LimeMicro:LMS6002D FAQ

Where can I find additional documentation for LMS6002DFN?
The LMS6002D-docs GitHub repository contains PDF versions of the documentation available on this wiki, along with additional documentation that has not made it onto here yet.

Graphical User Interface (GUI)
See the LMS Suite FAQ.

How can a 3GGP waveform be transmitted via the LMS6002DFN chip?
There are two ways to test 3GPP waveform:


 * 1) using digital IO, generated with a third party baseband processor, DSP or FPGA;
 * 2) using analog IQ signals, fed directly to LMS6002DFN. For example, Analog IQ input signals for Tx can be generated using signal generator with appropriate IQ modulator option.

What settings are required for operation with Analog IQ signals?
Using Tx Analog IQ inputs, DACs should be powered down. Using Rx Analog outputs, ADCs should be power down.

Do I need to use the ADF4002 option?
Yes. Synchronizing your test and measurement equipment with evaluation board reference clock will cure the frequency error in your setup. Especially useful in EVM and sensitivity measurement.

How should I synchronize measurement equipment with the evaluation/interface board?
Connect 10 MHz reference signal to J4 connection on the board. Program onboard ADF4002 with GUI on the tab called Board.

How do I select VCOCAP code for the desired frequency?
VCOCAP code have to be tuned when you change frequency. VCO Comparators are implemented on chip to monitor VCO Tune voltage. This gives a logic 0’s when VCOCAP code value is correct. See VCO and VCOCAP Code Selection Algorithm.

Which clock buffers have to be enabled for the normal operation of LMS6002DFN?
For full duplex operation only TX DSM SPI and RX DSM SPI have to be enabled. Other clock buffers have to enabled only for calibration procedures.

How should I connect the Lime evaluation board to the HSMC connector?
Lime supplies adaptor boards for HSMC connector. For more information please email [mailto:sales@limemicro.com sales@limemicro.com].

Does LMS6002DFN require a specific order for setting up the internal registers on power on? Should there be a time delay between every register set?
There is no particular order required. There is no need to define a time delay between every register set either.

Where are the spurs coming from around the carrier signal at +/-200 kHz?
On the evaluation board there is a -5V switching regulator for the differential buffer on the RX analog IQ outputs, which injects a 240kHz spur into the PLL, but is largely filtered by the 100kHz PLL loop filter. This spur can be removed by disconnecting the power to the -5V regulator.

What is the common mode voltage and voltage swing setting to drive Tx Analog (baseband) input ports?
LMS6002D Analog inputs require a 0.6V common mode offset voltage with a 300 mVp-p voltage swing for optimum performance.

What is the LO leakage calibration procedure for transmitter?
Set transmitter for the desire frequency. Set the gain for the required output power. Turn off DAC or set DAC to generate minimum DC. Make sure that there is no any signal applied at TX RF input during DC calibration Execute auto DC offset calibration. Minimize LO leakage by manually adjusting I/Q DAC LO leakage calibration registers. This calibration is described in LMS6002Dr2-Quick Starter Manual-EVB_5_r2.pdf (p.66).

What is the LO leakage calibration procedure for receiver?
Before calibrating, set the receiver for the desired frequency. Set VGA2 gain to maximum (30 dB). Execute calibration for RXLPF and RXVGA2.

Is LO leakage calibration required when frequency has been changed?
LO leakage calibration is required if only 3GPP band has been changed. Calibration remains stable over the entire frequency range within a given 3GPP band.

TX LO leakage is specified as -50 dBc. Can it be maintained over entire TX dynamic range?
This is specified for maximum transmitter gain. With LO leakage calibration you should be able to achieve -50 dBc for entire Tx dynamic range and frequency range.

How many clock cycles are required for DC offset calibration?
DC calibration function takes 64 clock cycles. DC calibration clocks are derived from the PLL reference clock as follows:


 * For RXVGA2: cal clock frequency is Fref/16;
 * For LPF: cal clock frequency is Fref/256;
 * 64 DC cal clock cycles are required per stage (TXLPF=2 stages, RXLPF=2 stages, RXVGA2=5 stages).

Can I use look-up table of LO leakage calibration over multiple frequencies?
Yes, there are on chip automated DC cancelation/calibration loops. We recommend to execute (trigger) these at each startup. Other calibration parameters such as TX LO leakage, IQ phase/gain imbalance can be stored and reloaded every time.

Why TXLPF, RXLPF, and RXVGA2 calibration routines return with DC_LOCK values of 0?
Please note that sometimes DC_LOCK is not a reliable indication. The reason for this is the fact that the DC offset compensation DAC step size is smaller around 0V DC, which corresponds to DAC code 31dec meaning "no need to compensate, DC level is good".

