/magnetic stirrer of the "MM4" type was obtained for suspiciously cheap.
The device was made in Czech Republic in 1996, making it 20 to 21 years old as of 2017.
A secondhand laboratory hot plate/magnetic stirrer of the "MM4" type was obtained for suspiciously cheap.
The device was made in Czech Republic in 1996, making it 20 to 21 years old as of 2017.
 Hotplate stirrer, top view |  Front panel |  Stir knob detail |  Back side |
 Back side |  Bottom |  Type plate |  Plate side view |
 Plate side view |
The device was opened, with four self-tapping screws on the corners holding together the aluminium bottom chassis and the plastic top cover, painted blue on black thermoplastic material. Four other M3 screws, in a square pattern, were holding the posts on which the hot plate assembly was mounted. These has to be removed as well.
The removal of top lid was not without issues. The speed pot knob turned out to be the block. A small screw had to be loosened so the clamp releases the pot shaft. However, significant force had to be used because the clamp corroded onto the shaft. The plastic top of the knob was removed by force from the aluminium clamp, and the clamp was then mostly cut along its length by a Dremel wheel. Its grip then could be released. The top shell then could be slid off the panel.
 Inside |  Plate view |  Electronics view |  Electronics view |
The plate contains two circuitboards, three switches, a speed regulating wirewound potentiometer,
and a shaded-pole motor with attached somewhat rusted C-shape permanent magnet.
 Stirring magnet |  Motor coil |  Switches |  Regulating pot |
The schematics was reverse-engineered from the boards.
The speed regulation system was found grossly lacking, with a lot of power dissipating uselessly on the series resistor and on the pot, and with too heavy demands on the power of the pot. A triac or a PWM regulator would be better.
The heating has only two power levels. Half-power, with the middle button, where the power is limited with a series diode to only one half-wave of the current, and full-power, where the diode is shorted. Primitive but easy. Again, a triac/PWM regulation would be finer.
 Schematics |
It turned out that the force used to remove the pot knob also cracked the wiper assembly. This was attached to the pot shaft with a small M2.6 screw. The wiper block, made from brittle phenol-formaldehyde thermoset resin, was cracked at the screw hole, and the original screw was bent and broken.
 Damaged pot |  Damaged pot |  Damaged pot parts |  Damaged pot parts |
The wires were disconnected from the ceramic two-pole connector block, and the heater plate assembly was removed. The steel standoffs with M3 threads in them were unscrewed from the bolts anchored in the top plate. The ceramic standoffs and the in-between round aluminium sheetmetal skirts were then taken off.
 Heating plate assembly |  Wiring detail |  Bare heating plate with skirt |  Detail of attached wires |
The top case was removed and thoroughly washed.
 Top case |  Top case |  Top case |  Top case, mount rod post |
The pot wiper was found to be more defective. The springy alloy used for the wiper was not springy anymore, probably due to overheating sometime in the device's tortured history. An attempt was done to fix part of the spring in place, recovering some of the stiffness. The attempt ultimately failed.
A replacement pot would have to be 3-watt 8.2 kΩ wirewound kind. These are difficult to find these days. It has reliable contact only in the slowest and fastest ends; for now, so be it.
 Pot wiper repair attempt |  Pot wiper repair attempt |  Pot wiper repair attempt |
At the slowest setting, where the magnet is rotating barely if at all, the voltage across the motor coil is a bit below 100 volts AC. At full speed the voltage is about 160V AC.
The device was assembled and tested and found non-heating. The hot plate was tested with the continuity meter, and, surprise surprise, it was found non-conductive. The resistive wire inside was apparently severed.
It was found the skirt can be removed from the plate, by rotating it slightly against the plate until the lock pins match the cutouts. The skirt then can be lifted off the plate.
 Hot plate, with removed skirt |  Skirt lock pin |  Removed skirt |  Skirt lock pin slot |
The heater is potted in some sort of fine-grained cement. The material is mercifully soft and amenable to being removed with a small chisel. A screwdriver and a little hammer were used.
So, let's go chopping.
