Handheld digital multimeters were just beginning to gain popularity in the late 1970s. Fluke was there of course, but so was Beckman. This is a Beckman 3020 3½-digit 2000-count model that dates to 1979. It has the classic Beckman ‘look’ and rotary switch design that persisted for many years, even after the Beckman Industrial division was absorbed by Wavetek circa 1992.
It arrives dirty and completely non-functional, with no display at all. The LCD has a disturbing area of discoloration in the middle, so it might be bad. This unit has an interesting green-and-gold color scheme. Unlike later models with a plastic tilt stand, this one has a stand made of stiff music wire. It can be removed easily, flipped over, and re-inserted to function as a hanger.
Opening the meter up reveals the usual Beckman two-board construction with a large rotary switch sandwiched in-between. Stamped on the inside of both the front and back parts of the case is proof of the unit’s age. It was manufactured in July 1979. Interestingly, the 10A shunt is soldered directly on the input jacks, which are nice machined tubes embedded in the plastic front cover. The 10A jack also has some sort of switching mechanism activated by a plastic plunger inside the jack. [In later observation, it was determined that inserting a plug into the jack switched the voltage sense line from the on-board current shunt to the 10A shunt.]
Seconds after removing the screws that hold the halves of the large rotary switch mechanism together, the first problem with this unit becomes obvious. The switch consists of 4 groups of thin metal leaves that are essentially just SPST switches, activated by protrusions on the big rotating disc in the middle. Two of the leaves in the right-hand group on the bottom board are mangled, indeed one of them appears to have melted completely in two. This is going to be interesting…
DMM Repair, Extreme Edition
Looking closely at the construction of the switch, it appears that if the leads are desoldered, the entire block can be lifted up and out for some possible extreme repair activity. Careful desoldering reveals that is is indeed the case, and the group of contacts lift out fairly easily, after all the solder is removed from the holes.
With full access to the leaves, it looks like the one that was mangled but not melted can be re-shaped. What is interesting about these leaves is that they are much more ductile than one would expect. They exhibit a little bit of ‘springiness’, but not much. This allows for bending the mangled one back in shape without breaking it, which you might expect with springy, brittle metal. On the other hand, it makes these contacts very, very fragile.
If you’re reading this because you have one of these meters, be very careful when working around these contacts, they will deform easily if you snag one or press on it too hard.
After searching through every drawer and box in the workshop for some kind of suitable metal to make a new leaf out of, I finally settled on a scrap from some Molex-type connector pins. When you buy a quantity of new pins, they’re often connected together on a ‘tree’, which is just the left-over material from stamping and forming the pin. This metal is soft enough to bend easily, but has just a bit of spring in it. I hope it works.
Turning Over A New Leaf
Cutting, trimming and shaping the tiny sliver of metal is a bit tedious. It has to be thin enough to move freely in it’s slot in the plastic frame that is still attached to the board. Eventually it looks about right. After applying flux to the ends of the new leaf and the ‘stump’ of the old leaf, the two are joined with solder. The entire contact block is then re-inserted into the switch frame and re-soldered to the PCB.
Good news and bad news. The meter powers up and appears to work, mostly. However the ‘b’ segment of the ones (least significant) digit never comes on under any circumstances. Voltage readings are a bit off, and the decimal point is flashing. A flashing decimal point is usually the low-battery warning on Beckman DMMs, but the battery measures 8.8V.
The LCD and main chip is part of a stack-up that includes 3 elastomeric connectors (zebra strips) and is often a source of problems. Disassembling this stack, cleaning and re-assembling does not improve the function of the meter this time however. While the stack was apart, I checked the LCD with a function generator (5Vpp bipolar square wave @50Hz). The segment in question does actually work, so it’s not the LCD itself. For some reason, it’s just not being turned on.
To see whether any particular segment is getting a signal to turn on, it will be necessary to compare it with the LCD backplane clock. ON segments will have a square wave that is out-of-phase with the backplane. OFF segments will be in phase with the backplane. It’s difficult if not impossible to probe any signals with the stack fully assembled, so to get the main chip to run but still provide access, a couple of plastic pony clamps were used to hold it to the PCB.
Probing the ‘b’ segment drive signal shows that it is always in phase with the backplane clock, no matter what function or range is selected, and even with the input shorted. It should be on to display a zero, so the conclusion is the main chip has a faulty gate in the path for this signal.
Scarecrow Needs a Brain
As mentioned earlier, Wavetek acquired Beckman Industrial and its DMM models. I happen to have a non-working Wavetek HD100 manufactured in 1994, with a design still very similar to this 3020. Looking at the main chip, I was surprised to find that even with a 15 year difference in age, the main chip in the Wavetek has the same part number (270-100) as the suspect one in the 3020! A quick transplant and it does appear that the older main chip is faulty, because the newer one works perfectly. The display is complete, voltage readings are correct and no more flashing decimal point.
During a full checkout, the 10A range did not work. It turned out to be the plunger switch on top of the jack not quite making contact. Bending the contact a little fixed the problem. The rotary switch repair seems to be holding up well, so far no further glitches in testing. With a complete scrubdown and quick lens polish, this 34 year-old DMM is looking and working well again.
DMMCheck calibration check results for this Beckman 3020:
|nominal for non-TRMS is 5.55
|nominal for non-TRMS is 1.11
|0.1 with leads shorted