Application Notes and
FAQ's
Application Notes
At the present time, we have no application notes available
on-line. Please contact Technical Support for more information,
or consult the FAQ list below.
Frequently Asked Questions (FAQ's)
The data sheet
for a device I'm interested in is not available online. How
can I obtain it?
Data sheets not available online may be faxed to you. Please
send a request to our Sales Department.
The device I'm
interested in isn't even listed on the web site. Is it obsolete?
At Data Delay Devices, we continue to support all of our
products. Contact our Sales Department for information regarding
such a device.
What's the difference
between a DDU4 and a DDU4F?
The DDU4F is a newer version of the DDU4, which uses a FAST-TTL
buffer internally instead of the older S-TTL buffer. This
results in faster output signal rise times and lower power
consumption. Otherwise, the two devices function identically.
The same can be said for any other device whose series ends
with the letter F. In applications requiring very short delays
and/or very high frequency signals, we suggest that you contact
Tech Support for additional information.
What's the difference
between a DDU11 and a DDU11H?
The DDU11H is a newer version of the DDU11, which uses a
10KH-ECL buffer internally instead of the older 10K-ECL buffer.
Otherwise, the two devices function identically. The same
can be said for any other device whose series ends with the
letter H. In applications requiring very short delays and/or
very high frequency signals, we suggest that you contact Tech
Support for additional information.
Are your parts
available in surface mount packages?
Nearly all of our parts are available in surface mount packages.
Consult the device data sheet for more information.
Do you manufacture
Industrial Range (-40C to 85C) devices?
All of our devices will operate over the Industrial Temperature
Range.
Do you manufacture
Military Range (-55C to 125C) devices?
Military devices are identified with the suffix M (e.g. DDU7F-100M).
All of the components in these devices meet their respective
MIL-Specs, and the internal IC's are screened to MIL-STD-883.
Our military products meet or exceed all the environmental
requirements of MIL-D-83531 (passive devices) or MIL-D-83532
(active devices). Furthermore, 100% screening is performed
for thermal shock (15 cycles, from -55C to 125C) and burn-in
(196 hours at 125C with power applied). Complete qualification
tests and test reports are available at cost.
Do your ECL parts
work at -5.00V or -5.20V? Can they be operated in PECL mode?
Our ECL parts work at both -5.00V and -5.20V, and can be
operated in PECL mode.
Are your delay
increment specifications guaranteed?
The delay increment specification is a design goal, and may
not be satisfied over the complete range of addresses (in
the case of PDU's) or taps (in the case of DDU's and passive
devices). Monotonicity is always guaranteed, however. We can
screen devices in order to guarantee the increment everywhere;
however, in most cases this would be considered a custom part.
The exception is with all of our standard monolithic products,
for which all increments are guaranteed to be within tolerance.
The total delay
tolerance and the increment tolerance for some units appear
to be incompatable. Why is that?
The delay tolerances published can often be confusing. Take
for example the 1507-100A, with an increment of 10ns +/- 2
ns and a total delay of 100ns +/- 5ns. According to these
specifications alone, there could exist (in theory) a unit
with increments of 12ns over the first six taps and increments
of 8ns over the last four taps. In this highly unlikely case,
the device would exhibit a delay of 104ns on the last tap
(within tolerance) and 72ns on the sixth tap (12ns in high),
and still be within spec. With the exception of our monolithic
devices, only the last tap is required to have a delay within
tolerance. We can screen devices in order to guarantee the
delay at all taps; however, in most cases this would be considered
a custom part.
How much jitter
can I expect from a digitally buffered delay line?
Data Delay Devices does not provide jitter specifications.
In general, however, there are two sources of jitter. The
first is generated by the buffers, and the associated jitter
figures can be obtained from standard IC documentation. The
second is delay-dependent and is associated with the rise-time
of the internal delay line. The longer the rise time, the
more sensitive the delay will be to external noise sources.
If your application is particularly sensitive to jitter, we
suggest that you contact Tech Support and describe your requirements
in detail.
What's the difference
between the Recommended, Suggested and Absolute minimum input
pulse width?
The Recommended minimum input pulse width is a performance
constraint. If it is satisfied, then the delay specifications
are guaranteed. If the input pulse width is less than the
Recommended width but greater than the Suggested width, certain
taps or addresses may be out of tolerance. In the case of
very small increments, monotonicity may even be violated at
isolated points. However, the signal will still pass through
the device with minimal distortion. Below the Suggested width,
the signal will begin to deteriorate significantly, and below
the Absolute width, no output should be expected.
It should be noted that the minimum pulse requirements apply
to both the width of the high and low portions of the signal.
As a result, the minimum input period is given by twice the
minimum pulse width, provided that the signal has a 50% duty-cycle.
Although the delays observed at these higher frequencies
may deviate from their values under low-frequency operation,
once a period and pulse width are given, the delay will not
change from one cycle to the next. In other words, there will
be no additional jitter due to operation at a fixed high frequency.
