Heathkit V7A

Heathkit V7A VTVM

As an electronic hobbyist, having multimeters and having lots of them seems like a necessity. Well, no, not really, but I tend to get a bunch of them for the heck of it. I got this full sized Heathkit V7A, originally sold as a kit, from a hamfest. It was probably introduced around 1958 making it more than half-century old technology.

Regular select ranges:

1500V AC/DC
500V AC/DC
150V AC/DC
1.5V AC/DC
R x 1ohms
R x 10ohms
R x 100ohms
R x 1Kohms
R x 10Kohms
R x 100Kohms
R x 1Mohms

This meter is analog with vacuum tubes (12AU7 dual triode and 6AL5 dual diode), and has just one semiconductor element in it - a diode for rectifying AC power. It still has a battery inside to deal with measuring resistance on its logarithmic (or is it 1/x, or is it something else?) scale. A single C battery is needed for measuring resistance. The power/mode select has OFF, AC, DC-, DC+, and ohms. Note that ohms and AC volts (or dB) use the two banana jacks where black is common, and red is volts or for the other terminal of the resistor. For DC, the phone plug tip is used with the black common banana jack. The DC+ and DC- modes will let you switch the polarity of the input, so two shielded phone probes are needed. This helps against introducing noise into the system as the input impedence is high enough to pick off signals off the air.

VTVM: The Vacuum Tube Volt Meter

The Vacuum Tube Volt Meter got its claim to fame due to extremely high input impedence compared to regular VOMs. With its 11MΩ input (10MΩ in meter, 1MΩ in probe), adding this load in parallel with resistances in the circuit being tested won't affect the circuit as much as one with lower impedence. A contrived case would be if you had a voltage divider with two 10KΩ series resistors across a 3V power supply. By circuit analysis you know the node between the resistors should be 1.5 volts. Using a VTVM, the parasitic resistance adds a 11MΩ across one of the resistors. As 11MΩ is larger than 10KΩ by three orders of magnitude, you can see that it will affect the reading very little. A regular VOM like my Eico, has much lower impedence. Trying to measure that same node in the 2.5V range, the input impedence is 2.5V*20KΩ/V = 50KΩ. Now you have the 10KΩ plus a 50KΩ resistor in parallel which is now 8.3KΩ. That in series with a 10KΩ resistor and a 3V source, you'll measure 1.36V instead of 1.5V. Not a large difference but it's significant - it's almost a 10% change in the voltage at that point. If the circuit had been balanced at 1.5V like for an op amp fake ground reference, you've just skewed the whole circuit!

Now most digital multimeters these days have input impedences at 10MΩ like my Fluke 77, no longer making the VTVM special.

Meter and movement

The meter that it uses is a 200µA. This means that it takes 200µA to deflect the meter full scale from its terminals. However none of the current from the device under test powers the meter, the vacuum tube amplifes the signal and displays it. What I found is that the meter and the 12AU7 amplification circuit, using a differential circuit, measures about 1.5V swing on the input grid. This 1.5V input voltage with virtually no current draw changes into a 200µA current to drive the meter.

The other strange thing is because this is powered, all the current draw is through the tubes and not much going through the battery even in ohms mode. I find the circuit pretty interesting: basically the 1.5V battery is hooked through the resistance tree straight to the input of the differential triode pair. So that means, when you have nothing hooked up, you have voltage going to the grid of the tube. As it's a grid, no current really flows. But this is the infinite resistance case - so when the full scale is shown on the meter, it's actually infinite resistance! This is the opposite of most regular VOMs; then again, keeping 200µA flowing will eat battery life. Well, how does this meter actually work for measuring resistance? With the unknown external resistance, you shunt the resistance of the divider to ground! So the meter drops toward 0µA (but not really - but the meter zeroing makes it look that way) as you reduce resistance to 0Ω. The meter scale is marked accordingly. However, as in most multimeters of the era, the internal 1.5V battery is passed through a low value resistor and you can get over 100mA going through your test probes. This is sufficient to turn on many diodes, and if it can't support 100mA, you can fry these diodes.
Here's a trick that works on old analog ohmmeters with Rx1 mode: you can use the ohmmeter's internal battery to light up an LED with the "Joule Thief" circuit - but a digital voltmeter won't work! Then again, I'm not sure you want to waste the battery in a VOM.


Someone abandoned this meter at a hamfest after not being able to sell it. Ebay prices for this meter is pretty low due to the large number of them out there, but still higher than a typical el-cheapo DMM from Harbor Freight. I grabbed it because, well, no hobbyist should be with a VTVM :). I suppose the reason why this meter was abandoned was mainly because it was, well, old and beat up. It was also not functional either, much like my Eico 555 which was also broken at acquisition (incidentally, I got both at the same hamfest, though different years). When I got the V7A, the meter was moving every which way making no sense of how it was supposed to work. So, it's a fairly broken meter.

Luckily it's a Heathkit and one of the most common VTVMs out there. But this just means you don't know who assembled the unit. And sure enough, there's plenty of cold solder joints in the meter. Also it looks like some liquid has met some components. It's in fairly sad shape.

The main cold solder joint that restored lots of functionality was resoldering the common point to the PCB. That fixed all of the meter flakiness that I saw. The next problem was that it was really inaccurate about the voltage, implying there's resistance where it shouldn't. I checked the highest DC measurement resistors in situ in pure laziness but the 7MΩ and 2MΩ resistors seemed a way off. Thus I had to stop being lazy and disconnect them to be sure.

They turned out to measure correct out of circuit. There must be a short elsewhere. After more disconnecting and measurement, I found that the DC ¼ inch phone plug was shorted and causing all the grief - I was shocked because mechanical devices like this usually don't short out! I replaced it with what I had on hand: a TRS stereo ¼ phone plug, and left the side connector open.

A third problem was that though the internal battery has its negative terminal on the chassis, the chassis was not connected properly to the rest of the circuit - it's open in places. I added a jumper wire to connect the two parts of the circuit. Fixing that, and now the ohmmeter works. (It remains to be seen whether I modify the unit to completely isolate the case. As designed, the negative common is hooked up to the case.)

I also replaced the #50 screw in bulb by ripping out the glass part and adding a small 6V incandescent lamp in the screw on base. Unfortunately the new bulb isn't as bright as the old one, but bright enough to indicate the unit is on. Old bulb was probably bright enough to backlight the meter face.

The DC probe itself was well worn out, at least on the phone plug end. It frayed out so badly that it was hanging on by a thread of stranded wire and could short out any time. The probe end had its 1M resistor fried, so I had to replace that too. This is not to mention the other two probes' insulation were dry rotting, cracked along the whole length of the wire. They need to be replaced, alas, I don't have any 1KV rated wire to use as probe wire.

After replacing the resistor, soldering a jumper, resoldering a cold solder joint, fixing the light bulb, replacing the phone jack, and then recalibrating, I now have a fully functionl VTVM. I still think I'd rather use the Fluke for most purposes, no power cord or warm-up time to deal with...