HP 43X series power meters - a progression teardown

[Tony_G]

Hi All,

As part of my travels in getting my project to work I ended up with one each of the HP 43X series power meters. I thought I'd open them up and share the images with you.

The 43X series of power meters are based around the concept of sending the RF power into a thermocouple which then produces a dc voltage proportional to the RF power. The sensors introduced at the same time as the first power meter are the 8480 series. They're still supported in the current Keysight power meter series.

That's 32 years of product support - Pretty awesome.

The defining characteristics of the meters were the swappable 8480 series sensors (to change dBm range, frequency, etc), an inbuilt 0.7% accuracy 0dBm power reference and the ability to enter calibration factors for the individual sensors in use. They're easily obtainable on EBay - The power meters that is, the sensors are pretty easy to blow up so you see a vast array of "for parts not working" sensors out there. As of today, I'd characterize the sensor market as:

• Less that $100 - Sensor really is dead
• More than $300 - Sensor is almost certain to be good
• $100-$300 - It's a crap shoot

This is actually how I ended up with the 3 power meters - It was several attempts to get a working 8481A sensor and still get a bargain - Sometimes you're the bat, sometimes you're the ball...

Anyway, I spent the last several days bringing these meters back to inside spec so that I could sell the ones I didn't need - I'm keeping the 437B and one 8481A - It's a nice combo.

The first meter I got actually was the 435B. While I'm not keeping this one, it's actually my personal favorite of the 3 - It's just a great example of that classic clean HP meter look. One the left you'll see a sensor controls (zero, cal factor and the connection port). In the middle you have the range setting. You can actually unscrew the outer ring and swap out the range card. Mine fortunately came with the 8481A range card as the only one. One the right is the power switch and the reference output:



If you take the top off you can see into the meter. Typical of units of this vintage the form of the analog meter display drives the size of the case resulting in a lot of open space. There is Option 001 which is a battery so it will take up some amount of that space - I didn't see it in my searching so I can't say how big it is.

In this image you can see the main board running horizontally across the bottom and the 1mW reference can at the front. The little blue trimpot is where you can control the amplitude of the built in 50MHz, 1mW reference. You are supposed to set this using a HP 432A & 478A Thermistor mount with NIST traceability. I didn't have that but I did have a HP 11683A Range Calibrator that had come from a calibration center. Using that as a transfer standard got me +0.5dBm of the actual value (how do I know? Well I actually ended up chatting to some guys that run an official Keysight cal lab and they measured the power reference for me on their NIST traceable calibrated gear at 1.11mW).



Most of the adjustments are made on the main board and from the right side. Here you can set things like the DC offset, full scale meter reading etc. While I was adjusting the meter I couldn't find A4R46 for the life of me. Eventually I checked the manual and in service sheet 5 there is a helpful shot showing where these adjustment points are. A4R46 is actually buried just to the bottom right f the gray can - You can just see it poking out in the image:



The HP 435B was introduced in July of 1974 - Here is the HP Journal Article on it - But it was quickly replaced by the HP 436A which introduced in Oct 1975 - HP Journal Article here

BTW I love trawling that HP Journal site - Tons of incredibly interesting articles there.

Anyway the 436A brought a digital view to the party with a red led display, push button selections and an interesting little analog meter in the upper left corner. From what I read, this was apparently a late addition to the 436A after prototype users complained that using the meter for peaking adjustments was too hard with the digital display. Another addition was the inclusion of a dB(Rel) mode. Now a user could measure an output, hit the dB(Rel) button and the meter would show all future measurements as relative to that:



Another big selling point for the 436A was the availability of Option 022 which added HP-IB to the meter:



Opening the meter up you can see that its full of stars..errr..boards. Running parallel with the display board there are 5 assemblies. The last one on the right is the Option 022 board and you can see it's daughter board in the upper right corner surfacing the HP-IB connector to the rear panel. Immediately below that is the power reference can and it has the exact same blue trim pot. I didn't open this up but I wouldn't be surprised is the reference was identical to the 435B. It certainly looks that way:



Removing the assemblies shows the basic motherboard layout for the meter:



Each assembly has this cool little swivel ring on it that you can swing out to grip in removing the card - Nicely thought out feature - Here are the boards without commentary:






I couldn't find when the 437B was introduced - Maybe my search fu wasn't up to task - but it marked a major change in the power meter series. It took the same look and feel as other instruments of that era I think - The one that leaps to mind is the HP 3478A Multimeter. This unit is much smaller than the other units, especially the 436A which was a monster, and included some great new features. You could store the cal factors for up to 10 sensors and then recall them from a simple menu. Then as you measured power you could just hit the Freq button, input the frequency and the cal factor would be applied. If you don't have a specific cal factor for that frequency then the meter would work it out for you. You could also set the duty cycle so that when measuring pulse power the value was more accurate.



When you take the covers off, you see a single double sided board. What you see on the top is a board that I think is well designed for service. There are a multitude of test points that you can access easily and they're spaced apart so you're not doing unnatural acts with your hooks/probes:



Flipping the unit over is where the adjustment points all live. The left most shielded can is the power reference and the right one houses the input gain circuitry:



In case you haven't seen what the power sensor looks like, here is an 8481A. In some of the previous images you can see the other side of this sensor and the cal factor chart that is on it. For some reason these charts tend to be removed from sensors on EBay. I don't know why and I asked the Keysight guys and they didn't specifically know either but the general consensus was that the charts get removed when the sensor goes out of the calibration period as a way of showing that this sensor shouldn't be used. There are two types of charts that I have seen. Ones which are a list, like my sensor, and ones which are an actual graph. Given the limited cal factor input on the 435B & 436A (basically .01 at a time) these sensors were probably made/released during that period. With later meters you can input the actual numbers:



Here is a view of the sensor cable end - You need a specific cable for these sensors, I'm using a HP 11730A 5' cable and they come in much longer lengths:



To wrap up, here is the HP 11683A Range Calibrator. This I got from an actual calibration lab so it still have stickers on it for the calibration and warranty so I can't pop the lid and let you have a look. You basically use this as a known value source (it replaces the sensor) and it allows you to simulate full value readings on each of the ranges:



Hopefully you found this interesting. To do a little post pimping, here are my other old HP gear posts:

• HP 415E SWR Meter
• HP 651B Test Oscillator

Till the next time.

TonyG