Hi guys,
as you all know, the HP 435A, 436A, 437A and so on are quite nice power meters and the meter itself is availabe at quite low prices. However, the power sensors are quite expensive. I own a 8481A? and a 8484A, both in nice and working condition, but I don't want to damage anything so I treat them like raw eggs and handle them extremely careful. So I always thought whether it would be possible to make a own diode power sensor which is compatible to the HP 436A. This would have the advantage that the 436A could be used (and therefore its HPIB capabilities), and if it works with the 436A, it will do so also with the other meters (and possibly others, I don't know). Further, the homebrew power sensor would be cheap and if it breaks, no matter, it can be repaired easily.
So today I spent a few hours on making my own power sensor. I checked the schematic of the 8481A power sensor, and I did basically the very same. I came up with this schematic: (I don't want to waste valueable web space for the group, so I uploaded the high resolution pictures to my webspace)
The letters on the right side are the connector pins for the HP 436A, of course. Then I made a quick 'dead bug' style prototype, as follows.... (please don't laugh at me :-) ).
I connected to the 436A and I also installed a SMA 30dB attenuator at the input and connected my power sensor to the power reference of the 436A. At first, the readings were quite off a bit, but using the little trim pot on my PCB, I was able to precisely adjust the gain of my sensor such that the reading on the power meter was exactly 0.00 dBm. I then connected a signal generator to the power sensor and tested various power levels between -10dBm and +20dBm. Incredible, but my power sensor was accurate to 0.1dB! I also tested different frequencies between 1MHz and 2.6GHz. The flatness was not extremely good, but between +/-1dB. So, basically, the concept seems to work. I also tested different ranges for the power sensor - there is a 'mount resistor' inside the HP power sensors (8481, 8482, 9494, ...) which tells the power meter which range the sensor uses; the meter then internally adjusts the gain according to the power sensor range.
It appears that it should be possible to home brew such power sensors. However, at the moment, I have two issues:
a) the auto zero function does not work - even if I run the autozero feature, the power meter afterwards displays -13dBm anyway, so I cannot properly zero my power sensor. At the moment, I do not understand why this is so. Has anyone on the list an idea?
b) I don't understand exactly how the amplifier using the NPN transistor works. When I was adjusting the gain of my sensor, I tried to change the collector resistor (330 Ohms), but this had almost no effect on the gain. The emitter resistor (1k Ohms) also had only little effect on the gain. Only when I placed the pot where it is now I was able to adjust the gain in a wide range. So it basically looks to me like this amplifier is a common emitter amplifier, with the gain being RC / RE, is that true? why then is a feedback from the emitter required to the meter?
Thanks for any hints, tips and so on,
Tobias HB9FSX
