Frequency Synthesizer using PLL

[yoboy]

Hi there, im a noob here.. i need help in designing and building a PLL frequency synthesizer using a PLL IC. I am planning on using 74HC4046, even though the only info i know about this ic is that it has superior phase comparator sections. The minimum required synthesizer’s specifications are as follows.

Frequency range: up to 10 MHz or better
Output voltage range: 0 to 5 V

Any form of help will be much much appreciated

 

[RadioRon]

First, some questions:

- do you want sine or square wave output?
- what is the low end of the frequency range that you need?
- what will it be used for?

Next, some tips. First of all, you may find that, while the 4046 is a good choice for a basic PLL operating at one frequency, your circuit will get somewhat complicated when you make it synthesize many frequencies. At the least, this will require several counter ICs. So, to make things easier, have a look at some synthesizer chips that include all the necessary logic into one part. You'll be happy to know that they all use the same structure of superior phase detector as found in the 4046.

A good starting point is an old part that I used back in the late 70's. Its old, but easy to use and quite reliable. Here it is:
https://www.electro-tech-online.com/custompdfs/2009/02/MC145151-2.pdf

It may be hard to get, so consider others too. I've had good luck with the National Semiconductor PLLatinum series. There are many, here is an example;
https://www.electro-tech-online.com/custompdfs/2009/02/LMX2310U.pdf

Some of the parts out there are tailored to use in phones and such, but many are good for general purpose use.

One of the main limitations in trying to go from 0 to 10MHz in a basic PLL synthesizer will be in trying to use a single VCO. We don't see many that span such a large percentage range. But there are clock synthesizers that get close so have a look at how those work too. Synthesizers for communications use rather than for logic clocks usually have to have low phase noise, which demands a different approach than is used by most clock synthesizers.

One simple approach that can work well is to synthesize a higher frequency, like 50 to 60 MHz for example (VCOs are much easier to get for such a range), and then simply mix the output of that with a 50 Mhz oscillator and then pass the result through a 10 MHz low pass filter. This will give you 0 to 10 Mhz.

edit:
Just want to add that my first working synthesizer was also built with a 4046 and an array of other chips. the 4046 is really good for learning how the phase detectors work. The entire design worked well in wire-wrap form, but mainly because I chose to use only 4000 series logic which has slow rise/fall times. I had built something similar previously using 74LS logic in wire-wrap form and it was a disaster of noise and glitches due to the lack of a good pcb layout. The one that worked was designed to replace a bank of crystals all around 7 MHz and 4000 logic was fine for that (with a few tricks added in to keep the frequencies low). So, you can use the 4046 if you are up to putting six or seven ICs together.

 

[yoboy]

hey there radioron, first of all, i would like to thank you for looking upon my problem. For the questions that u asked:

-do you want sine or square wave output?
Our lecturer did not specify any requirement for this, so whichever output which gives a higher output should do.

- what is the low end of the frequency range that you need?
the low end for the frequency range was not specified either, so I assume it is fine to use a frequency which can compromise with the IC that is being used. Ideal case will be from 0 to 10MHz.

- what will it be used for?
This is an assignment in my university, in which the primary objective is to design a Frequency Synthesizer ranging from 0 to 10Mhz using a phase locked loop IC with output voltage ranging from 0-5V. This is the minimum specifications given to us. Extra features to the synthesizer is welcomed.

I have also looked into MC145151 and LMX2310U datasheets. If im not mistaken, i find that both this ICs are frequency synthesizer ICs. Our task required us to use a PLL IC to design a frequency synthesizer. I hope I am not confusing you.

We have no knowledge in this field, so we are kind of completely clueless. Our lecturer suggested us using 74HC4046, so basically we know nothing else. But we are learning more and more by each hour. I find your last paragraph very helpful. We will have more research done on this part. Again, thanks a lot for spending your time to entertain our problem.

 

[RadioRon]

If the goal is to learn, then working with the 4046 is a good idea.

