My battery backup setup

[justMongo]

No, the cap will not affect it. The caps keep the 3.3V regulator from becoming a high frequency oscillator.

Being ancient, I've done some analog design for hybrids. Likely the blower unit has a resistor to ground.
There would be a reference voltage and a comparator such as an LM311 that compares the reference leg to the leg produced by the power supply 3.3V and 2.7k resistor and ground resistor in the blower circuity.

The LM311 is like an OP Amp; but the output is an open collector meant to interface with logic circuitry.
Sometimes we provided a little feedback in comparator circuits to create a little hysteresis so the circuit wouldn't jitter if the compared Voltages were very near each other.

Like this one:

[verbatim1]

No, the cap will not affect it. The caps keep the 3.3V regulator from becoming a high frequency oscillator.

Thanks Mongo for that information that the cap value isn't going to get in the way.

The analog circuitry must be cheap (based on the patent description), so what you came up with seems as good as any assessment.

What matters, I think, are whether it works, and whether EMI protection is needed, and then tricks on how to build the thing.

Given that:
A. A typical "12V" battery is only nominally 12Volts, and can be 13.8 (plus or minor a few volts). Do we need to regulate this power supply voltage?

B. The center-pin plug is likely going to be hard to find. Do we have a source for the plug?

C. The ResMed unit itself generates EMI and it may or may not be "sensitive" to EMI (how so?), and we know the up-converter unit has low-pass ferrite coils, but we also know that the up-conversion process is inherently filthy when it comes to EMI. If we use this solution of a 24V (nominal) battery, do we introduce any EMI at all with this Mongo solution?

In summary, open questions are:
1. How well regulated must the 24VDC be to work?
2. Where can we source that non-standard center-pin plug?
3. Is the ResMed A10 really so sensitive to EMI that it needs filters?

[sdb7802]

In summary, open questions are:
1. How well regulated must the 24VDC be to work?
2. Where can we source that non-standard center-pin plug?
3. Is the ResMed A10 really so sensitive to EMI that it needs filters?

1. I'll test an A10 with my variable power supply. I'll post the results.

2. This cable has a built-in ferrite bead for EMI suppression: google ebay for "AC-DC-Power-Jack-Plug-with-Cord-7-4-mm-x-5-0-mm"

3. The A10, with its internal filtering and the ferrite beads on the approved ac power supply and dc converter, has passed an EMI emission test. The ferrite bead on the proposed cable may suppress the radiated or induced EMI sufficiently. Another suitable bead could be slipped over the cable for more suppression.

[verbatim1]

1. I'll test an A10 with my variable power supply. I'll post the results.

Thanks Steve, for volunteering to test that out.

It may be that the ResMed is picky about the 24Volts (as two fully charged batteries can be as high as around 28Volts DC); or it may be that the ResMed A10 is not finicky at all about the voltage.

If it is finicky, we'll need to add a 24VDC voltage regulator.

Also, what do you think we should use for the inline fuse size?
I'm assuming a nominal draw of less than about 3 amps, but a 5 amp fuse may be too low (perhaps due to metal fatigue over time)?

What do you think it should be fused at? 10 amp fuse?

2. This cable has a built-in ferrite bead for EMI suppression: google ebay for "AC-DC-Power-Jack-Plug-with-Cord-7-4-mm-x-5-0-mm"

That is a wonderful suggestion!
I like that it appears that both HP and Dell use a similar center-pin Barrel-&-Tip connector so that it will be easy to source.

I found the measurements somewhere, so, I'm not sure if the inner Tip ring is 5.0mm or 5.2mm. Somewhere I saw it was 5.2mm so I hope it's 5.0mm for the sake of compatibility with the HP & Dell connectors.

3. The A10, with its internal filtering and the ferrite beads on the approved ac power supply and dc converter, has passed an EMI emission test. The ferrite bead on the proposed cable may suppress the radiated or induced EMI sufficiently. Another suitable bead could be slipped over the cable for more suppression.

