My battery backup setup

[sdb7802]

Happy Memorial Day to All!

Thanks to all who have sacrificed to protect and preserve our freedom and liberty.

What is the current sourcing need for the 3.3VDC voltage regulator? For example, the LD1117V33 or the LD1117-3.3 TO-220.

Will a typical 1/4 Watt resistor do?

The +3.3V regulator has to source approximately 1mA. But it needs to withstand the +24V input, which the LD1117V33 won't do. The LD1117V33 will only handle +16V.

I used a MCP1804T-3302I/DB which is rated at +28V.

A 2.7k ohm, 1/4 watt, 5% resistor will do fine.

[verbatim1]

The +3.3V regulator has to source approximately 1mA. But it needs to withstand the +24V input, which the LD1117V33 won't do. The LD1117V33 will only handle +16V.

I used a MCP1804T-3302I/DB which is rated at +28V.

A 2.7k ohm, 1/4 watt, 5% resistor will do fine.

Thanks Steve for the breakdown voltage advice (which I hadn't considered when I was looking up which voltage regulator to buy).

Here is the spec sheet (http://ww1.microchip.com/downloads/en/De...22200C.pdf ) for that Microchip MCP1804 "150 mA, 28V LDO Regulator With Shutdown" device.

As you noted, I see from the datasheet that it has a 2.0V to 28.0V "Input Operating Voltage Range:", which covers the nominal 24 volt input.

It certainly consumes almost nothing when not in use (50 μA of quiescent current), while the maximum output current is 100mA at less than or equal to 3 volts and 150ma above 3 volts, which is far more than the 1ma that you mentioned it needs to source.

Thanks also for confirming that a quarter-watt 5% tolerance resistor will do.

Since I'm brand new to CPAP, I wasn't aware of these non-standard power supply issues until yesterday, so, I guess I'm debating between building my own (as you did) or just buying (1) the ResMed 12V-to-24V up converter; (2) a 12V sealed lead-acid battery in the 35 Amp Hr or greater range; and (3) a 12V SLA battery tender

[verbatim1]

Other than the cynical reason, do we have any good engineering reasons why Resmed went to the hugely non-standard 24-volt-with-3.3volt-sense power supply?

Googling for the patent, I find this:
Power management in respiratory treatment apparatus
US 20140366876 A1

I see that they do the obvious, which is to prioritize power to the fan over the less immediate needs of the heater and humidifier, but that has nothing to do with the non-standard 24volt supply, nor the 3.3-volt sense resistor.

Deep in the patent description is the line "Optionally, a pull up resistor may be implemented to code information about a component coupled with the bus such that different resistors may indicate different components or different component functionality".

It goes on to explain that the pull-up resistor is in the power supply and that its value is sensed by the "analog sensing elements of the transceiver".

To me, the use of the word "transceiver" is strange, but they must be using it in a non-standard way, as I've only seen the word used in terms of radio transmissions. They seem to be using the word "transceiver" merely as the items sending and receiving signals on a non-radio-frequency communications bus.

Basically the patent explains, in words, the circuit used, as it says "if the pull-up resistor represents the type of power supply, a detection of 3.9 K ohms may be interpreted by a master controller that the power supply is an infinite supply (mains).".

This is interesting, because it intimates that ResMed machines can handle the 120Volt 60HZ mains, which, if true, I was wholly unaware of.

The patent goes on further to say "A detection of 2.7K ohms may be interpreted as a 90 Watt power supply.", which, incidentally, is the wattage of "my" 24-volt power supply.

Additionally, they outline that "A detection of 1.8K ohms may be interpreted as a 60 Watt power supply" and that "A detection of 1.0K ohms may be interpreted as a 30 Watt power supply".

I find it odd that these actual numbers are listed in a patent, simply because there is nothing inherent in the resistor values that determines the wattage, and patents are generally rather purposefully vague on such essentially-meaningless intrinsic values.

The patent does state the obvious, which is that "Other voltage levels and coding schemes may also be utilized to code information about the power supply unit in this analog manner. ", so, that bolsters my surprise that they bothered to list the actual resistor values.

As would be expected, in addition to the analog resistor sensing, the patent does vaguely mention that "Optionally, digital messaging may also be utilized to detect information about the power supply unit or other components of the bus. ", which simply means that the patent is trying to garner rights over putting "anything" in the power supply line to provide once-removed information about said power supply.

Of course, a digital signal presupposes a "smart" power supply, which is exactly what they hint at when they mention in closing that "a data message with information from circuits or memory of the power supply unit or based on signals from the detectors or sensors of the power supply unit may be sent from a processor of the power supply unit to a system level controller or processor.".

If anyone can make sense of this more so than I just did, I would be appreciative, as I really don't see why they have such a need to make the power supply non standard.

For example, a Lenovo Thinkpad W510 senses the power supply WITHOUT a pull-up resistor, so, certainly there are other ways of doing things that don't in and of themselves make the power supply non standard.