Under this condition, comparator responds as:


 * UP, UP, UP corresponding to DC_LOCK = "111" or
 * DOWN, DOWN, DOWN corresponding to DC_LOCK = "000" instead of
 * UP, DOWN, UP, DOWN, ... corresponding to "1010..."

In this case DC_lock is not reliable indicator of success or failure of DC offset cancellation algorithm. We suggest to use DC_REG_VAL instead of DC_LOCK as in the pseudo-code below.

Note that CHIP_NOT_OK condition should not happen as this is being checked in our production test.

How to execute calibration routines if there is no read back function in my baseband?
You can use the calibration sequences as shown below.

Can RX calibration be maintained over the entire Rx dynamic range?
Once the RX path is calibrated the DC offset will change more with RXVGA1 gain than with RXVGA2 gain. When you change RXVGA1 gain you have to recalibrate, so use RXVGA2 for AGC first before you start engaging RXVGA1. This behaviour can be improved upon a lot in the digital domain (BB/FPGA) by implementing an averaging filter. More information can be found in LMS6002Dr2-Improving transceiver performance using digital techniques-1.0r1.pdf.

What is total power consumption of LMS6002DFN
In full duplex mode at maximum gain the power dissipation is around 1.5W.

What are the maximum transmitter and receiver gain values?
Transmitter gain consists of: TX LPF Gain 6dB, TX VGA1 gain -4dB, TX VGA2 Gain 25dB, total gain is +27dB.

Receiver gain consists of: RX LNA Gain 12dB, RX VGA1 gain 30dB, LPF Gain 6dB and RX VGA2 Gain 30dB, total is +78dB

What is the receiver and transmitter PLL lock time?
Maximum PLL lock time is 20us.

How to design a new PLL Loop Filter?
Loop filter has already been designed and optimized for LMS6002DFN. Please refer to reference/EVB board schematic “REF6002-15 Schematics.pdf”, page 5 or “EVB6002-6_schematics_v1.pdf”, page 6.

What is the PLL Loop Filter bandwidth?
Loop filter is designed for 100 kHz bandwidth.

What is RF bandwidth of the Tx1 and Tx2 outputs?
Both transmitter outputs RF bandwidth is same from 0.3 GHz to 3.8 GHz. On evaluation board Tx1 is matched to 3GPP Band1 and Tx2 has an Broadband matching network.

What are the baseband filter bandwidths implemented in LMS6002DFN?
The baseband filter bandwidth can be selected from 0.75 – 14 MHz IF, giving an RF bandwidth from 1.5 MHz – 28 MHz.

What is the group delay for 5 MHz and 2.5 MHz filters in LMS6002DFN?
The group delay of 5 MHz filter is around 300ns. For 2.5 MHz group delay is 600ns.

Does the LNA need to be power down when the RF LOOPBACK is in operation?
For testing purposes the LNA does not need to be powered down. If this is to be done in the field, to avoid any interference, LNA must be powered down. To do this you will have to enable Direct signals in RxFE tab Direct Signals [0x7003] and deselect LNA Modules [0x7D01].

What is the maximum CW signal input level for the receiver?
The maximum CW signal level for the receiver LNA2, with maximum gain, at frequency of 1.95 GHz is -60 dBm. Input signal varies with the frequency.

What is the IIP3 for the receiver?
The RX IIP3 is frequency dependent. The IIP3 for the receive LNA2 is -1dB, at 1.95 GHz.

What is the RX Mixer P1dB?
The Rx Mixer Input IP3 = 22.5 dBm hence the Mixer input P1dB would be 11 dBm.

What is the TX OIP3 (or OIP1) at maximum and minimum gain settings?
TX OIP3 is frequency dependent. OP1dB is about +15 dBm at maximum gain and OIP3 is typically 10 dB higher. Reducing VGA2 gain by 25 dB, the OP1dB and OIP3 will fall by 25 dB.

What is RF bandwidth of the receiver inputs?
The RX LNA1 RF bandwidth is from 0.3 – 2.8 GHz, Rx LNA2 is 1.5 – 3.8 GHz, Rx LNA3 is 0.3 – 3.0 GHz.

What is the TX Noise Figure at maximum gain settings?
TX Noise Figure is largely defined by TXLPF and TXDAC. Low Pass Filter noise figure is about 35 dB relative to 50 Ohm.

