Topics

Modular uBitx - "Ex: Harmonics"

Dennis
 

Glad you're back Gordon.

Bet of luck with your recovery.

73
Dennis W7DRW

Gordon Gibby <ggibby@...>
 

I have come out of surgery now, and I think I’ll be home tonight and may be able to construct voltage divider on my Heathkit Cantenna to get a attenuated signal that I can further attenuate on toward a ICOM 718 S-meter.   I have about 4  fixed attenuator pads.  Using those, Setting the baseband signal about the same for each test I should be able to get a “baseline” on a mostly UNmodified version three, I think it has diode to protect Q90

Then I can build my little relay daughter board and repeat the measurements which should give relative indications of any improvement or degradation

If not tomorrow, probably Wednesday.

Gordon




On Aug 20, 2018, at 15:06, ajparent1/KB1GMX <kb1gmx@...> wrote:

Alan,

You hit the high and low points well. 

The biggest thing is we have a known fail.
We believe its due to layout.
Its not easily fixed for most.
 
Your not the only one that wrote that as well but your correct.

>>The LPF's work. The physical relay device chosen works, the problem was how it was all put together<<

Bottom line is all that said whats the fix.

My cut is bypass the filters on the board so their losses are not additive.
Create a better board layout that can mount over the existing filters.
Make it so the existing relay control signals are usable, simple wires from a to b.

Cheap as now one will fix it otherwise and many will not bother anyway.

Production boards have the best fix applied to prevent more with this.

Allison

ajparent1/KB1GMX
 

Alan,

You hit the high and low points well. 

The biggest thing is we have a known fail.
We believe its due to layout.
Its not easily fixed for most.
 
Your not the only one that wrote that as well but your correct.

>>The LPF's work. The physical relay device chosen works, the problem was how it was all put together<<

Bottom line is all that said whats the fix.

My cut is bypass the filters on the board so their losses are not additive.
Create a better board layout that can mount over the existing filters.
Make it so the existing relay control signals are usable, simple wires from a to b.

Cheap as now one will fix it otherwise and many will not bother anyway.

Production boards have the best fix applied to prevent more with this.

Allison

Alan de G1FXB
 

Thanks Gordon / Alison,

In the mists of time, along time ago now......

there was a suggestion of what about using 'this or that', off the shelf relay board,

I agreed with another contributor that it didn't look particularly "suited to RF" applications. General observation, not a full on analysis.

(Just what ever is ultimately chosen should be a best as possible a "cure all" (to which there is no such thing, we can only strive to become close)

(Some regions are governed by the -50dB rule, we cannot ask you NA / US solutioneer's  be guided by that.

It would be nice however not to have to reinvent the wheel to come up with another solution that satisfies "our legislation"?))


Now it's resorted to quoting Lecher Lines / Skin resistance and  waving smiths charts at dawn:-P


Can we agree the theory of short, direct traces is good

"Noodle soup" bad.

Well bonded top / bottom groundplanes are good

Input traces parallel to output traces, bad

Attention paid to layout, Good?

and 1001 things that that all can be read in a book, possibly EMFD all contribute to good practice.

(Sometimes even that is not enough,

and sprinkling 103's liberally everywhere at least shows willing, or weakness....  :-) )


It was what I was trying to portray?

looking back, at what I wrote.

The LPF's work. The physical relay device chosen works, the problem was how it was all put together


Alan


On 20/08/2018 15:02, ajparent1/KB1GMX wrote:
Alan,

With short lines the impedance is rarely critical unless they are parallel
or over a questionable ground.  The other is narrow lines are low in surface area
and even at HF that meas RF loss to skin resistance though small.

One thing I tried after ripping the relays off. I stripped all the interconnecting
paths top and bottom and left bare pads.  I reinstalled the relays from the
ground side (bottom)  then wired them up using a different scheme.  Since
there were only three relays  I did only three filters to prove wiring not componenets.
The filters used were 80/40/20 and it tested far better.  The wiring scheme
attached.

The layout is the entire issue.  Look under the relays and its long noodle soup.
 
