Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

max038 pll locked as an HF vfo?

Status
Not open for further replies.
neazoi

neazoi

Advanced Member level 6
Joined
Jan 5, 2008
Messages
4,122
Helped
13
Reputation
26
Reaction score
15
Trophy points
1,318
Location
Greece
Activity points
36,951
Hi, I want to make an all HF bands (1-30MHz) local oscillator for a direct conversion receiver, and I consider using the max038 pll locked with the mc145151.

Can It be done?

I see in the datasheet of max038 that internal or external pll locking is an option and in vact they have made a pll with the mc145151 in the last page.

I need only 1KHz steps and the mc145151 can do it in the HF to 30MHz. Also the max038 is specified to 40MHz actually in the datasheet so I guess if a 30MHz LPF is used at the output this will atenuate harmonics from 15MHz and above.

I consider using a simpler solution with fewer chips, with a varicap (20-500pF) connected either at the Range or the Fadj of the max 038, instead of using all these chips.

Will this work?
 

Maybe consider a slightly modern solution like DDS clocked from TCXO. I think that max038 went out of production several years ago...
 
  • Like
Reactions: neazoi

    neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
filip.amator said:
Maybe consider a slightly modern solution like DDS clocked from TCXO. I think that max038 went out of production several years ago...
Click to expand...

the dds is tiny, quite noisy for a direct conversion receiver and needs programming, which I do not know how to do.

I prefer the pll solution and I am trying to see if the inductor-free DIP max038 will be stable enough when pll locked.
 

MAX038 goes only up to 20MHz. In fact the one that I have on my desk works only up to 18MHz.

**broken link removed**
 
  • Like
Reactions: neazoi

    neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
vfone said:
MAX038 goes only up to 20MHz. In fact the one that I have on my desk works only up to 18MHz.

**broken link removed**
Click to expand...

The datasheet states maximum operating frequency 40MHz typical.
 

betwixt said:
... and you think DDS is noisy! Wait until you see what comes out of a MAX038!

Brian.
Click to expand...

I was going to start experimenting with it, but now you held me back.
I was going to experiment with the ad9851 and you held me back referring to the noise.
That's for good of course!
What is really considered as a low noise vco, always talking for the HF frequencies?

I have good VXO designs in mind, but no experience in low noise HF VCOs
How about the circuit in figure 6?
 

Attachments

  • Reproducible Low Noise Oscillators.pdf
    788.9 KB · Views: 127
Last edited:

Interesting article.... its nice that it focuses on practicality and not pages of calculations.

However, if I'm right about your intentions, the main issue won't be the internally generated noise from the oscillator but the noise on the tuning voltage. Consider you are looking for a tuning range of (almost) zero to 30MHz and are working from a supply of say 12V. If you want your tuning resolution to be suitable for SSB, lets say 100Hz is the largest allowable frequency step, that implies you need 30,000,000/100 = 300,000 steps in 12V or each step is tuned by a change of only 0.00004V (40uV) which would be very difficult to keep clean and stable. For good quality audio I would target far less, maybe 25uV and bear in mind I assumed 12V tuning range, in practice from a 12V supply it would be less than that.

You start to see why DDS, which produces the signal mathematically is a reasonable solution. It removes the need for extreme voltage regulation while retaining the stability of a quartz crystal.

The other method that might work better for you is to shift the VCO frequency much higher, say 50MHz to 80MHz which is still easy to achieve with home construction methods but the 30MHz range is no longer 100% of the frequency span so tuning becomes easier. You can down-mix with a crystal oscillator back to the frequency you want without adding significant extra drift.

I did look at the possibility of a simple synth using a PIC in an unusual way, I haven't go the time to experiment at the moment but the idea might be worth someone following up. I considered clocking a PIC at 30MHz using a quartz crystal, most of them will do that easily. The PIC produces a ramp-up or ramp-down signal from one of it's pins, it might be possible to tri-state it to give a 'hold' level as well. The ramp voltage is filtered then goes to a VCO tuning 30MHz to 60MHZ. The VCO output is mixed with the 30MHZ PIC clock frequency to produce 0-30MHz and 60-90MHz signals (subtracting and adding). Using the subtractive product through a suitable LPF, the signal is fed to the timer input on the PIC where it's frequency is measured and compared to the one you want. If it is not correct, the ramp is adjusted to bring it back on frequency again.

If it works, it makes a minimum component count synth that can tune 0 to 60MHz. 0-30MHz from the LPF and 30-60MHZ using as HPF in the same way. There would be an interesting side effect that the tuning reverses direction over half the range but that is very easy to fix in software.


Brian.
 
  • Like
Reactions: neazoi

    neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
betwixt said:
Interesting article.... its nice that it focuses on practicality and not pages of calculations.

