More Progress on the MHST

More progress on the MHST rig and thus some more sharing. I note that when there is nothing but boring words, the interest level drops. This often happens on the weekend posts. So, for this last day of March 2025, hopefully more interesting stuff.

The MHST project has two main elements and these are of course the hardware and the software. The more visible of course is the hardware and even then, not much of that. The not so much seen of course is the software. Both are needed to be working for the total rig to work.

I posted a video about some of the progress and problems if only to show it was alive. One comment -- hey you have the presence of the opposite sideband image. That was true as the video was done without any adjustments. Contained internally within the QUISK software is a means to compensate for not so perfect hardware where component tolerances do not create the perfect 90-degree phase shift. In fact, you have the capability to make these adjustments over the whole band of operation.

So, to that commenter here is very short clip of an SSB signal where the opposite sideband image has been dramatically reduced. There are adjustments for both transmit and receive.





With help from the VWS Maker's Group I now have both the SPOT and PTT buttons working in QUISK. Hit either button on the QUISK dashboard and the software goes into Transmit and the Relay that activates the T/R sequence works. 

What doesn't work is that I don't see any audio streams going into the ADE-1's on Transmit. Not sure in which pile is the problem. It has to be something in the settings that directs the signal from the Microphone Sound Card into an I and Q streams. More opportunities.

One more photo. I got 40M FT-8 receive sort of working and again a matter of the right settings.



Them that know can make it go.

73's
Pete N6QW




The MHST is now working on the RPi5

I am happy to report that I now have the Pi Pico able to work with the RPi5 SBC. The issue was that the package to recognize the Pi Pico has to be manually loaded into the QUISK directory. Once installed what should normally happen is that when QUISK is executed that radio should show up on the menu. It didn't on the RPi5 SBC.


It was a mystery as using essentially the same process the radio did show up on the Windows 10 machine. Then, I decided to closely read the instructions where I spotted something you needed to do differently for the Linux install. Dah RTFM! Suddenly the MHST rig is in the list. I still have a few warts to polish off but a huge step in reducing the size of the rig with the bonus of the features in the Linux form factor versus the Windows 10.

One really important Linux feature is the Pulse Audio Control Panel that is loaded in a terminal like this.

sudo apt-get install pavucontrol

This control panel gives you a whole series of adjustments and selections with regard to the two sound cards. This panel has the Mic Gain and overall Audio gain level sliders as well as the line levels for the I/Q channels on both transmit and receive. #11 is not always the best adjustment.

I did find with the package developed by the VWS that some sound items on the QUISK configuration menus do not show up automatically as a drop-down menu but must be added manually. I cannot fault their package as I am using their software for a completely different application and rig. The manual entry process did work.

Another issue was the Com Port. In the Windows 10 I was able to determine the 2nd Pi Pico is now com22. The 1st Pi Pico, I smoked, it was com21 and those com ports must be manually entered in the QUISK Configuration set up. The Linux com port is a bit trickier and is entered as /dev/ttyMACO

It is basically there but there are still some features requiring a bit more work and so we press on.

It is understood that some big thing will be announced from 1600 Pennsylvania Avenue on Tuesday -- how appropriate as that is April Fool's Day. T. S. Elliot, a noted American author and poet once said that April is the cruelest month. Maybe Tuesday's big news is more in line with T. S. Elliot. 

Where do all those who were RF'D go for employment. There is a timing disconnect between tariffs to create more US businesses and when those businesses would be in full production to absorb displaced workers. Two years is a very optimistic time horizon. Surely someone who attended Wharton would know that (from Econ101). So, what do us common folk do for two years?

Keep in mind those who were RF'D were not bad employees, yet they are categorized by one individual as Waste, Fraud and Abuse.

Them that know can make it go.

73's
Pete N6QW

The Ship is Righted.

A productive day yesterday on two fronts. I visited the XYL at the cemetery to bring some of her favorite flowers and was happy to see that the headstone was installed. I had been told it might take six months. Now no longer an unmarked grave.

The second front was the MHST is back on track. As I had suspected it involved the software. This requires a bit of explanation that deals with the two different approaches using the QUISK software. (Windows/Linux)




But let us start with some fundamentals of the I/Q implementation which is basically sound cards -- TWO of them. One sound card, typically external, involves the processing of the I/Q data stream to/from the computer. The second sound card handles the Audio output and Microphone input. This often is an external sound card built into a mic/headset dongle.