Allison

ajparent1/KB1GMX
 

Alan,

With short lines the impedance is rarely critical unless they are parallel
or over a questionable ground.  The other is narrow lines are low in surface area
and even at HF that meas RF loss to skin resistance though small.

One thing I tried after ripping the relays off. I stripped all the interconnecting
paths top and bottom and left bare pads.  I reinstalled the relays from the
ground side (bottom)  then wired them up using a different scheme.  Since
there were only three relays  I did only three filters to prove wiring not componenets.
The filters used were 80/40/20 and it tested far better.  The wiring scheme
attached.

The layout is the entire issue.  Look under the relays and its long noodle soup.
 
Allison

Alan de G1FXB
 

Hi Gordon,
Yes in HF terms the traces on the uBiTX board are like you say a fraction of a wavelength.

Never the less they have contributed to the downfall in reduction of unwanted products.
Good RF design practice doesn't just apply to beyond 10Mtrs.
Alison & Warren, probably others who are also brainstorming solutions?
have all taken the Relays & LPF components & verified that in isolation they achieve their design goals.
The problem exists in how they have being laid out, on board,
Alison has already explained concepts of RF circulating currents and the importance of stitching between copper pours.

There is an arduino auto ATU design on the web that uses an 8 or 12 relay version of the board being discussed.
Is it cheap & does it work, Yes.
Would it work "better" if attention was paid to good RF practice. I believe, Yes.
Would it cost more, Yes.

Sometimes "good enough" is OK,
other times we learn it takes a little extra?

Alan

On 20/08/2018 11:02, Gordon Gibby wrote:
So, I’m not an expert at this, but remember that impedance changes on transmission lines are dependent on what fraction of a wavelength you traverse.

In other words, On a Smith chart, when your connection wire between boards, or a pcb-trace,  is modeled as an unmatched transmission line, the impedance change from input to output depends on how far in arc you move, measured in terms of fractions of a wavelength.   Since the highest frequency this set is made for is 10 m,  A couple of inches of wire is a tiny fraction of a wavelength, and there won’t be much impedance transformation.

So I think the microstrip transmission line type designs become important when printed circuit board connection lengths become a sufficient fraction of a wavelength to allow Undesired impedance changes, if the transmission line doesn’t have the correct impedance.

My suspicion would be you want to use short wires, shielded if they are long, and keep signal wires apart to avoid cross talk.  Twisted ground wires could improve separation.  

All my Heathkit HF Rigs were built this way.... eons ago, and somehow or another they worked.  

More knowledgeable people than me might have better suggestions.   


On Aug 20, 2018, at 05:25, Alan de G1FXB via Groups.Io <g1fxb@...> wrote:

Agree with John,

the board looks to be designed with AC/DC switching in mind, the relay contacts are presented to the board edge with terminal screws.

With that in mind It's likely the track layout would not be optimal for minimal impedance changes & crosstalk @ RF ?

Alan


On 19/08/2018 15:21, John Brock wrote:

Is there a concern for how the contacts are routed on the PC board so far as RF is concerned? What about the ability of the contacts to reliably switch relatively low level signals?

John Brock

WA8US



Gordon Gibby <ggibby@...>
 

So, I’m not an expert at this, but remember that impedance changes on transmission lines are dependent on what fraction of a wavelength you traverse.

In other words, On a Smith chart, when your connection wire between boards, or a pcb-trace,  is modeled as an unmatched transmission line, the impedance change from input to output depends on how far in arc you move, measured in terms of fractions of a wavelength.   Since the highest frequency this set is made for is 10 m,  A couple of inches of wire is a tiny fraction of a wavelength, and there won’t be much impedance transformation.

So I think the microstrip transmission line type designs become important when printed circuit board connection lengths become a sufficient fraction of a wavelength to allow Undesired impedance changes, if the transmission line doesn’t have the correct impedance.

My suspicion would be you want to use short wires, shielded if they are long, and keep signal wires apart to avoid cross talk.  Twisted ground wires could improve separation.  

All my Heathkit HF Rigs were built this way.... eons ago, and somehow or another they worked.  

More knowledgeable people than me might have better suggestions.   


