CRT Selection & Sourcing

Everything else in this series — the high-voltage supply of Vol 3, the deflection and video amplifiers of Vol 4, the firmware of Vol 5 — exists to serve one component you cannot design or print: a small electrostatic cathode-ray tube, almost certainly built between 1945 and 1975, that you will buy used and on faith from a stranger. The tube is the one part of a scope clock that is genuinely scarce, genuinely fragile, and genuinely dangerous in two directions at once — it implodes if you crack the glass and electrocutes you if you touch the wrong pin while it is lit. Choosing it well is therefore the single most consequential decision in the whole build, because the tube dictates the socket, the filament resistor, the A2 ceiling, whether you need a kilovolt multiplier, and ultimately whether the clock looks like a crisp green instrument or a fuzzy amber blob. This volume teaches you to read a CRT spec the way a builder reads it — not the way a 1950s engineer wrote it — reproduces the TubeTime crt-driver compatibility list in full, gives the two pinouts the owner’s OSC4.4 documentation supplies (DG7-32 and 6Lo1i), and then walks the real-world sourcing minefield of surplus, eBay, hamfests, gas, and screen burn.

6.1 How to read a CRT spec for scope-clock use

A vintage CRT datasheet is dense with parameters that mattered to the original oscilloscope or radar designer and are irrelevant to you, mixed with a handful that decide whether the tube works in a clock at all. The trick is knowing which six numbers to extract and which forty to ignore. Below are the parameters that actually drive a build decision, in roughly the order you should check them when a candidate tube comes up for sale.

6.1.1 Diameter and the face geometry

The screen diameter — the leading digit(s) of the type number — sets the physical size of the clock and, more subtly, the deflection budget. A 3RP1 is a 3-inch round tube; a 5BP1 is 5 inches; a 2BP1 is 2 inches; the DG7-32 is a 7-cm (≈2.75-inch) European type. Bigger faces are more dramatic but need proportionally more deflection voltage to sweep the beam corner-to-corner, which stresses the deflection amplifiers (Vol 4) and, on post-deflection tubes, the high-voltage supply. The TubeTime driver set is explicitly untested above 5 inches (see § 6.2), so for a first build, 2 to 5 inches is the sane window. The owner’s OSC4.4 runs the small DG7-32 and 6Lo1i tubes, both comfortably under 3 inches, which is part of why that build is approachable: small tubes ask less of every supply rail.

6.1.2 Base type and the socket you will have to find

Every CRT terminates in a glass base with pins on a specific circle in a specific pattern, identified by a base code such as 11B, 12E, 14J, Loctal, B9A, or 5BP. The base code is not a footnote — it is half the sourcing problem. You must obtain a matching socket (or wire directly to the pins, which is fragile and not recommended for a tube you intend to keep). Diheptal (14-pin) bases, duodecal (12-pin), and the various 11-pin oscilloscope bases each need their own socket, and some — particularly the older Loctal and the 14-pin diheptal sockets — are themselves becoming collector items. The crt-driver repository ships a BaseDiagrams.pdf precisely so you can identify the base and order the right socket before the tube arrives. Confirm the base before you buy; a tube with an unobtainable socket is a paperweight.

6.1.3 Filament / heater voltage

The heater warms the cathode and runs on low-voltage AC. Almost every tube in the compatibility list is either 6.3 V or 2.5 V — the two standard receiving-tube heater voltages. This matters for two reasons. First, the driver must supply the right heater voltage from its filament winding, and getting it wrong either fails to light the cathode (too low) or burns it out (too high). Second — and this is the practical lever — the filament dropping resistor Rfilament is chosen to land the measured heater voltage in spec. The OSC4.4 instructions are explicit: a 4.7 Ω, 3 W resistor “works for many CRTs,” but some European types such as the DG7-6 want 10 Ω, 3 W, and you are expected to measure the voltage across the heater pins on first power-up and adjust Rfilament up if it reads high. Heater voltage and base together are what actually dictate the socket and the resistor (per-tube resistor selection is treated more fully in Vol 4).