In a Linux implementation using a RPi3, RPi4 or RPi5 there is an issue with using any internal sound cards on the SBC. In the earlier Pi Models, there is an on-board Headphone jack but no Microphone jack. The RPi5 does not have that Headphone jack. The ASUS Tinker Board has a combo headphone and microphone jack AND two separate internal sound cards associated with that single combo jack. So typically, with the Linux (aside from the Tinker Board) you have an external sound card for the Audio Out and Microphone In. 

There is a bonus with the Linux distro called pulse audio and a separate app installed as pavucontrol which is a dashboard to make all sorts of audio adjustment for the two sound cards. These include designating the Line In and its level for the I/Q, the Microphone gain level and the Audio output level. 

On the QUISK dashboard most times you can only adjust the Audio output level. (QUISK for use with the Hermes Lite 2 does have an extra control for input RF gain level.) When setting up the sound cards in Linux and on the QUISK dashboard you can actually see the sound cards by name and by specific designation similar to how hard drives are designated in Windows.

Summary for Linux QUISK: auxiliary sound control and ease of identification of which sound cards are which.

The Windows distro does not accommodate the pavucontrol app and when you call up the Sound Cards in QUISK, the choices are less and not as clear. You can gain access to some Sound controls by going into the Windows Control Panel and clicking on the Sound Icon. Here you must find the appropriate sound device and making it a default. There are some level controls available.

Summary for Windows QUISK: Clunky and limited.

So, what was the problem. I made a sound card in the Windows Control panel the default but not the one needed for Line Input. I would have easily spotted that in the Linux version with the Pulse Audio app but not in the Windows version.

Them that know can make it go.

73's
Pete N6QW


The Beat(ing) Goes On.

For some strange reason my mind works while I sleep and often, I will awake with an answer to a confounding problem from the day before. That didn't happen last night!

Instead, I did think of a new mantra to describe the situation.


Things simply do not aright themselves through spontaneous correction.

Yesterday was both a Good Day and Bad Day at the bench with the MHST project. First the Good News. I was able to get a HIGH on GP2 when you hit the PTT button on the QUISK panel. It all boiled down to adding a #define statement in the Arduino IDE that cites the Pin title and GPIO Pin Number. 

That worked but for as fast as the Pi Pico is (133 MHz) there is a noticeable delay (with no internal delay statements in the code) from button engagement until there is power on the pin and the same is true once the button is released. Thus, not like unkeying the Microphone and you are in receive. More opportunities for going down the rabbit hole.




But then the two Bad News Stories. I noticed a weird display on the spectrum. Two humps right in the middle. A quick check and there were no LO signals, and the Pi Pico board suddenly became the IRF510 test where you place your thumb on the heatsink and you get physically burned. 

Smoked is an appropriate term, but also inexplicable as I had not touched or tweaked anything. Recycling and restarting things did nothing and so my new mantra arose from the smoke.

I had a NIB Pi Pico and got that loaded with code and installed. Back working so that was good. I left the radio on and went in the house to check on my dinner that was simmering on the stove and came back out to the lab to find the screen saver on but now an issue with the radio is not receiving. The 1st check and the Pi Pico was not hot to the touch...

I turned things off and called it a day. Dinner was quite good and hopefully my new mantra is wrong, and the radio will once again work today.

Overnight I thought about any changes I had made but those involved software and not anything on the hardware side. I will look for any wires that might have been shorted or not connected. But the fact remains I smoked a Pi Pico and there is an anomaly afoot. In looking at the prior posts I am always fixing a problem with the Pi Pico -- is there a much larger issue that my simple design may need some attention.

If you know then you can make it go.

Listening to some 7/5 calls in a QSO on 40M, I got a run down on how several (YL) democratic representatives will be arrested and at least one will be deported -- simply a report of what they heard on Fox News. They actually believe what they were saying. News headlines from this morning are silent on any such actions. Maybe yet today. Stupidity and a ham -- that is a toxic combination!

73's
Pete N6QW




AI to the rescue...