On Aug 20, 2018, at 05:25, Alan de G1FXB via Groups.Io <g1fxb@...> wrote:

Agree with John,

the board looks to be designed with AC/DC switching in mind, the relay contacts are presented to the board edge with terminal screws.

With that in mind It's likely the track layout would not be optimal for minimal impedance changes & crosstalk @ RF ?

Alan


On 19/08/2018 15:21, John Brock wrote:

Is there a concern for how the contacts are routed on the PC board so far as RF is concerned? What about the ability of the contacts to reliably switch relatively low level signals?

John Brock

WA8US


Alan de G1FXB
 

Agree with John,

the board looks to be designed with AC/DC switching in mind, the relay contacts are presented to the board edge with terminal screws.

With that in mind It's likely the track layout would not be optimal for minimal impedance changes & crosstalk @ RF ?

Alan


On 19/08/2018 15:21, John Brock wrote:

Is there a concern for how the contacts are routed on the PC board so far as RF is concerned? What about the ability of the contacts to reliably switch relatively low level signals?

John Brock

WA8US


Tom, wb6b
 

Hi Gordon,

Your inductors and bypass caps are probably better as RF isolation than the way the opto-isolators are implemented on the generic 4 relay board. If the generic board is used with a good common ground (ideally a metal chassis as a big, low impedance, common ground) and, as you suggested, ground the un-selected filter outputs with the relays, these generic relay boards could work. The relay on each end of the filters arrangements seem to have more room for success without carefully taking into account coupling across the relays and layout as compared to the DPDT single relay per filter designs. I do need to put together a way to actually measure the harmonics from my uBITX with reasonable accuracy. 

Tom, wb6b

Randy Hall
 

Allison

Oh well, it was worth asking.

Randy, K7AGE

On Sun, Aug 19, 2018 at 10:38 AM, ajparent1/KB1GMX <kb1gmx@...> wrote:
Randy,

Its not a drop in as you need about 13dbm (20mw) of drive.  Many of
the ills noted are before that.  Using Q90 and predriver stages would
give you that but,  by then its already not gain flat by any means.

It would not solve many of the issues by a long shot because of that.
A driver and predriver would be needed, better output filters are still
a must.  In a modular component system the modules still have to 
perform well.  

Allison


ajparent1/KB1GMX
 

Randy,

Its not a drop in as you need about 13dbm (20mw) of drive.  Many of
the ills noted are before that.  Using Q90 and predriver stages would
give you that but,  by then its already not gain flat by any means.

It would not solve many of the issues by a long shot because of that.
A driver and predriver would be needed, better output filters are still
a must.  In a modular component system the modules still have to 
perform well.  

Allison

Randy Hall
 

From Han's posting in the QRP-LABs group -

Who will be the first to install this into a uBITX?

Randy, K7AGE

*****

I hope to write up some details of the PA board in the next couple of days. This PA board is so GOOD that it will be made available also as a standalone 10W HF Linear PA kit. It is ready for production and should be the first of the QSX modules ready for sale.

The requirement for expensive Mitsubishi RD16HHF1 etc transistors in a decent amp is an absolute misconception. In reality they are very very expensive and not necessary in HF applications such as this.

This amplifier is proud to use two inexpensive IRF510 transistors. A great deal of attention has been paid to a symmetric and compact layout for best performance. 

I am very grateful to Allison KB1GMX for her advice, encouragement, support and testing during the development of this Linear PA.

More details soon but briefly:

1. Two IRF510 in push-pull, two BS170 in push-pull as the driver.

2. Gain 26-28dB, better than 2dB gain flatness from 2-30MHz. Still only 4dB down even at 6m, and respectable even to 4m.

3. Can easily achieve 10W across 160-10m using 12V supply

4. 80m band tested harmonic levels 2nd -38dBc, 3rd -31dBc. Thats BEFORE a low pass filter is added - addition of the LPF will result in a very clean output.

