Topics

Solid-state T/R switching #ubitx


Ted
 

I've thought of doing away with the physical relay switching between TX & RX, by using a power transistor to activate the RX whenever TX is not otherwise invoked, and another one to activate the TX path (and drop the RX switch device) when called upon. 

It sounds plain enough. Has anyone else done this already?  I'm not necessarily trying to address anything that I'd call a shortcoming, other than that it IS a mechanical switch [and "because I think I can"].

For the rig in question - a V5 board - I use almost exclusively CW and I believe I can get better semi-qsk this way and lose the relay noise at the same time. More instantaneous reaction to Dits would be a
win, not that I think this build with its CEC firmware necessarily suffers so badly from that like the stock V6  does, but I'd like to see what effect such switching has on operation.

Comments?

Ted
K3RTA


Bill Cromwell
 

Hi Ted,

Your idea is not new and there are a lot of radios that use solid state switching. You can implement QSK with electronic switching. "Semi-breakin" really means "not breakin":) I have been considering that too but in a different radio that uses a mechanical bandswitch.

You will have to consider how you are are going to select the appropriate low pass filters if you want to eliminate all of the relays. That is not meant to say nay nor to discourage you. You might keep the relays controlled by the raduino for low pass filter selections and just use the electronic switching for the T-R changeover. I may do the same thing. But it is also possible to eliminate those relays. A lot more thought and mechanical work.

In the radio I am making plans for bandswitch is already designed, built. and operational. It is not needed for the T-R function. You could just do that with your uBitX. Keep us posted.

73,

Bill KU8H

bark less - wag more

On 10/17/20 3:36 PM, Ted via groups.io wrote:
I've thought of doing away with the physical relay switching between TX & RX, by using a power transistor to activate the RX whenever TX is not otherwise invoked, and another one to activate the TX path (and drop the RX switch device) when called upon.
It sounds plain enough. Has anyone else done this already?  I'm not necessarily trying to address anything that I'd call a shortcoming, other than that it IS a mechanical switch [and "because I think I can"].
For the rig in question - a V5 board - I use almost exclusively CW and I believe I can get better semi-qsk this way and lose the relay noise at the same time. More instantaneous reaction to Dits would be a
win, not that I think this build with its CEC firmware necessarily suffers so badly from that like the stock V6  does, but I'd like to see what effect such switching has on operation.
Comments?
Ted
K3RTA


Ted
 

Thanks, Bill.

No, I'm not trying to eliminate relays from the board, only the function of the T/R so that I can get the recovery for "semi" QSK a little faster.... and it is semi-QSK if I don't have to manually move a lever or a knob from TX to RX in between words.  If I can get that between letters, that would sure be nice as a sunny day.  

I'd like to volunteer that, in no way am I implying commentary upon the original design; I'm just carrying on with the intent of the kit by experimenting and modifying to suit.  

I've had one of the original T/R relays in a V3 die on me, and an Axiom T/R relay die on me in a V5, in far fewer than the tens of thousands of cycles these are typically rated for.  So, of course, accepting in advance that CW is not the primary role of these radios, I'm looking at making my portable box (and likely a Bitx-40 board I came across as well) into something that is more suited for the quick-recovering and wear-free picture of a true CW rig.   The reason for posting is that of wondering if I'm on the right track with something like FET's switching on in opposition to each other per TX/RX duty, rather than the usual PIN diode route, due to this design sharing so much of it's core for both TX and RX paths.  

Tnx, 

Ted 
K3RTA

 


Bill Cromwell
 

Hi Ted,

The changeover has to 'disconnect' the antenna from the receiver and hopefully ground the receive antenna terminal. The PIN diode scheme is really good for that. In the past I have used and active switch.

The T-R changeover also should apply/remove power (at least the bias) on the transmit finals. The VFO has to be applied to the transmitter or receiver as appropriate. The receiver should have it's gain drastically reduced if not completely muted. I think some of that is already done in solid state switching. I am too lazy to look right now.

Of course the uBitX is intended as a starting place for us to make our own radio. I think your project is worthwhile. I have taken my own measures to get usable CW operation from my V3. When I complete the work on my other transceiver (100 watt solid state) I will probably move to upgrade my V3 to QSK. What you call semi-QSK is more accurately automatic T-R changeover. In my Kenwood that works of the VOX system and uses the sidetone to trip the VOX. The sidetone is not transmitted as in some SDR radios.

My PFR-3 has real QSK:) I use it instead of the uBitX for the three bands it covers.

