Tag Archives: r&s

Rohde&Schwarz NRT Power Reflection Meter: tantalum issues

Recently, I got this very nice instrument – a Rohde&Schwarz NRT “Power Reflection Meter”. It is designed to measure transmitted and reflected power, a quite handy instrument used for the installation of cell phone antennas, etc.

This unit was received with a label “completely dead and burned”, but it turns out that the damage was not all that bad.

Some quick survey showed that there is a 12 VDC power supply installed, which is connected to the main board by a cable. On the main board, there are three capacitors in parallel for the 12 V bus. One of these had burned out, only pieces and black smoke remained. The board looked damaged, but after careful removal, there was no problem with the board found. Used some methylated spirits to clean the mess thoroughly. There one thin trace underneath one of these capacitors, so better don’t work on it with coarse tools or in a rush. I dissected it under the microscope. The ground connection pad of these capacitors has no heat relieve arrangement, but the caps are soldered to the heavy ground plane – a strong soldering tool will be needed to melt the solder in adequate time.

I desoldered not only the bad guy, but also the adjacent capacitor, to test its performance and value (there is no schematic). These are 100 µm low ESR caps. I found it appropriate to replace two of them by one 470 µF electrolytic capacitor, plus, a 10 µF multilayer ceramic capacitor in parallel. Saving money to buy 100 µm small-size tantalum that are difficult to hand solder without risking damage.

One intriguing part came up during repair, a 78ST105S regulator, which is there to provide a 5 V rail from the 12 V input. Quite a nice part, now obsolete, but first time a ever saw such design. It is essentially a high efficiency voltage regulator, easy to use, but it has a buck circuit installed to get higher efficiencies compared the customar 7805 regulators.

Some quick check of isolation resistance and the power supply (without anything connected) – it provided stable 12 VDC to a test load, so the shorted cap didn’t cause any damage to the supply. The moment of truth: switching on the NRT instrument. Working fine.

To complete the unit, I will need to see where to get an affordable NRT sensor. These don’t come cheap. List price is 3800 EUR each, used units go anywhere from 500-2000 EUR. Definitely, expensive.