Topics

Full QSK for a uBITX

John Seboldt K0JD
 

I thought I saw someone talk about a full break-in mod for a uBITX. I'm sure I could figure out at least the circuit changes to eliminate the relay, but then changing the firmware to let it happen would have to be addressed too.

Did I miss someone's posting of a way to do this?

By the way, the pop suppression stage in the v4 and v5 (Q74) works well when retrofitted to my v3 board. Photo attached.

John K0JD
Milwaukee WI

Gordon Gibby
 

What an interesting idea!

Toss in some PIN diode‘s (or even rectifier diode‘s diode’s according to Alyson ) instead of the crucial antenna switching relay, rewrite the code so the band switching portions of the relays stay put until you change bands, and you’d have it. (I think off the top of my head)

I have always bemoaned the fact that I let my HW 16 go away, this might be a very cool project when I retire in just a few months

On Aug 5, 2019, at 21:23, John Seboldt K0JD <k0jd-l@...> wrote:

I thought I saw someone talk about a full break-in mod for a uBITX. I'm sure I could figure out at least the circuit changes to eliminate the relay, but then changing the firmware to let it happen would have to be addressed too.

Did I miss someone's posting of a way to do this?

By the way, the pop suppression stage in the v4 and v5 (Q74) works well when retrofitted to my v3 board. Photo attached.

John K0JD
Milwaukee WI




<q74.jpg>

MadRadioModder
 

It can be done using diodes.

-----Original Message-----
From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of John Seboldt K0JD
Sent: Monday, August 5, 2019 8:24 PM
To: BITX20@groups.io
Subject: [BITX20] Full QSK for a uBITX

I thought I saw someone talk about a full break-in mod for a uBITX. I'm sure I could figure out at least the circuit changes to eliminate the relay, but then changing the firmware to let it happen would have to be addressed too.

Did I miss someone's posting of a way to do this?

By the way, the pop suppression stage in the v4 and v5 (Q74) works well when retrofitted to my v3 board. Photo attached.

John K0JD
Milwaukee WI






---
This email has been checked for viruses by AVG.
https://www.avg.com



--
…_. _._

Tom, wb6b
 

On Mon, Aug 5, 2019 at 06:28 PM, Gordon Gibby wrote:
Toss in some PIN diode‘s
It looks like PIN diodes have the advantage that the amount of current you need to keep them fully conductive through the full cycle of the RF current is significantly less. 

With a regular diode you have to pump enough current through the diode to counter the RF current at the "negative" peak of current. With a PIN diode the current does not need to counter the current of the negative half of the RF current flow because the nature of the PIN diode is, once forward based into the conductive state, to remain in the conductive state for microseconds or more. 

Here is an interesting PDF I ran across.
https://www.qsl.net/va3iul/RF_Switches/RF_Switches.pdf

One thing I wonder about is if some of the mosfet output solid state "relays" might work. I've seen the small lower current DIP sized "relays" used to switch analog signals up to several Mhz. Such as the low cost scopes being discussed in another thread. I looked at a 2 amp version and it had 260pf of off state capacitance. That is likely excessive for anything higher than 80 meters or so. But, maybe there are better parts.

Here is the part I looked at:
https://toshiba.semicon-storage.com/ap-en/product/opto/photocoupler/detail.TLP3555.html

Here is one of the parts being used as a analog switch in an 8Mhz scope:
http://www.ixysic.com/home/pdfs.nsf/www/CPC1017N.pdf/$file/CPC1017N.pdf

I think this is the one used in the $28 1Mhz scope:
http://www.ixysic.com/home/pdfs.nsf/www/CPC1008N.pdf/$file/CPC1008N.pdf

Tom, wb6b




Tom, wb6b
 

Jerry Gaffke
 

Here's an old discussion about some possible improvements to the uBitx for CW:
    https://groups.io/g/BITX20/message/36947

Jerry, KE7ER

Tom, wb6b
 

On Mon, Aug 5, 2019 at 09:35 PM, Jerry Gaffke wrote:
Here's an old discussion about some possible improvements to the uBitx for CW
Following your thread, this article listed was very interesting.

http://wa5bdu.blogspot.com/2017/03/pin-diode-tr-switch.html

Article implies some common rectifier diodes are PIN diodes in disguise. At 50 or 60 HZ a few extra microseconds of reverse conduction would be negligible. Wonder if being PIN diodes is actually the case. I shall commence googling...