What is transmitter to receiver noise isolation on LMS6002DFN?
Measured transmitter noise on receiver 3GPP Bands: Band 1 – -135dBm/Hz with TX output -2.2 dBm/3.84 MHz; Band 5 – -131.2 dBm/Hz with TX output 2.85 dBm/3.84 MHz;

What is the settling time of TX/RX gain blocks after they are set via SPI control?
Settling time of each gain block in LMS6002DFN is less than 100ns.

What is recommended gain table for the receiver?
Notes:


 * 1) * Sensitivity measured with WCDMA 12.2 RMC Signal. Processing gain 25 dB.
 * 2) ** Minimum Eb/No = 5.7 dB.

How can RSSI be used in LMS6002DFN?
There is no RSSI block in LMS6002DFN chip. RSSI can be calculated digitally using BB/FPGA as RSSI=SQRT(I*I+Q*Q).

How can the VGA1 code be converted into dB’s?
Please use this formula: G [dB] = 20*log10(127/(127-Code)), where Code is gain control word, 0 <= Code <= 120.

Can the envelope detectors within the LMS6002DFN be used for calibration?
On chip peak detectors are working however providing 30 – 40dB dynamic range only which may not be enough for most calibration requirements. In any case, the level of un-calibrated TX LO leakage is already low enough so cannot be detected by the internal detectors. We recommend to put peak detector after external PA and use it for calibration.

Can the internal LPFs be bypassed?
Internal TX/RX LPFs can be bypassed.

What is heat thermal resistance and junction temperature of the LMS6002DFN?
The LMS6002DFN package has a 22°C/W thermal resistance (θja) to free air. If this is soldered to the board, then the thermal resistance of the complete unit from junction, through the case, through the board and to free air (not forced) becomes a calculated θja = 19.7°C/W, developed as:

θja (junction to air, soldered to the board) = θca (thermal resistance from the case interface on the board to air) + θjc (thermal resistance junction to case) = 14.0 + 5.7 = 19.7°C/W.

These numbers are extracted using Lime recommended layout. Using these calculations show that with a typical application having 1.5W dissipation, the chip temperature will be 1.5*19.7 ~29.8°C above ambient.

How to set LMS6002DFN for TDD operation?
Modification of EVB hardware:


 * 1) Replace 150 Ohm resistor on pin 72 (TRXVDDDSM18) of LMS6002DFN with 0Ohm.
 * 2) Remove 22 Ohm resistor on pin 60 (TXVDDVCO18) of LMS6002DFN.
 * 3) Connect TXVDDVCO18 (pin 60) and RXVDDVCO18 (pin 84) via 22 Ohm to single 1.8V supply.

The procedure for LMS6002DFN register programming is described below:
 * 1) Program TxPLL and Rx PLL for wanted LO frequency while in FDD mode.
 * 2) Switch to TDD mode by altering register 0x0A [1].
 * 3) Set 0x0A [0] register to 0 when transmitter is in operation and to 1 when receiver is in operation.

How to improve receiver linearity?
Power down RXVGA2 and RXLPF DC calibration comparators after calibration. To do that:

Set 0x6E[7:6] = "11", power down RXVGA2 DC comparators Set 0x5F[7] = "1", power down RXLPF DC comparator

Please note, auto DC offset calibration has to be executed before comparators are powered down.

What is the difference between hardware TXEN and STXEN, similarly for hardware RXEN and SRXEN?
SRXEN/STXEN registers are equivalent to TXEN/RXEN pins. The internal control signals are constructed using logical AND in the following way:

iTXEN = TXEN and STXEN iRXEN = RXEN and SRXEN

What is the environmental rating for LMS6002DFN?
The LMS6002DFN chip is specified for industrial environment, based on JESD47G-01 standard.

What is the KVCO for the VCO frequencies from 4GHz to 8 GHz?
Please see the graph below.



Does the LMS6002DFN supports JESD207 interface?
Yes. Lime has developed VHDL code to support JESD207 interface. Please request through [mailto:enquiries@limemicro.com enquiries@limemicro.com].

What are recommended CLK_jitter characteristics for Rx_CLK, Tx_CLK and PLL_CLK?
TXCLK, RXCLK require the usual jitter specs for the 12 bit DACs/ADCs. Less jitter (better phase noise) of PLLCLK results in improved phase noise in PLL and overall system EVM.

Is it possible to use 2.5 V data signal and clock signal with the LMS6002DFN internal DACs and ADCs?
Yes. For more information please go to datasheet section “Implementation Low Voltage Digital IQ Interface”, page 9.

Is RX_CLK_OUT required for RXD sampling?
No, but can be used. Please follow the reference schematic of RX_CLK layout in document “REF6002-15 Schematics.pdf”, page  2.