6.1.4 Maximum A2 (accelerating) voltage

A2 is the final accelerating-anode voltage — the rail that determines beam energy and therefore brightness. Each tube has a maximum A2 it will tolerate before arcing or damage, and this is the number that decides whether your supply is adequate or excessive. The compatibility list gives an A2 ceiling per family: small tubes (the 2-inch class and the 1-inch 913/1DP* types) want A2 kept below ~1 kV; a 2BP* tolerates up to 2750 V; a 1EP* up to 1500 V. Critically, A2 is a ceiling and a sourcing fact, not a supply-design choice — you pick the tube, read its A2 limit, and then size the supply to sit comfortably under it (the supply side is Vol 3). The OSC4.4 runs its small tubes at a modest ~300 V top rail, far below any of these ceilings, which is exactly why it is gentle on the glass.

6.1.5 Deflection factor

Deflection factor — quoted in volts per centimeter (or volts-per-inch) of beam travel, per deflection-plate pair — tells you how hard the deflection amplifiers must swing to fill the screen. A tube needing 20 V/cm on a 5-inch face is asking for a large differential swing across the plates; a tube at 5 V/cm is easy. Higher A2 makes the beam “stiffer” and raises the deflection factor (a faster beam is harder to bend), which is the hidden cost of brightness: crank A2 for a brighter trace and you now need more deflection voltage to keep the face full. For clock use you want a tube whose deflection factor your chosen amplifier can satisfy at the A2 you intend to run; this is the parameter that quietly couples the tube choice to Vol 4.

6.1.6 Phosphor and persistence

The final character is the phosphor — encoded in the digit after the P in the type number. P1 is the classic medium-persistence green (the scope-clock default, and what gives the genre its look); P7 is a long-persistence blue-flash/yellow-decay phosphor that smears motion and is generally wrong for a clock (it ghosts); P31 is a brighter green common in later tubes. For a vector clock you want medium persistence — long enough that the refreshed face does not flicker, short enough that moving elements (a sweeping seconds hand) do not leave trails. P1 and P31 greens are the safe picks; avoid the radar long-persistence phosphors unless you specifically want the smear as an aesthetic.

6.2 The TubeTime crt-driver compatibility list

The single most useful sourcing document in the open-source ecosystem is the CompatibilityList.txt shipped with Eric Schlaepfer’s (TubeTime’s) crt-driver board set. It enumerates the electrostatic CRTs the driver has been confirmed (or expected) to work with, grouped by diameter, with base code, heater voltage, a compatibility verdict, and — where relevant — a maximum A2 voltage. It is reproduced verbatim below, followed by an explanation of the shorthand. Treat it as a shopping filter: when a tube comes up for sale, find it here first.

6.2.1 The list, reproduced

The table below is the CompatibilityList.txt content laid out as a Markdown table. Blank cells in the “Max A2” column mean the source list gave no figure for that family. The “Notes” column captures anything the source appended after the verdict (for example “Center face deflection electrode,” “Rectangular,” “Dual gun!”).

