Restoring the X68000 CRT Flyback Transformer with Epoxy Sealing

The flyback transformer, abbreviated FBT, sits at the heart of every X68000 CRT chassis. It generates the 25 kV anode voltage that lights the phosphor and produces the horizontal deflection amplitudes needed for the platform's signature 768×512 display modes. Sharp sourced these transformers from several OEM partners, and the units in early ACE machines are not always identical to those in later XVI revisions. What they share is a reliance on vacuum-impregnated polyester resin around the primary winding, and that resin is the part that ultimately gives up.

In coastal Australian cities such as Sydney and Brisbane, warm humid air accelerates the ageing of that resin. Moisture absorption changes the dielectric properties of the windings and produces micro-arcing inside the bobbin. Inland enthusiasts in Melbourne or Adelaide see the same failures, but usually a few years later in the unit's life. The visible sign is a faint ozone smell, a high-pitched whine from the deflection yoke, or a screen that collapses to a single horizontal line.

Epoxy sealing is one of the more pragmatic repair paths when an original replacement is unobtainable. New old stock X68000 FBTs trade for eye-watering sums on Japanese auction sites, and even then sellers cannot guarantee the part has been stored dry. A restorer instead removes the suspect transformer, bakes it gently to drive out moisture, and pots the vulnerable joint with a low-stress electronics-grade epoxy. The technique has been used for decades in the Australian ham radio community to rescue television line output transformers from landfill, and it transfers directly to the X68000 with only minor adjustments to the cure schedule.

What follows is the procedure used on several machines documented in the project notes. The article covers the diagnostic signals that point at the FBT, the materials you need, the actual potting workflow, and the high-voltage checks you should perform before powering the machine back up. None of this is a substitute for proper respect of the voltages involved, but it is well within the reach of a hobbyist with a multimeter and a temperature-controlled soldering station.

Recognising the symptoms of a failing flyback

The X68000 gives you surprisingly little warning before a flyback failure. The most common early indicator is a faint crackle from inside the chassis on a cold morning, accompanied by the screen taking an extra second or two to appear. This is the high-voltage winding arcing internally as the resin contracts overnight and leaves a tiny void. As the chassis warms up, the metal expands, the gap closes, and the arc disappears, which is why so many owners ignore the warning.

A second symptom is the high-voltage rail drifting. You can watch the +B supply on the main board rise and fall as the FBT struggles to maintain regulation, which shows up as the picture growing slightly wider when the machine is warm and shrinking again after a power-off. Owners in Perth have described this as the image "breathing", and it is a very reliable tell.

The third symptom is the dreaded single horizontal line. By the time you reach this stage, the FBT has usually failed to the point where the horizontal output transistor is also struggling. Replacing the HOT without addressing the transformer is a common Australian trap, because the new transistor will be destroyed within minutes by the same over-voltage spike that killed its predecessor. If you see this symptom, isolate the FBT first.

Why epoxy sealing works for this repair

The X68000 flyback transformer is potted in a wax-and-resin compound that was adequate for the climate of a Japanese electronics factory in 1988, but not for a thirty-year-old part sitting in a garage in western Sydney or a storage unit in suburban Geelong. The wax softens over time, the resin develops micro-cracks, and moisture finds its way into the windings. Once that happens, the dielectric strength of the insulation drops, and arcing begins.

Epoxy sealing works because a modern two-part electronics epoxy has far better moisture resistance than the original wax blend. It also has a higher dielectric strength, so once cured, the internal insulation is genuinely improved. The trick is to drive out the moisture first, because trapping damp inside a sealed block will simply move the problem from the windings to the core. A low-temperature bake, typically two hours at sixty to eighty degrees Celsius, takes care of that.

The other reason epoxy sealing works is that it provides mechanical support. The original transformer relies on the bobbin and the core to keep the windings in compression. When the resin shrinks, the windings can vibrate at the 15.6 kHz horizontal frequency and physically chafe against the core. Potting the vulnerable area locks everything in place and damps the audible whine that often accompanies the early stages of failure.

Tools, workspace setup and Australian sourcing notes

You will need a temperature-controlled hot air station or a small laboratory oven, precision screwdrivers, isopropyl alcohol, a small syringe, and a slow-cure electronics epoxy. None of these are difficult to source in Australia. Major electronics retailers stock the epoxy, and pharmacy-grade isopropyl works for the cleaning steps.

A bench in a garage or shed is fine, but avoid working anywhere with high humidity. Brisbane summer days above thirty-five degrees with eighty percent humidity will undo your drying step the moment you open the oven door. If your workshop is in a coastal suburb, run a small dehumidifier for a couple of hours before you begin, and keep the transformer in a sealed container with a sachet of silica gel between the bake and the pour.

