¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Wireless Connectivity to Ventilators via Android - Problems??


 

The UF CSSALT OS-Vent team is considering interfacing with an Android UI. What are some expected problems with wireless connectivity?


 

That just doesn't seem like a good idea since external devices like that are pretty easily hacked. Why do they want it?

Jack, W8TEE

On Thursday, April 16, 2020, 2:03:31 PM EDT, kevinahao@... <kevinahao@...> wrote:


The UF CSSALT OS-Vent team is considering interfacing with an Android UI. What are some expected problems with wireless connectivity?


 

I kind of was initially.
I have implemented BLE for several other projects and I was initially considering to use it to control our ventilator.
However, it is not very practical. Imagine 10~20 ventilators in the same hospital area. How to track which phone is connected to which ventilator.
What if a connected phone gets missing but still holding connection? The only way out to connect to another phone is by reseting the ventilator to force disconnection. Not very practical. Also, the Ventilator needs to have its own MPU to peek going even when the connection drops.

After some consideration I believe that we could provide some connectionless data via advertising package. It could provide some status and dynamic data for rendering some graphics. This is just an idea¡­ need to check if there is AT commands (on HM-10 module) is possible to dynamically change the Advertising (or beacon) packages. I do not have time to explore this further; Would require extra $$$ and I do not see much benefit.


 

¿ªÔÆÌåÓý

I am writing a serial 485 out of all data and showing on a big screen via RS485 for seeing the charts etc in testing and demos. Your code works great. I had to manage everything in cmH20 but you have done that recently too. I do not like the BLE idea either. For rapid deployments that can cause issues. I am using a different sensor as we do not have the fancy ones here. BMP280 + a cars MAF sensor + plus tomorrow will change solenoid to car DBW throttle for flow control as we do not have sprinklers for that sort. Thanks for a good job, I like your coding with the right abstractions. I think the sensor code was a bit over the top for making it generic. Best just make each return cmH20 and leave the raw out as too specific for each sensor.

?

From: [email protected] <[email protected]> On Behalf Of Marcelo Varanda via groups.io
Sent: Thursday, 16 April 2020 11:41 pm
To: [email protected]
Subject: Re: [VentilatorDevelopers] Wireless Connectivity to Ventilators via Android - Problems??

?

I kind of was initially.
I have implemented BLE for several other projects and I was initially considering to use it to control our ventilator.
However, it is not very practical. Imagine 10~20 ventilators in the same hospital area. How to track which phone is connected to which ventilator.
What if a connected phone gets missing but still holding connection? The only way out to connect to another phone is by reseting the ventilator to force disconnection. Not very practical. Also, the Ventilator needs to have its own MPU to peek going even when the connection drops.

After some consideration I believe that we could provide some connectionless data via advertising package. It could provide some status and dynamic data for rendering some graphics. This is just an idea¡­ need to check if there is AT commands (on HM-10 module) is possible to dynamically change the Advertising (or beacon) packages. I do not have time to explore this further; Would require extra $$$ and I do not see much benefit.


 

When did you clone the repo? A functional BMP integration was committed last night (in ET Canada). This morning I changed a bit BMP280 library to only compute pressure as we do not need the other values. Pressure unit is hardcoded to Pa to save some code space in the library. BTW... there is no library... but the cpp code brought from the library source.
And yes, there are room for improvement. I do not like floating point... I prefer to work in1000 times the target unit. Replace floats is in my todo list.


 

I have cloned since more than a week but take regular merges from upstream so I know the evolution of it. Yes, I have now removed my quick BMP280 implementation to use yours. I have car MAF still. It was working but regular merge resolution was an issue so now that will not happen. I have an additional valve and will use the stepper for managing flow control value (from car DBW). I think that will be minimum changes. I have yet to try out the pressure sensor and Flow control as busy with other substitute hardware (proportional valves not found), differential sensor not available here either.


 

Cool. Flow sensor has been left away for while as we are waiting for more clarifications. Now that pressure.cpp file is not being used despite being called by main.
Simulator is probably broken as well (in case you are using Qt).


 

I think we need to stay on track. The original goal was to build emergency ventilators to rapidly deploy in the pandemic to save lives.? Android connectivity does not accomplish that in the short term. It is an add on for data collection, and I understand the importance of data collection.? We have a USB port on the unit. If we need to collect data, let's interface through the USB instead of RS485.? Of course the FTTDI chip is simply a USB to UART translator. RS485 is pretty easy and good to use over long distances and for all practical purposes can be considered a differential RS232 line with the TX and RX tied together.? You can just as easily use another FTDDI chip to take the USB stream to RS485 but do it as a separate project.? On my board I have a three pin connector where a USB interface could be mounted as a daughterboard, so it can be done inside the box.? On the other hand cell phones are cheap. .Interface the cell phone to the USB port and then enable the wifi and use that to collect your data at a central point.? The cell phone doesn't need to be activated so a disposable 20 buck burner phone should do. Then we just need a basic streaming output of a data set from the UART to the FTDI chip which will USB to the phone. That would be easy to program.? It also would not compromise the integrity of the firmware if it is only transmitting.? Later on down the road you can inplement a command interpreter for remote control.?
73
Mike N8WFF


 

I actually designed my first prototype on a nodeMCU with some basic inputs and reports from the device to my phone over wifi, however, due to concerns that the devices could be hacked, I scrapped that plan. I don't know enough about network security to implement such a system in life critical hardware.


 

I've worked for NYPD and NCPD as a programmer?