One regular blog reader turned to chatGPT with my problem stated in yesterday's blog.  The answer was in the form if you were using Micro-Python or regular Python for coding. Indeed, the pin designations were different from what I had posted. I thought this is the answer. Well sadly none of that compiled in the Arduino IDE. But the exercise convinced me that indeed there has to be something different in the way the pins are named.

The ever-curious blog readers would ask -- why does the built in LED code work? In fact, the often-used example in how to code the Pi Pico is the Miyagi LED ON LED OFF tutorial. If that works making a Pin HIGH should work. One possible and likely THE reason -- the built in LED IS NOT a Pin on the Pi Pico and therefore no Pin is specified in the code.

This cannot be this hard. Perhaps another run at chatGPT only specifying only show code that can be used in the Arduino IDE. It all comes down to asking the right question to the problem at hand.

Well, I made a run using co-pilot and this is another answer. This is a blinking LED example using an external LED and there is a clue in the define statement which designates the GPIO Pin, Thank you AI. I loaded the code and that worked!







If you know you then you can make it go.

73's
Pete N6QW
 
PS: The photo is of nothing special just there; otherwise blog readers will pan any posting without a photo. Initially I had selected a photo of Ann Margaret on a Motorcycle and then realized she is my age, and most blog readers would not understand the selection.




A Programming Mystery...

 As hams we all love the challenge of a mystery especially when a process approach can help unravel that a malicious person put a pin in your coax. That is why you aren't hearing stuff and the SWR is through the roof. That didn't happen but is an example and something that is physical in nature -- you spot the pin, pull the pin and all back to normal. 

On the other hand, where the issue is software the obvious is not easily seen or detected. The off the wall SWR and nothing being heard were huge clues in our example. Those clues are not so clear in a software problem I have with the Pi Pico.

You hit the PTT tab on the QUISK software and thanks to KK4DAS the on-board LED lights green, so you have a visual cue that the MHST is in the transmit mode. But we need one more action in the code and the very same sequence of code that lights the on-board LED must also create a HIGH signal on one of the GPIO pins to trigger the TR relay system.

This sound so much like the Blinking LED tutorials on many devices. The Mr. Miyagi LED ON LED OFF code. (Teaching Daniel san how to wax on wax off.)




The 1st issue with the Pi Pico is that there is an On-Board Green LED but it is hardwired on the Pi Pico substrate and not accessible as a Pin. A real sledgehammer approach is to have a no-touch control (proximity mode IR photoelectric sensor) read the Green LED and it triggers the TR. What a KLUDGE!

The software aficionado would write code to affect that pin going HIGH. There are 4 places where code entry would be required.

1. Designating a Name and Pin Number 

int PTT = 4; This says Pin 4 is the chosen pin.

2. In the setup()

pinMode(PTT, OUTPUT);
digitalWrite(PTT, LOW);

This tells the code that Pin 4 is an Output pin and its initial state is LOW, or no output

3 & 4. In the loop void()

if (something happens like the PTT Button pushed) then

{

digitalWrite(PTT, HIGH);

}

So long as the PTT button is engaged then the designated Pin (4) will be high, and the On-Board LED will be lit Green.

Now if we do nothing else the 1st time you hit the PTT button the designated pin will be HIGH until you power off the Pi Pico.

Following that if statement you need to state the alternative using an else statement. 

else
{ digitalWrite(PTT, LOW);

}

When the PTT button is invoked then the LED comes ON and the designated Pin goes HIGH. When the PTT button is not engaged the LED is off and the designated Pin is low.

Placing this code in the very same locations in the sketch where the On-board LED statements are placed, we should see the LED Glow and the Pin is HIGH. 

Well, it doesn't as the LED lights, but no Pin goes HIGH. The first issue I saw was the Pin designation. This was observed with the Seed, Xiao RP2040. Or the example from Arduino Nano. Pins 2 and 3 are interrupt pins frequently the connection points for an encoder. But these are Pin designations not the real pin numbers on the device and I think are the 5th and 6th pins on the actual hardware. 

My conclusion is one problem is how the Pins are named in the sketch.

An internet search on the Pi Pico revealed others were having this same issue. Spoiler, no clear answer as one solution was to use a code form specific to the Raspberry Pi and non-Arduino -- that doesn't help if you are using the Arduino IDE.

