Measuring the HP 11720A Pulse Modulator and Coax Cables

Lew had 2 RF pulse modulators on the market: $25 for one, $30 for each, as a result of his spouse is
asking not how a lot he’s in a position to promote this lab tools for, however what number of tons of fabric
he has been in a position to eliminate.

As common, I’ve don’t have any want for this stuff, however the value was proper, and enjoying with
take a look at stuff is all the time enjoyable. They usually turned basically free by making a package deal deal
that included an HP 423A energy meter (with sensor!) and haggling the worth down by $40.
All people glad!

The primary pulse modulator is an HP 11720A supported vary from 2 to 18GHz. The second
is an HP 8731B. This one has a spread of 0.8 to 2.4GHz.

On this weblog put up, I’ll solely take a look at the HP 11720A.

(Click on to enlarge)

After that, I digress and begin measuring the standard of a bunch of
coax cables…

A pulse modulator is conceptually a fairly easy gadget. The primary enter takes in
a continuous-wave (CW)
RF sign, one which has a continuing amplitude and frequency.

A modulation sign enters by a second enter. It modifies the amplitude of the
RF sign. Within the case of a pulse modulator, the second enter is often digital, and
the modulation is totally on or off.

The output RF sign is incoming RF sign after it has been subjected to the
modulation sign.

Easy!

The
HP 11720A datasheet
mentions just a few potential functions: pulsed radar programs, and an adjunct for wideband
sweepers and synthesizers. It undoubtedly sounds very cool to construct my very own
radar, however likelihood is that I’ll by no means really feel comfy sufficient with RF to enter that
sort of discipline. We’ll see…

The HP 11720A has RF type-N connectors as enter and output, and a BNC connector
for a 5V TTL modulation sign, with 50Ω termination, although +3V is enough
to regulate the output. ^{On my gadget, the entrance panel says “+50 VDC MAX / -05 VDC MAX”,
however after additional inspection, it seems that the decimal level has competely pale away.
Oops!}

Both method, because it’s a pure digital enter, the modulation sign both passes the
incoming RF sign to the output, or it switches it off fully.

The frequency vary is a fairly luxurious 2 to 18GHz, however that makes it unattainable
to watch the output with an everyday oscilloscope.

The handbook suggests one resolution to check the output of the modulator: combine it with an
unmodulated further RF supply with the identical RF frequency, and the modulation sign
ought to come out. Since I lack a second supply, I’ll be verifying the output energy
as an alternative.

Talking of energy: the RF enter energy is listed as +20dBm (0.1W), however you may go as much as
+33dBm (2W) earlier than you injury the factor. None of my RF toys come near +20dBM, so I don’
t want to fret about that. The minimal RF pulse with is 50ns, which ought to to positive when
taking pictures a radar pulses to the planes which might be flying over our home on wet days.

There’s an insertion lack of lower than 6dB between 2 and 12.4GHz, and fewer than 10dB between
2 to 18GHz. That’s the quantity by which the RF output degree will probably be decrease than the enter.

When first launched, it bought for $2600. I’m unsure when that was, however the handbook was printed
in 1977.

(Click on to enlarge)

Like many different RF modulators, the 11720A makes use of a PIN diode
to do the modulation. PIN, or higher P-I-N, stands for P area, Intrinsic area, N area.
Opposite to common diodes than have P- and N-doped silion areas subsequent to eachother, a PIN diode has a
non-doped silicon area in between.

(© Georg Wiora – GFDL)

At low frequencies, a PIN diode behaves like an everyday diode, however at excessive frequencies
it behaves like a present managed resistor. When ahead biased, the resistance
is about 30Ω, however when reverse biased, the resistance is 1000’s of Ohms.
It’s the excessive frequency conduct that we’re all in favour of: within the case of the HP 11720A, it permits
the modulation sign to sink the RF incoming sign to floor or not.

As common for take a look at tools from that period, the
operating and serivce manual
comprises the complete schematic of the gadget, in addition to a Ideas of Operation part.

