3D prints · No. 02

Three gel blaster parts from phone photos

Three parts for a toy gel blaster, measured with a soft tape and a phone camera. The catch I rebuilt was broken at the end I least suspected.

12 October 2026 · updated 15 October 2026 · Designed in CadQuery, printed on a Creality K1 Max · Short version ↓

PLA and PETG
2 h 48 min printing
about 40 g
11 prints
3 parts

The whole gel blaster lying on a wooden table: black body with orange and blue splatter paint, curved magazine, skeleton stock, and a small red printed trigger behind the magazine. — The blaster with the red printed trigger in, behind the magazine. The battery caps are inside the stock. The catch is printed but not fitted yet.

My first printed trigger was thicker than the one it replaced, its pivot pin stuck out of the wrong face, and it fit nowhere.

Two black triggers side by side on a white surface. On the left the slim original with a pink foam pad on its back plate. On the right the first printed copy, chunkier and textured. — **Fig. 01** · The original on the left, glued back together after it snapped. My first copy on the right.

The blaster is an AK-style gel blaster, a battery-powered toy that fires soft gel balls. Two things on it were broken, and one thing I wanted to add. The trigger had snapped, and gluing it back hadn't made it work. The magazine catch had lost the paddle you press to drop the magazine. And I wanted it to run on an 18650 cell, a rechargeable lithium cell the size of a finger, tucked inside the stock.

Close-up of the blaster's black body split open: the motor, a small white switch, and red and black wires running out at the bottom towards the stock. — **Fig. 02** · Opened up for the battery job: the switch the trigger's hook presses, the motor, and the wires that run back into the stock, where the cell goes.

The battery

The battery came first, because it looked easy: a little cradle to hold the cell in the channel inside the stock. I measured the channel with the tape curved around its inside and got three centimetres. A curved tape is a rough way to measure a channel, so before the real thing, a test print of three cradle slices, 24, 26 and 28 mm wide. Eight minutes.

Close-up of the opened stock: the grey 18650 cell lying in its channel, and a soft tape measure held across the channel by a thumb, reading 5 to 7 cm. — **Fig. 03** · The second measurement, the one that found 2 cm: the tape held flat across the channel with the cell in place. The first time I had curved it round the inside and got three.

None of them went in. The channel is really 2 cm wide and the cell is 18.5 mm, which leaves no room for a wall on either side.

Pale blue render of a long U-shaped cradle with the cylindrical cell lying in it. — **Fig. 04** · The cradle I drew first, with the cell in it, and the four caps that replaced it. Drawn in CadQuery, the Python CAD library I draw everything here in.

Pale blue render of four short round caps in a row, each with a pocket in its face. — **Fig. 04** · The cradle I drew first, with the cell in it, and the four caps that replaced it. Drawn in CadQuery, the Python CAD library I draw everything here in.

So the walls went away and only the ends stayed: two round caps that plug into the channel, one at each end of the cell, each holding a contact plate, with a hole for the wire. Four caps, so two spares, in 14 minutes. I super-glued them into the stock and fitted the contacts.

The trigger

I'd started with three numbers: 3 cm long, 1 cm for the top part that presses the switch, and half a centimetre for the part below it. The first design read them as thicknesses, a 10 mm top half and a 5 mm bottom half, put the pivot pin standing off the face, and went to the printer before I'd seen a picture of it. Two copies, 21 minutes, on a Sunday afternoon. That's the one on the right in Fig. 01.

The small black original trigger lying on a white floor next to a soft tape measure, its curved blade pointing left and its thin plate and hook along the top. — **Fig. 05** · The original beside the tape. This photo, two like it, and my three numbers were everything the first design had.

My numbers meant something else. The 1 cm was the depth of the thin plate along the trigger's back edge, the plate whose hook presses the switch. The half centimetre was the thickness of the body. The plate itself is only 2 mm thick, because it slides in a slot inside the gun. Three photos looking at the trigger end-on settled the real shape.

