Tube Meter
A 6E2 magic-eye tube as a level meter: a green bar that closes when it hears you · By Onion Madder · Mess O' Pedals
A 6E2 is a tuning indicator from the age of valve radios: a little triode and a fluorescent target in one glass bottle, and a green bar on the side of the glass that closes up as the grid goes negative. Feed the grid the rectified envelope of a signal and it becomes a level meter that glows, the way a VU needle never will. This board is everything around the tube: a transistor to lift a line-level signal up to something the tube can see, a diode and two resistors to turn that into a negative voltage, a regulator for the heater, and the high voltage, made from scratch. No bought module: a 9 V AC wall adapter drives a small mains transformer backwards, its 120 V winding gives about 90 V AC, and a voltage doubler turns that into about 250 V DC. It is the old way, back-to-back transformers, and it means there is no mains anywhere in the box and nothing switching next to the audio.
Before you start
tap to expand / collapse
This build has about 250 volts DC in it, and that can kill you. The wall adapter is 9 V and is not the danger. T1's 120 V winding, the doubler and everything on strips a to h are, and so are the three 400 V capacitors, which hold their charge after the power is off. R10 and R11 are there to drain them, and you still do not trust them. The rules, none of them optional:
The adapter must be AC, 9 V. Read its label: it says AC OUTPUT, or shows a sine wave. A DC adapter puts T1's 12 V winding straight across a DC supply, which is very nearly a short circuit, and something gets hot fast.
T1 must be a 120 V transformer. A 230 V one run backwards makes nearly twice the voltage, and doubled that is past every capacitor on the board.
Measure the high voltage before the tube goes in. Power up with the tube out of its socket, meter on TARGET to GND, one hand. It should read no more than about 250 V; with the tube in it will sit a little lower. If it reads high, R2 is the part to raise. The 6E2's own limit is 250 V.
Never touch the board, the tube socket or the transformer with the power on. Not to adjust, not to probe with a finger, not to steady it. Switch off, then wait, then measure.
After switching off, measure the TARGET pad to ground with a meter rated for 300 V before you touch anything. If it reads more than a few volts, discharge it: a 100K resistor on insulated leads from TARGET to GND, held for ten seconds, then measure again. Do this every time. Not most times.
Set the trimmer with an insulated trim tool, one hand, never a metal screwdriver. VR1 sits on the high voltage. It is set once, with the lid off, and then left alone.
One hand. When you must work near it live - you should not need to - keep the other hand in your pocket, so a shock cannot cross your chest.
Everything high-voltage is insulated and mounted. The board on standoffs, never loose in a metal box. The two high-voltage wires to the socket and T1's two 120 V leads in insulation rated for it, with heat-shrink over every joint. The transformer bolted down. No bare copper anywhere near the lid.
Strip i stays empty. That is what keeps the 250 V a full strip away from the low-voltage side and from your fingers on the solder side. Do not put a part or a wire there because it looks convenient. Keep the solder side clean: a whisker of solder between two of the top eight strips is a whisker at 250 V.
Drill before you solder, and blow the box out. Aluminum swarf is conductive, and this is the one board on the site where a stray chip can arc.
The tube gets hot. Not dangerously, but you will not want your thumb on the glass, and neither will the paint on the box near the window.
What you'll need
tap to expand / collapse
- 1× small mains transformer, 120 V to 12 V (T1), about 2 VA or bigger, PCB or chassis mount - spot it: a little laminated block with its two windings marked 120V (or 115V) and 12V. It runs backwards here: the 12 V winding is fed and the 120 V winding is the output. A 120 V part only - see Before You Start. It bolts to the box floor and its four leads come to four pads: the 12 V winding to AC_A and AC_B, the 120 V winding to HVAC and HVAC_G
- 1× 500K trimmer (VR1) - spot it: a small square block with a screw slot and three legs in a row, marked 504. Check its voltage rating covers 250 V; if the maker does not say, the fixed 470K the tube data asks for is the safe fallback in its place
- 3× 4.7 µF electrolytic, 400 V (C8, C9, C10) - spot it: about 8 mm across, with 400V printed on the can. Not a 16 V or 50 V part, however much it looks the same. The stripe marks the negative leg
- 2× 1N4007 (D6, D7) - spot it: black body, gray band, 4007 on the side. The doubler's diodes. A 1N4001 looks identical and is NOT rated for this: read the number
- 4× 1N4001 (D1 to D4) - the bridge for the low-voltage side. Any 1N400x works
- 1× 1N4148 signal diode (D5) - spot it: tiny glass body, black band. The clamp that makes the envelope negative