What is the frequency
response of a passive delay line?
A passive delay line is essentially a low-pass filter with
extremely linear phase response in the passband and a very
sharp cutoff. Theoretically, the frequency response of a delay
line (in MHz) is given by 350 divided by the rise time of
the line (in ns). In some cases, the rise time of each device
is listed explicitly on the data sheet. More often, however,
only the delay-to-rise-time ratio (TD/TR) is given, from which
the rise time can be determined. For example, the 2211-200B
has a delay of 200ns and a TD/TR of 10. Therefore, the rise
time is 20ns and the frequency response is 17.5MHz. It should
be noted that for very small delays the theoretical bandwidth
is often not quite attainable due to the non-ideal properties
of the inductors and capacitors used in manufacturing, as
well as parasitic effects associated with circuit layout.
The signal coming
out of a passive delay line looks awful. What am I doing wrong?
Chances are, the device is not terminated properly. The output
of the line must see an impedance equal to the characteristic
impedance (Z0) of the line itself. This impedance must be
constant over the bandwidth of the line. In most applications,
the output of the line is connected to a very high impedance
input stage. In such cases, it suffices to connect a fixed
resistor equal to Z0 from the delay line output to ground.
The signal fidelity can be improved even further by impedance
matching the input of the line. The line should see an impedance
of Z0 looking back into the driver. Typically, if a low-impedance
driver is used, a series resistor of value Z0 is placed between
the driver and delay line input. It should be noted that impedance
matching at the input results in approximately 6dB of signal
attenuation. If this cannot be tolerated, then the delay line
should be driven directly from a low-impedance source.
If a tapped delay line is being used, only the output of
the line is to be terminated, and all taps must go to high-impedance
inputs. The same hold for mechanically variable lines - the
fixed output must be terminated, while the variable tap should
see as high an impedance as possible.
How much signal
distortion/attenuation should I expect from a passive delay
line?
Distortion is defined as the amount of ringing present on
the output of a delay line in response to a step input, and
is caused by impedance mismatch. A delay line exhibits constant
impedance over its 3dB bandwidth. Above this frequency the
impedance becomes unpredictable. Therefore, even a properly
terminated line will generate reflections for frequencies
above its bandwidth, which will appear as ringing. It is therefore
imperative that the input be band-limited in order to minimize
this effect.
The low-pass-filtering effect of a passive delay line also
increases the rise time of signals passing through it. If
the input rise time is known, the output rise time may be
computed as the square root of the sum of the squares of the
input rise time and the delay line rise time. Thus, a signal
with a 5ns rise time passing through a delay line with a rise
time of 8ns will emerge from the line with a rise time of
approximately 9.4ns.
The signal attenuation of a delay line is determined by its
DC resistance (R) in relation to its impedance (Z0). When
the device is terminated on its output only, the output signal
is scaled by a factor Z0/(Z0+R). When it is terminated on
both input and output, the signal is scaled by a factor Z0/(2Z0+R).
These hold for all frequencies within the passband of the
delay line. Beyond cutoff, the signal is attenuated extremely
sharply.
Can I connect
passive delay lines end-to-end to get more delay?
It is possible to connect the output of one line directly
to the input of another, provided they have the same characteristic
impedance. It that case, terminate (with a resistor to ground)
the output of the second line only. In general, this is done
when a larger delay is desired without sacrificing bandwidth.
For example, two 1514-50B devices give 100ns total delay with
a bandwidth of 35MHz, while a single 1514-100B device cuts
off at 17.5MHz. However, the same performance (35MHz) can
also be achieved using a 2211-100B. In general, do not use
multiple devices when a single one can be found that will
meet the required specifications.
It is always possible to increase delay without sacrificing
bandwidth by placing an isolation buffer between the lines.
In this manner, any number of lines (with arbitrary impedances)
can be placed in series.
Do you manufacture
delay lines in the millisecond range?
No, our devices all function in the ns/us range, with the
maximum delay being about 5us.
I'm interested
in a dash number that doesn't appear in the product data sheet
table. Would this be a custom unit?
If the dash number lies between two numbers listed in the
data sheet, then the part can be specified as a standard unit.
If, on the other hand, it lies outside of the range of values
listed, it will probably have to be a custom unit. In that
case, please consult Tech Support.
How do I order
a custom unit?
To order a custom unit, we ask that you fax a request for
quote (RFQ) to 973-773-9672. In the RFQ, please include all
requirements, such as delay accuracy, input conditions and
size restrictions. Also, indicate the quantity of units needed.
We will respond via fax with a price/delivery quotation and
a unique custom part designator.
How can I obtain
samples?
Samples can be obtained by contacting our Sales Department
by phone at 973-773-2299 or by contacting one of our distributors.
Do you have a
minimum order?
Data Delay Devices handles any order of $75 or more.
How can I get
a price/delivery quote?
Send a request via fax
or e-mail to our Sales Department.
|
