[hpsdr] Mercury during transmit.

[David McQuate]

Looking at Mercury's schematic and BOM, it seems that Mercury's 
resistance to "high" applied power may be
due to the bidirectional clamp diode, D1,  part number CBLC03C,
but other components may nevertheless be at risk.

The datasheet for D1 shows a graph of peak pulse power vs pulse duration. 
Two points on the straight line of the log-log plot are
350W for 20 microseconds, and
    35W for 2 milliseconds.
Extrapolating to longer pulse duration,
3.5W for 200 milliseconds, and
0.35W for 20 seconds.

The diodes start conducting (1 mA) at 4 volts, and clamping voltage is  
7 volts at 1A,
or  19 volts at 20A (20 microsceonds max).

At eight volts peak-to-peak at Mercury's input, D1 will be just barely 
conducting.
The input power from a 50 ohm source would be 160mW.
If the 20dB attenuator is in the circuit ("preamp off") its input 
resistor, R50, will absorb 99% of the
incident power.  It is a 0603 size SMT resistor.  Typical power ratings 
are 60 to 100 mW.
So the limiting component seems to be the attenuator, unless it is 
switched out.

If the attenuator is out of the circuit ("preamp on"), the LTC6400-20 
preamp will receive whatever
power makes it past D1.  The preamp has its own built-in clamp diodes 
connected to the +3.3v
supply and ground.  These diodes are rated 10 mA max.  Input voltage 
just slightly larger than 3.3 + 0.7,
(that is,  +- 4 volts) will result in rms current larger than 10mA, so, 
again, 160 mW seems the upper safe limit.

While we have stories of survival at higher power, it seems to me 
prudent to never apply more than
60 to 100 (maybe 160) mW.

Other than relays, it might be possible to use transistors (BJT or FET) 
or forward- and reverse-biased diodes
to provide the needed protection.

73,
Dave
wa8ywq





Richard Ames wrote:
> ***** High Performance Software Defined Radio Discussion List *****
>
> Mercury is reported to be a bit more rugged than implied below; see
> http://openhpsdr.org/wiki/index.php?title=Mercury_-_Frequently_Asked_Questions and the references quoted.
>
> Of course the relays are still good / necessary.
>
> Regrds, Richard.
>   
>

 1251762701.0