[moved] mc145151 in all bands HF PLL oscillator

[neazoi]

Hi, I have found this PLL circuit with vfo for HF, which seems straight forward and does not require any programming (important to me) and the output is sinewave.

Here is the description from the author:
"Electronic dial uses 3x up / down counter (CD4516 binary counter), CD4013 (flip-flop), and CD4011 (clock, NAND gate).
Each CD4516 IC will control the status of 4 bits MC145151 divider, then these three IC will control the 12 bits divider. By simply taking the 10bit divider (1024), then the other 4 bit dividers (MSB) are used to determine the initial VCO work.
For example, if the status of RA0, RA1, and RA2 is high (open), then the crystal 8.192Mhz will be divided 8192 and give a step of 1Khz and split range from 3-1024 Khz. By setting the status of the last four divisors (N10, N11, N12, N13), the initial frequency can be found). With N10 = 1 (1 = open, 0 = ground), N11 = 1, N12 = 0, N13 = 0; Its initial frequency is 3,072 Mhz. The electronic dial will change the frequency every 1khz up / down to 1024Khz. So the VCO range will vibrate from 3.074Mhz - 4.096Mhz"

I would like to use it in a direct conversion transceiver for all HF bands and I wonder about some points:

1. How to order this rotary encoder, what type is it, how to ask for it at the electroncis shop?

2. In the example, the step size is 1KHz. What configuration of switches and crystal must I choose, so that this step size can be decreased? I only need 200KHz for each HF band.

3. Can I just use the 1KHz step example and manually trim the 8.192Mhz crystal oscillator frequency with a capacitor, to fine tune the VCO? In that case, an estimation of the range of the fine tuning?

 

[d123]

Hi,

**broken link removed**

(It appears in Mouser as "End of Life", not stocked, call to ask. There will likely be a few places online and on dusty shelves that stock them, don't you think? There also appears to be a Motorola version of the same IC.)

 

[betwixt]

They have been obsolete for many years.

The step size in a PLL is determined by the divided reference clock. With 8192 KHz clock and RA0, RA1 and RA2 all high (switches all open) the divisor is 8192 so it gives the 1KHz steps. As the switches are closed, the division ratio is smaller so the steps increase in size which is the opposite of what you need. With that device, the only way to do it is to drop the crystal frequency or provide it with a clock at a lower frequency, the problem that creates though is the number on the 'N' inputs has to be made correspondingly higher to get the same frequency from the VCO.

Basically:
The step size is the crystal frequency divided by 8192.
The VCO frequency is the step size multiplied by the 'N' number.

Brian.

 

[BigBoss]

You don't have to stick to this circuit.MC145151-2 parallel loadable PLL so you can use DIP switches in order to program the PLL. ( read carefully programming guide and especially PLL Loop Filter calculation chapter )

Xtal Frequency= Step Size ( Hz ) x R ( Decimal Value of Reference Divider )
I don't recommend you to shift the Xtal value, just change the Reference Divider Value as shown in the circuit schematic.

 

[neazoi]

You don't have to stick to this circuit.MC145151-2 parallel loadable PLL so you can use DIP switches in order to program the PLL. ( read carefully programming guide and especially PLL Loop Filter calculation chapter )

Xtal Frequency= Step Size ( Hz ) x R ( Decimal Value of Reference Divider )
I don't recommend you to shift the Xtal value, just change the Reference Divider Value as shown in the circuit schematic.

I have thought it so. This would avoid the whole complexity and only the pll chip would be needed. But because I am planning to use it in an all HF bands transceiver, this would make it terribly impractical when one wants to scan the band.

- - - Updated - - -

They have been obsolete for many years.

The step size in a PLL is determined by the divided reference clock. With 8192 KHz clock and RA0, RA1 and RA2 all high (switches all open) the divisor is 8192 so it gives the 1KHz steps. As the switches are closed, the division ratio is smaller so the steps increase in size which is the opposite of what you need. With that device, the only way to do it is to drop the crystal frequency or provide it with a clock at a lower frequency, the problem that creates though is the number on the 'N' inputs has to be made correspondingly higher to get the same frequency from the VCO.

Basically:
The step size is the crystal frequency divided by 8192.
The VCO frequency is the step size multiplied by the 'N' number.

Brian.