5. Very large heatsink for the 10W power rating. Ran continuous 100% duty cycle at 10W output for 1 hour without failure or degradation

6. Tested to 20V supply voltage; tested open load; tested shorted; tested with 20W 100% duty cycle for 10 minutes... all with no failure or degradation

7. Quite easily delivers a lot more than 10W too...

Above results are for the PA module alone, not integrated within the QSX. I think it is very unlikely that that you would be able to replace the PA module in the QSX with anthing that could deliver as good performance, even assuming you are willing to spend a lot more money.

Anyone who thinks IRF510s aren't good for the job, will need to reconsider their view after this.

73 Hans G0UPL



On Sun, Aug 19, 2018 at 8:07 AM, Eric Flores <floreseric1981@...> wrote:
The desired signal at Q10 is always right around 45mhz.
Has to be, or it won't go through the 45mhz crystal filter


Ah, I mixed that up then. I had glance through the circuit description and assumed the first LO was 45Mhz (but that was the desired IF). I went back and saw this: 

the first oscillator tunes from 45 MHz to 75 Mhz

So in the RTL, you would always have to account for a variable frequency offset when tuning the RX.

Eric

On Sun, Aug 19, 2018 at 07:47 Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
The desired signal at Q10 is always right around 45mhz.
Has to be, or it won't go through the 45mhz crystal filter

On Sun, Aug 19, 2018 at 01:37 AM, Eric Flores wrote:
Tom, you’re thinking right up my alley. I was also considering a RPi to handle digital modulation modes. 
 
But, I am also considering using it to run an RTL-SDR for all of the RX work and tap the signal from the 1 mixer (just off C10 and the base of Q10, and completey removing R10, R11, Q10) and send the RX strait to RTL w/ small piece of coax. The IF should be 46Mhz-75MHz (for 160-10m) right there, which is well within the range of the dongle so you wound not need a hamitup upconverter. The one issue that might prevent this is I can’t seem to get rid of a 1/2 second delay in the RTLs. 
 
I was also considering having a headless ubitx so I can have the radio in a shack and operate from the house, which could end up being a few hundred yards away. It would be controlled via a network connection to either another RPi or a computer (possibly via rtl_tcp for RX control, a REST API for general radio control and an RTP stream for TX voice). 
 
As an added plus, raspian supports Docker so each of these little services can be their own container making upgrades a breeze. 


Eric Flores <floreseric1981@...>
 

The desired signal at Q10 is always right around 45mhz.
Has to be, or it won't go through the 45mhz crystal filter


Ah, I mixed that up then. I had glance through the circuit description and assumed the first LO was 45Mhz (but that was the desired IF). I went back and saw this: 

the first oscillator tunes from 45 MHz to 75 Mhz

So in the RTL, you would always have to account for a variable frequency offset when tuning the RX.

Eric

On Sun, Aug 19, 2018 at 07:47 Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
The desired signal at Q10 is always right around 45mhz.
Has to be, or it won't go through the 45mhz crystal filter

On Sun, Aug 19, 2018 at 01:37 AM, Eric Flores wrote:
Tom, you’re thinking right up my alley. I was also considering a RPi to handle digital modulation modes. 
 
But, I am also considering using it to run an RTL-SDR for all of the RX work and tap the signal from the 1 mixer (just off C10 and the base of Q10, and completey removing R10, R11, Q10) and send the RX strait to RTL w/ small piece of coax. The IF should be 46Mhz-75MHz (for 160-10m) right there, which is well within the range of the dongle so you wound not need a hamitup upconverter. The one issue that might prevent this is I can’t seem to get rid of a 1/2 second delay in the RTLs. 
 
I was also considering having a headless ubitx so I can have the radio in a shack and operate from the house, which could end up being a few hundred yards away. It would be controlled via a network connection to either another RPi or a computer (possibly via rtl_tcp for RX control, a REST API for general radio control and an RTP stream for TX voice). 
 
As an added plus, raspian supports Docker so each of these little services can be their own container making upgrades a breeze. 