CRT type	Base	Fil V	Compatible?	Max A2	Notes
0.75″–1″ diameter
913	913	6.3	No	500 V
1CP*	Loctal	6.3	Yes	1000 V	HC, Untested
1DP*	B9A	6.3	Yes	600 V	HC, Untested; requires power-supply mod
1EP*	11V	6.3	Yes	1500 V
2″ diameter
902A	8CD	6.3	No
2AP*	11B	6.3	Yes	1100 V
2AP*A	11L	6.3	Yes	1100 V
2BP*	12E	6.3	Yes	2750 V
3″ diameter
906 / 3AP*	7AN	2.5	No
3AP*A	7AN	2.5	No
908A	7AN	2.5	No
3BP*	14A	6.3	Yes
3BP*A	14G	6.3	Yes
3CP*	11C	6.3	Yes		HC
3DP1	14C	6.3	Yes		Center face deflection electrode
3DP*–3DP*A	14H	6.3	Yes		Center face deflection electrode
3EP*	11N	6.3	Yes		HC
3FP*	14B	6.3	Yes		PDA
3FP*A	14J	6.3	Yes		PDA
3GP*	11A	6.3	Yes		HC
3GP*A	11N	6.3	Yes		HC
3JP*	14J	6.3	Yes		PDA
3KP*	11M	6.3	Yes		Untested
3MP*	12F	6.3	Yes
3QP*	9D	6.3	Yes		HC, Untested
3RP* / 3RP*A	12E	6.3	Yes
3SP*	12E	6.3	Yes		Untested
3WP*	12T	6.3	Yes
3XP*	Loctal	6.3	Yes		HC, Poor performance; rectangular
5″ diameter
905	5BP	2.5	Yes		HC
905A	5BR	2.5	Yes		HC
907	5BP	2.5	Yes		HC
912	912	2.5	Yes		Untested
5ABP*	14J	6.3	Yes		PDA
5ADP*	14J	6.3	Yes		PDA
5AQP*	14G	6.3	Yes		Very poorly
5AQP*A	14G	6.3	Yes		Very poorly
5AP*	11A	6.3	Yes		HC, Untested
5BP*	11A	6.3	Yes		HC
5BP*A	11N	6.3	Yes		HC, Untested
5CP*	14B	6.3	Yes		PDA
5CP*A	14J	6.3	Yes		PDA, Untested
5DEP*	12E	6.3	Yes
5GP*	11A	6.3	Yes		HC, Untested
5HP*	11A	6.3	Yes		HC, Untested
5HP*A	11N	6.3	Yes		HC, Untested
5JP*	11E	6.3	Yes		HC, PDA, Untested
5JP*A	11S	6.3	Yes		HC, PDA, Untested
5LP*	11F	6.3	Yes		HC, Untested
5LP*A	11T	6.3	Yes		HC, PDA, Untested
5MP*	7AN	2.5	No
5NP*	11A	6.3	Yes		HC, Untested
5RP*	14F	6.3	Unlikely		PDA, Untested but PDA is probably too high voltage
5RP*A	14P	6.3	Unlikely		PDA, Untested but PDA is probably too high voltage
5SP*	14K	6.3	Yes		PDA; dual gun!
5UP*	12E	6.3	Yes		Fuzzy
5VP*	11N	6.3	Yes		HC, Untested
5XP*	14P	6.3	Unlikely		PDA, Untested but PDA is probably too high voltage
5YP*	14Q	6.3	Yes		Untested

The source closes with the blanket caveat: the circuit has not been tested with tubes larger than 5 inches.

6.2.2 Decoding the notes

The compatibility verdicts are terse, and three pieces of shorthand in particular carry real wiring and safety consequences:

• HC — Heater/Cathode tied. On these tubes the cathode is internally bonded to one leg of the heater. The crt-driver instructions are emphatic about which leg: that heater leg must connect to pin 5 of J2, not pin 4 — “or you will short out the filament transformer.” In practice the HC note means you cannot treat the heater as a floating two-wire load; one heater pin is also your cathode connection, and miswiring it dead-shorts the filament winding. Always confirm orientation against the base diagram before power-up.
• PDA — Post-Deflection Acceleration. A PDA tube has an extra accelerating electrode after the deflection plates, fed from a very high voltage to add brightness without raising the deflection factor. The list’s note is direct: “Power supply needs optional voltage multiplier installed to get 3 kV+.” A PDA tube is therefore not a beginner choice on the TubeTime board — it requires the Cockcroft–Walton multiplier section described in Vol 3 to reach its rated 3 kV-plus, which is more high-voltage exposure and more to go wrong. Several of the “Unlikely” 5-inch entries (5RP*, 5XP*) are flagged precisely because their PDA voltage is probably too high for the driver.
• Untested / Poorly / Fuzzy. A “Yes (Untested)” is an expected compatibility based on the base and gun architecture, not a confirmed one — treat it as a calculated risk. “Very poorly,” “Fuzzy,” and “Poor performance” are exactly what they sound like and are reasons to keep looking if you have a choice.

Two additional rules ride along the top of the list. For CRTs 2 inches and under, keep A2 below 1 kV to avoid damaging the tube — the small-tube ceiling is a hard limit, not a guideline. And the 1DP* family specifically requires a power-supply modification (its B9A 9-pin miniature base and low A2 ceiling of 600 V put it outside the stock supply’s comfort zone); be especially careful with these.

6.3 The tube families seen in this hub

The source folders in this project cluster around a handful of families. Knowing their character lets you match a tube to a driver and a difficulty level.

6.3.1 DG7-32 and 6Lo1i — the OSC4.4-documented tubes

The owner’s OSC4.4 documentation supplies measured pinouts and operating voltages for two specific small tubes: the German DG7-32 (a 7-cm round electrostatic tube, 12-pin) and the Russian 6Lo1i (6ЛО1И, a 14-pin tube). These are the worked examples because they are exactly what the OSC4.4 was built and measured against, and both run on the board’s gentle ~300 V top rail with a 4–6 VAC heater.