I find this confusing: "...using a phase locked loop IC with output voltage ranging from 0-5V". Does he mean that the phase detector output range should be 5V, or does he mean that the AC output from the entire synthesizer should have a 0 to 5V peak to peak swing? I suspect the former as it makes more sense to specify that. Perhaps you should confirm this with your teacher. In either case, though, it means you will use a 5V power supply voltage.

Since you are not making a communications synthesizer, it would be fine to use the VCO that comes inside the 4046. You can tune the range of this VCO to suit your design and it will indeed be able to tune down very close to 0Hz. There are different versions of the 4046. The oldest is the CD4046, which is built in the original 4000 series logic. It is fairly slow and its VCO probably won't oscillate above around 1 MHz. A more modern part would be the 74HC4046 whose VCO will oscillate well beyond 10MHz.


There are many different ways to structure a synthesizer. The simplest structure is where the VCO is at oscillating at the output frequency, and the feedback divider is an integer type. You will probably want to use this basic structure. You will have to add the divider chips to the 4046 to make it work. Study the data sheets for the 74HC4046 as well as for the CD4046 as they explain things a bit differently and this gives you more insight into the parts operation. The key design decision that you face is to choose what the frequency will be going into the phase detector. This choice will determine how fast your synthesizer will lock and how many individual frequencies that you can synthesize between 0 and 10MHz.

 

[yoboy]

Hey there again radioron, its always a pleasant to see a reply from you. First of all, after confirmation from my lecturer, i found out that ur assumption about the output voltage range is right. 0-5V is the is the AC output peak swing from the entire synthesizer.

We have studied alot about 4046 IC and how it can be used to form a frequency synthesizer. We found out that, by itself, 4046 serves as a PLL IC, and by adding programmable divider and a reference frequency, a frequency synthesizer can be formed (correct me if i am wrong). You have also mentioned about the divided chip. We have some doubts though. What can we use to form this divider block? Is there a special IC that can be used as a programmable divider?

Looking forward ur reply, thanks is advance

 

[RadioRon]

You are correct, that you can take a PLL IC like the 4046 and add external dividers to make a synthesizer. Before I suggest some divider chips, read the information in this link under "PLL Basics" and "Synthesizers".
**broken link removed**

This next link explains things a little bit differently, but you may learn something from it. You don't need to read the whole thing because he goes into more advanced topics about the synthesizer than you care about at the moment. The key to understanding the function of the external divider is in the section called "Higher Frequencies" so read up to and including that one at least.
**broken link removed**

The best counter type to use would have these terms in the description:
"presettable" , "BCD counter". There are several types that could work, some that count up and some that count down. The main thing is that you want the counter to put out one pulse for every xx input pulses and you get to program the value of xx using parallel BCD inputs. An example of a part that can be used is the 74HC160.

 

[yoboy]

Hey there, sorry for my late reply. I have read the useful websites that you've given. I have few questions in my mind though.

First, from the 2nd link that you gave, there was a LP filter constructed. Can i use the same values for the component to construct mine? I do not know if it will be sufficient/insufficient to support my frequency range.

Next, I have lots n lots of doubts about the dividers. Please correct me if i'm wrong. In our case, a divider is a pre-programmed IC which will divide the frequency right? That also means, the output radio frequency from the VCO will be the reference frequency multiplied by the divided value, right? What confuses me is, why is there a range given to me by my lecturer? Wouldnt the output from the VCO will always be the same for a given programmed divider?

Lets say im using a 1kHz reference frequency, and a programmable divider which divide by 1000. By using this, my theoretical VCO will always be 1Mhz right? But if i want to change it into 2Mhz, i still do have to reprogram my divider, yes? Is there a way to alter the divider value without reprogramming it?

Looking forward ur reply. Thanks alot in advance.

 

[RadioRon]

The author in the second link shows three loop filters in that section. The first one is too simple, but the second one, the one that has damping, is a good choice. The component values in the filter are a function of the loop bandwidth and the phase detector comparison frequency which are usually chosen by the designer. If you use the same or similar comparison frequency, like 1 KHz for example, then you can use the same values as a starting point. But once you have your loop working, you will find that tuning the values of the loop filter is an interesting and useful experiment.


