
1.134. RSSI=-114
1.135. RSSI=-114
1.136. RSSI=-114
1.137. RSSI=-114
1.138. RSSI=-119
1.139. RSSI=-119
1.140. RSSI=-119
1.141. RSSI=-119
1.142. RSSI=-119
1.143. RSSI=-119
1.144. RSSI=-119
1.145. RSSI=-119
1.146. RSSI=-119
1.147. RSSI=-119
1.148. RSSI=-119
1.149. RSSI=-119
1.150. RSSI=-119
1.151. RSSI=-119
1.152. RSSI=-118
1.153. RSSI=-118
1.154. RSSI=-118
1.155. RSSI=-118
1.156. RSSI=-118
1.157. RSSI=-118
1.158. RSSI=-118
1.159. RSSI=-118
1.160. RSSI=-118
1.161. RSSI=-118
1.162. RSSI=-118
1.163. RSSI=-118
1.164. RSSI=-118
1.165. RSSI=-118
1.166. RSSI=-118
1.167. RSSI=-118
1.168. RSSI=-118
1.169. RSSI=-118
1.170. RSSI=-118
1.171. RSSI=-118
1.172. RSSI=-118
1.173. RSSI=-118
1.174. RSSI=-118
1.175. RSSI=-118
1.176. RSSI=-118
1.177. RSSI=-118
1.178. RSSI=-118
1.179. RSSI=-118
1.180. RSSI=-118
1.181. RSSI=-118
1.182. RSSI=-113
1.183. RSSI=-113
1.184. RSSI=-113
1.185. RSSI=-113
1.186. RSSI=-113
1.187. RSSI=-113
1.188. RSSI=-113
1.189. RSSI=-113
1.190. RSSI=-113
1.191. RSSI=-113
1.192. RSSI=-113
1.193. RSSI=-113
1.194. RSSI=-113
1.195. RSSI=-113
1.196. RSSI=-114
1.197. RSSI=-114
1.198. RSSI=-114
1.199. RSSI=-114
1.200. RSSI=-114
1.201. RSSI=-114
1.202. RSSI=-114
1.203. RSSI=-114
1.204. RSSI=-114
1.205. RSSI=-114
1.206. RSSI=-114
1.207. RSSI=-114
1.208. RSSI=-114
1.209. RSSI=-114
1.210. RSSI=-114
1.211. RSSI=-118
1.212. RSSI=-118
1.213. RSSI=-118
1.214. RSSI=-118
1.215. RSSI=-118
1.216. RSSI=-118
1.217. RSSI=-118
1.218. RSSI=-118
1.219. RSSI=-118
1.220. RSSI=-118
1.221. RSSI=-118
1.222. RSSI=-118
1.223. RSSI=-118
1.224. RSSI=-118
1.225. RSSI=-116
1.226. RSSI=-116
1.227. RSSI=-116
1.228. RSSI=-116
1.229. RSSI=-116
1.230. RSSI=-116
1.231. RSSI=-116
1.232. RSSI=-116
1.233. RSSI=-116
1.234. RSSI=-116
1.235. RSSI=-116
1.236. RSSI=-116
1.237. RSSI=-116
1.238. RSSI=-116
1.239. RSSI=-116
1.240. RSSI=-114
1.241. RSSI=-114
1.242. RSSI=-114
1.243. RSSI=-114
1.244. RSSI=-114
1.245. RSSI=-114
1.246. RSSI=-114
1.247. RSSI=-114
1.248. RSSI=-114
1.249. RSSI=-114
2.0. RSSI=-114
2.1. RSSI=-114
2.2. RSSI=-114

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #70 on: February 06, 2016, 02:49:49 PM »
Which is also why I think adding a small offset for AFC when using low beta signals where most of the enegy is concentrated around the carrier frequency gives better sensitivity, there is an effective  'notch' in the filter due to the DC cancelling roll-off.

Which DCC have you settled on using?

I don't know how thoroughly John K2Ox tested, or whether he had AFC on-off, but John K2Ox reported improved sensitivity with smaller percentage DCC settings.  He also found improvement by using a different DCC setting when senseboost was enabled than when senseboost was disabled:  https://lowpowerlab.com/forum/index.php/topic,222.msg955.html#msg955

On the other hand, I notice Joe Lucid is using DCC=000b (equals 16%):  https://lowpowerlab.com/forum/index.php/topic,982.0.html

So, two smart guys but seemingly opposite conclusions from one another.