However, if I'm right about your intentions, the main issue won't be the internally generated noise from the oscillator but the noise on the tuning voltage. Consider you are looking for a tuning range of (almost) zero to 30MHz and are working from a supply of say 12V. If you want your tuning resolution to be suitable for SSB, lets say 100Hz is the largest allowable frequency step, that implies you need 30,000,000/100 = 300,000 steps in 12V or each step is tuned by a change of only 0.00004V (40uV) which would be very difficult to keep clean and stable. For good quality audio I would target far less, maybe 25uV and bear in mind I assumed 12V tuning range, in practice from a 12V supply it would be less than that.

You start to see why DDS, which produces the signal mathematically is a reasonable solution. It removes the need for extreme voltage regulation while retaining the stability of a quartz crystal.

The other method that might work better for you is to shift the VCO frequency much higher, say 50MHz to 80MHz which is still easy to achieve with home construction methods but the 30MHz range is no longer 100% of the frequency span so tuning becomes easier. You can down-mix with a crystal oscillator back to the frequency you want without adding significant extra drift.

I did look at the possibility of a simple synth using a PIC in an unusual way, I haven't go the time to experiment at the moment but the idea might be worth someone following up. I considered clocking a PIC at 30MHz using a quartz crystal, most of them will do that easily. The PIC produces a ramp-up or ramp-down signal from one of it's pins, it might be possible to tri-state it to give a 'hold' level as well. The ramp voltage is filtered then goes to a VCO tuning 30MHz to 60MHZ. The VCO output is mixed with the 30MHZ PIC clock frequency to produce 0-30MHz and 60-90MHz signals (subtracting and adding). Using the subtractive product through a suitable LPF, the signal is fed to the timer input on the PIC where it's frequency is measured and compared to the one you want. If it is not correct, the ramp is adjusted to bring it back on frequency again.

If it works, it makes a minimum component count synth that can tune 0 to 60MHz. 0-30MHz from the LPF and 30-60MHZ using as HPF in the same way. There would be an interesting side effect that the tuning reverses direction over half the range but that is very easy to fix in software.


Brian.
Click to expand...

No I was thinking of the max038 WITH the "band" switching capacitors (NP0), so the tuning range will be far less.
My yesterday tests were disappointing, as I could not generate +-5v at the current required from a single 741 opamp. I gave up, despite the absence of inductors or varicaps...

Very interesting idea is yours. The thing that comes in mind is the "step size" of that circuit, is it of the order of KHz?

A thing about dds, even the small ad9851 dds consumes too much power from my battery handheld device. Since it operates on RX all the time, I mention it as a side effect.
 
Last edited:

It doesn't have a step size as such because it doesn't use a PLL and therefore no fixed division. If done with PWM for example, it would be restricted by the available resolution but the idea I had doesn't have that limitation. The PIC would continuously sample the input frequency (from the down-mixer) and compare it with one you selected by some other means (keypad, encoder, serial link etc.) then decide if the error was plus, minus or none. The rate at which it made that decision would depend on the processing speed and as that would be based on the 30MHz clock, would be very fast (potentially 7.5 million adjustments per second).

For example, if a higher tuning voltage meant higher frequency, as would be normal for varicaps:

You set the desired frequency to say 21.000MHz
1. The input frequency is sampled if it is 21.000MHz tri-state the drive to the ramp generator (a single PIC instruction!)
2. if the input frequency is below 21.000MHz, set the drive to the ramp generator high so it increases the VCO frequency
3. if the input frequency is higher than 21.000MHz, set the drive low so it decreases the frequency.
4. go back to step 1.

The input frequency here is (VCO - 30MHz) so for 21.000MHz, the VCO would actually be running at 51.000MHz.

The idea isn't without pitfalls, the tuning is still by rapid up or down digital signals so you have to filter them but that isn't difficult (RC network) and large frequency jumps might take a while to stabilize but there are ways to get around that too:
"if (abs(actual-desired) > sensible tuning time) then increase ramp current"

Brian.
 

Another solution is to use a programmable divider to obtain a band from a VCO that works much higher frequencies.
For instance , if we intend to obtain 1-30MHz band, our oscillator can work 960MHz with a programmable divider 32-960.
 

BigBoss said:
Another solution is to use a programmable divider to obtain a band from a VCO that works much higher frequencies.
For instance , if we intend to obtain 1-30MHz band, our oscillator can work 960MHz with a programmable divider 32-960.
Click to expand...

It is hard to make stable VCOs. Much harder to make them on higher frequencies!
 

The PLL is there to stabilize the VCO so the problem is much reduced. However, the object is to produce 1 - 30MHz in steps I would guess of no more than 100Hz so the division ratio isn't 32 - 960, it's 320,000 to 9,600,000 - that is slightly more problematic 8-O

Brian.
 

Another solution is to mix two PLL controlled VCO to obtain 1-30MHz signal.
Let say first VCO works at 100MHz and fixed, second one is programmable ( as desired) and works between 101-130MHz.When these 2 VCOs are mixed, f1+f2 and f1-f2 products are obtained then summing products are filtered, extracting products are used.
Otherwise designing single VCO that will work between 1-30MHz is almost impossible.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top