On the threat of creating a idiot of myself, right here’s my very own summarized and simplified understanding
of the way it works:

• The RF sign is available in from the underside.
• It first goes by a biasing circuit that provides a DC bias sign to the RF
  sign, which switches the sequence diode additional up on or off.
• A low-pass filter prevents the incoming RF sign from leaking into the motive force circuit.
• The PIN modulator itself has a sequence diode and a bunch of shunt diodes. (I’m
  unsure if the sequence diode is a PIN diode as nicely.)
• A management sign provides one other bias present to change the shunt diodes on or off.
• When the modulator passes by the sign, the shunt diodes are switched off and
  the sequence diode is swap on, and vice versa when the modulator is obstructing the
  enter sign.
• There are two management sign: the one which goes to the PIN modulator itself take care
  of retaining the modulator on or off. The one which goes into the BIAS circuit
  generator sturdy pulse when the modulation switches on or off: this makes the
  switching motion should sooner, and thus permits for quick edges.
• Lastly, a excessive go filter makes positive that a few of the management currents are blocked
  earlier than sending the sign again out.

Right here’s what the principle RF actors appear to be in the true world:

(Click on to enlarge)

The management logic is comparatively easy: a handful of logic TTL buffers, some timing logics,
and drivers. Try the handbook for the schematic.

I created a measurement setup to check if the unit works, and the way it behaves throughout the the frequency vary.

From backside to prime:

• Wiltron SG-1206/U sweep generator

  It might probably generate RF indicators from 10MHz all the way in which to 20GHz, with an output of
  15dBm most. The frequency isn’t as rock stable as you’d usually need, it may be off
  by 0.1% or so, but it surely’s greater than enough if all you want it a sign with a frequency
  that’s within the common vinicity of what you need.

• HP 8656A RF sign generator

  Not used on this case: it’s most frequency is barely 990MHz which is just too low for the
  pulse modulator. It’s within the stack as a result of eradicating it will be an excessive amount of problem. These
  issues are heavy…

• Advantest R3273 spectrum analyzer

  It might probably course of enter indicators from 100Hz to 26.5GHz. On this image, it’s simply exhibiting a
  30MHz the calibration take a look at sign, however it may be used to do energy measurement as nicely.

• HP 33120A LF sign generator

  The gadget that was raised from the dead.
  Right here, it’s used to generate the 5V modulation sign for the heart beat modulator.

• HP 11720A pulse modulator

  RF enter is related to the sweep generator. Modulation enter comes from the LF sign
  generator. RF output is related to the sensor of an influence meter.

• HP 432A energy meter

  The ability meter has a thermistor-based energy sensor that’s plugged in straight into
  the pin modulator.

When the LF sign generator is sending out a 0.4Hz TTL sq. wave, the ability meter sees
a sign happening and off:

The pin modulator is working!

On this case, the sweep generator is sending out a 3GHz/10dBm RF sign, and we’re measuring
a 6.8dBm sign popping out of the pin modulator.

Right here, the HP 432A unit is doing the ability measurement, related straight to the output of the
pin modulator.

I’m utilizing a 4ft RG-214 coax cable with N-type connectors on each side, to attach the sweep
generator straight to the pin modulator. These sort of cables very sturdy, and rated as much as
11GHz, however I’m going all the way in which as much as 20GHz.

Let’s see what occurs:

The primary column exhibits the take a look at frequencies, the second is the ability output with the ability detector plugged
straight into the output of the generator. Third is the ability on the output of the cable. The fourth column exhibits
the ability after going by the pin modulator, and the final column exhibits the loss contained in the pin modulator,
the distinction between the thirds and the fourth column.

(Click on to enlarge)

We’re seeing the next:

• For all measurements, the ability begins to drop above 12GHz.

  That is regular: the detector of the ability meter is barely rated till 10GHz.
  Nevertheless, even above 10GHz, the meter is measuring one thing and we ought to be
  in a position to evaluate outcomes relative to one another.