Gordon Gibby <ggibby@...>
 

Hi John Brock I’m not sure exactly which project you were referring to, but if the external low pass filter relay project:

1. You basically want to do the wiring so the RF signal don’t get mixed into HIGHER low pass filters.   I do that by grounding the higher low pass filters when no longer in use.  On the Little board that someone pointed out the relay contacts are right there for you, so you can wire them with short or shielded wires and minimize cross talk.  You’re trying to reduce capacitance between circuits that shouldn’t know about each other

2.  You also want to avoid blow by getting around through the coil connections which are wired on both input and output of the filters and communicate by way of the DC wiring.  Decoupling bypass capacitors or  the optical coupling are good solutions for that

3.  Relays have been used to switch transmit/receive/ filter signals for a bunch of years.  My heathkits  had them.  Sometimes the Contacts get a little oxidized and less signal gets through, a bit of paper used to burnish them and they work better.   That is, if you can even get to them.  




On Aug 19, 2018, at 10:22, John Brock <brock71@...> wrote:

Is there a concern for how the contacts are routed on the PC board so far as RF is concerned? What about the ability of the contacts to reliably switch relatively low level signals?

John Brock

WA8US

John Brock
 

Is there a concern for how the contacts are routed on the PC board so far as RF is concerned? What about the ability of the contacts to reliably switch relatively low level signals?

John Brock

WA8US

Jerry Gaffke
 

The desired signal at Q10 is always right around 45mhz.
Has to be, or it won't go through the 45mhz crystal filter


On Sun, Aug 19, 2018 at 01:37 AM, Eric Flores wrote:
Tom, you’re thinking right up my alley. I was also considering a RPi to handle digital modulation modes. 
 
But, I am also considering using it to run an RTL-SDR for all of the RX work and tap the signal from the 1 mixer (just off C10 and the base of Q10, and completey removing R10, R11, Q10) and send the RX strait to RTL w/ small piece of coax. The IF should be 46Mhz-75MHz (for 160-10m) right there, which is well within the range of the dongle so you wound not need a hamitup upconverter. The one issue that might prevent this is I can’t seem to get rid of a 1/2 second delay in the RTLs. 
 
I was also considering having a headless ubitx so I can have the radio in a shack and operate from the house, which could end up being a few hundred yards away. It would be controlled via a network connection to either another RPi or a computer (possibly via rtl_tcp for RX control, a REST API for general radio control and an RTP stream for TX voice). 
 
As an added plus, raspian supports Docker so each of these little services can be their own container making upgrades a breeze. 

Gordon Gibby <ggibby@...>
 

Tom— well, that 4 relay module certainly is cheap!   And it looks like it has optically isolated controls of the relay coil’s.   That would get around the coupling back through the coil’s.   Probably even better than the inductors and bypass capacitor is that I stuck on. That little eight dollar board runs (it seems ) on 5 V, but it wouldn’t be hard to step down the TX switched voltage to 5 V to run this.   You might could just switch the input to a 5 V regulator and it might be fast enough

I would think that that’s a pretty elegant solution. Had I known about it I might have done it.

Cheers!!
Gordon
 


On Aug 19, 2018, at 05:36, Tom, wb6b <wb6b@...> wrote:

Using the RTL_SDR sounds interesting. I'd been concerned about audio delays even with using software audio filters/equalizers. Was running this by another local ham, and he suggested that some delay would likely not be an issue, as many times Hams are using the webSDR site to help hear people in a net that they can't hear directly.

I guess if I can respond to my call in a net roll call before the net control calls the next call, all is well. So, maybe a half a second is acceptable. But, at some point it will be troublesome. When I use OpenWebRX on my local network, it still will try to buffer up several seconds of audio, that is something I'd need to improve for a headless radio arrangement.  

I looked in my notes the web protocol I was looking into is called WebRTC. Supposedly it is supported by a number of browsers, now.

Docker is cool, I use it with some of my consulting clients. I just made a clone of my Pi's SD card, so when I start stringing more audio software applications together and Jack Audio stops working and it is too perplexing to get back to a known working state, I can just start again from a known system that has WSJT working and try again. At least to me, Linux has one of the most confusing audio chains imaginable. FLdigi seems to use a different audio backend than WSJT so getting both to work is requiring some work. The logic of the Linux sound support may strike me as I play with it more. 

