Finally I’ve settled with a ‚preliminary final‘ setup.
It integrates all three approaches:
- snubber across the motors,
- MOVs,
- and SSR zero crossing switching
which might be some overkill.
There is one china clone module with
4 Omron G3MB-202P SSRs switching the motors at zero crossing, one for each direction of the platter motor. They stay cool with the ‚bigger‘ 0.5A pump motor and seem to have no problems switching the motors.
I put all of the
snubbers and MOVs in parallel with the loads now as described below in more detail. Because place is no issue, I put all of them on a breadboard pcb.
Unfortunately, taking the snubbers from the switches does not fix the hum issue. Because the
SSRs do not switch off completely but leave 7V RMS on the load, especially the platter motor is humming and you can feel that hum mechanically when touching the motor.
Voltage across the platter motor when SSR is switched ‚off‘:So I stuck to the idea to have an
additional mechanical relay switching off L for all motors with 1s delay after all SSRs are ‚off‘ (see schematics) and switching on with the start of the first motor. This with a delay to allow the mechanical contact to settle in ‚on‘ position before the SSR starts the motor. It works fine.
For security reasons the complete deivce is switched (L and N) with a small but serious major brand
contactor (see schematics). A second mechanical relay (not in schematics) sits normally connected in the contactor‘s coil circuit and engages to switch off the machine completely after an idle time of 15 minutes.
Also not in the schematics are four
fuses, one in the power connection for the complete machine as well as one for the 12 and 5V PSUs, one (1A slow) for the motors, MOVs and snubbers, and one for the contactor coil circuit.
In more detail:
All snubber capacitors are Wima X2 rated ones. The snubber resistors are 5W metal oxide resistors now.
The MOVs are the 275V ones discussed above. I did some measurements to verify the effect but was not able to get any significant differences with versus without MOV. As Tim stated above, because of the zero Volt switching of the SSRs there seems to be no overvoltage to engage the MOVs. I decided to leave them in place anyhow. But - as they are oversized and seem not to be triggered into action normally, I did not put a e.g. 10 Ohm resistor in series. In case they end shorted, there is the 1A motor circuit fuse in place.
Motor 1This is the
line synchronous clock drive motor that drags the suction arm across the record with L=9.1H.
It's the motor that had killed the original snubber resistor. From the last test setup, the capacitor now is increased to 100nF, R decreased to 100 Ohm, because the snubber no longer bridges the switch, which in the last setup led to motor rattling.
Switching off:
C2/magenta: SSR input signal (active high now with the new four SSR module)
C4/green: voltage across the SSR switch/load side.
Motor 2This is the 80W vacuum pump motor, dragging 0.5A.
The snubber here is 470nF/330 Ohm now.
Switching off:There is some +/-450V at the beginning. I assume it’s somehow caused by the motor inertia, maybe modulated by pressure inside the three pistons. But honestly I have no idea. Anyhow, the 450V are not enough to trigger the MOV into action.
Motor 3:This is the two directional permanent 10W split capacitor motor that drives the platter of the record cleaning machine.
I did not chage the values. Originally the calculated ones were higher, but as 22nF/200 Ohm worked in the last setup, it’s 22nF/220 Ohm (new 5W resistor) now.
Switching off for both directions: There is a small overshoot of 370V, but even less critcal than the 450V at the pump motor.
One question left:
Does the mechanical relay see any kickback when switching off the contactor’s coil?Obviously not. There is some bouncing, I’d say, but nothing to worry about. Even when switching exactly at max voltage as it is the case here.
During switching,
the ATmega shows no problems at all which was the major objective of this endeavor. I had it counting tenth of seconds on a I2C-display and there were no hickups at all.
Tim and all, thanks for your input. It’s far from being a professional solution now, but for me it’s a status to work with, and I do now understand parts of what’s going on at least.
I will now print a case for the display unit and start to write some more software to make the appropriate LEDs blink. 😊
As always, any comment appreciated.