NEVER use VOX operation when using RF Power Amplifiers.

It should be obvious why.

The transmitter, and the PA will switch on after audio modulation is detected.
You will certainly have a hot-switch condition with the PA T/R relay when using VOX on the transceiver.

It is assumed serious operators will always use proper keyed PTT lines between computer, transceiver and PAs.

 
Multi-Send Design Considerations

There is a minimum 29mS delay from when the PTT is keyed, before the analogue band voltages can be read by the microcontroller.
It actually takes 32-33mS for the 9V TX rails to fully ramp up to 9V, but given the windows I have designed in, we can measure these @ 29 mS after the PTT is activated.

msend_rise
Above: Yellow = PTT line, Green = VT9V rail. 1 mS per minor Hor division

The COTO reed relays used in Multi-Send have a 0.5mS actuation specification (including bounce time).
They’re very quick!

Therefore, the appropriate SEND relay output has switched by 29.5 mS after the IC-910’s SEND line is activated.

 
Timeline Captures

As I only have a 2-channel DSO, the following captures are:
green trace = VT9V rail (consistent with above capture)
Yellow trace = 144 MHz RF output envelope

Transmitting @ 144 MHz full power (100W) into a Cellwave DC-2 GHz 250W dummy load, with a coupled detector output.

Ideally, a 3-channel DSO would be nice to show PTT in the timeline as well!
Bear in mind, the PTT would have been pressed 32 mS before the green VT9V trace fully ramps up.

wspr_vox
Above: Running WSPR using transceiver VOX to key.
VOX is not a good idea when using RF Power Amplifiers.

 

wspr_dtr
Above: WSPR using DTR (serial port) to key
RF out appears 1050 mS after the Multi-Send relay switch

 

wsjt_dtr
Above: WSJT using DTR (serial port) to key.
RF out appears 1530 mS after the Multi-Send relay switch

 

Cudos to Joe Taylor for the generous allowance for T/R switch times in his software.
There is plenty of time for preamplifier and power amplifier switching to take place in an orderly and sequenced manner.

 

digipan_dtr
Above: DigiPan using DTR (serial port) to key.
RF out appears 32 mS after the Multi-Send relay switch

 

cwtype_dtr
Above: CW Type using DTR (serial port) to key. (The “C” in CQ is clearly seen).
RF out appears 250 mS after the Multi-Send relay switch

 

paddle_bkin
Above: CW using a Bencher Paddle, IC-910 set to Break-In: on.
Note time scale: 400 uS per minor Hor division.
RF out envelope starts to ramp up ~ 0.5 mS after the Multi-Send relay switch

For a longer delay before any CW RF appears, operators can activate the Transmit button before using their favorite paddle, otherwise the IC-910 in CW tends to behave like SBB using VOX.

 

For SSB contacts, we all know not to start talking until after the PTT has been pressed – so no capture required!

 

FAQ

Q)
What’s the advantage of using the ICOM Multi-Send Kit, vs using the voltage in the coax that switches the ICOM preamps.
As far as I can tell the voltage it timed to prevent hot switching of the preamps relays, so tapping this off for a amp should work.

A)
That question came up at GippsTech.

If you want to retain the ability to selectively use pre-amplifiers as well, then you cannot go down that path.
Also, it would be an extremely undesirable situation to be switching a RX Preamp and PA off the same control line, irrespective of where that change of state was in the timeline.

The Preamp switch should occur earlier in the timeline, whereas the PA enable should occur after, like it does here using the Multi-Send.
From the DSO timing captures above, we can see that a PA can be enabled xx milliseconds after entering TX mode, and before any RF comes out…. and well and truly after the RX Preamp volts have disappeared.