I have decided that the complexity of this pll is not worth it for my simple all bander. However, it is the first time I found a complete all HF bands PLL that I could modify to my needs, since there is no PCU programming involved.
For that same reason (complexity) I am more towards using my all bands vxo circuit **broken link removed** combined with a 1-chip 7-segment (lightweight) frequency counter. I have tested it by controlling it with two digital potentiometers DS1669 one for coarse and one for fine tuning of a varactor. This works nicely (although not locked), but the only thing is that ceramic resonators are hard to find at the higher bands. So this tends to be a "hunt for crystals and ceramic resonators" for the bands of interest The DS1669's could be avoided, but I found it convenient and they gave a "digital feeling" in this analogue tuning.

I have tried to find circuits for PLLs, si570 and si5351, DDS chips, max038 and even ltc1799 resistance controlled circuits. The simplest of them should drift and the more complex, are quite complex, compared to the rest of the transceiver.
A DDS would solve many of tha problems in this all bander, but even the smallest DDS projects are complex.

I would be happy to discuss any other solutions you might think of. Note that this is a handheld simple DSB all bander.

 

[betwixt]

I think all the problems hinge on your unwillingness to use an MCU. They are not complicated to program and cost less than your resonators.

If I was to design a low-cost hobby all band receiver (or transceiver) I would use an MCU to handle band switching, programming a PLL and displaying the frequency and make the PLL tune in small steps over say 500KHz. That makes the range wider than most HF bands, the 28MHz band could be treated as four smaller bands.

Then by mixing, under MCU control, you could 'shift' all the HF bands to a common IF range where any demodulator you like could be used.

DDS modules also require some digital control signals but you can buy them on a well known web site for almost nothing. As they stand, their output isn't particularly 'clean' but it is adequate for a simple receiver and you can make a VCO covering up to 30MHz with continuous coverage for less than 20 Euros.

If you really want to use digital potentiometers, remember they are only a variation of DAC where the input signal is used as the reference and the output is Vref/(resolution * input value). You may get better control using a real DAC which typically have 1024 steps instead of the 64 you get with the DS1669.

Brian.

 

[neazoi]

I think all the problems hinge on your unwillingness to use an MCU. They are not complicated to program and cost less than your resonators.

If I was to design a low-cost hobby all band receiver (or transceiver) I would use an MCU to handle band switching, programming a PLL and displaying the frequency and make the PLL tune in small steps over say 500KHz. That makes the range wider than most HF bands, the 28MHz band could be treated as four smaller bands.

Then by mixing, under MCU control, you could 'shift' all the HF bands to a common IF range where any demodulator you like could be used.

DDS modules also require some digital control signals but you can buy them on a well known web site for almost nothing. As they stand, their output isn't particularly 'clean' but it is adequate for a simple receiver and you can make a VCO covering up to 30MHz with continuous coverage for less than 20 Euros.

If you really want to use digital potentiometers, remember they are only a variation of DAC where the input signal is used as the reference and the output is Vref/(resolution * input value). You may get better control using a real DAC which typically have 1024 steps instead of the 64 you get with the DS1669.

Brian.

If I knew how to program an MCU I would definitely use a pll or a dds and I could also drive a display and add some more features to the whole circuit.
So I tend to think more on a classic analogue circuit, but adding some digital "tricks" to it (it still remain analogue).
In that context, I am trying to find the simplest option that gives more features, it's a poor man's project anyway.
I am sure you understand what I mean.
If I can't do it simple enough, I do not think I have another option than to use a DDS project from the web.

 

[betwixt]

I fully understand but this project is best designed with a mix of hardware and software, for example, it is easy to replace all the circuitry in the lower half of the schematic in post #1 with a single inexpensive MCU and have it drive an LCD frequency display as well.

With a little care, the PLL can be incorporated into software as well. The timers on most PIC processors for example can measure frequencies up to 50MHz with no additional circuitry.

Somewhere on the internet (I 've lost the link) is an interesting VFO using a frequency locked loop rather than a PLL. Basically, you tell it what frequency you want, it measures the frequency from the VFO and it produces a 'go higher' or 'go lower' voltage to the varicap to adjust the frequency.

Brian.

 

[BigBoss]

In fact it's so simple to change the frequency by using binary UP/DOWN counter.First of all, you should fix the comparison frequency by fixing R counter.
Let it be 200kHz and then every changes in LSB will increase or decrease the VCO frequency by 200kHz.
So by using 4bits binary UP/DOWN counter will change the VCO frequency 16 steps so 800kHz deviation can be obtained.
Note : 4bits binary counter with Up and DOWN push-buttons can be found on the internet.Quite simple circuit..

 

[neazoi]

I fully understand but this project is best designed with a mix of hardware and software, for example, it is easy to replace all the circuitry in the lower half of the schematic in post #1 with a single inexpensive MCU and have it drive an LCD frequency display as well.

