Some months ago, I purchased an Agilent 34970A Data Acquisition unit.
Together with the instrument there were also two Agilent 34901A Multiplexer cards.
In the advert it said that the unit reported error 614. Unfortunately, I did not pay it much of attention because I usually solve most of the errors. This time the error was a bit difficult.
When I received the unit, I could see that it reported Error: 612, 613, 615, 617, 618 and 619. After looking in the service manual and Google for possible solutions it looked like I had bought a unit that can’t be repaired. At many places the signal switching ASIC U101 was mentioned as much likely to fail and that this component is not available anymore.
Luckily, I also have an Agilent 34401A DMM. That unit use an identical DMM implementation as the 34970A so it was just to start comparing signals under different conditions.
The good thing was that I did not see much of difference around the U101 between the two instruments. That encouraged me to continue the measurements.
One setup that I used was to use the instruments (34970A or 34401A) and measuring Ohms in the 100 Ω range and test the input resistance on an Keysight 34465A DMM set to DCI, 10 mA.
With the 34401A I measured 2.5 Ω at the same time as 1 mA was measured by the 34465A. That was expected. With the 34970A the measured resistance was 1.67 Ω and the current 0.67 mA.
Now I could dig into the signal chain and measure the voltages at every node on the two instruments. (See attached pictures.) It turned out that the voltage after Q202 was much lower on the faulty unit.
Also, the voltage at Q211 was very low. And the PNP transistors also showed different voltages on the faulty unit.
Measuring the PNP transistors PN junction drop did not show any fault there.
After replacing Q211 with a new (PN4117A in TO-92 replaced by a MMBF4117 om SOT23) things started to look “better” (or different). The current jumped up to 2.5 mA and the measured resistance to 6.5 ohm.
After replacing Q202 with a new (2N5461 in TO-92 replaced by a 2N5461 in SOT23) the current jumped back to 1 mA and the measured resistance to 2.6 Ω (close enough).
So, the problem was just two transistors! Easy fix.
I don’t know if there are other faulty 34970A DMM cards out there that can be fixed in similar way?
Is it true that the U101 can fail? It is easy to think that U101 is broken when looking at all the error codes and try to figure out a common route cause. But these transistors are also common.