Tom, wb6b

MadRadioModder
 

It’s easy.  Look at this example of the front end of the Elecraft K3.  You just need to DC bias the diodes above the lowest AC cycle plus the diode voltage drop (0.6V in the case of the BAV99’s and the 64-04’s they use here).  Note the 7T and 7R (7 volts for Transmit or Receive) used for biasing the diode switches… and the PRE_ON and PRE_OFF voltages (again, 7 volts) to pass or block the  preamp Q4.

 

 

 

 

 


Virus-free. www.avg.com

--

…_. _._

MadRadioModder
 

Oh yeah! For selecting the crystal switches too…

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of MadRadioModder via Groups.Io
Sent: Tuesday, August 6, 2019 12:56 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Full QSK for a uBITX

 

It’s easy.  Look at this example of the front end of the Elecraft K3.  You just need to DC bias the diodes above the lowest AC cycle plus the diode voltage drop (0.6V in the case of the BAV99’s and the 64-04’s they use here).  Note the 7T and 7R (7 volts for Transmit or Receive) used for biasing the diode switches… and the PRE_ON and PRE_OFF voltages (again, 7 volts) to pass or block the  preamp Q4.

 

 

 

 

 

 

Virus-free. www.avg.com


--

…_. _._


--

…_. _._

Tom, wb6b
 

On Mon, Aug 5, 2019 at 10:56 PM, Tom, wb6b wrote:
I shall commence googling...
Interesting. My eyes are blurry from staring at my computer screen. Evening looking at old datebooks from the 60s and 70s (on webarchive), when they still may cared about something as quaint as a diode. Finally found this one could be accurate little trinket that has been gathering digital dust for 16 years in the corner of some hard drive on the web. Surely there must be more, but this is what I've found so far. 

---- Cut ----
https://elecraft.mailman.qth.narkive.com/jwRdNb1r/1n4007-diode-question#post9

Hi
 
The 1n4000 series of diodes are general purpose rectifier diodes designed
for low frequency use primarily 60 Hz power supplies. The construction is
the same for the series except the 1n4007. The lower voltage diodes are
basic PN junctions and would not work as a RF switch. The 1n4007 however has
an extra intrinsic (undoped) layer forming the P-I-N junction. This is, I
would guess, is to get the 1Kv PIV rating but it also allows the 1n4007 to
work as a RF switch like a true PIN diode The RF characteristics are not
specified on this diode because it is not designed to do this type of work
it just happened to work out that way and they are very cheap. True PIN
diodes work better than the 1n4007 but are more expensive and not as readily
available. Elecraft has designed the RF switching to work with the 1n4007
and making substations of different diodes may not work. Other 1n4000 series
defiantly do not have the PIN structure so they will not work.
 
Don Brown
KD5NDB
---- Cut ----

Tom, wb6b

John Seboldt K0JD
 

A lot of you are posting some good things, some of which I remember. The 1N4007 working as a PIN diode is well known to me - I duplicated such a circuit from a Ten Tec Argosy which is 50 watts. I got a bunch of 200 volt PIN diodes to try to do something for my 100 watt amplifier project - hope that will be enough.

Obviously the circuitry is easy to apply - and we can learn to change the firmware.... Even without the firmware change it might be nice to have the relay removed for a nice silent TR switch.

On 8/5/2019 20:23, John Seboldt K0JD wrote:
I thought I saw someone talk about a full break-in mod for a uBITX. I'm sure I could figure out at least the circuit changes to eliminate the relay, but then changing the firmware to let it happen would have to be addressed too.

Did I miss someone's posting of a way to do this?

By the way, the pop suppression stage in the v4 and v5 (Q74) works well when retrofitted to my v3 board. Photo attached.

John K0JD
Milwaukee WI


Jerry Gaffke
 

It's easy to think up a solution that should work.
But this sort of thing can take weeks to get to work well.
Or it could just work.

What's easy is to get discouraged from playing with this stuff
when others with more experience or perhaps hubris
tell me how easy its is. 
And it turns out to be a bit more complicated than assumed.

