[time-nuts] Practical considerations making a lab standard with an LTE lite
saidjack at aol.com
Sun Nov 23 22:44:22 EST 2014
Any buffer options added to the board would have caused either additive phase noise or added power consumption, and possibly yet another low noise LDO to be required.
On the 20MHz units there is already a strong buffer that can drive 50 Ohms terminations so adding a buffer in front of the coax connector on that version would have just added unnecessary phase and AM noise, parts count and cost, and power consumption, and would have resulted in a product with worse performance than we have now. That configuration is the "normal" one so we did not add unnecessary circuitry that would have decreased product performance.
On the 10MHz boards with external DIP-14 TCXO there is no buffer, and adding one would have required to possibly add yet another low noise supply regulator and possibly another MMCX connector. Since this is the "optional" configuration, we optimized for highest performance for the "standard" configuration.
Adding this many features to the board required some trade offs to be made, and we have to keep in mind the initial goal of the entire effort: to provide an easy way to evaluate the performance of our LTE Lite module - hence its called the LTE Lite Evaluation board. Everything else was a bonus.
But in the end it should be fairly trivial to put a 50 Ohms driver and low pass filter together using either a CMOS gate or a simple emitter follower. We also need to keep in mind that generating a Sine Wave output would have consumed 200mW to 250mW additive power and thus would have more than doubled the total power consumption.
Lastly we have three outputs on the board so we would have required three additional buffers and their support circuitry, all that for a questionable improvement.
Or instead of adding a bunch of buffers one can use somewhat short cables and 1M input impedance on the target hardware and that will work perfectly too without any changes..
Sent From iPhone
> On Nov 23, 2014, at 16:22, Charles Steinmetz <csteinmetz at yandex.com> wrote:
> Said wrote:
>> The 10MHz units have a different RF output than the 20MHz units. The
>> 20MHz units have a 50 Ohms series-terminated and buffered RF output, while the
>> 10MHz units have the TCXO output drive the MMCX connector directly without
>> series impedance matching. Both drive the line with 3.0V CMOS levels. This
>> means the cable on the 10MHz unit should be kept as short as possible, and
>> that impedance matching for maximum power-transfer is not required nor
>> desired. The suggestion that Charles made for checking the impedance by
>> progressively loading the output more and more is valid for Sine Wave outputs, but
>> not for CMOS outputs as implemented on the LTE Lite.
> Absolutely correct -- I did not anticipate that anyone would make unbuffered logic levels available to the external world.
> In that case, I'd put a logic-level line driver immediately at the unit (by immediately, I mean with a small breakout card that plugs directly onto the LTE's MMCX connector with no intervening cable). For example, all 6 outputs of an HC14 or AC14 hex inverter connected in parallel, or a dedicated line driver chip like an HC365/366 or AC240/244/540/541.
> The buffer should be inside the enclosure with the LTE, and I would also add a T-network filter to convert the logic-level square wave into a sine wave. This would confine all of the fast logic transitions inside the shielded box, where they can do the least mischief.
> For the T-network, I like 10uH/50.5pF/10uH, others like 1.5uH/310pF/1.5uH. Both draw ~ +/- 35mA from a 5v logic output. Make sure your buffer can supply this current, and feed the T-network through 10nF and 50 ohms in series. You'll get a 1Vrms (13dBm) sine wave into 50 ohms (675mVrms with 3v logic). H3 is down 40dBc with the 1.5uH network and 60dBc with the 10uH network. [Note that the apparent source impedance is > 50 ohms, so the open-circuit voltage is more than double.]
> Best regards,
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts