Date   
Re: BITX40 output power at 36 volts to the IRF510 #bitx40 #bitx40help

Timothy Fidler
 

Jerry .. bit  of confusion. He is talking re uBitx40 and you are not.

Re: End Fed antennas w/ uBITX #ubitx

iz oos
 

I wired my first bad copy of a bad copy 1:49 transformer. I have to say that I am puzzled. I used a high permeability NiZn core. Measured inductance without the capacitor is less than 8uH, and 380uH at the secondary. Please I would be interested in knowing these values of a good copy. Then I installed a 100pF high voltage capacitor at the primary and then I measured swr of  with the antenna analyzer without any load. Surprisingly the SWR was between 3.5 to 7 across the ham bands. All other transmission line transformers I had previously built show infinite or not measurable SWR as it should be. What are the SWR values you find in your good copy without any load?


Il 04/ago/2018 16:33, "ajparent1/KB1GMX" <kb1gmx@...> ha scritto:
Iz oos,

TLTs can be made with low loss.    However making a matched set may be harder
as they have to have the exact same phase response.

Also the 1:9 is 450 ohms  (50*9).

Allison

Re: One question only...

Jerry Gaffke
 

Arv,

I like it, nice and simple.

I'd consider a 50mhz LPF up front, another mixer, mmic amp, and one of these 86.85mhz SAW filters:
    https://wireless.murata.com/RFM/data/px1002.pdf
Then no worries about those harmonics.

Fully analyzing the audio tones using an FFT on an ARM processor should give
sufficient resolution for stuff like evaluating phase noise.

The simple direct conversion mixer means you get both sidebands.
But that software on the ARM could figure out which audio tones go up and which go down
as your VFO moves, and thus separate out high from low sidebands.

I've been thinking of building something like this for a couple years,
have not yet tried.
 
Jerry, KE7ER


On Sun, Aug 5, 2018 at 12:54 PM, Arv Evans wrote:
The attached drawing shows the general idea from what I can remember.
 

Re: End Fed antennas w/ uBITX #ubitx

ajparent1/KB1GMX
 

Pretty good sign the core loss  is very high.  Its likely it does not match well at ~2500 ohms

I have never measured any for inductance as its like measuring it with a DC ohmeter,
not very informative.

Those I have tend be very high SWR no load high is greater than 20:1 at the low end
and at least 10:1 at the high end for a low power unit near me.  The problem is that's a
marginally informative as the loss can still be high or the ability to match is poor.

I can build a monoband L network in minutes and be far more sure it will be low
loss. 

Allison

Re: One question only...

Jerry Gaffke
 

Of course, my suggestions make it not so simple.
And somehow anything capable of phase noise measurements
must also deal with a very powerful carrier, not trivial.



On Sun, Aug 5, 2018 at 01:19 PM, Jerry Gaffke wrote:
Arv,

I like it, nice and simple.

I'd consider a 50mhz LPF up front, another mixer, mmic amp, and one of these 86.85mhz SAW filters:
    https://wireless.murata.com/RFM/data/px1002.pdf
Then no worries about those harmonics.

Fully analyzing the audio tones using an FFT on an ARM processor should give
sufficient resolution for stuff like evaluating phase noise.

The simple direct conversion mixer means you get both sidebands.
But that software on the ARM could figure out which audio tones go up and which go down
as your VFO moves, and thus separate out high from low sidebands.

I've been thinking of building something like this for a couple years,
have not yet tried.
 
Jerry, KE7ER

Re: Harmonic performance - SSB vs CW

Timothy Fidler
 

AJ Parent.. did you see my request for the  Msg on short form descrip of  work necessary to get 12 W out of Bit x 40 sent back to me .. Ie I had no record (silly me) as I was going to send it on. You may have missed that email from me in all the noise.

2. Resistors R97 and 98 near the IRFs : 47 ohm each. Has someone like you measured the VD across these at quiescent and back calculated the bias current.?? The word on the street - Hendricks kits build book (IRF510 x 2, 12W out and Plenty of drive too via BS170s ) as prev quoted - says it needs to be 50mA min per device. My suspicion is it is set  light on to ensure the Vreg does not give up.. perhaps at 35-40 mA. It might need to be higher than even 50 mA  for this SMPS Mosfet above 20 Mhz to stop clipping at change over during the RF cycle. If the bias is light on and can't be improved.. (obviously the light on bias regulator is an issue).. then the RD06HHF1## may be the only solution.  The Pinout does not work as a replacemen but the only pin that gives real trouble is the Source pin,  which would have to go directly  the the RF transformer winding in each case. This part sells for about $US  5 a pair on Ebay out of PRC. (and is not hugely popular so may well be not worth faking).  RF Parts in Calif has them for around $4.70 each.  NB There is nothing on the Mitsi datasheets re bias requirements for this in class B operation.

##  Caution this part may be very skittish in setting up the bias - and (2) it might also be over driven by the current driver set . Latter issue could be fixed with a through hole resistor wired Gate to drain. I note in passing that the IRFz24 which is known as  proven RF kludge mosfet (SMPS is the intended use) has 3x the transconductance of the IRF 510 so it may respond better to light-on  bias. The downside -   but also 3 x the nominal input capacitance so it is going to be harder to drive. So with a change out of driver transistors to say 2n2219A metal case and reduction in the emitter resistors in those four drivers " etc ". However it is cheap,  and it goes pin for pin with the incumbent. My minor and uneducated thoughts for this fine morning - Nz time.

Re: rotary encoder without detents

Dennis Yancey
 

Jim is correct on those encoders. I have tried the ones with and without detents. Both work, but I prefer the ones without derents for the uBitx. I have also used these successfully. 

https://www.ebay.com/itm/Rotary-Encoder-Module-Brick-Sensor-360-Degree-Rotation-Encoder-Module/202017629993?hash=item2f09306729%3Ag%3AyfIAAOSwiqtZjTJ~&_sop=1&_sacat=0&_nkw=rotary+encoder&_from=R40&rt=n


72 and God bless
KD4EPG

Re: rotary encoder without detents

Lee
 

I already had some of those cheap eBay encoders for playing with arduinos.  I took it off the board, opened it up, and removed the detents.   I tried it with the detents first and did not like it. My radio layout does not allow connecting a USB to the raduino unless I take it out.   Bad planning.  I made up an encoder on 6" wired terminated in a DuPont connector so I can check out the raduino while it is connected to my computer.
--
Lee - N9LO  "I Void Warranties"

 

Re: Harmonic performance - SSB vs CW

Gordon Gibby
 

Excellent excellent information thanks, Allison!!!


On Aug 5, 2018, at 15:39, ajparent1/KB1GMX <kb1gmx@...> wrote:

The fix may be a unobtainium (in India) relay.  Only about 2$ each.

I already mentioned there were Omron relays that are far better....

However what the board lacks that would help greatly is top pour that
are grounded areas on the top with many vias "stitching" them to the
bottom to prevent coupling to adjacent lines.  Also the line are not
"50 ohm characteristic over .060 g10/fr4 epoxy glass.  Why is that
important coupling to adjacent lines.  That and general routing like the 
PA output line going past the 10M filter without a copper pour.

For testing cutting the top return line and using a wire on the backside
is as good as coax... it helps but the contacts internally don't help with
the stock relay.  Did that for one filter as that can prove it and it helps
a little (1-2db at 80m. So I added a layer of Cu foil to the around and top of
each relay and grounded that, only a small DB change at 80m but
nearly 6db at 10M.  Small things add to help.

I broke the DC (TX+) line to the realys and added a choke and that 
improved things by a DB... not much but shows interaction.  The
relays share a DC common so bypass at each one underneath
1DB improvement.  Small but also shows interaction.

I keep saying this requires a  is a system level fix.  Many small
things combined not any one part.  you cannot treat it like you have
a defective relay and replacing it makes the world better.

Allison

Re: One question only...

Arv Evans
 

Jerry

This thing started out as an attempt to see if a simple receiver could be made to work
without any tuned front-end.  It does work, but received a lot that I did not want to hear.
That experiment then morphed into a fixed-gain receiver for evaluating received signal
levels.  Eventually rampant modifications brought it to this stage of development.  There
is still much more that can be done with the idea.

Your idea of directionalizing frequency movement is good, and is already being used to
determine the N-value of harmonics (2nd moves at twice frequency changes, 3rd at
3-times, and so on). 

Method of operation at present is to note level of spurious signal, then add attenuation
until the primary signal is at the same level.  Reading on the attenuator is then assumed
to be the level difference between primary and spurious.

Arv
_._


On Sun, Aug 5, 2018 at 2:19 PM Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
Arv,

I like it, nice and simple.

I'd consider a 50mhz LPF up front, another mixer, mmic amp, and one of these 86.85mhz SAW filters:
    https://wireless.murata.com/RFM/data/px1002.pdf
Then no worries about those harmonics.

Fully analyzing the audio tones using an FFT on an ARM processor should give
sufficient resolution for stuff like evaluating phase noise.

The simple direct conversion mixer means you get both sidebands.
But that software on the ARM could figure out which audio tones go up and which go down
as your VFO moves, and thus separate out high from low sidebands.

I've been thinking of building something like this for a couple years,
have not yet tried.
 
Jerry, KE7ER


On Sun, Aug 5, 2018 at 12:54 PM, Arv Evans wrote:
The attached drawing shows the general idea from what I can remember.
 

Re: rotary encoder without detents

Jack Purdum
 

I slide the other way. There's a net I check into every once and a while that meets on a specific frequency. Using my FT1200, it takes me several tries to nail the precise frequency since the tuning dial is so sensitive to the touch. With my µBITX, I nail it on the first try. I have good quality (Bourne) encoders without detents and several optical encoders, but I really like the encoders with the  "speed bumps" cuz I can set it more easily.

Jack, W8TEE

On Sunday, August 5, 2018, 4:36:40 PM EDT, Dennis Yancey <Kd4epg@...> wrote:


Jim is correct on those encoders. I have tried the ones with and without detents. Both work, but I prefer the ones without derents for the uBitx. I have also used these successfully. 

https://www.ebay.com/itm/Rotary-Encoder-Module-Brick-Sensor-360-Degree-Rotation-Encoder-Module/202017629993?hash=item2f09306729%3Ag%3AyfIAAOSwiqtZjTJ~&_sop=1&_sacat=0&_nkw=rotary+encoder&_from=R40&rt=n


72 and God bless
KD4EPG

Re: One question only...

F1BFU - Fr - 79
 

Many thanks Arv.
I will try in this way and I will keep you informed.

Gilles F1BFU / FR

Le dim. 5 août 2018 à 21:54, Arv Evans <arvid.evans@...> a écrit :
Giles  F1BFU

I am not close to my hamshack to check for accuracy. 
The attached drawing shows the general idea from what I can remember.
Sweepy.png
Looking at this reminded me that with very little change it could be made
into a CW or swept frequency signal generator (Si5351a out through mixer
and attenuator, with audio into mixer to modulate carrier).  Every time I look
at this it seem to suggest yet another modification. 

Arv
_._



On Sun, Aug 5, 2018 at 12:13 PM Gilles Delpech <gilles.f1bfu@...> wrote:
Hi Arv

your experimentation interests me. I'm finishing building QRPLabs VFO / SigGen with Relay Card to switch LPFs. I also have a K5BCQ attenuator. I would like to understand your use of the QRP Labs Signal Generator.
Please can you give me a little diagram of your editing ?

Many thnaks Arv

Le dim. 5 août 2018 à 20:05, Arv Evans <arvid.evans@...> a écrit :
Jack

The design won't stay still long enough to write an article.    8-)

It is actually pretty simple.  I used a 50 ohm resistor as the input to the balanced mixer.
That lets me use the step-attenuator (also home-brewed) for input calibration.  Plug-in
filters can be added to the front-end but normally I just use a tunable series LC to suck
out unwanted frequencies (2nd or 3rd or other harmonics of the LO).  Output of the mixer
goes through a termination (stolen from the BITX BFO mixer design), followed by an
emitter follower 2N3904 to feed audio to a voltmeter, or oscilloscope, or fldigi on Linux.

A PCB would be nice if I could ever settle on not making any more modifications.  To do
that I would have to draw a schematic, which has not happened yet.  It is just one of those
things that sets on a shelf above the workbench and gets used whenever needed.

Arv
_._


On Sun, Aug 5, 2018 at 11:51 AM Jack Purdum via Groups.Io <jjpurdum=yahoo.com@groups.io> wrote:
That would make a great article!

Jack, W8TEE

On Sunday, August 5, 2018, 1:46:23 PM EDT, Arv Evans <arvid.evans@...> wrote:


Gordon

Several years ago Hans G0UPL started marketing his SI5351a based VFO kit.  With that as
a very stable LO and a balanced mixer to LPF to Audio voltmeter it is possible to make a simple
receiver for chasing spurs, harmonics, and gremlins.  Of course it was impossible to resist
adding oscilloscope display and rewriting the software to scan specific frequency ranges.  Now
I have a direct-conversion spectrum scanner that can be calibrated and used to track ghosts
and goblins emanating from my home-brew builds.  Calibration is done by use of a fixed-value
step attenuator for 1, 2, 4, 5, 10, 20, 40, and 50 db. 
This has been a work-in-progress for several years and just keeps getting better as I find ways
to modify and to use it. 

Arv
_._

On Sun, Aug 5, 2018 at 10:07 AM Gordon Gibby <ggibby@...> wrote:
How could we make measurements of signal level accurately? Compared to fundamental?   Note  I would probably be using a 2nd uBitx as the receiver.


Gordon



On Aug 5, 2018, at 11:53, ajparent1/KB1GMX <kb1gmx@...> wrote:

On Sun, Aug 5, 2018 at 07:55 AM, Doug W wrote:
97.313(a)
A KW is the minimum power according to some. ;)

Snark:  For me 100W is serious QRO for HF.  At VHF I'm just warming up the driver.

Allison

Re: One question only...

Jerry Gaffke
 

Hmm,  
Any tone from the second harmonic would move twice as fast as the fundamental.
And 3x for the third harmonic etc.
So distinguishing that stuff in software, we might get by without that 86.85mhz IF filter.
And no front end LC.

So same as your schematic, but a $2 Blue Pill ARM board for the processor?
Something with a 16 bit ADC to listen to the audio with?

Jerry


On Sun, Aug 5, 2018 at 01:29 PM, Jerry Gaffke wrote:
Of course, my suggestions make it not so simple.
And somehow anything capable of phase noise measurements
must also deal with a very powerful carrier, not trivial.

On Sun, Aug 5, 2018 at 01:19 PM, Jerry Gaffke wrote:
Arv,

I like it, nice and simple.

I'd consider a 50mhz LPF up front, another mixer, mmic amp, and one of these 86.85mhz SAW filters:
    https://wireless.murata.com/RFM/data/px1002.pdf
Then no worries about those harmonics.

Fully analyzing the audio tones using an FFT on an ARM processor should give
sufficient resolution for stuff like evaluating phase noise.

The simple direct conversion mixer means you get both sidebands.
But that software on the ARM could figure out which audio tones go up and which go down
as your VFO moves, and thus separate out high from low sidebands.

I've been thinking of building something like this for a couple years,
have not yet tried.
 
Jerry, KE7ER

Re: One question only...

Kees T
 

Yes, Allison, Trust But Verify.

I have not had the opportunity to test a number of STOCK uBITX Transceivers to see what the spurs look like. You have apparently had a chance to look at a number of them so I defer to your assessment.  It does surprise me a little.

Guess the only clean solution is to just add a switchable Low Pass Filter (LPF) to the 5-10W output for the "culprit" frequencies you are going to use, maybe just the higher frequencies ?.....can't hurt, unless you forget to switch it to the correct band. W8DIZ used to offer a kit and I'm sure there are others out there now.

What would you suggest ? Do you know of any switchable LPFs out there ? I was fiddling around with a LPF/BPF unit years ago.  The LPFs were based on a W3NQN article from Feb 1999 QST. They worked well when tested in a professional lab.

73 Kees K5BCQ

Here is an excerpt from my website........ schematics are on my website too

  • These boards all have a 2.5" x 0.6" form factor and come in two varieties, BPF or LPF. Redundant header pins are located at both ends (or you can solder directly), The designs are symetrical so the boards can be reversed or flipped without effect. The boards can be used standalone and plugged one at a time or used with a switcheable host board of your design or mine (available later). I have the bare boards only and header pins available for $2 per board plus postage in the USA. Toroids are listed below (the capacitors are all 630V or 1KV NPO 1206/1210 size SMT. You would use one of the free programs like Jim Tonne's, ELSIE, or other filter program (AADE, etc) for the appropriate values for your needs.


Re: One question only...

Jerry Gaffke
 

Ah, you beat me to it!
Jerry


On Sun, Aug 5, 2018 at 01:56 PM, Arv Evans wrote:
Your idea of directionalizing frequency movement is good, and is already being used to
determine the N-value of harmonics (2nd moves at twice frequency changes, 3rd at
3-times, and so on). 

Re: WTB uBITX v3 or 4

WD0AKZ
 

Thanks to those that wrote me with offers.  I did buy a v3 and am looking forward to building it.

73's to all

de George
WD0AKZ
dit - dit

Re: Harmonic performance - SSB vs CW

Gary Anderson
 

Allison,
I'm not arguing with you.  Like I tried to emphasize, I left reading this thread early Sat morning.
While you mentioned better relays indeed, they just happened to be the same brand for which I found part numbers ....
All I remember from reading this morning is that the relays you would choose for RF wouldn't work. 
So since I found part numbers that _may_ work slightly better, I decided to share and not just keep in my personal notes.
It was trivial, although time consuming, to go through datasheets.  Why should someone else have to do that again, unless finding a better alternative?

I'm not trying in anyway to undercut any of the work you have done and shared.

Agreed that is a system level fix.  Adding up incremental improvements to reach the goal. 

Agreed that the PCB artwork can be greatly improved, without changing the board size for compatibility.

Your experience tells you that a wire on the backside is as good as coax.  My experience tells me that if I am going to route a signal, using coax has no negative effects.
These are 50 ohm filters and should be matched with 50 ohm transmission lines.  Or am I completely in left field here?

Could you look at the relay datasheets and, based on your expertise, make a recommendation if the relays would be better, adding to an improvement, albeit not the single point solution?

Regards,
Gary













Re: Harmonic performance - SSB vs CW

ajparent1/KB1GMX
 

Got it and deleted it.  Hint!

I'll answer here as your running a lot of noise and most of it very mistaken.

The bias for the IRF510 in many flavors of BITX are low to hold the gain down in an attempt
to avoid oscillation.  Most of the amps I use run from 100 to near 250ma depending on design.
For those its not class A more like AB or AB1 as the peak current is easily 2-3A.
For example the KD1JV recent 20M SSB design the sweet spot for output matching was
100ma and it didn't want to take off even at 250ma.  Nice simple amp doing 7W at
12.6 and more like 11 at 15v using IRF510.  I've used them at 6M too at high power.

The regulator has very little current loading for the gate bias is nearly zero current.
What you say about bias is not very useful or accurate.  MOSFETS are voltage
driven and the DC input resistance is high.  So a LM78l05 can easily provide
the voltage to hit several amps of drain current.   Your whole supposition on
biasing MOSFETS is faulty.

Generally the Drain pin goes to the transformer.  The source goes to ground.

The datasheet gives a rather poor amount of data on the part.  Compare it to
the MRF, MRFE, or BLF parts.  However class B unlike bipolar parts requires a
lower bias but no so low that impedance drift off with increasing drive.  And single
ended amp class B is not so clean were more current for class AB1 is usually used.

The RD parts are not magic.  Its a 12V mosfet and at higher voltages it will die.
About the only advantage is the tab is the source rather than drain and an
insulator kit with pad and shoulder washer solves that.  I find some of the
RD parts handy at Upper VHF and UHF.  

Allison

Re: Harmonic performance - SSB vs CW

Timothy Fidler
 

any post TEF wrote suggesting bias current to gate needs to be 50mA is wrong. I think it went direct to AJP . The delta drain on application of bias voltage needs to be plus 50mA per device for this type of mosfet at this power level.  I have to rush and cannot find the post. Beware.

Re: One question only...

Kees T
 

Jerry,

We probably need another thread for Spectrum Analyzers since this one is "Does the uBITX have spurs and how to get rid of them".

Relative to a SA, I would consider using a EU1KY VIA as a base, Jack probably has more info ?

73 Kees K5BCQ