Can the sampling clock/rate for internal DAC and ADC be changed on the evaluation board?
The sampling clock can be changed depending on the clock source. That will require some 0 Ohm link soldering/disordering. All clock distribution options are described in the Quick Start Manual (section “3.4 TCXO Frequency and Data Clocks Distribution”, page 17).

What is the latency of the data converters within the LMS6002DFN?
DAC and ADC latency is around 10 TXCLK/RXCLK cycles and is not changing with the setup.

What is the maximum sampling rate for DAC’s and ADC’s?
Maximum sampling rate for DAC’s and ADC’s is 40MHz. Sampling rate is defined by the rate at which the pins of TXCLK (pin 19)  and RXCLK (pin 17) signal lines can be clocked. These are twice the data converters sampling rate. For more information please refer to the data sheet.

What is the recommended signal level of reference clock to drive the internal PLLs?
Recommended level for PLL reference clock is 3.3Vpp, CMOS type signal.

What type of coupling should be applied for reference clock to the internal PLLs?
By default, PLL clock input buffer is set to AC coupling mode. The same setting can also be used in DC coupling mode

How to improve ADC spectrum?
RXVGA2 CM voltage code has to be set to 13 which corresponds to 0.82 V (register 0x64[5:2] = '1101'). Also set ADC reference gain adjustment to 1.75 V (register 0x59 [6:5] = '01').

What is the recommended footprint for the LMS6002DFN?
Lime offer two types of footprints for LMS6002DFN, via in pad and via off pad. The latter version is a cost reduced option by avoiding the in-pad vias. Both footprints are tested and verified as reliable to be used in production. Please refer to “LMS6002Dr2 PCB Layout Recommendations-1.0r0.pdf” document. See also the Component Libraries for KiCAD.

What is the power up, down and the reset sequence for LMS6002DFN?
There is no particular power up sequence required. As usual, it is recommended ESD (3.3V) supplies to come up first and go off last.

'''However, there is no issue even if this timing is violated for short period. To identify ESD pins, see pin description in the data sheet.'''

A low pulse (10ns min) on RESET pin is recommended.

What are the recommended power supplies for the LMS6002DFN?
Switcher can be used for 3.3V. LDO is recommended for 1.8V to ensure a clean VCO supply.

Is it necessary to connect the Pin #42 ATP (Analog Test Point)?
Analog test point is made for production test. It should be left open. Please refer to reference schematic “REF6002-15 Schematics.pdf”, page  2.

Can the ADC be left unconnected if it is not used in my application?
Yes.

What is the purpose of the 22 Ohm resistors on pins 60 and 84?
The purpose of 22 Ohm resistor on pins 60 and 84 is to improve IQ phase imbalance. This also reduces VCO current from 40 mA down to 20 mA.

What is recommended metal mask size and depth for solder paste?
The design is constrained by the 14x14 pads on the inner row. This will not print well on a 5 mil foil so we reduced the foil to 4 mil.

What is the range of operating moisture condition in %?
After 168 hours at < 30C, 60% relative humidity.

What is package warp after reflow soldering?
Package warp should not exceed 0.05mm.

=== According to “LMS6002Dr2 PCB Layout Recommendations-1.0r06.pdf”, page 2, Figure 5, solder paste for GND pattern (center pad) is split into grid of 13x7（0.3mmx0.7mm）rectangles. What is the reason? Is solder volume too much if this GND pattern is not split? ===

The grid serves two purposes - to reduce volume so that it does not float too high and yet has good coverage to dissipate heat, and to avoid vias as much as possible. 40% is a good rule of thumb commonly used. Although it is typically done with a simple window pane. 25% is a typical minimum to ensure coverage and yet not sit too low.

Using a solid block would be bad as it would cause open circuits on the outside pads and may cause the device to spin during reflow. Lots of little pads will typically all reflow at the same time whereas one large pad will reflow as the heat hits it and not all at the same time. This can cause the device to rotate during reflow.

Our assembly house has special design software which automatically calculates the ratio of stencil paste surface area in contact with the pad to the side wall surface area of the stencil, referred to as the "print area ratio", all designs are passed through this checking software. As a result a 3mil stencil was chosen to make sure the pads could print OK.

Other questions
Questions concerning Myriad-RF hardware, such as the Reference Development Kit (original Myriad-RF RF module and interfaces), other Myriad-RF boards and general system development, should be posted to the appropriate category on Discourse.

Questions concerning the LMS6002D and other Lime Microsystems devices should be posted to the open source support group.

Document Version
Based on LMS6002D FAQ v1.0r13.

Changes since document generation:
 * Updated 1.1. to note location of PDF documentation
 * Added links to other documentation on this wiki