For example, this from post   https://groups.io/g/BITX20/message/32099

#####################s
I'd like to do a uBitx like rig without relays as they seem large and unreliable, use analog switches and PIN diodes instead, perhaps 1n4007's for the TR switch.   Lots of folks claiming success with the 1n4007 in that position, including "ElectronicAntenna Switching" by Wes Hayward in QEX, May 1995.  But worth looking at some of the rants here:    http://forums.qrz.com/index.php?threads/looking-for-a-pin-diode.147454/page-2

Specifically this one:

"Since the 1N4007 has been made by probably fifty different companies over the years, and is offered in epoxy, glass, ceramic packages with square, round, and hexagonal die, and with die geometry from less than .040 square to larger than .090 round (and everything between), and may be fabricated by single diffusion, double diffusion, epitaxial, suprataxial or just about any other method a manufacturer chooses, nobody having a 1N4007 really has any idea what he has.The 1N4007 is also available as a die alone, without any package, in several geometries and diffusions.Since its Cj and trr are unspecified, they vary a lot from fabrication to fabrication. The really good 1N4007s made 30 years ago by Motorola in Scottsdale are long gone and now they're all made offshore, to absolutely no standards.This is why I wouldn't generalize about using a 1N4007 as a PIN substitute: Some might work great this way, others absolutely won't."
########################


On Mon, Aug 5, 2019 at 10:56 PM, MadRadioModder wrote:
It’s easy.

Gordon Gibby
 

​so I'm a little ignorant here -- if you know an inexpensive and useful PIN part no., this would be great to publicize here for the ignorant like me....


cheers

gordon



From: BITX20@groups.io <BITX20@groups.io> on behalf of Jerry Gaffke via Groups.Io <jgaffke@...>
Sent: Tuesday, August 6, 2019 11:26 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Full QSK for a uBITX
 
It's easy to think up a solution that should work.
But this sort of thing can take weeks to get to work well.
Or it could just work.

What's easy is to get discouraged from playing with this stuff
when others with more experience or perhaps hubris
tell me how easy its is. 
And it turns out to be a bit more complicated than assumed.

For example, this from post   https://groups.io/g/BITX20/message/32099

#####################s
I'd like to do a uBitx like rig without relays as they seem large and unreliable, use analog switches and PIN diodes instead, perhaps 1n4007's for the TR switch.   Lots of folks claiming success with the 1n4007 in that position, including "ElectronicAntenna Switching" by Wes Hayward in QEX, May 1995.  But worth looking at some of the rants here:    http://forums.qrz.com/index.php?threads/looking-for-a-pin-diode.147454/page-2

Specifically this one:

"Since the 1N4007 has been made by probably fifty different companies over the years, and is offered in epoxy, glass, ceramic packages with square, round, and hexagonal die, and with die geometry from less than .040 square to larger than .090 round (and everything between), and may be fabricated by single diffusion, double diffusion, epitaxial, suprataxial or just about any other method a manufacturer chooses, nobody having a 1N4007 really has any idea what he has.The 1N4007 is also available as a die alone, without any package, in several geometries and diffusions.Since its Cj and trr are unspecified, they vary a lot from fabrication to fabrication. The really good 1N4007s made 30 years ago by Motorola in Scottsdale are long gone and now they're all made offshore, to absolutely no standards.This is why I wouldn't generalize about using a 1N4007 as a PIN substitute: Some might work great this way, others absolutely won't."
########################


On Mon, Aug 5, 2019 at 10:56 PM, MadRadioModder wrote:
It’s easy.

Ashhar Farhan
 


I just wanted to clear up on a concept. The PIN diodes that are used for T/R are the ones that CANNOT switch at RF frequencies. That's precisely why they are used. So, let me explain this ...
An RF charge across it travels through precisely because the junction cannot switch on and off so rapidly and hence it remains on (or off). A PIN diode is like that, it can handle large currents, but it cannot change the current fast enough. That's why for switching 100 watts of RF, you don't need a few amps of bias current. a few milliamps will do.
There are ways of switching T/R with even 1N4148s (for QRP work). W7ZOI has, as usual, been there and done that. Read this paper to understand the concept : http://w7zoi.net/tr-qrp.pdf

- f

On Tue, Aug 6, 2019 at 9:09 PM Gordon Gibby <ggibby@...> wrote:

​so I'm a little ignorant here -- if you know an inexpensive and useful PIN part no., this would be great to publicize here for the ignorant like me....


cheers

gordon



From: BITX20@groups.io <BITX20@groups.io> on behalf of Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io>
Sent: Tuesday, August 6, 2019 11:26 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Full QSK for a uBITX
 
It's easy to think up a solution that should work.
But this sort of thing can take weeks to get to work well.
Or it could just work.