In the Seed RP2040 the pin designations used in the code are like 26u, 27u, 28u and not pin numbers like 1, 2 or 3.

So, while the sketch will compile it does not work. Actual Pin 4 is GPIO Pin 2 or GP2, use of this causes the sketch to abort. Other suggestions in the Internet search were to use D4 or p4. Those compiled but did not work!

This is just a Mr. Miyagi wax on wax off code (from the Karate Kid movie). It can't be that hard. Any suggestions? 

I have already found a company who will provide me a free sample of a no-touch control as that may be the means du jour to TR the rig.

Them that know can make it go.

On the recent SolderSmoke Podcast #258 it was clarified as to what Crank it in Robert and Crank it in Ralph actually means. I also want to thank the many podcast listeners who offered their condolences and support on the passing of my XYL.


73's
Pete N6QW


MHST Repair Details

The Fix Is In! 




The MSHT project started with a board found in the junk box. Like Mary Jo, it had been around the block and in the back seat of the 57 VW Beetle more than a few times. Since building that board I have improved how I design PCBs like more generous pad sizes and using better components. These older boards are subject to frequent repairs.

Repairs included replacing the I/Q stereo jacks at the lower LH corner, replacing the two 10K pots (left edge below the LO SMA connectors) and finally a small circuit board for right side ADE-1. The pads for the ADE-1 were really small and a touch up with the soldering iron lifted the pads off the board.

But this is a good thing as the ADE-1's sold by https://www.mostlydiyrf.com can easily be integrated into a design of a radio-built Manhattan style similar to what I did for the repair.


Fixed.

This video is a bit long but on a 28-inch screen sure lets you see the signals. Keep in mind this is a minimalist SDR. Look and listen and that all is coming from a beat up 4X6 chunk of PC board.

It is still not working on transmit and so that is the next challenge. It's there but I have not found the key to unlock it as yet.

Them that know can make it go. 

Time to dump you know who's stock. When you employ amateurs (non-ham type amateurs) you have a clown show.

73's
Pete N6QW


MHST is Fixed and Back Working!

 



Two issues: Only one USB port recognizes Com21 and an ADE-1 had to be replaced.

The ADE-1 being bad showed up when the Spectrum showed DSB.



Two sidebands, same signal.

When you know stuff, you can do stuff!

Someone suggested to me to dump any Tesla stock now -- its headed for the crapper.

73's

Pete N6QW

Just When You Think ...

The following statement contains Religion, a Mystery, a Failure and a Proclamation of Frustration which best sums up the current state of the MHST project.


Oh my God, the MSHT stopped working and I have no clue as to why and have not been able to isolate the root cause.

After yesterday's euphoria of having the rig launch it does no more. The software boots up properly with no error messages and the settings in QUISK have not changed. I see a spectrum on the monitor but no signals. The Pi Pico is generating quadrature LO signals.

I even swapped in a different RADIG board -- nothing changed and that was after swapping in new ADE-1's on the original board. I was able to test the 2nd Board using the original lash up with a non QUISK controlled LO, and it works FB. 

So, possibilities -- corrupted software, a bad sound card (a second sound card shows the same issue). Maybe there is a problem with the Pi Pico board, and I have in hand a 2nd Pi Pico so we can test that possibility. But wisdom says something happened and no sense smoking a second Pi Pico should that truly be the issue.

The Windows 10 Quisk in comparison to the Linux Distro is an impediment as I am more familiar with the Linux version and the tricks to evaluate problem areas. Likely they are there, I just don't know how to access them in Widows 10. Incidentally, Linux is showing itself to be far superior to any windows for these types of applications. It is so good it won't accept the Pi Pico software in QUISK.

So, nothing more will happen until I get the mystery solved. It is sure hard when the whole engineering team is comprised of but one person.

IF you know, you can make things go.

73's
Pete N6QW


SUCCESS! I got the R Pi Pico to interface with QUISK and directly Control the Frequency of the MHST!

With a huge Tip of the Cap to KI3P, Oliver, from the Vienna Wireless Society who wrote the software to have the Si5351 talk to QUISK and interface with a SDR Radio!  Needless to say, I was inspired by the work of the Vienna Wireless Societies Maker Program. Thank you, guys, for making something very possible.

We now have the frequency of the MHST being controlled by QUISK. This is HUGE!!!!!!!!!



