Forums :: NEW! Web Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Jan Fri 22, 2021 2:17 am


All times are UTC [ DST ]





Post New Topic Post Reply  [ 10 posts ] 
Author Message
 Post subject: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Sun 13, 2020 11:17 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2495
Location: Saskatoon
I'm not sure if this thing qualifies as a frequency synthesizer or not. I guess it depends on how broad the definition of frequency synthesizer is. This circuit tunes from about 550kHz to 1700kHz in 5kHz steps. I don't have any specific purpose in mind for it at the moment. It may end up being the master oscillator for a Part 15 transmitter, or it may be a local oscillator for a receiver project. Mainly, this was an exercise to see if such a thing could be built, and to keep me out of trouble during the lockdown.
Attachment:
Layout04s.jpg
Layout04s.jpg [ 372.43 KiB | Viewed 480 times ]

The circuit is a variant of a Phase Locked Loop. Unlike the normal variety of PLL frequency synthesizers that use a digital counter to divide the frequency in the PLL feedback loop, this one uses a sampling circuit running at 5kHz. The output of the phase comparator is an aliased signal in the 0-2.5kHz range. This variety of PLL is known as an Alias Locked Loop. Otherwise, the circuit is a fairly standard PLL, except that it uses tubes.

The first job was to come up with a precise 5kHz reference oscillator. The lowest frequency crystals I could find were 20kHz. So, I used a 6U8A triode/pentode, with the pentode section as the crystal oscillator, and the triode section as a divide by 4 blocking oscillator, running at 5kHz and synchronized to the crystal oscillator.
Attachment:
MainCct_R3s2.png
MainCct_R3s2.png [ 67.57 KiB | Viewed 480 times ]

The phase comparator uses a 6JH8 sheet beam tube as a balanced mixer, which acts as both the subsampler and phase comparator. The plate signals drive the 4 diode charge pump to create the DC control signal for the VCO. The first version of the circuit used 1N4148 solid state diodes, but I switched to 6AL5's when I built the PC board. I found no difference in performance between the solid state diodes and the tube diodes.

The VCO uses another 6U8A. The pentode section is connected as a reactance tube to act as a voltage controlled variable inductance in parallel with the fixed inductance of the oscillator coil. The triode section is a standard Hartley oscillator.

Main tuning is done with the variable capacitor, but when the frequency gets close to one of the 5kHz increments it snaps to that frequency and then stays locked on until the tuning cap is moved far enough to lose lock.

Fortunately, all of the circuit components were readily available. The 20kHz crystal and the blocking oscillator transformer (actually a pulse transformer) are from Digikey. The oscillator coil is the universal replacement type from Antique Electronic Supply, Part No. P-C70-OSC.

As I got the various sections of the circuit debugged I made PC boards, and transferred those sections of the circuit from the breadboard to PC boards. The first was the crystal oscillator/blocking oscillator frequency reference (upper left in photo). Next was the phase comparator and charge pump (upper right in photo). The last section was the VCO (bottom right). The loop filter will also be on the VCO board once the component values are finalized. For now, the loop filter is on the breadboard.

The meter in the upper right of the photo is used as a centre tune meter and has a high impedance two MOSFET (temporary) differential preamp on the breadboard. (It's connected to points M1+ and M1- on the schematic.) Also on the breadboard are a series string of four 33V zeners used as a 132V voltage regulator for the blocking oscillator. I added this when I was having stability problems, but it turned out not to be a power supply problem. So, I think I can probably go with an unregulated supply. I'll test this later. If it does need to be regulated, I'll use a 0B2 regulator tube. The only solid state component I don't have an easy replacement for (yet) is the pair of protection zeners in the crystal oscillator circuit.

At this point, the only thing missing from the circuit is a reliable lock indicator. The centre tune meter can be used for detecting PLL lock, but it's not intuitive. The meter is mainly intended as an aid for adjusting the tuning cap so that it's in the centre of the lock range. For detecting actual lock acquisition, the simplest thing I found, so far (aside from having a scope permanently connected), is to take the two signals from the 6JH8 plates (connection points M2+ and M2- on the schematic) and feed them into a differential amplifier and then send the output of that to a speaker. The differential amp removes the 5kHz sample frequency because it's common mode, leaving only the error frequency which will be in the audio range (0-2500Hz). When it's off sync, it generates a squeal that gets lower in frequency until lock is acquired, at which point it goes silent.