• the loss within the RG-214 cable begins at 0.7dBm and steadily drops, to 4.4dBm at
  10GHz, after which the loss will increase quickly.

  Not sudden for a cable that’s rated as much as 11GHz.

• Lastly, the pin modulator loss is 4.2dBm at 1.5GHz, however that’s exterior the desired vary of
  2 to 18GHz. From 2GHz to 14GHz, the loss is nicely under 4dBm, but it surely will increase quickly
  at 16GHz.

Conclusion: the HP 11720A works! I can swap the output on and off, and loss contained in the
modulator is inside spec as nicely.

One factor that becames rapidly clear whereas experimenting is the significance of the kind of
coax cable. I measured the ability traits of 4 cables:

• the 4ft RG-214 coax with N-type connectors that I used above.
• a 6.5ft RG-58 coax with BNC connectors. That is principally your commonplace coax cable.
• a 5ft RG-142 coax with SMA connectors.
• a 12ft RG-142 coax with SMA connectors.

They’re rated at 11GHz, 5GHz, 8GHz and 8GHz, respectively.

(Click on to enlarge)

We will see how the 5ft RG-142 cable is behaving higher than the specs. On the excessive vary,
it’s doing even higher than the RG-214! The 12ft cable has the next loss than its 5ft
sibling, but it surely’s nonetheless doing a suitable job. The RG-58C has no enterprise doing something above
6GHz.

The R3273 spectrum analyzer has the flexibility to measure energy, so I gave {that a} attempt as nicely, however there’s two variations:

• I lowered the sweep generator output from 10dBm to 0dBm, to keep away from out-of-range errors.
• It’s good to zoom in the principle frequency to measure energy, in any other case the spectrum analyzer doesn’t decide up the sign
  more often than not. It is a elementary distinction between an influence meter with a thermistor, thermocouple or diode detector
  sensor head and a spectrum analyzer: a sensor head is wide-band and can decide up the principle sign in addition to harmonics,
  and the whole lot in between, a spectrum analyzer doesn’t. I selected a span of 50MHz round the principle sign frequency for
  the spectrum analyzer measurements.
• There’s no energy sensor you can plug in straight into the output of the output of the generator. So that you’re all the time
  measuring the cable as nicely.

Due to the final level, the info under doesn’t present the cable loss (the distinction between direct output and cable
output), however the absolute energy that’s popping out of the cable.

(Click on to enlarge)

The graphs are kind of the alternative of the earlier one, which is what one might count on.

One bizarre outcome are the numbers for 5GHz and 10GHz, the place the spectrum analyzer exhibits a major dip
in energy for all cables.

That is much more apparent after we subtract the output of the ability meter and the spectrum analyzer, and modify for the 10dB distinction
in energy setting:

(Click on to enlarge)

The ability meter is barely rated for frequencies as much as 10GHz. Under that, the distinction stays inside cause for the
RG-58 and RG-142 cables, apart from 5GHz and 10GHz. Above 10GHz, the spectrum analyzer remains to be able to measure energy,
however the energy meter drops off, and the outcomes begin to diverge. For the RG-214, the outcomes begin to diverge
at 6GHz.

Along with studying about PIN diodes, I expertise first-hand the problem of make dependable
measurements with RF indicators. It’s very simple to get incorrect outcomes resulting from take a look at setup situation.
For instance, when measuring the RG-214 cable with the spectrum analyzer, I obtained energy numbers
that had been 30dB too low, even after tightening the cable adapters. Solely after unscrewing the
adapter utterly after which screwing it again in, did I get affordable outcomes. This occurred a
variety of instances. It in all probability doesn’t assist that each one my cables and a lot of the equipement are sourced
from Craiglist, the electronics flea market, and AliExpress. And, in fact, even the extra
latest items, such because the spectrum analyzer, haven’t been correctly calibrated in years.

So all the outcomes right here have to be taken with a really thick grain of salt: they’re exhibiting tendencies, however
there undoubtedly not correct.