"Next, I have lots n lots of doubts about the dividers. Please correct me if i'm wrong. In our case, a divider is a pre-programmed IC which will divide the frequency right? Yes. However, they are preprogrammed by you, not by the manufacturer.

That also means, the output radio frequency from the VCO will be the reference frequency multiplied by the divided value, right? Exactly right, yes.

What confuses me is, why is there a range given to me by my lecturer? Wouldnt the output from the VCO will always be the same for a given programmed divider? You are misunderstanding the key point of the divider. It is not preprogrammed by the manufacturer, rather it is preprogrammed by you during your design. A "pre-settable" bcd decade counter is a type that gets pre-set to a specific count every time it puts out an output pulse. Each output pulse cycle, the presettable counter reads in a pre-set value as a starting point and begins counting down from that number. It reads the preset value from parallel inputs which you must configure as either ground or 5 V (either zeros or ones). Most designers use a row of dip switches to make this easy to do for many different values. This is how you can program any value into your counter, and therefore change the output frequency of your synthesizer.


Lets say im using a 1kHz reference frequency, and a programmable divider which divide by 1000. By using this, my theoretical VCO will always be 1Mhz right? But if i want to change it into 2Mhz, i still do have to reprogram my divider, yes? Is there a way to alter the divider value without reprogramming it? It is quite easy to alter the divider value as mentioned above. You must read the data sheets for some dividers and maybe do some research to see exactly how the schematic will look.

 

[yoboy]

Hi radioron, I would like to applogise for my misleading question. I do understand that the divider is not preprogrammed by the manufacturer. What i was trying to ask was, is there a way to change the divided-by value on the go. For example, when we are doing a demonstration/presentation for the hardware once completed, we might have to show that the device can synthesize other frequency other than the one which was being showed to the lecturer. To do this, do we have to reprogramme the IC or is there a way to do it on the go? Sorry again..

We have gathered lots and lots of information, thanks to you. We will plan on designing the hardware by next week, then shop for the components required. We hope you dont mind guiding us more in the advanced stage of our project in the future. Thanks again for being very helpful.

 

[yoboy]

Sorry to bother you.. W have been cracking our heads on how to make the divider tunable.. Our synthesizer ought to be tunable on the spot. One suggestion that we come up with is to use a microcrontoller. Is there a way to do this?

 

[Nigel Goodwin]

What do you mean by 'tunable'? - you could use a microcontroller to set the division ratio, or you could use logic chips, or simple switches.

Any remotely modern synthesiser is likely to be set by software, and would certainly require a micro.

 

[yoboy]

Hi Nigel goodwin..

I learned that the divider ICs must be programmed to set the ratio. But we cant be using this forms of dividers, since we should be able to adjust the output of the synthesizer by tuning the device (which we suppose to build) on the spot. Can you fill me more about how the frequency divider can be formed using a microcontroller?

 

[Nigel Goodwin]

Hi Nigel goodwin..

I learned that the divider ICs must be programmed to set the ratio. But we cant be using this forms of dividers, since we should be able to adjust the output of the synthesizer by tuning the device (which we suppose to build) on the spot. Can you fill me more about how the frequency divider can be formed using a microcontroller?

It doesn't - you use the micro to set the division ratio of the programmable divider.

'Programmable' doesn't mean it needs programming, just that it has a number of pins that you set high or low to set the ratio you want.

 

[yoboy]

wow, thats new.. Does that mean, by using a micro controller, i can input high or low to the divider to set the divison ratio? Radioron suggested 74HC160 divider IC. Can i be using this IC for this purpose? What about the micro-c, any suggestion on which one will be perfectly suited for the job? We dun have any other functions to our design, so a bulky controller will really be inefficient since its only role now is to control the divider. Appreciate your help.