Fingers holding the original trigger seen from its back: a narrow black plate with a pink foam pad, a small round boss near the top and the pivot pin sticking out of the end. — **Fig. 06** · One of the three end-on photos, and the rebuilt design from the same side, drawn in CadQuery. The round boss near the top is the hook that presses the switch.

Pale blue render of the rebuilt trigger from the same side: the narrow plate, the round hook boss near the top, the pin at the end. — **Fig. 06** · One of the three end-on photos, and the rebuilt design from the same side, drawn in CadQuery. The round boss near the top is the hook that presses the switch.

Since that day, two rules. Nothing prints until I've seen a picture of it. And a part that has to fit gets a coupon first: just the piece that matters, small and fast.

The trigger's coupon was the back plate, the hook and a wider foot at the end of the plate, 8 minutes of printing. It fit. The one change: the foot had to go, so the plate ends flat.

Fingers holding a small dark printed test piece, a thin plate with a hook, with a red circle drawn around a wider head at its top end. — **Fig. 07** · The coupon. It fit the slot. The circled foot at the end of the plate had to go.

The final trigger printed two copies in red PLA in 15 minutes. Neither was quite right: the start of the plate was a bit too long and the end that presses the switch a bit too short. I cut the long end back with pliers and fixed the short one myself. Then it worked perfectly.

Fig. 08 · The final trigger, turning. Drawn from the file that printed: the 2 mm plate with its hook, the 5 mm body, and the pivot pin at its end, on top in this view.

Close-up of the blaster's trigger area: the red printed trigger sits in the trigger guard behind the black magazine, in front of the grip. — **Fig. 09** · The red trigger in the gun, trimmed with pliers. It works.

The wrong end

The magazine catch was harder, because I didn't know what the whole part looked like. What was left is a small lever. At one end is the ring, which pivots on a pin through its hole. Higher up, the wedge sticks out and hooks the magazine. Beside the ring sits a small stub. At the opposite end, the far end, the lever looked snapped off. The paddle you press had broken off somewhere, and I knew it needed about 1 cm more to press on.

Fingers holding the broken black magazine catch lever: a ring at the bottom, a wedge sticking out to one side higher up, and a short stub. — **Fig. 10** · What was left of the original catch, the day I measured it: the ring, the wedge, the stub and the far end. I read the far end as the break.

This time the thickness came from photos of the lever's edge: 9.5 mm. The new paddle went on the far end, and it printed in 11 minutes. Only once it was off the bed did I see it: the paddle was on the wrong end. It belongs at the ring end.

Fingers holding the black lever end-on, showing its far end: a flat, smooth face. — **Fig. 11** · The far end was never broken; the stub beside the ring is where the paddle snapped off.

Blurry zoomed crop of a tape-measure photo: the lever's ring with a short, ragged stub sticking out beside it. — **Fig. 11** · The far end was never broken; the stub beside the ring is where the paddle snapped off.

The far end wasn't broken at all. It was flat and smooth, with a moulding line on it. The real break was the stub beside the ring, so small it had read as a tab. The paddle moved there, 10 mm long and the full 9.5 mm thick, with three grip ridges.

The new catch design drawn over a photo of the original lying next to a tape measure: the copied part in red with a yellow outline, the new paddle in blue at the ring end. — **Fig. 12** · The new catch over a photo of the old one. Red is the copy of what was left: the ring, the wedge, the stub and the far end. Blue is the new paddle, growing out of the stub at the ring end.

Before the corrected catch printed, I spotted a weak point at the wedge: it would snap there, easily. The model had a 0.4 mm gap cut right across the lever, so only the wedge held the two halves together. I closed it, and now every part gets its cross-section checked along its length for thin spots before it prints. The catch printed in red PETG in 11 minutes, at a quarter past three on Monday morning, 12 October.

