Hi,Noel!

I had a similar problem. I solved it with the help of a virtual com-port splitter.

My setup is FTDX3000+AIRSpy+SDRConsole+(Any Contest Log or HAMLog included OmniRig) over VSPE 64

73! Vit. RT6M

�

Interested to hear about you replacing the input transformer Bernard.

It�s something I was planning to do and even bought a couple.

I looked at the original transformer and found it wasn�t exactly as described in the Lime circuits.

If you check the manufacturer�s data for the original transformer you�ll see it�s not a classic transformer with primary and centre-tapped secondary as shown in the drawings.

I hadn�t realised this until I measured the RF voltages going into the LMS7002 and found very peculiar readings.

�

Changing the three input capacitors will improve the 100KHz performance but best to use 100nF. You can get narrow types that will fit over the originals.

However, I actually gave up on the Lime�s LF and HF performance and made a simple upconverter.

�

Now for something different�.

I did some experiments today though on 23cms

�

Listening on 1296.796� ish tuning to the G8MBU IOW beacon about 20km away using a broadband vertical aerial about 15 feet above ground

Mode: CW and loads of gain

Signals were fading very deeply but recognisable with clear morse code.

Lime:

L input modified by removal of the 8nH choke, noise floor -106dBm, signal -98dBm

H input noise floor -104dBm, signal -98dBm

W input noise floor -104dBm, signal -92dBm

SDR Play

Noise floor -101dBm, signal -93dBm

To my ears, the SDR Play gave the cleanest signal.

The Lime had a huge spurious spike in the centre of the display measuring -80dBm for L, -82dBm for W and -84dBm for H

SDR Play had a completely flat noise floor with no spurious spike

�

The SDR Play works down to a few KHz with long wave stations at 77KHz and 60KHz very clear in the waterfall.

I get equal results to this on the Lime with the upconverter but without this it�s unusable at LF.

Allan G3PIY

�

�

---

From: sdr-radio-com@... [mailto:sdr-radio-com@...]
Sent: 16 May 2017 11:05
To: sdr-radio-com@...
Subject: RE: [sdr-radio-com] LimeSDR PCB SPICE analysis
Importance: High

�

�

Very nice work of approach

�

On the input RX1_L

I removed the inductance MN18

I replaced the T3 transformer with Coilcraft WBC8-1LB transformer

Replaced the MN17 and MN20 capacitors with 470 pF

On C14 put a pair of parallel diodes BAV99W to protect the input from the surge

So far no improvement toward the 100 KHz

To be continued

�

Best regards

Bernard

�

De�: sdr-radio-com@... [mailto:sdr-radio-com@...]
Envoy�: lundi 15 mai 2017 21:51
��: sdr-radio-com@...
Objet�: [sdr-radio-com] LimeSDR PCB SPICE analysis [5 Attachments]

�

�

[Attachment(s) from numeric included below]

The attached files contain a spice ac analysis of the LimeSDR LOW input receiveRX2_L 700 MHz - 900 MHz (with and without the 8.2nh MN46 inductor) and the TX2_1 30 MHz - 1.9 GHz transmit output. This is basically a black box analysis derived from data already supplied. No PCB was harmed during the analysis! No PCB was actually tested. No soldering iron, no test clips used, no test probes, no instruments used, no scopes, no generators and no power applied.
Data is derived where possible from the data sheet and from S parameter TC1-1-13MA+ transformer data downloaded from Microcircuits. The transformer S parameter data was converted to a Spice model using s2spice.exe command line utility found on the Yahoo groups LTspice https://groups.yahoo.com/neo/groups/LTspice/files/%20Tut/S-Parameter/S-Parameter%20to%20SPICE/
The transformer data is for 4 ports with a common reference or ground. This is utility is useful for up to 8 ports.

LTSpice can be downloaded from: http://www.linear.com/designtools/software/#LTspice
LTspice is a high performance SPICE simulator, schematic capture and waveform viewer with enhancements and models for easing the simulation of switching regulators. It is free to use.

Transmit summary:
The transmit frequency plot is basically flat from 100Khz to 3G. Assumed 50 ohm balanced source generator, 25 ohms source for each half, derived from data sheet. According to the analysis no change is needed to the pcb. The schematic range of 30 MHz to 1.9 GHZ is misleading or I screwed up in the analysis, please check!

Receive summary:

The frequency plot clearly shows why HF reception is very poor with the MN46 8.2nh inductor. Response is 75 db loss at 1 MHz and 30 db loss at 30 MHz.
With the 8.2nh inductor removed the loss is dramatically improved -1.3 db at 1 MHz and -0.36 db at 30 MHz. Its basically flat from 718 KHz up to 1G and useful to 3G within a few db.
Assumptions. I used the input impedance for the A/D convertor. The data sheet listed the input to be 4K. The A/D input may not be the actual AN5 and AP6 input pins. Again please check. Also note, the TC1-1-13MA+ S parameter data is listed from 0.3 MHz to 4000.0 MHz. Data was interpolated for 100 KHz performance.
Comments, suggestions, criticisms welcome.

 

Sigi,

That's just what he's got the scope zoomed in to.  The lower left corner says "50.000 MHz".

Best,

Mark  KF5VQY