 

[RadioRon]

A microprocessor is not needed and is actually a distraction and waste of time. I'm not sure if you understand how to make a synthesizer tuneable on the spot. Let's review some basics first.

As the VCO is oscillating, its output is feeding the input of the programmable divider constantly. The divider is constantly dividing the VCO frequency by some ratio to get down to the phase detector comparison frequency. As long as the VCO is oscillating there is an output from the programmable divider.

The programmable divider goes through a cycle where it re-loads itself every so-many (10 or less) cycles of input. This re-loading is done from the parallel inputs automatically by the divider chip. So, you can program the dividers using switches that pull each parallel input to either high or low voltage. The divider chip will continuously read these inputs as it is operating. So, the moment that you change one of the switches, the divider will read that in and begin to divide with this new ratio. The new ratio forces the synthesizer to push the VCO to a new frequency.

So changing the frequency of the synthesizer is as simple as changing the parallel inputs to a different bcd number by changing the switches. Everything else happens automatically.

 

[yoboy]

The programmable divider goes through a cycle where it re-loads itself every so-many (10 or less) cycles of input.

By reload, do you mean reload a new divider ratio?

This re-loading is done from the parallel inputs automatically by the divider chip.

What are these parallel inputs? are they just simply high voltage levels?

So changing the frequency of the synthesizer is as simple as changing the parallel inputs to a different bcd number by changing the switches.

What do you mean by parallel inputs to a different bcd number?

I can understand the working principle of the synthesizer now, and how to make it tunable. But im still kinda vague about the divider part. For now, wat I understand is the divider's ratio can be increased or decreased by supplying voltage in to certain pins (am i right?). What I don't understand is about BCD number..

Will it be wise for me to buy the 74HC160 divider IC to run a few tests? This IC is capable of dividing like how you described right?

 

[RadioRon]

By reload, do you mean reload a new divider ratio?

Yes.

What are these parallel inputs? are they just simply high voltage levels?

Look at the data sheet for the 74HC160. Pins 3, 4, 5 and 6 are called D0 to D3 or otherwise "data inputs". These are the parallel inputs. You present a binary number to these inputs (where in binary a 0 means 0 volts and a 1 means 5 Volts) with D0 being the least significant digit of the number.

What do you mean by parallel inputs to a different bcd number?
BCD means Binary Coded Decimal. It means a binary number of 4 bits that is limited to range from 0 to 9 decimal.

I can understand the working principle of the synthesizer now, and how to make it tunable. But im still kinda vague about the divider part. For now, wat I understand is the divider's ratio can be increased or decreased by supplying voltage in to certain pins (am i right?). What I don't understand is about BCD number..
In general you are right, but at this point, I think you need to do some googling on your own to answer these basic questions. Alternatively, you can study a text on basic electronic logic, and especially on sequential logic including flip flops as these are the basic building blocks of dividers.

Will it be wise for me to buy the 74HC160 divider IC to run a few tests? This IC is capable of dividing like how you described right? As you are a beginner I think it is necessary for you to understand basic NAND, NOR type logic gates and also to understand basic flip flop operation first. If you are going to buy an IC, you should be prepared that you will destroy a few by making mistakes and so you may consider buying many of these, perhaps up to 4 spare parts for each one you think you need. You may find it a good idea to buy a small number of basic gates and a flip flop IC simply for experimentation to understand their operation.

 

[yoboy]

Thank you very much.. I will post it here again on my updates later on.. U have been very helpful..

 

[Nigel Goodwin]

Perhaps you might consider buying a book about PLL's?

I have an excellent one, and it's here on Amazon

Amazon.com: Design of Phase-Locked Loop Circuits With Experiments: Howard M. Berlin: Books

I recommend it highly.

 

[yoboy]

Perhaps you might consider buying a book about PLL's?

I have an excellent one, and it's here on Amazon

Amazon.com: Design of Phase-Locked Loop Circuits With Experiments: Howard M. Berlin: Books

I recommend it highly.

Thank you Nigel. I will definitely consider buying one