Googling, I see that they list "clip on ferrite core noise suppressors" in strange (to me) designations.

• Ferrite Core 1/4 Cord Noise Suppressor
  
• Ferrite Core 1/2 Cord Noise Suppressor
  
• Ferrite Core 3/8 Cord Noise Suppressor
  

What is the design decision we make to select the ferrite core?
Does it matter "where" the ferrite core goes?

PS: The circuit is getting larger and more complicated! (This is a good thing, because we're ironing out real-world details.)

[TiredInSoCal]

I don't know much about electronics, but would love to buy a cable or maybe a kit of the supplies + instructions from one of you when this is ready

[justMongo]

Googling, I see that they list "clip on ferrite core noise suppressors" in strange (to me) designations.

Ferrite Core 1/4 Cord Noise Suppressor
Ferrite Core 1/2 Cord Noise Suppressor
Ferrite Core 3/8 Cord Noise Suppressor


What is the design decision we make to select the ferrite core?
Does it matter "where" the ferrite core goes?

These are split down the middle and can be snapped around an existing wire or cable.
The ferrite choke goes close to the EMI source one wishes to suppress. It keep the distal wire/cable from acting as an antenna.

1/4, 1/2, 3/8 are the diameters of cable they will snap closed around.

[verbatim1]

I don't know much about electronics, but would love to buy a cable or maybe a kit of the supplies + instructions from one of you when this is ready

I don't think anyone here is considering "commercializing" this.

The goal, I think, is to list the parts that are needed, and to draw the diagram, and after that, it's up to you to solder them together property.

Have you ever soldered resistors, capacitors, and transistors together?

These are split down the middle and can be snapped around an existing wire or cable.
The ferrite choke goes close to the EMI source one wishes to suppress. It keep the distal wire/cable from acting as an antenna.

Thanks for that answer.

I'm really confused about why we need/want these ferrite beads.

I understand your words, which are that:
a) Any EMI generated by the 3.3 volt regulator with pull-up resistor can be "suppressed", and,
b) The few feet of wire acts as an antenna.

Both statements confuse me for slightly different reasons:
1. Does this circuit GENERATE EMI? I don't see where, or how. EMI means oscillation, and this is a DC circuit, so, where is the oscillation?

2. Even so, I do agree that a length of wire is an antenna to radio waves that happen upon it; but why do we care if the wire acts as an antenna?

Since the CPAP machine is basically a fancy fan, with a bit of electronics (probably far less than what it in a typical cellphone), why do we even CARE if the wire acts as an antenna?

The ResMed motor must generate more "noise" than anything, and it can't be "that" sensitive to AM/FM signals picked up by the wire on a 24VDC line, can it?

So, fundamentally, I would worry more about a lightning strike on the equipment than the EMI it generates or receives. But, that's just because I'm confused why we care about the puny amount of radio energy induced in the wires, or about the puny amount of oscillation in the sense circuit.

1/4, 1/2, 3/8 are the diameters of cable they will snap closed around.

Thanks for that explanation.
I appreciate your patience.

BTW, I would like to correctly attribute the diagram. Did both Mongo and Steve design and build the thing? Who designed it? Who built it?

I just want to get the attribution correct.


To add further reference value to this thread, here is an earlier version (apparently) of the ResMed patent application:

• Patent application title: POWER MANAGEMENT IN RESPIRATORY TREATMENT APPARATUS
  
• Patent application number: 20110162647
  
• Respiratory method or device means for mixing treating agent with respiratory gas control means responsive to condition other than user's airway pressure
  
• http://www.patentsencyclopedia.com/app/20110162647

A key paragraph, with respect to the circuit depicted in the OP, is:

For example, 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). A detection of 2.7K ohms may be interpreted as a 90 Watt power supply. A detection of 1.8K ohms may be interpreted as a 60 Watt power supply. A detection of 1.0K ohms may be interpreted as a 30 Watt power supply.