The last photo shows the VCO output in the upper scope trace synchronized to the reference signal (lower trace), and the frequency counter below shows the frequency to be 850.000 kHz. (The frequency counter was calibrated to WWV, and was used to trim the reference oscillator.)
Attachment:
ScopeFrqCtr_s.jpg
ScopeFrqCtr_s.jpg [ 389.32 KiB | Viewed 480 times ]


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Sun 13, 2020 5:58 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 956
Location: Burke, VA 22015
Nice work!
I really like how you sup-sample the VCO output for comparison with the 5kHz reference. That allows you to do away with the frequency divider.
Can you explain how that reactance tube works?

_________________
Binh


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Sun 13, 2020 11:02 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2495
Location: Saskatoon
The reactance tube circuit uses a resistor/capacitor phase shifter: the resistor is between plate and grid, and the capacitor between grid and ground. The signal voltage appearing at the plate will be shifted 90° at the grid due to this network. Since plate current is proportional to grid voltage, the plate current is also shifted 90° relative to plate voltage. Thus, from plate to ground (or plate to B+) it will appear to be a reactance. In this case the reactance is inductive (swapping R and C makes it capacitive). The amount of inductive reactance varies with tube gm (transconductance) which varies with grid bias. So, the amount of inductance is controlled by grid bias. The two components in this circuit that set the reactance characteristic are Cr and Rr, which are somewhat intermixed with the loop filter components. This was to avoid having to add a bunch of additional DC blocking capacitors and bias resistors (or a buffer stage). For this reason, loop filter capacitor C3 is located where it is in order to block DC between the plate and grid, and since it's much larger than Cr, it doesn't affect the reactance characteristics too much and still performs its loop filter function. This also allows a DC path for the VCO control voltage from the loop filter to the grid.

On the oscillator side, the variable inductance from the reactance tube is in parallel with the 22pF capacitor and the secondary winding on the oscillator coil. This works to adjust frequency in the same way as when you have a bandspread variable capacitor connected to a tap on the oscillator coil. In this case, instead of tapping down on the coil, the separate winding is used, and the inductance is variable while the capacitance is fixed. This arrangement gives smaller adjustment range than if connected directly to the main oscillator coil winding. In my initial experiments, I found that the reactance tube had way too much effect on frequency when connected to the main winding, and it needed to be tamed down. Using the separate winding also solves the problem of feeding B+ to the reactance tube.


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 12:54 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4311
I LOVE THIS.

This has to be a true original invention, and I'd take the bet that it's not been done like this before. I love that analog treatment that snaps into a stable frequency.

Glad to know there are different kinds of PLLs. (I wonder how many kinds there are.)

It's a bit off beat (negative pun intended) which is always refreshing to see. Kind of like my digital clock radio that the segments are little windows with mechanically operated shutters that open and close to make the segments glow. A legitimate and novel way of doing things.

WAY TO GO, BOB!


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 1:18 am 
Member
User avatar

Joined: Mar Sun 22, 2020 5:56 am
Posts: 1160
Location: Arvada, CO, 80004
WUNDERFUL!

I've got a hot idea for it. If it is possible, replace the variable capacitor with a varactor diode. Then connect the output to grid 1 of a 12BE6 or 6L7, with a LC tuning circuit on grid 3. Instead of a variable capacitor, a varactor diode again. Create a sawtooth oscillator with a 6Q5G/884/885 or a 6D4, and link it to the varactors, along to a X scale on a CRT tube (906/3AP1). The output of the 12BE6 or 6L7 then goes through the normal AA5 radio stages, up the amplifier. That signal is sent to the 3AP1's grid. The AGC control is amplified, as if it was to be used for a magic eye tube. But it would go to the Y scale of the 3AP1 tube. What this whole thing does, is it plots the whole AM band on a Picture of a CRT tube, with all of the modulation plotted. A separate AA5 radio can be implemented, and it's dial pointer in front of the CRT tube to show station's strength, modulation, and relative position.

Anyway WHAT A WELL DESIGNED CIRCUIT BOB!

_________________
Electronics are filled with smoke. It’s my job to put the smoke back in when they fail.
Cheers,
Jay


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 2:32 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2495
Location: Saskatoon
Thanks for the comments guys. Much appreciated.
Macrohenry wrote:
This has to be a true original invention, and I'd take the bet that it's not been done like this before.