What's easy is to get discouraged from playing with this stuff
when others with more experience or perhaps hubris
tell me how easy its is. 
And it turns out to be a bit more complicated than assumed.

For example, this from post   https://groups.io/g/BITX20/message/32099

#####################s
I'd like to do a uBitx like rig without relays as they seem large and unreliable, use analog switches and PIN diodes instead, perhaps 1n4007's for the TR switch.   Lots of folks claiming success with the 1n4007 in that position, including "ElectronicAntenna Switching" by Wes Hayward in QEX, May 1995.  But worth looking at some of the rants here:    http://forums.qrz.com/index.php?threads/looking-for-a-pin-diode.147454/page-2

Specifically this one:

"Since the 1N4007 has been made by probably fifty different companies over the years, and is offered in epoxy, glass, ceramic packages with square, round, and hexagonal die, and with die geometry from less than .040 square to larger than .090 round (and everything between), and may be fabricated by single diffusion, double diffusion, epitaxial, suprataxial or just about any other method a manufacturer chooses, nobody having a 1N4007 really has any idea what he has.The 1N4007 is also available as a die alone, without any package, in several geometries and diffusions.Since its Cj and trr are unspecified, they vary a lot from fabrication to fabrication. The really good 1N4007s made 30 years ago by Motorola in Scottsdale are long gone and now they're all made offshore, to absolutely no standards.This is why I wouldn't generalize about using a 1N4007 as a PIN substitute: Some might work great this way, others absolutely won't."
########################


On Mon, Aug 5, 2019 at 10:56 PM, MadRadioModder wrote:
It’s easy.

Jerry Gaffke
 

My point is, good luck to the casual experimenter with a DVM
trying to debug a TR switch using the wrong flavor 1n4007.
What you really need is a good fast scope and a fair idea of what
a PIN diode should be doing.

My advice: 

When you get stumped, try asking questions in the forum
after giving a full description of what you are doing.
Could well be that somebody else here has been down that same road.
Persistence often pays, and I've seen some amazingly persistent people
here solve some difficult problems through mostly force of will.
But if it's no longer fun, it's ok to put it on the shelf and move on
to another puzzle.

It's a hobby!

Jerry


On Tue, Aug 6, 2019 at 08:26 AM, Jerry Gaffke wrote:
What's easy is to get discouraged from playing with this stuff

Jerry Gaffke
 

Many experimenters report success with the 1n4007, still true today.
Perhaps somebody who knows what they are doing and who has had success
using the 1n4007 can report the vendor, manufacturer, and part number they used.
Preferably not something bought by the bucket off of ebay.

If you don't mind spending a few dollars, there are fully spec'd PIN diodes out there
that you could use in a T/R switch.  


Many times this stuff can be easy to get working.
But when it doesn't work, just keep in mind that many
awfully smart people get hung up for weeks trying to get
what looks like a really simple analog circuit to work correctly.

Jerry,  KE7ER


On Tue, Aug 6, 2019 at 08:39 AM, Gordon Gibby wrote:

so I'm a little ignorant here -- if you know an inexpensive and useful PIN part no., this would be great to publicize here for the ignorant like me....

 

Tom, wb6b
 

On Tue, Aug 6, 2019 at 08:26 AM, Jerry Gaffke wrote:
including "ElectronicAntenna Switching" by Wes Hayward 
I was able to find the Wes Hayward article on-line. 
https://archive.org/stream/QEX19812016/QEX%201995/QEX%201995-05#page/n3/mode/2up

Tom, wb6b

Tom, wb6b
 

On Tue, Aug 6, 2019 at 08:49 AM, Jerry Gaffke wrote:
What you really need is a good fast scope and a fair idea of what
a PIN diode should be doing.
Maybe a uBITX into a dummy load (possibility 100 ohms at the UBITX and another 100 ohm load after the rectifier-under-test) at 80M could be used as a "high" power signal generator to power a test fixture where you could look at the RF waveform with a scope and compare how much current is needed to feed into candidate rectifier diodes, for full conduction through the RF cycle, to determine their "PINness".