- 1× 7806 voltage regulator (U1), TO-220, plus a clip-on heatsink - spot it: the big three-legged tab package, 7806 or L7806 on the face, legs IN, GND, OUT left to right with the face toward you. The heater draws 0.3 A, so this sheds about 2 W and will run warm. It stands on its legs on the board
- 1× 2N3904 NPN transistor (Q1), TO-92 - spot it: 2N3904 on the flat face, legs E, B, C left to right with the flat toward you
- 2× 220K (red-red-yellow) (R10, R11) - the bleeder, which empties the high-voltage capacitors when the power goes off
- 2× 100K (brown-black-yellow) - R1, the fixed half of the anode load, and R3, the top of the transistor's bias
- 1× 22K (red-red-orange) (R2) - the high-voltage filter, and the part to raise if the supply reads over 250 V
- 1× 10K (brown-black-orange) (R5) collector load, 1× 8K2 (gray-red-red) (R4) the bottom of the bias, 1× 470R (yellow-violet-brown) (R6) emitter, 1× 1K (brown-black-red) (R9) which, with C7, keeps the ripple off the transistor
- 1× 1M (brown-black-green) (R7) and 1× 2M2 (red-red-green) (R8) - the envelope's attack and its bleed
- 4× 100nF ceramic (104) (C1 to C4) - input coupling, collector coupling, the envelope hold, and the regulator's input
- 1× 1000 µF electrolytic, 25 V (C6) - the bridge's reservoir. About 10 mm across; the layout leaves it room
- 2× 100 µF electrolytic (C5, C7) - the heater rail and the transistor's rail
- 3× wire links (J1 to J3) - J1 and J3 carry ground down the right edge, J2 doubles the anode strip
- 1× stripboard, 19 strips × 26 holes or larger
Off the board
- 1× 6E2 magic-eye tube and a noval (9-pin) socket, chassis mount - spot it: a glass tube about 22 mm across with a flat green window down one side and nine pins in a circle with a gap. Usually from China, often with a socket. The socket is the thing you wire to; the tube plugs into it
- 1× 9 V AC wall adapter, 1 A, and a barrel jack to suit its plug - AC, not DC. The label says AC OUTPUT or shows a sine wave
- 1× 100K linear pot (B100K) - the sensitivity. All three lugs
- 1× toggle for power, in the adapter lead, 1× 3.5 mm jack for the input
- Insulated wire rated for the voltage (ordinary hookup wire of 300 V rating is fine), heat-shrink for the socket and T1's leads, standoffs for the board, a 100K resistor on insulated leads for discharging, an insulated trim tool, and a meter rated for at least 300 V DC. These are not optional
- An upright 125B with a window filed in the face - the template further down. T1 has to fit in it too; measure yours against the box before you drill
The tube, pin by pin
These started as the EM84's base connections, the tube the 6E2 is usually compared to, and a 6E2 schematic has since agreed with every pin this board uses: grid on 1, cathode on 3, heater on 4 and 5, the target on 6 straight to the high voltage, and pins 7 and 9 tied together through the anode load. That schematic also ties pins 2 and 8 to pin 6; leaving them unconnected, as here, is also fine. Still, check the sheet that comes with your tube before you wire the socket - if it disagrees with this table, the sheet wins. Pins count clockwise from the gap, looking at the pins.
| Pin | What it is | Wire it to |
|---|---|---|
| 1 | Grid | The GRID pad. The negative envelope |
| 2 | Internal connection | Nothing. Do not connect |
| 3 | Cathode | GND |
| 4 | Heater | The HEATER pad, 6 V |
| 5 | Heater | GND |
| 6 | Target, the fluorescent screen | The TARGET pad. High voltage |
| 7 | Deflection electrode | The ANODE pad, with pin 9 - they are tied together at the socket. High voltage |
| 8 | Internal connection | Nothing. Do not connect |
| 9 | Anode, the triode plate | The ANODE pad. High voltage |
The numbers the design leans on. From the 6E2's own data: heater 6.3 V at 0.3 A, and 250 V as the most the tube should be given. From the EM84's, because a 6E2 sheet with them in has not turned up: a 470K anode load at 250 V; the bar going from fully open to fully closed as the grid goes from 0 to about −22 V; and at least 150 V on the target before it glows at all. The trimmer is there because those borrowed numbers are borrowed.
The layout file is the guide
Like the 555 theremin and the Opto Duet, the build document for this one is the file itself: ↓ Download Tube_Meter.json - import it into Copper Bottom and the editor gives you the board picture, the placement walkthrough, the electrical checks and a printable build sheet, all from the same file. The editor draws the 7806 at the size a TO-220 really is, legs IN, GND, OUT, and it draws T1 beside the board with the other panel parts, a wire from each of its four leads to its pad. If you took a copy of this file before September 23 2026, take it again: that board took its high voltage from a bought module, and on it Q1's bias was wrong - a 22K where the 8K2 is now held the transistor hard on, so it could not pass a signal.
The netlist
Quoted verbatim from the layout file. It was written from the board, so checking the board against it can only pass - what it is for is carrying the circuit somewhere else, and telling you what each strip is supposed to be doing while you stare at it.
