Electronics > Repair
Metex M-3850 high battery current draw repair (+teardown)
(1/1)
hanakp:
Korean-made Metex multimeters used to be quite popular in my country during 1990s and 2000s, because they had slick design and advanced features like autoranging for much lower price than "reputable" producers. Two weeks ago, I got an unused Metex M-3850 made in 1994, basically it sat in a warehouse for the last 30 years. It used to be top model at the time, you can check out its specs in its manual at the Internet Archive. It has RS-232 interface and comes with a 5.25" floppy disk with a MS-DOS program to display measured values (it was useless even back then, because you couldn't control the multimeter remotely). I decided to keep it as a backup meter, so I tested it. All ranges were still well within spec, but it drained the 9-volt battery very quickly, even when turned off (in standby). The standby supply current fluctuated over time, usually it hovered around 200 uA, but could go as high as 900 uA. Since the meter comes from pre-Google era, all my searches about this problem came up empty. Fortunately, its service manual is available at Elektrotanya and I started poking around the usual suspects, electrolytic caps. But they all had negligible leakage current. In the end, I found 2 separate problems that caused the high standby draw:
1. The multimeter uses a J/K flip-flop (IC9, CMOS 4027) to auto-standby after a few minutes. It's permanently connected to the battery (VCC). But only one flip-flop in the package is used, the other one has all inputs floating. Apparently their input voltages drifted over time which cased the supply current to fluctuate. It's odd Metex designers would do such an amateur mistake, but they did. I suspect it may affect other Metex instruments, too. Anyway, I connected all inputs (pins 9 to 13) to ground (pin 8 ) with a thin wire. Note IC9 is surrounded by display and button PCB pads, so you must keep the wire within its footprint. That reduced the standby current to about 60 uA - better, but still way too much for a battery-operated device.
2. PNP transistor Q5 switches the battery votage (VCC) to the rest of the circuits (V+) when the meter is turned on. But the original transistor had high C-E leakge. After I replaced it with BC807-40, the supply current dropped below 0.1 uA. Unortunately, Q5 is partially hidden under C35, so you have to desolder it first.
I made some more photos of the multimeter's internals, but it has pretty standard construction for the era. Only several THT resistors around MAX134 stand out, they look like they were added manually because the original resistor network wasn't available. It was a real chore to put the meter all back together properly, so I hope this was the only repair it will ever need...
coromonadalix:
this meter has some Metrix look a like in the MX-5x series i tough they could share the same tech inside, obviously not
Navigation
[0] Message Index
Go to full version