A Case of Design Obsolescence - Sunbeam Electric Blanket Controller Model BL0700 Type 591
Apparently nothing lasts forever and eventually stuff just plain wears out through fair wear and tear. Electronic stuff in particular has a ‘use by’ date even if it is not written in the manual. Electrolytic capacitors in particular tend to loose capacitance over time and as a consequence can stress other components. There are repair shops that will replace electrolytic capacitors as a matter of course during servicing of equipment that is a few years old, and bill you appropriately. But I am digressing...
This particular story is about a Sunbeam electric blanket controller Model BL0700 Type 591 connected to a Sunbeam BL2987 super king electric blanket. The blanket is about 7 years old. It gets intermittent use for about three months a year. During these periods it is left plugged in and powered, but the blanket itself is only turned on for a couples of hours a night.
Figure 1. Sunbeam Electric Blanket Controller Model BL0700
Over time the LCD displays have been getting dimmer on both controllers until a few weeks back when one of them stopped working at all and both the power and on LEDs came on dimly. That’s not good. I unplugged both controllers from the mains.
I set about dissembling the dead controller, ensured that everything was safely discharged, cut the leads to the electric blanket close to the circuit board so I didn’t have to lug the whole blanket around, and put the controller on the test bench.
Disassembly is quite straight forward. First gently prize off the off-on-timer slide switch knob (it has been adhered to a slide switch on the circuit board and it must be removed before the board can be freed from the case - otherwise the switch will fall to pieces). Now undo the four Phillips screws on the underside of the case and the case will lift free. Once the unit is disassembled be careful handling the LCD to avoid chipping or cracking it.
Figure 2. Remove the On Off Timer Switch to Disassemble
There was nothing obviously damaged so I set about sketching out the power end of the circuit. Note that there is no need to draw the microcontroller or the LCD drive as if these have failed then the controller is beyond economic repair. From the component values printed on the components I already have an idea of why both controllers have failed over time. But let’s not get ahead of ourselves and jump to premature conclusions.
Figure 3. Power Supply Schematic Diagram
In essence the power supply is a capacitive AC voltage divider with some protection including a thermal fuse and a metal oxide varistor connected to a half wave diode rectifier followed by a Zener regulated capacitive filter. Note that there is a 560 K bleed resistor across the 1 uF capacitor to dissipate any residual charge when the power is turned off. This is an essential component to reduce the risk of accidental electric shock or damage to your test equipment.
I have made quite a few of these supplies over time for other projects. I have a data sheet somewhere that describes a capacitive AC LED driver but there are lots of other data sheets such as Microchip Application Notes AN954 http://ww1.microchip.com/downloads/en/appnotes/00954a.pdf that describe the basic principles.
These supplies are small, light and low-cost (no transformer required), dissipate hardly any heat (the current and voltage are 90° out of phase across the capacitor) and can supply up to about 100 mA with sensible capacitance values. Their biggest drawback is there is no mains isolation.
With the schematic sketched out I set about testing the diodes, resistors and capacitors with a simple Ohmmeter. Nothing was obviously wrong (but note that in-circuit values will be influenced by other components and cannot necessarily be measured in circuit). When prodding about with your Ohmmeter it is good form to observe correct DC polarity when you can to avoid damaging stuff.
Figure 4. Top of PCB
Okay, time to take some measurements. The Zener diode is 4.7 Volts so we should expect around 4.7 Volts across the associated 470 uF filter capacitor. I connected a hand-held DC voltmeter (isolated from ground) across the Zener I plugged in the controller. The voltage was measured at 2.0 Volts! This less than half of what I was expecting. So what is causing this? Either the down-stream electronics are drawing excessive current, or the filter capacitor is open circuit, or the main supply capacitor has lost capacitance.
Figure 5. Bottom of PCB
With the mains disconnected and everything discharged I connected my 5 volt DC bench supply through a 10 Ohm resistor directly to cathode of the diode rectifier and turned on the power. The resistor was just a precaution to limit the current in the event of an down stream or accidental short circuit. The controller immediately sprang into life with the full LCD contrast restored and both LEDs functioning properly.