After careful removal of some material, ceramic beads were exposed. These are used as insulation and mechanical protection of heating spirals made of resistive wire. The spiral is laid in the aluminium casting in two loops.
 Partly exposed ceramic beads |  Partly exposed ceramic beads |  Partly exposed ceramic beads |  Partly exposed ceramic beads |
A small piece of the bead skirt was chopped off, exposing access to the bare spiral. A continuity tester was used to find which end of the wire was connected, and where the spiral is intact. More spots then were exposed to locate the fault in an iterative way. The second such spot was of course made in the wrong direction.
 Exposed heating spiral |  Exposed heating spiral |  Exposed heating spiral, multiple spots |
Gradually, two spots were exposed that were connected to the different wire ends. The fault had to be located in between.
 Exposed heating spiral, fault spot located |  Fault spot exposed |  Fault spot exposed |  Microphotograph of failed wire end |
The wire was pulled out and one end was microphotographed, showing a pattern of local overheating and mass loss by oxidation, creating a self-accelerating condition. The ends of the wire were pulled out of their beads-chain channel and twisted together.
 Wire pulled out |  Wire pulled out |  Wire twisted together |  Wire twisted together |
 Wire twisted together |  Wire twisted together, microphotograph |  Wire twisted together, microphotograph |
The earlier-tested resistance wire brazing
method was used.
Copper was used as the brazing filler metal, due to its higher melting point than brass.
A test using borax as a flux was performed and went well. The heater
wire itself was however more stubborn; flux or not it was forming a thick layer of oxide and the
copper melted but failed to wet the wires properly, usually just forming a little ball in the
molten boron glass. A more aggressive sodium tetrafluoroborate flux had to be used, with the
risk that its remains will be corrosive and shorten the life of the repair.
 Copper brazing, test |  Wires twisted together |  Copper braze added |  Fluxed, melted |
 Brazed joint after partial flux removal |  Repaired spiral |  Repaired spiral |  Repaired spiral |
The broken beads were covered with pieces of Cerablanket-class high-temperature insulation wool,
then sealed with Kittfort-brand Rudokitt
high-temperature cement for fireplaces and let dry.
 Bead detail |  Bead detail |  Exposed spiral covered with ceramic wool |  Ceramic wool covered with high-temperature putty |
 Ceramic wool and high-temp putty |  Covered holes |  Covered holes |  Covered holes |
The power cable was found to be annoying. An IEC power socket, used in all sorts of devices these days, is more useful; it allows storing and carrying the devices without having to bother with the snake nest of attached cables, removing a significant source of frustration from the technician's life. It was decided to perform a fairly standard modification.
The catch here was the close presence of potentially conductive liquids. The socket had to be protected against liquid influx at accidental spills. A mounting adapter was designed and 3d-printed from less-flammable ABS material. The connector was sunk a bit into the assembly, to protect it, and tilted down about 5 degrees to inhibit gravity-facilitated flows. A two-part shroud was printed and attached to the top case.
 Power connector adapter |  Power connector adapter |  Power connector adapter |  Power connector adapter, attached |
 Power connector adapter, attached |  Power connector adapter, attached |  Spill skirt parts |  Spill skirt parts |
 Spill skirt cutout |  Spill skirt cutout |  Spill skirt mounted |  Spill skirt mounted |
 Spill skirt and connector in place |  Spill skirt and connector in place |  Spill skirt and connector in place |
With a spill test, the shroud was found to be slightly underperforming; it worked but without the desired margin. Its next modification will involve a bit longer "roof" with raised far end, to stop the liquid flow from reaching the top of the connector and to deflect it to the sides.
New shroud for the power connector was designed, with longer "roof" with slightly raised end, and drip tips at the bottom sides. The parts were also hot-melt glued in place in order to seal them against seeping liquid.
 New shroud design |  New shroud design |  New shroud design |  New shroud design |
A spill test shown a preexisting weakness, the possibility of the liquid to splash over the raised edge under the hotplate skirt and enter the electronics. This has to be addressed later.