[sdb7802]

Other than the cynical reason, do we have any good engineering reasons why Resmed went to the hugely non-standard 24-volt-with-3.3volt-sense power supply?

I guess since it's a medical device, ResMed wants to make sure the attached power supply has the power capability to run the machine properly. To check the power capacity you would have to measure the output voltage under maximum load current. You could attempt to max out the current by maxing out the blower, heater plate, and tubing heater. For the blower to draw max current it would need some type of switchable flow limiter for the test. Or you could use a switchable dummy resistive load on the power input. Either way costs money and could possibly fail and inhibit normal operation.

So, ResMed apparently choose "the resistor to 3.3V" method to "check" the power rating. If the the voltage level or resistor value is not correct the machine will not work.

My tests on two units showed the voltage can range from 2.4 volts to 3.5 volts. The resistor can range between 2.3k to 3.55k ohms.

I guess some of the stuff in the patent doesn't necessarily apply to the A10.

Some older Dell laptops use a coded signal (one-wire bus) to check the power supply. But they just issue a warning and don't shut down the machine.

I guess we should be grateful ResMed made it so easy to work around.

As for using 24 volts, I guess they wanted to save on copper and component costs. For a given power requirement (in this case, driven by mechanical and thermal function), using a higher voltage means lower current. Smaller wire, smaller pc board traces, and cheaper electronic components are required. There are some downsides to using a higher voltage. Obviously, I'm just guessing.

[eseedhouse]

As for using 24 volts, I guess they wanted to save on copper and component costs. For a given power requirement (in this case, driven my mechanical and thermal function), using a higher voltage means lower current. Smaller wire, smaller pc board traces, and cheaper electronic components are required. There are some downsides to using a higher voltage. Obviously, I'm just guessing.

Doesn't sound right to me. Twice the voltage at half the amps still delivers the same amount of power, and it's power that heats, not volts or amps separately. The heating and stress on the circuits should be identical. The power cube on my Resmed machine is rated 90 watts at 24 volts, which is 3.75 amps. At 12 volts that would require 7.5 amps, hardly a difficult load for any of your bog standard printed circuit boards.

[Sleepster]

Doesn't sound right to me. Twice the voltage at half the amps still delivers the same amount of power, and it's power that heats, not volts or amps separately.

Let's take as an example a 96 watt power brick. At 24 volts it puts out 4 amps of current. Somewhere downstream and in series is a device with a resistance of 0.5 ohms. It's part of a circuit that draws 4 amps, so the voltage drop across it is 2 volts and the power it consumes is 8 watts.

Now again suppose we have a different 96 watt power brick. At 12 volts it puts out 8 amps of current. That same device that's downstream still has 0.5 ohms of resistance so the voltage drop across it is 4 volts and the power it consumes is 32 watts.

To put it simply, more current through a device means a greater voltage drop across that device and more power consumed by that device. This is why a tea kettle in the UK will heat water four times faster than one in the US. Twice the voltage gives you four times the power output.

[justMongo]

You need a highly responsive, low inertia blower motor and impeller for a CPAP to respond in milliseconds.
I postulate that using 24 Volts makes engineering that motor "better." (Or perhaps easier.)

[srlevine1]

I am curious what generator you are using that doesn't kick your APC offline due to "dirty power." I tried a small Honda that was basically for power tools on remote sites and the power was so dirty, my APC was cycling like an oscillator. Thanks.

[verbatim1]

To check the power capacity you would have to measure the output voltage under maximum load current.

I agree that it's easy for ResMed to "check the power capacity" simply by cheap analog circuitry that senses a resistor value which is essentially 900K Ohms per 30 Watts as in the table below:

• 3.9K=mains
  
• 2.7K=90W
  
• 1.8K=60W
  
• 1.0K=30W
  

However, I'm not sure WHY it matters WHAT the power supply is. I may have missed something in this thread, but I don't understand what the ResMed A10 would do differently if it were connected to a lower power supply.

QUESTION:
Q1: What does the ResMed machine do differently if it's connected one day to a 90W power supply, and to a 30 Watt power supply the next day?

NOTE: I can "guess" that it lowers the power requirement of the circuitry, perhaps shutting off non essentials such as the heating element; but that's just a wild-butt guess.

[surferdude2]

If you connect a 90 watt load to a 30 watt rated power supply, the power supply will lose regulation of the output voltage and will overheat internally to the point of component failure unless some means of protection has been provided in the design. All the while, the load in question will not receive the proper voltage and will either not operate at all or will exhibit strange behavior.

If you connect a 30 watt load to a 90 watt power supply, all devices will work normally. The point is, having too high of a power supply rating has no effect on the load but having too low of a rating is damaging to equipment.

It sounds like Resmed has designed their power supply units to have this 3.3 volt line as a means of overload protection. Bypassing it would put you on your own for the consequences. I wouldn't hesitate to do it but would warn others to consider the risks carefully.

I have often considered the old rule of never "substituting copper for engineering" as the possible reason for some of the strange things that Resmed does but I'm never quite so sure that they are doing anything for that reason.

Dude