This confirmed my earlier suspicions about the cause of the failure. I desoldered the 1uF 400 VDC Mylar capacitor (mounted beside the slide switch) and measured its value. As expected it had dropped to just 0.4 uF. So it’s reactance had increased and the supply current to the rectifier had dropped from about 75 mA rms to 30 mA rms.
Figure 6. Errant Capacitor (C1)
Now we come the interesting part of figuring out why this component had failed over time. On the basis of the symptoms I expected (and it was later proven) that the same capacitor on the other controller was also on its way out, with a measured value of about 0.53 uF.
This controller operates from nominally 240 Vrms 50 Hz AC and is essentially connected directly between phase and neutral. While the particular component was rated at 400 V DC this IS NOT ADEQUATE for the AC duty. A quick check on any Mylar capacitor data sheet will confirm this.
So the capacitor in question has an inadequate voltage rating for the intended use. In essence this means that these Sunbeam controllers have designed obsolescence (that is to say they are designed to fail). I have contacted Sunbeam and I got an automatic reply saying I would be contacted in three days, but they haven’t bothered.
The repair should have been as simple as getting a replacement 1 uF Mylar capacitor with an appropriate AC Voltage rating (630 V DC is appropriate). The problem is that I cannot find a 1 uF Mylar capacitor with an appropriate voltage rating that will fit in the controller case.
The solution was to reduce the capacitance to 0.82 uF. I didn’t have any of these in stock but replacements were ordered and arrived in less than a week.
With the replacement capacitors fitted in both controllers, and the blanket wiring reconnected to the circuit board, both units are working as good as they day they were first plugged in. And when they do eventually wear out it will not be because of an under-rated capacitor and the blanket will not be replaced with a Sunbeam product.
Another Example of Sunbeam Design Obsolescence
I thought I had reached the end of this saga, but I was interested to hear from a reader with a Model 71660, Part Number 116766-001, Sunbeam electric blanket that had failed after just one year of use (hey, just outside of warranty).
The problem this time was not with the controller, but with a small circuit board, Part Number 101191-202, embedded in the blanket proper. The board is a safety feature designed to detect open or short circuits in the wiring to the elements. It contains just four SMD components (three resistors and a PNP transistor) and is housed in a plastic casing inside the seam of the blanket.
Figure 7. Defective Safety Circuit
The blanket elements and controller were determined to be serviceable by the reader with a few simple resistance measurements of the element windings using an Ohm-meter and testing the controller with another blanket.
I was able to ascertain the function of the board with a bit of reverse engineering and some educated guesswork, confirmed when the reader found the patent for the circuit (US 006,355,912 B2, and in particular Figure 2).
While I cannot determine the manufacturers’ of the SMD components used in this board without a significant forensic examination, I have determined that the probable cause of this failure is another underrated component. This time it is an SMD resistor that, on the basis of typical manufacturer’s ratings for similar resistance’s in an identical packages, is not appropriately rated for the applied peak voltage. As a consequence both the resistor and transistor had failed on the board.
I don’t mean to pick on Sunbeam, but this appears to be another example of a Sunbeam electric blanket designed to fail. Their design team are probably having a party at your expense because the failure occurred right after the warranty expired. Shame on you Sunbeam.
The additional cost of an appropriately rated resistor would have been a fraction of a cent! The cost of the repair is less than a dollar in parts and maybe an hour to determine the fault, remove the board, replace the components, reinstate and test everything. However, but for the perseverance of the reader, this fault would have gone unreported. Most folk would have consigned the blanket to the landfill and Sunbeam would be anticipating another sale!
Interestingly, in studying the associated patent, US 006,355,912 B2, one can see that the circuit fails to achieve the primary utility of detecting element faults without power applied. Further, it does not detect a raft of element failures that could be easily determined from the controller without the need for a safety circuit in the blanket. So not only is their circuit implementation faulty, but so is the basis for the associated patent.
I’m sure you will draw your own conclusions but I won’t be purchasing a Sunbeam ‘anything’ in the future.