As a more crude measurement, possibility measuring the RF voltage or using some other "power" meter, like using to "forward" power reading of an SWR meter connected to the second (after the possibly PIN diode) would work. You would measure how much current you need to feed into the possibility PIN diode to get the same reading as with a jumper wire replacing the rectifier diode under test.

Tom, wb6b

John Seboldt K0JD
 

This is quite interesting, with a nice sophisticated full-tilt switch. Interestingly, the Ten Tec Argosy (50W) derives the high voltage back bias from the RF itself with some silicon diodes and filter capacitors, so providing a full high-voltage power supply is a bit of an overkill.

John K0JD

On 8/6/2019 15:01, Tom, wb6b wrote:
On Tue, Aug 6, 2019 at 08:26 AM, Jerry Gaffke wrote:
including "ElectronicAntenna Switching" by Wes Hayward 
I was able to find the Wes Hayward article on-line. 
https://archive.org/stream/QEX19812016/QEX%201995/QEX%201995-05#page/n3/mode/2up

Tom, wb6b

Jerry Gaffke
 

Here's a manual for the Argosy, T/R switching (and SWR metering) is on page 51:
    http://www.tentec.com/wp-content/uploads/2016/05/Argosy_II_Model_525_Manualf933.pdf
The full transmit RF waveform goes through C1, diode D4 ensures the bottom of that
waveform is at DC ground, diode D3 then rectifies it to create a DC voltage across C2 that
is equal to the peak-to-peak voltage of the transmitted RF.  The notes then say that this
relatively high DC voltage available during transmit is used to reverse bias diodes D1 and D2
such that  no transmit signal goes through them to the RCV connector (which then goes into
the receiver front end).

I'm still staring at it, almost have it figured out.
But D1,D2 look to me like they would be forward biased during transmit,
which would not work at all.

Note that the PIN diodes at D1,D2 are called out as 1n4007's.
And that this transceiver works up through 30mhz.

Another thing to note, the final remains connected to the antenna during receive,
shunting some of the receive energy to ground.
Some of the other T/R switches use additional PIN diodes to isolate the final
from the antenna during receive.

Jerry, KE7ER


On Wed, Aug 7, 2019 at 01:37 PM, John Seboldt K0JD wrote:

This is quite interesting, with a nice sophisticated full-tilt switch. Interestingly, the Ten Tec Argosy (50W) derives the high voltage back bias from the RF itself with some silicon diodes and filter capacitors, so providing a full high-voltage power supply is a bit of an overkill.

John K0JD

Vince Vielhaber
 

According to the voltage chart on the next page, that junction is going to have a large negative voltage on it. That would stop the transmit from going to the receiver.

Vince.

On 08/07/2019 07:35 PM, Jerry Gaffke via Groups.Io wrote:
Here's a manual for the Argosy, T/R switching (and SWR metering) is on
page 51:

http://www.tentec.com/wp-content/uploads/2016/05/Argosy_II_Model_525_Manualf933.pdf
The full transmit RF waveform goes through C1, diode D4 ensures the
bottom of that
waveform is at DC ground, diode D3 then rectifies it to create a DC
voltage across C2 that
is equal to the peak-to-peak voltage of the transmitted RF. The notes
then say that this
relatively high DC voltage available during transmit is used to reverse
bias diodes D1 and D2
such that no transmit signal goes through them to the RCV connector
(which then goes into
the receiver front end).

I'm still staring at it, almost have it figured out.
But D1,D2 look to me like they would be forward biased during transmit,
which would not work at all.

Note that the PIN diodes at D1,D2 are called out as 1n4007's.
And that this transceiver works up through 30mhz.

Another thing to note, the final remains connected to the antenna during
receive,
shunting some of the receive energy to ground.
Some of the other T/R switches use additional PIN diodes to isolate the
final
from the antenna during receive.

Jerry, KE7ER


On Wed, Aug 7, 2019 at 01:37 PM, John Seboldt K0JD wrote:

This is quite interesting, with a nice sophisticated full-tilt
switch. Interestingly, the Ten Tec Argosy (50W) derives the high
voltage back bias from the RF itself with some silicon diodes and
filter capacitors, so providing a full high-voltage power supply is
a bit of an overkill.

John K0JD