With a little care, the PLL can be incorporated into software as well. The timers on most PIC processors for example can measure frequencies up to 50MHz with no additional circuitry.

Somewhere on the internet (I 've lost the link) is an interesting VFO using a frequency locked loop rather than a PLL. Basically, you tell it what frequency you want, it measures the frequency from the VFO and it produces a 'go higher' or 'go lower' voltage to the varicap to adjust the frequency.

Brian.

I wish I had more tome to learn how to program an MCU... many great things could be done in minimum of hardware then.

Here is an extensive reference about FLLs. http://hanssummers.com/huffpuff.html I have looked all of them in the past, but as I can recall I could not find a circuit to operate at all HF bands. The best I could find started from 8MHz and upwards. Indeed, it could be an option as well, however, again, band switching inductors would be needed. The only benefit compared to band switching crystals I described earlier, is the greater range (if crystals are used). However, if ceramic resonators are used, the frequency should be quite stable without an FLL and they can be pulled to 80KHz at 40m (tested).

To avoid band switching coils a dds would be probably be the best option. However, do you have any ideas of how one could lock an LTC1799 with an fll or pll? This might do for all bands without inductor or crystal switching and in a simpler manner than the dds.

- - - Updated - - -

In fact it's so simple to change the frequency by using binary UP/DOWN counter.First of all, you should fix the comparison frequency by fixing R counter.
Let it be 200kHz and then every changes in LSB will increase or decrease the VCO frequency by 200kHz.
So by using 4bits binary UP/DOWN counter will change the VCO frequency 16 steps so 800kHz deviation can be obtained.
Note : 4bits binary counter with Up and DOWN push-buttons can be found on the internet.Quite simple circuit..

BigBoss, unfortunatelly 800KHz step is not useable in a 200KHz or 1MHz band range.

 

[betwixt]

It probably isn't worth trying to lock an LT1799, the conversion of error voltage to controlled resistance is more hassle than it's worth and it produces a square wave output which would need selective filtering to 'clean up' before you could use it as a receiver LO. Using it without a frequency locking mechanism would be pointless as it is only specified to be accurate to +/- 300Khz at 20MHz making it quite likely to be so far off it would be out of band altogether.

The FLL reference I found before is here: http://www.next.gr/microcontrollers/pic/Frequency-Locked-Loop-for-HF-l7223.html a bit closer to home!

It doesn't show the VCO part but it demonstrates how simple a control system can be, the connections on the left side go directly to a standard LCD display module. The 16F628 is now obsolete but there are pin compatible and cheaper replacements available which could for example allow you to add a keypad or band switching circuit as well.

There is another trick you could consider but again it requires some software assistance, my HF225 receiver uses a PLL to generate large steps in frequency (1KHz I think) then a DAC to generate a 'fine tuning' voltage, increasing the resolution to about 10Hz. The DAC is adjusted until it has covered enough small steps then it is reset and the PLL makes the next big jump. It gives, in theory, continuous coverage but the HF225 suffers from slight temperature drift in the DAC which means there is sometimes a noticable jump in CW/SSB tone as it goes from one PLL step to the next.

Brian.

 

[neazoi]

It probably isn't worth trying to lock an LT1799, the conversion of error voltage to controlled resistance is more hassle than it's worth and it produces a square wave output which would need selective filtering to 'clean up' before you could use it as a receiver LO. Using it without a frequency locking mechanism would be pointless as it is only specified to be accurate to +/- 300Khz at 20MHz making it quite likely to be so far off it would be out of band altogether.

The FLL reference I found before is here: http://www.next.gr/microcontrollers/pic/Frequency-Locked-Loop-for-HF-l7223.html a bit closer to home!

It doesn't show the VCO part but it demonstrates how simple a control system can be, the connections on the left side go directly to a standard LCD display module. The 16F628 is now obsolete but there are pin compatible and cheaper replacements available which could for example allow you to add a keypad or band switching circuit as well.

There is another trick you could consider but again it requires some software assistance, my HF225 receiver uses a PLL to generate large steps in frequency (1KHz I think) then a DAC to generate a 'fine tuning' voltage, increasing the resolution to about 10Hz. The DAC is adjusted until it has covered enough small steps then it is reset and the PLL makes the next big jump. It gives, in theory, continuous coverage but the HF225 suffers from slight temperature drift in the DAC which means there is sometimes a noticable jump in CW/SSB tone as it goes from one PLL step to the next.

Brian.

The PIC circuit is recommended to only 20-25MHz max

Have a look at this circuit I have found. He claims to operate fine to 30MHz. I have found the HCF4517BE chip already at low price.