Then there's Semtech, which seems to be suggesting DCC of 010b (=4%) for RxBw as the default, but 100b (=1%) as its default for AfcBw.

These numbers are all over the map!

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Real time RSSI measurement broken RFM69CW?
« Reply #71 on: February 06, 2016, 05:38:17 PM »
I use the ones given in the Freescale app note for AFC of 0.5% for both RxBw and RxBwAfc, but I'm now also using a modulation index of around 1 as does this app note..

Let's think about this, remember the DCC cutoff frequency is effectively half the width of the 'notch' in the full receiver filter bandwidth (because the full receiver bandwidth is two 'single side' RxBw bandwidths back to back).

In high index modulation systems the energy of the signal is concentrated around the two modulation frequencies and not at the centre frequency (i.e at Fc+Fdev and Fc-Fdev) so it would make sense that increasing the DCC cutoff would increase sensitivity, because you are widening the notch in the middle which removes some noise but there's not a lot of signal there. The Semtech datasheet says "The cutoff frequency of the DCC can however be increased to slightly improve the sensitivity, under wider modulation conditions". This is therefore consistent.

On the other hand, for low index modulation most of the signal energy is not at the modulation frequencies but instead concentrated around the centre carrier frequency, so widening the notch in that case would have the opposite effect, you'll be removing energy from your signal and decrease sensitivity. Hence the Freescale AFC app note uses a small DCC cut-off of 0.5%.

So I think the answer depends on whether your are using a high or low modulation index. At least that is how I understand it.

Mark.
« Last Edit: February 06, 2016, 05:41:04 PM by perky »

WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #72 on: February 06, 2016, 07:26:35 PM »
That does shed light on things.  However, for low beta, that app note says:
"Settings to be used:
RX LO offset should be set to between 10% and 15% of Fdev.
RX BW should be greater than the single sideband signal bandwidth + RX LO offset.
DCC should be about 10% of the RX LO Offset"

That would imply DCC of 1.0-1.5%.  So that probably means DCC = 100b, because the nearest settings on either side of that are 0.5% and 2%.

I don't know how that app note arrived at the 0.5% figure (Table 2 of the appt note), unless maybe that's what it should be if AFC is disabled (?).

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Real time RSSI measurement broken RFM69CW?
« Reply #73 on: February 07, 2016, 02:40:58 AM »
DCC cutoff frequency is relative to RxBw, not Fdev. If anything the app not sets it too high:

RxBw = 156.24kHz
LowBetaOffset = 3.416kHz
so 10% of LowBetaOffset = 341.6Hz, which is 0.22% of RxBw.

But then is does say 'about 10%' and also says you should fine tune it by looking at the sensitivity as you do it.


WhiteHare

  • Hero Member
  • *****
  • Posts: 1296
  • Country: us
Re: Real time RSSI measurement broken RFM69CW?
« Reply #74 on: February 07, 2016, 09:55:01 AM »
DCC cutoff frequency is relative to RxBw, not Fdev.
You're right.  Now it makes sense.  I finally get it.  Thank you! 

...and also says you should fine tune it by looking at the sensitivity as you do it.
I do wish the appt note had elaborated on that.  Just how does one "look at the sensitivity as you do it"?  It reads almost like some kind of screw adjustment.  One idea that I'm toying with is to perhaps average a large number of the RSSI's that we're now getting, say 10,000 or 100,000 or 1,000,000 of them.  Then, since there are only eight possible DCC values, just index through all of them, and for each one print the average RSSI for that setting.  Then perhaps repeat that cycling again a few more times just to ensure that the results are consistent and not thrown off by some intermittent interference source.  Assuming that shows a consistent pattern, then the prefered DCC would be the one with the lowest average RSSI (i.e. experiencing the least noise).  Is that how you're approaching it, or do you see a better way?

I suppose the proper way to do it isn't that but rather to have test equipment that can be wired directly to the receiver antenna so as to reliably send packets at known power levels to the receiver, and then to ratchet down the power to the spec-sheet levels for receive sensitivity.  Then fiddle with the DCC values to find the best seting to receive at.  However, since I don't have the equipment to do it that way (do you?), I'm just trying to figure out what would be the next best thing.
« Last Edit: February 07, 2016, 10:38:42 AM by WhiteHare »