My first post here.
I’ve been reading the posts related with LimeSDR 1.4 and I hear lot of complains without definitive answers from Line Micro. I have already purchased the board and eventually will get it in 1 month.
I don’t have experience in hum reception but I am staring to worry about my investment
Can someone clarify to me:
Is LimeSDR 1.4 significantly worst in terms of signal to noise figure compared with LimeSDR 1.2?
Comparing the schematics I see that 1.4 has SPDT Switchs SKY13323 and some MOSFEts for a RF loop back (not available in 1.2). Are those components to blame for the not good RF performance in 1.4?
RX1_L of LimeSDR 1.4 doesn’t have other then simple matching network on the RF path. If we have bad signal to noise for this port it means that the LMS7002M performance is not good or we have software issue.
Any comments on this?
Or am I completely confused and all issues people discuss here are software related?
RX1/2_L and RX1/2_W are using TC1-1-13MA+ transformer [4.5 1GHz]@1dB insertion loss
RX1/2_H is using TCM1-63AX+ transformer [10 6GHz]@1.3dB insertion loss
so it seems the transformers are wide band enough to explain the issues at 100MHz
I have tried to simulate the ‘matching network’ for the RX1_L (same for RX2_L) and if I am not mistaken we seems to have significant attenuation for some regions. I hope somone clarify that I am doing something stupid? (markers are on the magnitude curve)
Hmm in fact I think I have to add the transformer model.
I am not sure if we should draw some conclusions without the transformer in the simulation.
I will try to include the transformer model based on the Mini Circuit datasheet tomorrow.
The question no one from Lime as answered so far, what was added to the 1.4 board compared to the 1.2 board, if it was the matching network, who decided to start it at 700Mhz rather than 100Khz…
I think a lot of people would like to know, specially the first 500 that now have the boards in hand…
Well I would not agree with that. Lime Micro has released the complete schematics so actually they implicitly gave this answer. But it will be nice if they are more active here indeed as we all only guessing.
Anyway I have compared the 1.2 and 1.4 RF part of the schematics and the only difference I have noticed is the those loopback solid state relays and mosfet switches.However this change do not affect RX1/2_L and RX1/2_W. So for those ports as per the RF part of the schematics the RF performance should be the same 1.2 and 1.4. If this is not like that (please confirm) I guess we have a software or other hardware issue (but not in the RF path).
LimeMicro come on, you are the only in a position to clarify what is going on.
Hello Dimitar,
did you manage to get the Transformer (TC1-1-13MA+) in to the simulation in LTspice? If so could you please tell me how? I found the S-Paramters of that transformer, but no way to put that into LTspice. Is there a way, or is that no possible?
Thanks
Erik
You are doing something “stupid”, but easily overlooked for someone with less experience - You have resistors in front of your AC signal sources. True signals don’t have a real source resistance, they exist as a signal traveling on a transmission line with a specific impedance (not resistance). To simulate this correctly, bump your sources signal value up to 1.0 to account for the resistor losses.
Also, C1 isn’t present in the real schematic. There is going to be some capacitance at the LMS7002 inputs, but I haven’t seen the S-parameter files to know what that is.
Unfortunately i still have not spend a time to include the transformer… Yes having the insertion loss and the reflection from the transformer datasheet should be enough to model the transformer as a dual port network. I hope I will find some time to do it.
I have not studied the LMS7002M datasheet. I am still wondering why we need the low pass behavior on the RF path. Anyone willing to shed some light?
@Kryan92 yes what I did is a quick and simple check. The resistors are to model the 50 Ohm impedance of the missing transformer. I agree that we need to add the transformer in the simulation.
C1 I took from the schematic. I don’t know if it is populated on the LimeSDR boards. Note however that 1.2pF has no significant effect for the frequencies below 100MHz for the above circuit.
@dpenv “I am still wondering why we need the low pass behavior on the RF path. Anyone willing to shed some light?” Do you mean high-pass behavior (from the shunt inductor and 510pF coupling caps?). Why that exists, I believe, is that the Lime team was trying to match on the _L receive channels for optimum performance on the cellular bands at 700-900 MHz. It’s inherently either infeasible or impossible to match for best performance on a wider frequency band without other trade offs.
Regarding the resistors, yeah, I’m saying that you can use the resistors there to make it accurate enough, but you’re going to get 6dB less signal across the than reality because those resistors in the sim are attenuating what in reality doesn’t exist (transformers transform a source impedance, which in this case is a 50 ohm ideal transmission line from microstrip and antenna).
Yeah high pass, the big drop in the insertion loss at low frequency I mean. (due to shunt inductance and serial capacitors)
Why do we need this shape in the pass band?
See my follow-up on the matching network - it tries to compensate the input impedance seen at the pins of the receiver input _L, which are non-ideal at certain frequencies (this is available in the S-parameters file which I haven’t seen, and don’t particularly care about), to some ideal amount (typically 50+j0 or 100+j0 ohms). If the receiver pins are capacitive (FET gates for the LNA usually are), you inherently have to pick a frequency point at which to focus on for maximum matching and performance, and tradeoff performance at other frequencies. This is what the shunt and coupling caps help do.
There are other options for matching that make a less severe mismatch or attenuation at lower frequencies, but this network is about as simple and as cheap as it gets for their targeted cellular band. Actually, the inductor here is doing most of the matching @ 900MHz, the caps are really just for AC coupling with the benefit of providing additional high-pass filtering for lower frequencies you’re already not matching for. If you’re doing cellular comms and go near an FM radio tower or other low-frequency, ultra-high power transmitter, you really don’t want the radio station to saturate and completely screw over your receiver if you antenna can pick it up at all.
Can someone point me to the impedance characteristics (s-parameter file Kryan92 is talking about).
Up to now I have assumed that the receiver input _L is 50 Ohm resistive.