Pi Pico Controller


Proof of Life

Right now, it's just working on Receive and the code needs some work for transmitting. Once I get that fixed, I will be happy to share my work.

HUGE caution here, you need more than a passing skill at software installation and the ability to troubleshoot. I have only been successful getting this to work on Windows 10 but the Linux R Pi5 has not worked for me.


VWS Shows up on the Windows 10 List
Serial Port com21

The Windows / Linux Problem. Despite the very same process the VWS SDR does not show up on the same list when using the Raspberry Pi Linux form. I have a machine with Linux Mint20, and it does show up on that drop down. Any clues?


The default in the VWS software is Com5. After several hours of nothing happening a dim light bulb said you need to change the com port. When I loaded the software on the Pi Pico, it told me the RP2040 com port was Com 21. That worked!

There is light at the end of the tunnel.

Them that know can make it go.

73's
Pete N6QW


How to Make a Seismograph

 For the last couple of weeks, we have been having earthquakes here in SoCal and two of them were 10 miles from my QTH and thusly I have more than just a casual interest. So, an Internet search turned up this Seismograph and yesterday a key part was delivered by Amazon.



Yes, it is not a lab quality device but likely would indeed register a 4.1 Quake (that is what we had). It does involve an Arduino so in my wheelhouse and is a one and done type project. If anything, a great party curiosity or just plain time killer to build and killing time spent watching the screen. 

The project uses a simple 4 wire OLED for a display. But what would be cool is a 3.5-inch Color TFT and it could be programmed so that anything greater than some set value to flash on the screen "Run, Duck and Cover". All sorts of possibilities.

Them that know can make it go.

Temperature check -- this might be a great time to pick up a slightly used Tesla EV. It is all over the internet that these vehicles are being dumped on the market for obvious reasons and when that happens -- an opportunity to make a deal. I wouldn't own one but for those with a strong stomach -- time to deal.

73's
Pete N6QW


QUISK Software Installation.

We are now moving to the arcane part of the project, the software to run the MSHT. Software is elusive and its installation is made even more difficult if like me, you type with only two fingers!

The QUISK software was developed by Jim Ahlstrom N2ADR, and it has been around for some time. It can be configured via a menu selection to operate with some very popular radios like the Hermes Lite 2 and if you own an old Soft Rock V6.3 it will do that as well. 

It comes in several varieties and can be installed on both Linux and Windows based machines. Installation on Linux machines is a bit more involved than on a Windows 10 machine. We are at a junction in that the Linux approach has been beset with some issues with trying to use a Pi Pico to control the frequency via QUISK. After loading QUISK on a big box Windows 10 64Bit machine the Pi Pico seems to respond better. 

So, I may have to initially use a Windows 10 Machine to get the prototype working and then try to port that over to a small form factor SBC.

To show the difference in the installations, Linux requires about a dozen files for the install and the Windows about 6. The Windows based QUISK seems more automatic to install.


Quisk Control Panel

The Windows 10 version requires you to 1st install Python3 using versions in the range 3.8 to 3.13 as the QUISK only works now with versions in that range. On one Windows machine I was able to install version 3.12 yet another would only let me install version 3.10. 

Open up Windows Power Shell and load the 6 files and top off with a desktop ICON. This lets you open the GUI front end where you must add parameters for your specific rig, like the sound cards and the type of rig. You must have two sound cards with one dedicated to the I/Q signal streams and the 2nd for the audio output and microphone input. For this 2nd sound card, I have a headset mic that has the sound card as a plug into the USB port and plug the headset into the USB adapter.

Them that know can make it go.

73's
Pete N6QW

Quadrature LO Module.

The heart of the I/Q SDR is to provide Quadrature LO signals to the ADE-1's. For the initial development work I was able to springboard off of the work of W9RAN and wrote code to provide the two Quadrature signals from a Si5351 using the Seed Xiao RP2040 as the controller.




The Seed Xiao is half the size of the Arduino Nano, and its clock speed is 8X faster with 2MB of program memory. For testing and development this works well. But this is unsuitable for the final configuration as it is not under direct control of the QUISK software.

The ultimate final configuration is to have a Raspberry Pi Pico interface with the QUISK software so there is direct control. That is currently in work and maybe the long pole in the tent.