Screenshot of a pale blue view of the catch design with a red circle hand-drawn around the narrow neck beside the wedge. — **Fig. 13** · The weak spot, circled on the drawing before it printed: a 0.4 mm gap cut across the lever beside the wedge, so only the wedge held the two halves together.

Twenty-five balls

The balls were a different kind of mistake. There was a file of 25 balls, 5.9 mm each, on a thin plate you snap them off: printed ammunition. Hard plastic balls in a blaster built for soft gel balls can jam it and hit a lot harder, and a round ball sitting on a flat bed touches it at one point. I was sure you could print them, so I did, in PETG, at ten to four that same morning.

Eight minutes in, the filament was everywhere.

Low-resolution printer camera frame: red PETG strings and blobs scattered on the bed in front of the print head. — **Fig. 14** · The ball plate through the printer's camera, eight minutes in. Every ball has let go of the bed and the nozzle is pushing the pile around.

Each ball starts as a dot smaller than the nozzle, and its whole lower half leans out over nothing. PETG stays soft and sticky for longer than PLA, so the nozzle dragged blobs around and knocked the small balls loose, and my PETG profile cools gently, where tiny parts need the fan at full. I stopped it and dropped the idea.

For a day and a half. On Tuesday evening the same 25 balls went back on the bed in pink PLA+, and this time the print ran to the end, 27 minutes. In the blaster they fired weakly. My guess is the size: a gel blaster's barrel is made for gel balls of 7 to 8 mm, and these were 5.9 mm, so the air went past them. So a size test followed the same night, 24 balls in four strips of 6.6, 7.0, 7.4 and 7.8 mm, 46 minutes, each strip marked with pips so I can tell them apart.

Where it stands

The trigger works. The catch and the battery haven't been tried yet, and the balls are waiting for their size to be picked.

Now I find the break before I draw the replacement. The end that looked broken was the one end that wasn't.

The short version

What went wrong, and what fixed it.

A battery cradle with walls didn't fit a 20 mm channel around an 18.5 mm cell.Two end caps that hold the contacts, nothing along the sides.
The first trigger was too thick, with its pin on the wrong face.Read each number as the part it measures, a picture before printing, a coupon first.
The coupon fit but had a wider foot at the end of the plate.Removed it. The final trigger needed a trim and then worked.
The new paddle went on the end that wasn't broken.Found the real break, a stub beside the ring, and moved the paddle there.
A 0.4 mm gap across the lever left one weak spot.Closed it, and check the cross-section along every part now.
Tiny PETG balls dragged into strings 8 minutes in.Dropped, then printed in PLA+ a day and a half later. They fired weakly, so a size test is waiting for its verdict.

How it was printed

Printer: Creality K1 Max, rooted, Klipper
Slicer: OrcaSlicer 2.4.2, K1M profiles
Measuring: A soft tape measure next to the part and phone photos. No calipers.
Battery end caps: 4 caps (2 spares), 14 min, 5.46 g, no supports
Trigger: Red PLA, 0.20 mm layers, 100 % infill, 2 copies in 15 min, 5.1 g
Magazine catch: Red PETG, 0.16 mm layers, 100 % infill, 11 min, about 2 g
Balls: 25 at 5.9 mm, red PETG, stopped after 8 min. The same 25 in pink PLA+, 27 min, then 24 in four sizes, 46 min

This is a toy. Its exact brand and model aren't known, so the post doesn't name one. 11 prints: the ten in the story, plus a false start of the catch cancelled after 37 seconds. Grams are what the printer fed over all 11, the failed PETG balls included.

The files

The trigger, the magazine catch and the battery caps as STLs, with previews and a README. They fit one toy of unknown make, so check yours before you print. The trigger has been fired; the catch and the caps have not been tried in the gun yet.

Parts, STLblaster.zip817 KB

License: CC BY-NC-SA 4.0. Credit: Abdallah El Belkasy, abdaishere.com/prints.

Off the bed