I think you're half right. Alias locked loops have been around for awhile, and they've seen more interest lately as communications technology gets into the higher GHz range, where digital counter performance starts to suffer. As for no one having done it this way before (using tubes), that's probably correct, and the reason would be due to a shortage of sufficiently misguided individuals that are allowed access to a soldering iron.
Mr. Highlander wrote:
If it is possible, replace the variable capacitor with a varactor diode. Then connect the output to grid 1 of a 12BE6 or 6L7, with a LC tuning circuit on grid 3. Instead of a variable capacitor, a varactor diode again. Create a sawtooth oscillator with a 6Q5G/884/885 or a 6D4, and link it to the varactors, along to a X scale on a CRT tube (906/3AP1). The output of the 12BE6 or 6L7 then goes through the normal AA5 radio stages, up the amplifier. That signal is sent to the 3AP1's grid. The AGC control is amplified, as if it was to be used for a magic eye tube. But it would go to the Y scale of the 3AP1 tube. What this whole thing does, is it plots the whole AM band on a Picture of a CRT tube, with all of the modulation plotted. A separate AA5 radio can be implemented, and it's dial pointer in front of the CRT tube to show station's strength, modulation, and relative position.

Hmmmmm... Well, this is the most tubes I've ever used in a project. I think maybe I'll leave the vacuum tube mega-projects up to you. That's more your specialty. Should be fun though.


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 2:55 am 
Member
User avatar

Joined: Mar Sun 22, 2020 5:56 am
Posts: 1160
Location: Arvada, CO, 80004
Just an idea really. Not much thought into it. I do tend to really puff circuits up with tubes. However, I have seen your other projects like the Stereo AM tube transmitter, and the 2V tube radio. It really is amazing.

_________________
Electronics are filled with smoke. It’s my job to put the smoke back in when they fail.
Cheers,
Jay


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 3:18 am 
Member
User avatar

Joined: Oct Sat 03, 2020 10:41 pm
Posts: 78
Wow, this is an incredible piece of engineering. Bravo!

_________________
Faith is ... a belief in things not seen. (Heb 1:1)


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 3:27 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2495
Location: Saskatoon
My philosophy is that no idea is too crazy to dismiss. If you're stubborn enough not to give up, persistence eventually pays off.

BTW, the devil is in the details, as they say. One of the things that I felt was critical for this to be a useful circuit is that the output be repeatable over a power cycle.

If it's operating, locked on some frequency, and then the power is shut off, left overnight, and then powered up the next day, it should come back on and lock to the same frequency without any fiddling of controls. This is the reason for the two trimmer potentiometers in the loop filter and reactance tube cathode. With a 1k cathode resistor, the cathode biases itself to about 6 volts. The best operating point was found to be with the grid at about -2.2 volts relative to the cathode. The cathode pot and 1 Meg. resistor are used for setting this. The pot is adjusted to get the -2.2 volts on the grid (a one time adjustment). However, the charge pump tends to add some further negative bias to the grid. If the AFC is switched off, then the lack of charge pump bias shifts the operating point, and the oscillator exhibits a sudden shift of frequency. It may shift far enough that it won't pull back into lock if the AFC switch is moved back to the ON position. That is the reason for the second pot and 1 Meg. resistor connected to the OFF contact on the AFC switch. This resistor/pot combination puts the same amount of bias on the reactance tube grid that the charge pump does. Thus, when the AFC is switched off there is very little change in frequency, and when switched back on, the oscillator immediately locks.

As a side note: the cathode of the top diode in the charge pump diode stack is connected back to the cathode of the reactance tube. The cathode voltage acts as the upper reference voltage, and prevents the loop filter output voltage from taking the grid positive relative to the cathode.

Another side note: Originally, I was planning to use a 6AL7GT tube as a centre tune indicator. I'd bought it from a reputable tube dealer, and it was supposed to be NIB. It turned out to be a dud. It appears that someone had stuck a dead tube in the box, and the tube seller must not have tested it before I got it. The phosphor was completely gone. :(


Top
 Profile  
 
 Post subject: Re: My Vacuum Tube Frequency Synthesizer
PostPosted: Dec Mon 14, 2020 4:46 am 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 663
Location: Equinunk PA 18417
Wickedly clever!

_________________
-Richard


Top
 Profile  
 
Post New Topic Post Reply  [ 10 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: Retrovert and 7 guests



Search for:
Jump to:  


































Privacy Policy :: Powered by phpBB