Transparent mode: CC1200 much worse than CC1120?

Sorry, this request was forgotten.

I did a quick test yesterday where I used a SG to send a '01' sequence to a CC1200 EM. I tested up to 200kbps without seeing any problems except for the known jitter. Have you tried with a known good sender?

TER,

we tested CC1200EM <-> CC1200EM, using SmartRF and a SG for binary 0101 injection. I did not yet use a dedicated external sender, as I assumed that CC1200 should work against another CC1200..

What's quite obvious: if  (Symbol Rate / Signal frequency) is not an even number prime, it's getting even worse.

Please give us some more hints to tackle: as TI currently is on 20 weeks lead time with CC112x, I need to decide by next week about next years volume. My purpose was to go with CC1200, but as long as this issue is not solved I can't take that risk.

regards

Addendum (ALWAYS REMEMBER THAT WE TALK TRANSPARENT SERIAL MODE):

Tried a SG as sender. SRF RX default setting 434-first default line (1.2kbps, etc.). Set the SG accordingly.

Result: only at exact 600Hz output is somehow OK, if I change to eg 580 oder 630Hz, it becomes dirty.

Same if I use our own setting (found in weeks of tests with CC1125): 16kBW, 8k symbol rate, 5k dev: as long as I use an even divider like 1kHz/2kHz etc, output is OK. As soon as I leave these "even" frequencies, output becomes unusable -> this is not jitter and one-to-one tests to our 1125/1120 shows, that this is CC120 specific.

You write Symbol Rate / Signal frequency: Signal frequency, is that RF frequency?

600Hz output, is this the datarate from the SG?

even divider equal to 1kHz etc, is this the datarate of the incomming signal?

When reading your two last posts it is not clear for me exactly what you apply as your testsignal. 

Signal frequency is meant as the data rate (modulation) I inject by the signal generator. So e.g. 600Hz/2kHz/etc.

Meanwhile I did a pure 1:1 test: Signal generator with antenna plus

a. CC1125 EM plus antenna

b. CC1200 EM plus antenna

Test modi:

1. Default SMartRF 434MHz, 1.2kbps, SG set accordingly. Then I tune SG modulation frequency: at 600Hz output both EM look the same. As soon as I tune slightly to e.g. 620Hz CC1200 output starts flickering (like shown in my initial message). CC1125 remains stable up to 800Hz.

2. 'our' range setting: 16k BW, 5k dev, 8k symbol rate: as long as I stay at 1/2/4kHz modulation, both EM look OK. As soon as I leave by a few Hz to e.g. 2.1kHz, CC1200 output signal becomes unstable (see above)

Meanwhile I know the CC good enough to conclude, that CC1200 sampling in transparent mode does have an issue (compared to CC1125): in transparent mode you can't predict the exact data rate, but the RF circuit need to deal with that.
I still hope that this is not a design issue ...

I been working with another case all day so I have not been able to check your finding. I will look into this to try to find out why you get the results you get but since you have to decide shortly for next years volume I would recommend ordering CC1120.

Understand, I'll proceed accordingly:

Nevertheless, also for some future developments, it's vital to tackle this issue.

Hopefully the problem is described in detail enough to be understood. I'm happy to provide more information or, even better, show it to you via e.g. Netmeeting.

I'm very keen about your findings - when do you expect to have time to dig into?

I looked some at today and saw the same as you; programming a datarate equal to 1.2kHz on the chip and and using 1.24kbps on the generator it looks like the receiver skip every n bit. In transparent mode what you program the datarate to on the chip should not matter but we have to investigate further to find out why.

TER,
honestly I'm not glad that you confirm our findings - hopefully this isn't something in the design that would make CC1200 unusable for us.

You say it looks like it would skip bits - it's worse: look at the duty cycle of the output signal, there is no single phase like the original.

When I did test in the lab it looked like that I got the same result as you (It looks like v1.13.0 of Studio does not enable transparent mode I found out now)  but I got one of the software people to convert some code from CC1120 to CC1200 and with this code everything are working just fine.8078.cc120x_transparent_mode.zip

We did 2 tests; with the 50bps settings and 1.2kbps. For the first one we sent with 11.3kbps and the second 647bps without seeing any problems.

thx;
as I did not run own code yet on TRxEB, please give me a hint how to use/load your code examples within the EM and I will try to follow.

But: is this then a bug in Studio's parameter settings? You can't setup a communication with different symbol rates between two CC1200, this is for sure. So please try to setup Studio with something like that with uneven sender/receiver symbol rate ratio - this doesn't work.

TER,
I've got your software example but as we are not so deep in direct programming TrXEb plus lacking IAR we hadn't a chance to verify.

My reference is still SmartRF, so my wish is if somebody could try to setup transparent mode with unequal symbol rate using CC1200 ..

It is very few registers that are set in the code:

  CC120X_IOCFG3=0x11
  CC120X_IOCFG2=0x09
  CC120X_MDMCFG1=0x06
  CC120X_MDMCFG0=0x65
  CC120X_AGC_CS_THR=0x16
  CC120X_PKT_CFG2=0x07

Carrier sense is used in the software example but not required. Have you tried with these settings?