If you are trying to adjust the phase of a pre-existing signal, there are a couple of options. However, none of them will cover such a wide frequency range. AD8340 can adjust the gain and phase of signals in the 700 MHz to 1 GHz range. The adjustment control is continuous cartesian (i.e. X,Y) . So if you set X and Y to, say, 1 and 1, you get a phase shift of 45 degrees (setting them to -1 and -1 woudl give you a phase of 225 degrees).
For other frequencies, you can use the ADL5390. However, to operate as a phase shifter, this device must be combined with an external 90 degree (fixed) phase splitter. Off-the-shelf 90 degree phase splitters generally work over at-most one octave. So, this might allow you to operate from, say, 500 MHz to 1000 MHz. However 1 MHz to 1000 MHz constitutes 10 octaves! You could add programmable dividers at the output of the ADL5390. However, these dividers, being digital in nature, would probably destroy the character of your signal by turning it into a square wave.
What is the application?
One other question. Are you looking to synthesize and control the phase of a sinewave? Or is it the case that you already have the sinewave and you want to pass it through a black box which has a programmable phase delay?
Yes, I see - one way is the synthesis - but it seems too expensive decause of the very high freguency - 1 GHz - it is too much for digital, i need 2.5 GSPS DDS - it is not anywhere. Perhaps it may be FPGA + DAC, but I'm afraid it will be more expensive. Besides, can I set the phase precisely at frequencies at 0.5 - 1 GHz with such digital method?
The other way - I need only to delay the phase, but for this way - I need to construct that black box. Or do you mean there is ready-to-use IC for that?
So Yes, let's assume that I already have sine wave and I need programmable phase delay.
Sorry, I don't know the application exactly. It will be some research of multichannel phase distribution system. As far as I know, the system must set phases of channels output signals in user-defined order. The power of one channel is very small - current is 1 up to 1000 uA, active load is 0.1 up to 0.2 Ohm, inductive load is 0.1 up to 1 uH.
Thank you very much for reply!
I thought I would mention the capabilities of our DDS parts, even though they may not be the best solution for this application. Almost all of our DDS parts have the ability to adjust the phase offset of their outputs in sub-degree increments. However these values are a division of a power of two, so it is not possible to exactly get 1° increments. The general equation for phase offset is PHASE = POW * 360 / 2^N where POW is the phase offset register value and N is the bit size of that register. For example our AD9910 part has a phase offset control that is 16-bits wide, so the phase tuning resolution is 360 / 2^16 ~= 0.0055°. Most other DDS parts have a 14-bit or 12-bit phase offset resolution.
The issue with a DDS, as you've already mentioned, is that it is a synthesizer so it is trying to generate a sine-wave using a DAC so the output frequency is limited by nyquist. You could follow a DDS with a PLL or a mixer to get higher frequencies, but that's adding more parts and more cost. It is also possible to capture images of the fundamental output frequency that are greater than nyquist by using a band-pass filter (amplitude will be reduced by the sinc response), but getting close to nyquist is still a problem. Below is a figure showing what I'm suggesting which is from application note AN-908 "Super-Nyquist Operation of the AD9912 Yields a High RF Output Signal".
Another concern when trying to use multiple DDS' to control the phase difference between them is making sure all of the chips are synchronized. There are quite a few application notes describing how to sycnhronize them depending on which part you decide to use.
If you have any DDS specific questions, please post them to the DDS forum.
Does your output need to be a sine wave, or is it going to be made into a clock signal, aka square wave? If creating clock signals, we do have some clock distribution parts that have programmable delays in them, but I am less knowledgable about those than I am about our DDS parts.
Kevin, thank you for reply. Our signal is principally sine wave, only sine wave.
I have a similar application, where I have an sinus RF signal in the range between 150MHz and 450MHz.
After splitting it, I need to control the relative phase between them with 1 degree accuracy.
I have been looking into some of your IC, and I have the following questions:
1. Do you have any out-of-shelf 90 degree phase splitter for this freq range?
2. Instead of using a the vector multiplier ADL5390, I could use a quadrature modulator as the ADL5385, but the problem is that the LO is divided-by-2 and phase shifted. Any Qmod without divided-by-two?
3. Then looking to the datasheet of the ADL5386, is not clear if the LO is divided by two or not. Could you please clarify that?
Thanks and best regards,
Retrieving data ...