New for May 2013! Our TDU page is now sponsored by Nuvotronics! Their PolyStrata copper conductor/air-dielectric technology is coax for the twenty-first century, born from a Darpa program. Be sure to check out our page on rectangular coax. If you need nanoseconds of delay in very small footprint, give them a call. Welcome aboard, Nuvotronics!
Click here to go to our main page on phased arrays
Click here to go to our page on AESAs
Click here to go to our page on PESAs
Click here to go to our main page on phase shifters
Click here to go to our page on group delay
Also new for May 2013! How many times have you had to convert electrical degrees to length or delay? Here's a simple spreadsheet that does the calculation three ways. If you are designing a time delay unit, you know what we are talking about... while you are thinking about this, commit the following to memory so you can do the calculation in your head: 300mm (~one foot) is one nanosecond delay in free space, and is one wavelength (360 degrees) at 1 GHz. Go to our download area and grab the spreadsheet!
New for December 2012: here are two downloads of papers on TDUs, which are public domain and courtesy of the Air Force (thanks, Tom!)
Also new for December 2012: check out our analysis of how much time delay a phase array needs here.
A time delay unit is a structure that provides a specific time delay, or programmable time delay, using a multi-path structure. It is similar to a phase shifter, but different. A phase shifter usually provides a fixed insertion phase difference between two states (flat phase over frequency). The two states are only slightly different in time delay, and certainly the two states differ in path length by less than a wavelength. A TDU can provide many, many wavelengths of phase shift, and the phase shift is ideally exactly proportional to frequency such that the group delay difference between the two states is flat over the required bandwidth. Got that?
James Darren and Robert Colbert star in Irwin Allen's Time Tunnel TV show
TDUs are used in phased arrays. Phase shifters at each element steer the beam, but they do not provide true time delay. Without true time delay, the beam will distort (squint) over frequency. TDUs are used at the sub-array level to mitigate (but not eliminate) this problem. Across a one meter phased array, TDUs with several nanoseconds of delay are needed.
Watch a video about air-coax time delay technology here. This was filmed at IMS2012 in Montreal. Thanks, Nancy!
Time delay can be accomplished in many ways. The most natural way is to use a length of coax; the rule of thumb one nanosecond per foot applies to air coax, PTFE filled coax (ER=2.2) will provide eight inches. Yes, we jump back and forth between the metric system and English units here at Microwaves101, but get used to it, the United States military uses ridiculous terms like "knots" and "Mach numbers" instead of meters per second, so you need to be able to convert lengths in your head or you will look like a dummy at some meeting in the future.
Another way to provide time delay is using fiber optic delay line. Here you will have to convert microwaves to light and back again. Perhaps an attraction is that the delay line can be made very small, but the TDU will be quite lossy and require amplification. Amplification means the delay line will be non-reciprocal, and you will need separate TDUs for transmit and receive, or some means of turning the amplifiers around.
Delay lines can also be made using microstrip, stripline, or coax, however, you will need to be careful in folding them up, as coupling between meandered lines is your enemy and will surely wreck the flat phase response your system needs.
There are many ways to provide switched paths for multi-bit TDUs. The most obvious is to cascade two-state bits, like all commercially available MMIC phase shifters. A pair of SPDT switches back-to-back with the delay and reference lines connecting them provides the TDU two states. However, if you consider that SP3T, SP4T, etc. switches can be used, there are many other options.
Specifications for TDUs
The following are some relevant specifications for TDUs:
Total time delay (expressed in nanoseconds or picoseconds)
Time delay flatness over frequency (expressed in %)
Number of bits
Amplitude matching over delay states (could be expressed as RMS amplitude in dB)
Amplitude tracking over frequency
P1dB at input
Noise figure (if amplifiers are used to boost the signal)
Problems with TDUs
Different lengths of transmission lines will have losses that increase with length as well as frequency. Here, fixed attenuators can be used to increase losses of short path lengths, and gain equalizers can be used to increase the loss at lower frequencies to straighten out the response.
We'll try to post an example of a time delay unit soon. We'll take a commercially available MMIC SPDT, integrate two of them back to back with coax or microstrip delays and a bypass path with an equalizer.