Tom, wb6b

Tom, wb6b
 

Using the RTL_SDR sounds interesting. I'd been concerned about audio delays even with using software audio filters/equalizers. Was running this by another local ham, and he suggested that some delay would likely not be an issue, as many times Hams are using the webSDR site to help hear people in a net that they can't hear directly.

I guess if I can respond to my call in a net roll call before the net control calls the next call, all is well. So, maybe a half a second is acceptable. But, at some point it will be troublesome. When I use OpenWebRX on my local network, it still will try to buffer up several seconds of audio, that is something I'd need to improve for a headless radio arrangement.  

I looked in my notes the web protocol I was looking into is called WebRTC. Supposedly it is supported by a number of browsers, now.

Docker is cool, I use it with some of my consulting clients. I just made a clone of my Pi's SD card, so when I start stringing more audio software applications together and Jack Audio stops working and it is too perplexing to get back to a known working state, I can just start again from a known system that has WSJT working and try again. At least to me, Linux has one of the most confusing audio chains imaginable. FLdigi seems to use a different audio backend than WSJT so getting both to work is requiring some work. The logic of the Linux sound support may strike me as I play with it more. 

Tom, wb6b

Eric Flores <floreseric1981@...>
 

Tom, you’re thinking right up my alley. I was also considering a RPi to handle digital modulation modes. 

But, I am also considering using it to run an RTL-SDR for all of the RX work and tap the signal from the 1 mixer (just off C10 and the base of Q10, and completey removing R10, R11, Q10) and send the RX strait to RTL w/ small piece of coax. The IF should be 46Mhz-75MHz (for 160-10m) right there, which is well within the range of the dongle so you wound not need a hamitup upconverter. The one issue that might prevent this is I can’t seem to get rid of a 1/2 second delay in the RTLs. 

I was also considering having a headless ubitx so I can have the radio in a shack and operate from the house, which could end up being a few hundred yards away. It would be controlled via a network connection to either another RPi or a computer (possibly via rtl_tcp for RX control, a REST API for general radio control and an RTP stream for TX voice). 

As an added plus, raspian supports Docker so each of these little services can be their own container making upgrades a breeze. 

Eric

On Sat, Aug 18, 2018 at 21:57 Tom, wb6b <wb6b@...> wrote:
The reason Raspberry Pis are so inexpensive is they take advantage of the price advantage of using basically the same processor that is used in hundreds of millions of radio transceivers, already. (smartphones) 

So, a Raspberry Pi finding its way back into another transceiver (uBITX) does not seem too far of a stretch. Not necessarily for SDR I/Q processing, but running digital modes in a small radio package is a good thing. Also, there are some audio/microphone compressor/equalizer software packages that can run under Linux that might improve your SSB transmit audio without needing to add an additional analog compressor board (albeit a second USB sound adaptor).

On the RFI noise front, I found a metal case for the Pi that will go inside the bigger transceiver case. 

I like Allison's idea of using Android tablets as a controller for the uBITX. I'm, also, looking into running a web server on my Raspberry Pi so I can operate my uBITX from either an iPhone or Android (without installing an App). Possibility borrowing a little from the web interface of OpenWebRX.  Some browsers may be capable of sending realtime audio back to the uBITX. I'm looking at the protocol that supposedly allows this. 

I've seen some commercial HF radios that let users send/receive text messages from their Smartphones. That, on Ham Radio, would certainly attract a new younger group of Amateur Radio operators to the hobby.

Tom, wb6b

 

Tom, wb6b
 

I like Gordon's add-on second set of relays idea and board. 

Just wondering if something like this relay board might do an acceptable job for someone wanting something ready made. Would follow Gordon's suggested uBITX board cuts and then wire one side of the filters to these relays. https://www.amazon.com/dp/B06XRYGM2C/

I've seen some SPST relays where the center conductor is connected to the frame (don't know if that is the case with these). That could possibility be a big issue. But, as this would only be connected to one side of the filters, and both sides are switched, due to the original relays still switching the other side, with careful choice of wiring, these might work.

---

---

Tom, wb6b