If you would like a copy of the Quadrature Seed Xiao Sketch, drop me an email to my QRZ.com address.


Them that know can make it go.

73's
Pete N6QW

Main MHST Module Board Layout

 


Starting at the 12 O'clock position we have the Band Pass Filter and the Orange wire at the top connects to one of the amplifier steering relays. On receive this leaf actually is connected to the output side of one of the steerable relays. On transmit the orange wire is now connected to the input side of the steerable RF amp.


As we move around the clock the splitter is connected to the BPF. This splitter consists of 8 Bifilar wound turns on a FT-37-43 core. The end of one winding connects to the start of the 2nd winding which in combination connects to the band pass filter and the two free ends are connected so that one lead goes to the input of one ADE-1 and the other end to the 2nd ADE-1 input. These two windings are bridged with a 100 Ohm resistor. Why you ask the resistor well that is a good question whose answer likely involves elements like providing a constant load and balancing signal levels.

At the five o'clock position is our sideband inversion relay which is not needed if you check the block in QUISK. At 7 o'clock is the two stereo jacks for I/Q input and output. At nine o'clock we have the two 600:600 Ohm modem transformers. These units cost about $6 each. However, you can buy 10 Pieces from Amazon for $10 and they work perfect. Finally, we have at 10 o'clock the Quadrature LO which is run through 10K pots so that the LO level is exactly the same for each ADE-1. (The Blue squares in the photo.) 



The output from each ADE-1 is run through a 100uFd cap and that input goes the modem transformers. The Modem coupling is to reduce any noise pickup -- a small incremental improvement.

The only power needed to this board is voltage to the J310 amp which is powered on at all times and power to engage the three board mounted relays which is only needed on transmit. Two of the relays steer the J310 amp and the third relay (lower right-hand corner) is part of the T/R operation and directs the antenna to the steerable amp. 

Them that know can make it go. 

If things are this bad in less than 60 days wait until year three. Consumer spending is down and the real impact of all of the terminations is yet to be fully felt. But hey when you are a billionaire, no impact to you.  


73's
Pete N6QW

Two More Modules for MHST: The BPF and the Steered J310 Amplifier

Last night at 8:17PM we had a 3.9 Magnitude " shake, rattle and roll" event and Bill Haley and the Comets were nowhere in sight. About 10 miles from my QTH.

Happy St. Patrick's Day!

The ever-skeptical ham wants to see the plots and the circuits. Today we will look first at the all-important Band Pass Filter. (This gives pause that maybe every ham has Thomas as a first name.)

Credit goes to VU2ESE for the BPF used in the Bitx40 and the 1st item is just a lift from that project starting with the plot and then the circuit. The 50 Ohms resistance is for simulation purposes and not installed when placed in the circuit.

 

Now that is a flat Band Pass!

 

This uses standard values!

  

 
We next turn our attention to the J310 Steerable Amp stage that is used as the Rx RF Amp on receive and the Tx Pre-Driver on transmit. This circuit is for 40M and provides 10dB of Gain. We must add the steerable relays so that is in essence the direction of amplification is steered for incoming on Receive to outgoing of Transmit.
 




While the steerable concept is shown above for the P3ST the two SPDT relays control the direction of how the amp works. On transmit the switched connection input is connected to the BPF and the output side goes to the follow-on RF stages. On receive the R contact is connected to the antenna via the TR and the output side goes to the Band Pass Filter. I am looking to make this a single BFR106 stage as it would involve less parts. The 13.3K resistance shown on the schematic is actually a 10k pot wired as a variable resistor in series with a 3.3k 5% fixed resistance. This gives a range of gain adjustment.

A source for the J310's, BFR106 and the ADE-1's all mounted on small circuit boards is from Todd K7TFC . Scroll down to the very bottom of the webpage to find these components!

Them that know can make it go!

Going shopping for vittles today but must stop 1st at the bank to get a home loan so I can afford to buy groceries this week. It is really sad where we are headed -- and some have the audacity to call this on the road to great prosperity. BS as only a few will prosper, and they are already billionaires. What about the rest of us?

73's
Pete N6QW

 

Neutralization.

The word neutralization has many interesting meanings. For Mafia Dons, like the Tony Soprano types, the word has the definition of simply el...