I am not sure about the clock frequency that must be used to lock to 30MHz. Probably a canned oscillator will used for convenience, but at what frequency (based on the max frequency specs of the various ICs in the circuit. (I already have a 74F74 /120MHz clock).

It seems quite a simple circuit, so I would like your advice on this.

 

[BigBoss]

BigBoss, unfortunatelly 800KHz step is not useable in a 200KHz or 1MHz band range.

800kHz is not frequency step, it's total useful bandwidth with "200 kHz frequency step"
You can change the step size by changing R Divider value but there will be needed bit size of the n-bit counter will be higher ( for instance 5,6 bits )

 

[betwixt]

It seems quite a simple circuit, so I would like your advice on this.

That isn't the complete circuit, just an add-on stage to the original huff-n-puff stabilizer. I understand how it works but I'm not sure how it's claim to lock faster is true as the speed the tuning voltage can change is decided by the 2M and 2u2 RC filter.

Regardless, there is another more practical way of speeding up a PLL based VFO, if the new frequency is more than a few steps away from the present frequency you increase the charge current to the voltage pump by partially or completely shorting the resistive part of the loop filter. It makes the jump in frequency much faster until almost in lock then reverts back to the slower time constant to hold it there.

Brian.

 

[neazoi]

That isn't the complete circuit, just an add-on stage to the original huff-n-puff stabilizer. I understand how it works but I'm not sure how it's claim to lock faster is true as the speed the tuning voltage can change is decided by the 2M and 2u2 RC filter.

Regardless, there is another more practical way of speeding up a PLL based VFO, if the new frequency is more than a few steps away from the present frequency you increase the charge current to the voltage pump by partially or completely shorting the resistive part of the loop filter. It makes the jump in frequency much faster until almost in lock then reverts back to the slower time constant to hold it there.

Brian.

I know I have mentioned different ways to create this oscillator needed. I apologize, to be thinking back and forth on the topic, but I am trying to find the best way (and simpler for me) to do it.

I read BigBoss post #4 and the DIP switches option does not seem too bad, apart from the fact that it is difficult to scan the band.
You see quite a lot of switching must be used for switching the 9 coils in the oscillator and possibly some front end filter switching (for the 9 bands) anyway. So If I have two 12-DIP switches, I could use their combination (24 positions) for frequency setting. The first 14 switches will switch the pll frequency and band and the rest will switch the band. Since there is a counter in the transceiver, the operating frequency will be displayed, so you do not have to worry whether you have made a mistake in the binary switching.

This isn't proffessional, but for hobby use, this might be more than enough.
Later on, I could replace the lots of switches with a single micro (big one) to do the job, with up/down buttons, if I find out how to program it.

I am now looking to find an all bands vco for the project that uses a coil without a tap, so that I could use molded chokes as inductors and simplify the design.

You told me not to do it, but I wonder how much could I pull the frequency by trimming the 8.192MHz reference oscillator in the pll?

Also, which replacements can I use instead of the 2sc2053?

 

[betwixt]

The 8.192MHZ is divided down by 8192 times so pulling it will have almost no usable effect after division. For example, if you pulled it 1KHz the final tuning would only move by 1Hz!

You can use almost any small RF transistor as a replacement for the 2SC2053, it has no special properties.

Bear in mind that if you do go down the MCU route, even the cheapest ones have many I/O pins you can use to feed the PLL. Even the humble PIC16F628A has 16 IO pins while only having 18 pins in total! Personally, I would go for a slightly larger device with a few extra pins and an ADC so you can implement full band switching and auto-scan facility so it automatically sweeps the band then stops when a signal is found. Writing the software is VERY easy and at least for PICs the assembler/compiler is free - don't let that side of things hold you back.

Brian.

 

[neazoi]

The 8.192MHZ is divided down by 8192 times so pulling it will have almost no usable effect after division. For example, if you pulled it 1KHz the final tuning would only move by 1Hz!

You can use almost any small RF transistor as a replacement for the 2SC2053, it has no special properties.

Bear in mind that if you do go down the MCU route, even the cheapest ones have many I/O pins you can use to feed the PLL. Even the humble PIC16F628A has 16 IO pins while only having 18 pins in total! Personally, I would go for a slightly larger device with a few extra pins and an ADC so you can implement full band switching and auto-scan facility so it automatically sweeps the band then stops when a signal is found. Writing the software is VERY easy and at least for PICs the assembler/compiler is free - don't let that side of things hold you back.

Brian.

The thing I worry about the pll, is that this particular pll IC can accept an input frequency of no more than 25MHz. This leaves the 10meter band out.