73,

Bill KU8H

bark less - wag more

On 10/18/20 8:22 AM, Ted via groups.io wrote:
Thanks, Bill.
No, I'm not trying to eliminate relays from the board, only the function of the T/R so that I can get the recovery for "semi" QSK a little faster.... and it is semi-QSK if I don't have to manually move a lever or a knob from TX to RX in between words.  If I can get that between letters, that would sure be nice as a sunny day.
I'd like to volunteer that, in no way am I implying commentary upon the original design; I'm just carrying on with the intent of the kit by experimenting and modifying to suit.
I've had one of the original T/R relays in a V3 die on me, and an Axiom T/R relay die on me in a V5, in far fewer than the tens of thousands of cycles these are typically rated for.  So, of course, accepting in advance that CW is not the primary role of these radios, I'm looking at making my portable box (and likely a Bitx-40 board I came across as well) into something that is more suited for the quick-recovering and wear-free picture of a true CW rig.   The reason for posting is that of wondering if I'm on the right track with something like FET's switching on in opposition to each other per TX/RX duty, rather than the usual PIN diode route, due to this design sharing so much of it's core for both TX and RX paths.
Tnx,
Ted
K3RTA


Curt
 

Indeed solid state switching would be nice.  An example of this is the K2 (the schematic is available).  The K2 does this so well because it employs a -90 volt bias on the switching diodes.  How does it do that?  Interesting - it takes an xtal oscillator (at an out of ham band frequency), rectifies it and multiplies it up in voltage using a lot of diodes.  It may take some effort to find good PIN switching diodes or perhaps use 1n4007 like the QCX PA.  I imagine somewhere there is a ubitx that already has solid state switching.  BTW, many pricey appliance rigs do use relays. 

73 Curt


Bob Lunsford
 

In the N7DDC 100W ATU, the switching transistor used is a 2N7002.

Here is the schematic...

Inline image



On Sunday, October 18, 2020, 9:44:07 AM EDT, Curt via groups.io <wb8yyy@...> wrote:


Indeed solid state switching would be nice.  An example of this is the K2 (the schematic is available).  The K2 does this so well because it employs a -90 volt bias on the switching diodes.  How does it do that?  Interesting - it takes an xtal oscillator (at an out of ham band frequency), rectifies it and multiplies it up in voltage using a lot of diodes.  It may take some effort to find good PIN switching diodes or perhaps use 1n4007 like the QCX PA.  I imagine somewhere there is a ubitx that already has solid state switching.  BTW, many pricey appliance rigs do use relays. 

73 Curt


MadRadioModder
 

I’ve been using BAP64’s for RF switching… they are super cheap… about 27 cents and only take about 7-8 volts to switch.  They are good to about 24 dbm or so (250 mW).  But you can even use 1N4004’s for RF switching at moderate power…

 

From: BITX20@groups.io <BITX20@groups.io> On Behalf Of Ted via groups.io
Sent: Sunday, October 18, 2020 7:23 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Solid-state T/R switching #ubitx

 

Thanks, Bill.

No, I'm not trying to eliminate relays from the board, only the function of the T/R so that I can get the recovery for "semi" QSK a little faster.... and it is semi-QSK if I don't have to manually move a lever or a knob from TX to RX in between words.  If I can get that between letters, that would sure be nice as a sunny day.  

I'd like to volunteer that, in no way am I implying commentary upon the original design; I'm just carrying on with the intent of the kit by experimenting and modifying to suit.  

I've had one of the original T/R relays in a V3 die on me, and an Axiom T/R relay die on me in a V5, in far fewer than the tens of thousands of cycles these are typically rated for.  So, of course, accepting in advance that CW is not the primary role of these radios, I'm looking at making my portable box (and likely a Bitx-40 board I came across as well) into something that is more suited for the quick-recovering and wear-free picture of a true CW rig.   The reason for posting is that of wondering if I'm on the right track with something like FET's switching on in opposition to each other per TX/RX duty, rather than the usual PIN diode route, due to this design sharing so much of it's core for both TX and RX paths.  

Tnx, 

Ted 
K3RTA

 


Virus-free. www.avg.com

--

…_. _._


Ashhar Farhan
 

With a 2N7000 you can handle upto 10 watts.


On Tue 20 Oct, 2020, 9:12 AM MadRadioModder, <madradiomodder@...> wrote:

I’ve been using BAP64’s for RF switching… they are super cheap… about 27 cents and only take about 7-8 volts to switch.  They are good to about 24 dbm or so (250 mW).  But you can even use 1N4004’s for RF switching at moderate power…

 

From: BITX20@groups.io <BITX20@groups.io> On Behalf Of Ted via groups.io
Sent: Sunday, October 18, 2020 7:23 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Solid-state T/R switching #ubitx

 

Thanks, Bill.

No, I'm not trying to eliminate relays from the board, only the function of the T/R so that I can get the recovery for "semi" QSK a little faster.... and it is semi-QSK if I don't have to manually move a lever or a knob from TX to RX in between words.  If I can get that between letters, that would sure be nice as a sunny day.  

I'd like to volunteer that, in no way am I implying commentary upon the original design; I'm just carrying on with the intent of the kit by experimenting and modifying to suit.  

I've had one of the original T/R relays in a V3 die on me, and an Axiom T/R relay die on me in a V5, in far fewer than the tens of thousands of cycles these are typically rated for.  So, of course, accepting in advance that CW is not the primary role of these radios, I'm looking at making my portable box (and likely a Bitx-40 board I came across as well) into something that is more suited for the quick-recovering and wear-free picture of a true CW rig.   The reason for posting is that of wondering if I'm on the right track with something like FET's switching on in opposition to each other per TX/RX duty, rather than the usual PIN diode route, due to this design sharing so much of it's core for both TX and RX paths.  

Tnx, 

Ted 
K3RTA

 


Virus-free. www.avg.com

--

…_. _._