It's hard to summarize the patent because of the 97 separate lawyer-included claims, but what the patent does is allow fundamentally two "claims" that we actually care about when designing our own tent camping battery based power supply for the ResMed A10:

• The analog sense circuitry senses the pull-up resistor value
  • Hence, it "knows" the power supply's ability to source power
    
  • At least with the granularity of (a) Mains, (b) 90W, © 60W, and (d) 30W (at roughly 30Watts per 900K Ohms)
• The machine prioritizes gas flow over accessories (such as heaters and humidifiers)
  • For example, the controller synchronizes de-powering of the heater during an increase in a speed of the blower
    
  • Fundamentally, that means the accessories are powered during exhalation, and de-powered during inhalation

In addition, we get some insight into how ResMed may be organizing their power and communication bus inside the A10 when they claim:

In some embodiments, the system components may be coupled together for electrical communication by a common system bus...For example, the system bus system may be implemented with a power supply line (e.g., a 24 volt supply line) a communication line (e.g., a 3.3 volt VCC logic high level) and a ground line.

Fundamentally, that implies (but does not decree) that the 24VDC is used internally to power the components while the 3.3VDC reference voltage is used to indicate a "logic 1" on the communication line.

Where that matters greatly is if someone tries to connect "just" a 24VDC power supply to the ResMed A10, the communications line will have a "logic 0", which, almost certainly would prevent the machine from running.

So, our main goal is simply to provide two things to the ResMed A10 in our tent-camping power supply:

• A logic 1 on the communication line (across a 2.7K Ohm pull-up resistor)
  
• A well-regulated 24VDC on the system bus (e.g., two nominally 12VDC batteries in series)

[justMongo]

Actually that circuit should not generate EMI. It's a linear series regulator; not a switcher. We see them right close to the resmed power bricks because they are switchers. I've used the snap-on type to suppress RF on coax shields from entering my HAM shack; I've used them on telephone wiring to keep RF signals from interfering with phones. I have a junk box full of them.

I've just be kibitzing in the thread -- so, no credit needed for me.

Here's one on a telco interface coming into the house.

[verbatim1]

Actually that circuit should not generate EMI. It's a linear series regulator; not a switcher. We see them right close to the resmed power bricks because they are switchers.

Thanks Mongo for the answer about the EMI and for the photo of the cores in use.

I couldn't figure out WHERE any EMI could be coming from, and, even if it was there, how would it affect the ResMed machine, which is, essentially, just a fancy fan?

I guess EMI can affect the cellular modem, but, even so, the ResMed device itself, having a 3-phase brushless motor inside, certainly generates more noise than does this linear power supply.

I updated the graphic, and corrected the attribution to reflect that it's Steve's baby. We're helping out on the graphic making it easy for others to assimilate. I think the fact we drew something that "looks" like discrete components is nice in that a schematic would turn some people off (who don't know how to read them anyway).

On the fuse, it looks like the ResMed supplies max out at 3.75 amps, so, how does a 10 amp quick-blow fuse sound to you and to Steve?

Steve: Any update on the testing of the voltage?

The MOST IMPORTANT thing is the voltage on the ResMed power bus.
If we use batteries, they can fluctuate between, oh, I don't know, maybe 10 volts each to 15 volts each? That's 20 to 30 volts, which is ok for the regulator (just barely); but what about the ResMed device?

[justMongo]

Couple of points. Inside the Blower unit electronics -- you have a comparator. It's output should not connect to the power (central pin) like that. The power (central pin) connects to Vin. The comparator output supplies a logic signal that shuts down the blower if the wrong power supply is indicated. Best just remove the comparator circuit from the diagram.

If you really want to get fancy, you idiot proof where the batteries connect by using a 4 schottky diode bridge such that any way the batteries are connected; the positive side is always steered to the positive leg of the circuit (same for negative leg.)