The DG7-32 maps to the OSC4.4 board terminals and measures as follows. Pins 1 and 12 are the filament/heater (4–6 VAC across the pair); pin 5 and pin 11 are no-connect.

DG7-32 pin	Board terminal	Voltage (ref. ground)
1	f (filament)	4–6 VAC across pins 1 & 12
2	G (grid)	≈ −300 VDC
3	K (cathode)	≈ −295 VDC
4	F (focus)	≈ −150 VDC
5	nc	—
6	Y2	≈ 150–250 VDC
7	Y1	≈ 150–250 VDC
8	A (anode)	≈ 150–250 VDC
9	X2	≈ 150–250 VDC
10	X1	≈ 150–250 VDC
11	nc	—
12	f (filament)	4–6 VAC across pins 1 & 12

The 6Lo1i is a 14-pin tube with a similar electrode set. Its measured map (ground reference taken at the tab of the 7805 regulator, per the instructions):

6Lo1i pin	Board terminal	Voltage (ref. ground)
1	f (filament)	4–6 VAC across f–f
2	K (cathode)	−295 VDC
3	G (grid)	−300 VDC
4	F (focus)	−150 VDC
5	NC	—
6	NC	—
7	X2	150–250 VDC
8	X1	150–250 VDC
9	A (anode)	150–250 VDC
10	Y1	150–250 VDC
11	Y2	150–250 VDC
12	NC	—
13	NC	—
14	f (filament)	4–6 VAC across f–f

Note the polarity logic shared by both tubes: the grid sits most negative (≈ −300 V), the cathode just above it (≈ −295 V) so the small grid-to-cathode difference biases beam current, focus is at ≈ −150 V, and the five deflection/anode electrodes float in the +150 to +250 V band. The instructions stress these are approximate but should be close — they are a sanity check during bring-up, not calibration targets.

Each tube also dictates two of the per-tube driver resistors on the OSC4.4 (the full deflection-resistor treatment is Vol 4):

• DG7-32: R99 & R4 = 330 KΩ, R6 & R8 = 220 KΩ.
• 6Lo1i: R99 & R4 = 220 KΩ, R6 & R8 = 180 KΩ.

(The OSC4.4 instructions also list a 2BP1 option for R99/R4 at 220 KΩ.) And Rfilament is the heater dropping resistor: ≈ 4.7 Ω, 3 W is the typical starting value, with 10 Ω, 3 W used for some European types such as the DG7-6 — set it by measuring the heater voltage on first power-up and raising the value if it reads above the tube’s 4–6 VAC window.

6.3.2 The 2AP1 / 2BP* family

The 2-inch tubes are the friendliest large-enough-to-enjoy option. The 2AP* (base 11B, 6.3 V) tops out at 1100 V A2; the 2BP* (base 12E, 6.3 V) tolerates up to 2750 V, the highest A2 of any 2-inch entry, which makes it bright for its size. The OSC4.4 instructions explicitly accommodate a 2BP1 by offering the 220 KΩ value for R99/R4. These are non-PDA, well-documented, and widely available — a strong first-clock family.

6.3.3 The 3RP1 / 3RP1A, 3MP1, and the 3-inch field

Among 3-inch tubes the standouts for a clock are non-PDA, non-HC types. The 3RP* / 3RP*A (base 12E, 6.3 V) is a clean “Yes” with no caveats — a classic oscilloscope tube with a P1 green phosphor in the 3RP1/3RP1A variants, and a frequent, affordable surplus find. The 3MP* (base 12F, 6.3 V) is likewise a plain “Yes.” Avoid the 3-inch entries marked PDA (3FP*, 3JP*) for a first build and the 2.5 V 7AN-based types (906/3AP*, 3AP*A, 908A) marked “No.” The 3-inch class is the sweet spot of size versus drive difficulty.