High-voltage safety gear is not optional. A proper discharge probe with a ten-megaohm resistor and a grounded clip is essential, because the X68000 anode can hold a lethal charge for hours after the machine is powered off. If you do not already own one, a local ham radio club in Melbourne or a Maker group in Adelaide can usually lend you one, or you can build it from a salvaged TV service probe and a length of thick silicone wire.

A multimeter with a high-voltage probe is also useful, though not strictly required. Even a basic ten-megaohm input meter will let you confirm that the +B rail is present and roughly correct before you reconnect the yoke. Anything fancier, such as a portable oscilloscope, is a luxury you can do without for this particular repair.

Comparison of suitable potting compounds

Several epoxies are commonly recommended for this kind of work. Below compares the three formulations most often discussed in retrocomputing circles, with notes that reflect their availability in the Australian market and the realities of postage from Sydney or Melbourne to the rest of the country.

Product	Mixed viscosity	Cure time at 25 °C	Dielectric strength	Notes
Loctite EA E-30CL	Low (~3 500 cP)	24 h full / 4 h tack	500 V/mil	Clear, UV-traceable for inspection.
MG Chemicals 832HD	Medium (~6 000 cP)	24 h full / 6 h tack	410 V/mil	Thicker, better gap filling.
JB Weld ClearWeld	High (~10 000 cP)	24 h full / 4 h tack	320 V/mil	Available at Bunnings, convenient local pickup.

For X68000 work the Loctite E-30CL is the preferred option when it is available, because the low viscosity lets it wick into the micro-cracks before gelling. The MG Chemicals product is the better choice when the transformer has visible gaps or where the original potting has crumbled away. JB Weld is the fallback when shipping times from the specialty suppliers are too long, which is a real consideration in regional Australia where standard postage from Sydney or Melbourne can take the better part of a week.

The step-by-step sealing procedure

Begin by discharging the anode and removing the chassis from the case. Lay it on an anti-static mat with the FBT facing up. The transformer is usually held in place by a metal clamp and a single screw at the base, and it is connected to the board by three or four heavy-gauge soldered pins. Desolder those pins with a temperature-controlled iron at around 380 °C, taking care not to lift the pads.

Once the FBT is free, label the orientation with a marker pen before you do anything else. Then clean the exterior with isopropyl alcohol and a soft brush, paying particular attention to the area where the original resin has cracked. This is where the new epoxy will be working.

Place the transformer in your laboratory oven or under a low-temperature hot air source at sixty to eighty degrees Celsius for two hours. This is the moisture-driving step, and it should not be skipped. After the bake, transfer the unit directly to a sealed container with fresh silica gel and allow it to cool to room temperature before you mix the epoxy.

Mix the epoxy strictly according to the manufacturer's ratio, stir thoroughly for the full recommended time, and draw it into a syringe. Apply a thin bead around the suspect joint, then add a slightly thicker fillet over the cracked resin. Do not flood the entire transformer; you are sealing specific weak points, not potting the whole assembly. Allow the epoxy to cure for the full twenty-four hours, then inspect the seal under a magnifier and backfill any remaining voids with a second thin layer before refitting the transformer.

Reassembly, HV testing and long-term care

When the FBT is back in place and the board is resoldered, perform a static resistance check before any power is applied. The primary winding should read a few ohms, the high-voltage winding should read essentially open, and there should be no shorts between any pin and the core. A reading of zero on any pin to the core means the seal has bridged something it should not have, and the transformer should be removed and cleaned before you proceed.

Power up the chassis through a current-limited variac or a dim bulb tester, watching the +B rail and listening for arcing. The first ten seconds are the most informative: a healthy FBT will let the horizontal oscillator lock immediately, and the screen will flash on within a couple of seconds. A weak unit will either fail to oscillate, or it will produce the high-pitched whine that you were trying to cure. If you see either of those symptoms, power down immediately and re-examine the seal.

Once the chassis is running cleanly, leave it on for an hour to confirm thermal stability. Then power it off, wait five minutes for the anode to bleed down, and discharge it manually with your probe. Open the case and feel the FBT; it should be warm but not too hot to touch. Anything beyond that suggests the windings are still overloaded, and you should reseal or replace.

Long-term, the best thing you can do for a freshly sealed X68000 is to keep it in a dry environment. A small silica gel canister tucked behind the case and replaced annually costs almost nothing and will extend the life of every dried electrolytic and sealed transformer in the machine. Combined with an annual inspection of the anode cap, that single habit will keep most restored X68000s running for another decade or more.

If you want to read the original session where this work was first sketched out, the notes live at diary. And if you tackle your own flyback repair, share the results with the wider community; there are few enough X68000s left that every saved machine genuinely matters.