My battery backup setup

[surferdude2]

Thanks for that info sdb7802, Interesting that they would locate all the components of the detection system inside the power supply. Corporate decision or necessity? Flies in the face of the patent data but perhaps they have revised that?

Since you have done the research, we don't need to guess any further. We have all we need to know to defeat it, if need be.

Dude

ps. That is not to say that defeating it won't bring forth a surprise that we may not like, since reverse engineering can often lead to disappointment.

[Marthajoy7]

Thank You so much. That means it is not a sense line like used on regulated power supplies. Just something to keep most people from using other sources of power, Trying to figure out why the regulator vs voltage dropping resistor. Could it be the regulator circuit uses less power to drop the voltage than a 47K ohm resistor uses to burn off the excess voltage? It would be in the ma range but maybe better use of battery power?
Or maybe the circuits are built so cheaply in china, that why not.
As far as voltage in the CPAP they would be the same, 1.8v across the same internal resistor. which means current flow from supply to machine on the pin line is identical. Only difference I could see is that with the regulator, the current on the pin remains the same, where as with the resistor the current drops slightly as the battery drains. Maybe I am agreeing now that the regulator is better. Battery voltage can vary from 27v to about 22v. So that is about a 20% variance. That means the Pin current can also vary by that amount. It must want that current to be constant no matter what the battery condition. Maybe I need to order some parts.

[PoolQ]

completely consistent with the patent. This is a voltage divider, the low side resistor is in the CPAP and has a fixed value and there is a voltage detection circuit, both of these MUST be in the CPAP in order to control the CPAP operation.

Now this detection circuit is most likely an IC powered either by 5 VDC or 3.3VDC, again most likely 3.3VDC. This means that the voltage it needs to detect must always be below 3.3 VDC.

If you used the 24VDC to generate this voltage, you would need a very high value resistor in the power supply and a very low value in the CPAP. This make getting all the different voltages that represent the different power supplies quite difficult because the resistor values available as standard at the high end have large gaps between them ie 47K and 56K and this makes a very small change in voltage across the low value resistor in the CPAP (it has to be a low value to keep the resulting voltage below 3.3VDC)

if you use 3 volts it gets easier as a single step change in the resistor causes a significant change in voltage

[surferdude2]

We are still left wondering how that makes the system any better or how it protects anything or in fact how it does anything at all to justify its existence. A fixed value regulated d.c. voltage can't convey anything other than on and off unless they superimpose some a.c. signal on it and clip it off at the cpap end.

We still don't know what we don't know and probably never will until we can see a schematic of the entire system.

[Perchas]

My tests on three A10 units shows they require at least 22.3V to power up. Once you start the blower, it will run until you lower the voltage to 20.5V.

Post a high res pic and I will try to identify the component. Which machine?

Both you and ve7ltd have performed tests without modifying the 24 V, have reached the same conclusions, notice the bolded 1.8 V, that is the key, the resistor in series with 24 V delivered to the Zener diode those 1.8V, it enables power supply.

So my insistence on trying powered with 12 V and resistance put half the value, instead of 47 k, 20 to 25 K.

The component maybe these by N or M SMD Zener? , I could not follow the printed circuit is four layers.

I did all the probing with the factory AC adapter, and determined what the resistance was between the 3.3V volt supply and the center pin. By grounding the pin, I determined that the resistance is right about 2.5 k ohms for the adapter I have.

I then plugged it in and watched the voltage when the machine booted up, and recorded the voltage to determine what the voltage it was looking for on boot. It was right around 1.8V (I will have to probe again for the two values). The machine seems to have a fixed resistance, and then switches on another resistor to test the adapter, then goes back to the fixed, and then starts some sort of data stream.

I first tried to build a voltage divider to lower the voltage on the pin. I could never get it to work quite right. I then just hooked up a 22K resistor in series, and used a 47K variable to find the range at which the machine would boot with all its features. I found that using a single 47K between 24V and the center pin yielded exactly what I needed - about 1.8V on the startup.

PRS One SE has implemented the central pin, I fed with a battery cable of a Dreamstation with the resistor 20k, testing a battery of 12000 mAmp that reached me today I managed to run for 9 hours 8 cmH2OInspiration at 0.360 mAmp (4W ) with a supply voltage of 5.9 VDC when the battery was almost empty and stop 5.5 V. Respironics are 12,0 Vdc

This test confirms that example's implementation respironics with central pin works with 55% less volts

A test picture, zoom, (sorry for the empty cup of coffee, provides support for battery)

[Marthajoy7]

I just had an idea. I went to ebay and found this cheaply.

Waterproof-DC-7-50V-To-5V-2A-USB-Female-Step-Down-Converter-Regulator-For-Car-QT

By using this I can also charge phones ect. It still can provide signal to pin with a different resistor. If I can remember ohms law.
A 2700 ohm drops voltage 1.5 volts - Current is 0.56 ma
To drop 5v to 1.8 would be 3.2 so the resistor should be in the vicinity of 5.8 k or computes as 5760.
5.6K or 5.9K would be standard values. If needed could use 2 resistors in series to get exact. How critical is the resistor value? 1% 20%
using same figures for 24 v says resistor should be 40k instead of 47k. But at 25.2 volts it would be 45.3k
Current values might not be quite right, just using figures provided.
lets try again 24 volts to 1.8v - 22.2 volts 222/470000 = 0.4723404 ma
With that current for 3.2v drop would be 6774 ohms or about 6.8k

[sdb7802]

My tests on three A10 units shows they require at least 22.3V to power up. Once you start the blower, it will run until you lower the voltage to 20.5V.

Just to clarify: These tests were done by feeding the center pin with +3.3V thru a 2.7k series resistor (as Resmed does).

[Marthajoy7]

Yes, I understand. Just question accuracy of readings as it seems the measurements were made in circuit. Capacitors and output impedance could affect measurement of resistance value. 5 volts should work just as well and will serve to allow recharging of usb devices. Once the current is known for sure, The resistance can be figured. The current and voltage at the pin will be exactly the same as using a 3.3v converter. The voltage will remain constant regardless of 24 volt batteries. For less than $3 I cannot buy parts to build your converter. I save money, works just as well, and allows another use as well. I have no qualms about just using a resistor, except that the current would change at the pin, based on battery condition. If it were a sense line, that would be preferential method. Since you say Resmed is not treating it as a sense line, then why should I, at least until schematic becomes available. It appears that the pin is going to a comparitor which is being used to hold high or low on a inhibit pin on their cpu. Half a ma is not going to burn out any electronics. That is what we are talking about as being sent through the pin on the connector.

[sdb7802]

Welcome to the forum: ve7ltd, Marthajoy7, and Perchas

I found that using a single 47K between 24V and the center pin yielded exactly what I needed - about 1.8V on the startup.

I have run this setup on two machines, and used it for 6 nights hooked to a 250aH 12V battery in my trailer. Worked fine every time.

Great find, ve7ltd!

I will test the 47k ohm resistor scheme over the allowable input voltage range of the A10 on my three units.


Martha, I will test your 5V + resistor scheme on my A10s.

I'll let you know the outcome.

BTW: here's an old scope trace of one of my A10s on power up:

   

I ran the data stream (not shown) through a protocol analyzer. Couldn't really make much sense of the data.

[OMyMyOHellYes]

I will keep using my PR System 1s that don't need no stinkin' third line! Just a plain battery that I know how to run.

OMMOHY