6.3.4 The 1EP and 5xP families

At the small end, the 1EP* (base 11V, 6.3 V) is the standout sub-inch tube — a clean “Yes” with a generous 1500 V A2 ceiling and no HC/PDA caveat, unlike its troublesome 1CP*/1DP* neighbors. At the large end the 5-inch field is mostly compatible but riddled with caveats: many entries are HC (cathode tied to a heater leg), many are PDA (need the multiplier), several are merely “Untested,” and a handful (5RP*, 5XP*) are flagged “Unlikely” because their PDA voltage is probably beyond the driver. The clean 5-inch picks are the non-PDA, non-HC types such as 5DEP* (base 12E). Remember the driver is untested above 5 inches — that is the practical upper bound for this board set.

6.4 Sourcing — the real-world reality

You will not buy a new electrostatic CRT; the last ones were made decades ago. Sourcing is therefore an exercise in surplus archaeology, and the channels and risks below are what separate a glowing clock from a box of dead glass.

6.4.1 Where the glass comes from

The standard channels, roughly in order of price and risk:

• eBay and online surplus dealers. The dominant channel. You will find both bare tubes and tubes pulled from scrapped oscilloscopes and radar sets. Filter by the type numbers in § 6.2; cross-check the base before bidding. The OSC4.4 folder even carries a “Buy CRTs Here” bookmark pointing at a known seller — surplus sellers who specialize in scope tubes are worth bookmarking.
• Hamfests and electronics swap meets. Often the cheapest source and sometimes the only place to find a matching socket. The downside is no returns and no testing — you are buying on visual inspection alone.
• Donor instruments. A dead 1950s–70s oscilloscope (a Heathkit, a Tektronix, a Waterman pocket-scope) is frequently cheaper than the tube inside it and arrives with the socket, the heater wiring, and sometimes a deflection-amplifier reference design already attached. The hub’s Waterman material exists for exactly this reason.

6.4.2 What “NOS” means — and does not

NOS = New Old Stock: a tube that was manufactured decades ago but never used, ideally still in its original box. NOS is the gold standard because the cathode has not been depleted and the screen has not been burned. But “NOS” is a claim, not a guarantee — a tube can sit unused for fifty years and still develop a vacuum leak (go “gassy”) or suffer cathode poisoning in storage. NOS lowers the odds of trouble; it does not eliminate them. A tested used tube with a photo of a live trace can be a safer buy than an untested NOS tube in a sealed box.

6.4.3 The two failure modes that matter: gas and low emission

Two age-related failures dominate, and both are partly checkable before you commit:

• Gas / loss of vacuum. Over decades, gas can leak in or outgas from the internal structure. A gassy tube shows a purple or blue glow inside the envelope when lit (ionized gas), a defocused or “soft” spot, and erratic brightness. Many electrostatic CRTs have a getter — a silvery mirror deposit inside the neck. A bright silver getter is a good sign; a white or milky getter has been consumed by air ingress and signals a dead vacuum. Always ask for a photo of the getter.
• Low emission / weak cathode. A cathode that has lost emission gives a dim trace even at full brightness and proper A2, and may need excessive heater voltage to show anything. There is no fully cooking it back; a weak tube stays weak. A seller’s photo of a bright, sharp trace is the best single piece of evidence of cathode health.

6.4.4 Screen burn

A tube pulled from an oscilloscope or radar that displayed a fixed pattern for thousands of hours can carry phosphor burn — a permanently darkened region where the beam dwelt. For a clock, a centered baseline burn or a graticule ghost will sit right under your clock face. Inspect the unlit screen in raking light for discoloration, and prefer tubes from instruments that drew varied images over those from fixed-pattern radar indicators.

6.4.5 Testing a tube before you trust it

If you can, energize a candidate tube on a bench supply before committing it to the clock — but only with the full high-voltage discipline of Vol 12 in force. The minimum useful test: apply the correct heater voltage and confirm the cathode glows orange; apply A2 well under the tube’s ceiling and a modest grid bias; and look for a single sharp spot that focuses cleanly and is green/white, not purple. A spot that will not focus, glows the wrong color, or needs the brightness wide open to appear is telling you the tube is gassy or weak. And the practical rule that ties this whole volume together: the filament voltage and the base, read off the spec, are what dictate the socket you order and the Rfilament you fit — settle both before the tube is in your hands.

6.5 A decision aid — matching a tube to a driver

The three drivers covered in this series ask different things of a tube, and the compatibility list’s caveats map directly onto difficulty. Use the guidance below to pair glass with board.