The off-board wiring
Twelve pads leave the board, two of them test points. HVAC, TARGET and ANODE carry the high voltage and are marked.
| Pad | Goes to |
|---|---|
| AC_A / AC_B | The 9 V AC adapter's jack. One lug through the power toggle to AC_A, the other straight to AC_B. Neither is ground - the bridge makes the ground. T1's 12 V winding goes to these same two pads, P1 to AC_A and P2 to AC_B |
| HVAC / HVAC_G | T1's 120 V winding, S1 to HVAC and S2 to HVAC_G, which is the high-voltage side's ground strip. HVAC is about 90 V AC. High voltage |
| GND | The input jack sleeve, the pot's lug 3, and the tube's pins 3 and 5 |
| TARGET | Tube pin 6. High voltage |
| ANODE | Tube pins 9 and 7, tied together at the socket. High voltage |
| HEATER | Tube pin 4. Pin 5 goes to GND |
| GRID | Tube pin 1 |
| SENS_W | The sensitivity pot's wiper. Lug 1 to the input jack tip, lug 3 to GND |
| +V / +12F | Nothing. Test points on the bridge output and on Q1's filtered rail, about 12 V and 11 V, and they tell Copper Bottom's checker those strips are supplies |
The four high-voltage wires are the ones to fuss over - two to the socket and T1's two 120 V leads. Heat-shrink over every joint at the socket, no bare wire anywhere, and route them away from the low-voltage side and the lid. A 300 V rated hookup wire is the normal stuff; what matters is that none of it is bare.
How it works, in one breath
The tube's bar is controlled by its grid: at 0 V the bar is open, and as the grid goes negative the bar closes, fully by about −22 V. So the job of the board is to turn a signal into a negative DC voltage that follows its loudness. A line-level signal is a volt or two, which is not enough to move the bar far, so it goes into a 2N3904 common-emitter stage first: 100K and 8K2 set the bias so the collector rests near the middle of its rail, 10K in the collector and 470R in the emitter give a gain of about twenty, and it can swing about 5 V either side of where it rests. Its rail comes through R9 and C7, which keep the bridge's ripple from reaching the grid as a flicker.
That swing goes through a 100n to a node with a 1N4148 clamp on it, anode on the node and cathode to ground. The diode will not let the node rise above about 0.6 V, so every positive peak is held there and the whole waveform is pushed down: its negative peaks now sit at roughly minus the peak-to-peak swing. A 1M and a 100n average that into a DC level and a 2M2 bleeds it away when the signal stops, so the bar closes fast and opens slowly, which is what a meter should do. The sensitivity pot sets how loud is loud. That comes to about −5 V at full swing, well short of the −22 V that closes the bar all the way, which is part of why the anode load is now a trimmer.
The high voltage is two transformers back to back, except that one of them is the wall adapter. The adapter turns the mains into 9 V AC; T1 is an ordinary 120 V to 12 V mains transformer wired the wrong way around, so the 9 V goes into its 12 V winding and about 90 V AC comes out of its 120 V winding. That winding is floating, so its other end simply goes to ground. C8 and D6 clamp the AC so it swings from 0 up to twice its peak, D7 and C9 catch the top of that, and the result is about 250 V DC with nothing switching anywhere. R2 and C10 smooth it, and R10 and R11, two 220K in series, drain all three capacitors when the power goes off.
The target goes straight to that supply. The triode anode goes to it through R1 and the VR1 trimmer, 100K fixed plus up to 500K, so the load runs from 100K to 600K and takes in the 470K the tube data asks for. More load pulls the anode, and the deflection electrode tied to it, further below the target, which widens the shadow; the trimmer is where you set how wide it sits at rest, for your tube and your supply. The same 9 V AC goes through a bridge, D1 to D4, into C6 for about 12 V, and the 7806 makes the heater's 6 V from that - a hair under its 6.3 V rating, well within what a heater tolerates, and the regulator earns its heatsink at 0.3 A.
Unbuilt, and these are the bench numbers. How far the bar travels at a given input, which is the transistor's gain against the tube's sensitivity, and the reason the pot exists; what T1 really gives run backwards, which depends on the transformer and on how far your adapter sags under load, and so where the supply actually sits; where the trimmer wants to be; whether 6 V is enough heater for a bright display; and how warm the 7806 runs. None of it is dangerous to find out if the rules above are followed. All of it is unknown until someone does.
Print at 100% / “Actual size”, never “Fit to page” - check the bar measures exactly 4″ before you touch a drill. The face is 66 by 121 mm but only 42 mm of it is drillable once the lid lip is clear.
Nobody has drilled this one yet. Measure your own tube's glass before you file the window; not every 6E2 is the same bottle. Offer the sheet up to the box and check it against your parts before you make a hole.
Tube data from the 6E2's own listing at Radiomuseum where it has the figure, and from the EM84's datasheet where it does not. The trimmed anode load and the pin connections were checked against a 6E2 level-meter schematic, which runs off the mains; this board makes its high voltage from a 9 V AC adapter instead. The clamp detector is the way every magic-eye meter has driven its grid since the tubes were new, and back-to-back transformers are an old trick for a small tube supply. The transistor stage, the guard strip, the layout and the box are Onion Madder's, drawn in Copper Bottom.