6.5.1 By driver

• OSC4.4 (Build B, Vol 8). Built and measured against the small DG7-32 and 6Lo1i tubes on a gentle ~300 V top rail. It is happiest with small, low-A2, non-PDA tubes; the documentation hands you the pinouts, the per-tube R99/R4/R6/R8 values, and the Rfilament starting point. If you want the owner’s exact, supportable path, source a DG7-32, a 6Lo1i, or a 2BP1.
• Dutchtronix AVR Scope Clock (Build A, Vol 7). The classic AVR design is tuned around the small-to-medium electrostatic tubes common in the kit community (2-to-3-inch class). Match it with a clean non-PDA tube — a 3RP1A, a 2BP1, a 3MP1 — and you are on well-trodden ground.
• TubeTime crt-driver (Build From Scratch, Vol 10). The most capable analog front end: a stiff ~800–1200 V supply, true differential deflection amps, a 6 MHz video amplifier, and an optional voltage multiplier for PDA tubes (Vol 3). This is the only one of the three that can properly drive a PDA tube to 3 kV+, and the only one whose compatibility you can check directly against the § 6.2 list — because that list is this board’s list.

6.5.2 Beginner-friendly vs. PDA-complex

The single sharpest filter for difficulty is the PDA flag. A non-PDA tube needs only the few-hundred-volt to ~1 kV rail your supply already produces; a PDA tube needs the high-voltage multiplier section, more shielding, and more respect. Sort your shortlist this way:

• Beginner-friendly (no PDA, no HC, confirmed “Yes”): DG7-32, 6Lo1i, 2BP1, 3RP1/3RP1A, 3MP1, 1EP*, 5DEP*. These run on a modest A2 rail, have plain two-wire-plus-cathode heaters, and are confirmed in the field or by the owner’s own build.
• Intermediate (HC — workable, but mind pin 5 of J2): the various 5BP*, 3CP*, 3EP*, 3GP* HC families. Perfectly usable, but you must wire the heater/cathode leg correctly or short the filament transformer.
• Advanced (PDA — needs the multiplier and the TubeTime-class supply): 3FP*, 3JP*, 5ABP*, 5CP*, 5SP* and the like. Save these for a second build once you are comfortable with kilovolt work, and only on a driver with the multiplier fitted.
• Avoid for a first clock: anything marked “No,” “Unlikely,” “Very poorly,” or “Fuzzy,” the 2.5 V 7AN-based 3-inch types, the 1DP* (power-supply mod), and anything over 5 inches (untested).

The honest summary: pick a small-to-medium, non-PDA, non-HC tube with a P1 or P31 green phosphor, confirm its base and heater voltage off the spec sheet, order the matching socket and a 4.7 Ω Rfilament before the tube arrives, and you have removed nearly every way a scope-clock build goes wrong at the glass.

6.6 References (Vol 6)

• TubeTime (E. Schlaepfer), crt-driver open-hardware repository — CompatibilityList.txt (reproduced in full in § 6.2) and BaseDiagrams.pdf. Held in 02-inputs/A - Open Source/crt-driver-master/.
• TubeTime, “Electrostatic CRT Driver Design,” tubetime.us/?p=183. PDF held in 02-inputs/A - Open Source/.
• OSC4.4 — AssemblyInstructionsForOSC4.4.txt (DG7-32 and 6Lo1i pinouts, measured operating voltages, per-tube R99/R4/R6/R8 resistor values, and Rfilament guidance — source for § 6.3). Held in 02-inputs/OSC4_4/.
• OSC4.4 supporting CRT documentation: DG7-32.pdf, CRT Data 2AP1-7JP1.pdf, National Union 2002.odt, and the “Buy CRTs Here” surplus-seller bookmark. Held in 02-inputs/OSC4_4/.
• CRT datasheets: 3MP1.pdf, 3RP1A-rca.pdf (RCA), waterman_s11a_manual.pdf (Waterman pocket-scope donor reference). Held in 02-inputs/CRT Data Sheets/.
• Cross-references within this series: Vol 3 (high-voltage supply, A2 ceilings, and the Cockcroft–Walton multiplier for PDA tubes), Vol 4 (per-tube deflection and filament resistor selection), Vol 8 (the owned OSC4.4 build, wiring, and bring-up sequence).