|
Re: FDA Approval
Thanks Gordon! - I've amended my text in By the way it just occurred to me that if your flow sensor is bidirectional and was placed at the Y connect it would be able to sense both inspiration and expiratory flow. This would really enhance the safety by allowing the controller to detect a failure of many components (notably the pneumatic inspiration valve. Is this where you were planning on putting the flow sensor? Erich Schulz,?mbbs, mba, fanzca 0410 277 408 On Sun, 12 Apr 2020 at 12:02, Erich Schulz via <erichbschulz=[email protected]> wrote:
|
|
Re: High cycle Rainbird and passive valve testing
Hit 1,000,000 cycles this evening. The two Rainbirds are still opening and closing every 0.5 second.
I replaced the PANDApeep and my designed PEEP with two copies of the UF designed PEEP that I 3D printed here from their geometry files. They have 75,000 cycles on them. ????????? Bob Benedict?? KD8CGH. |
|
Re: P Sensor mounting
I used JB Weld KwikWELD to seal the pressure sensor wires on my valve test system. I just brought them through a hole in the cap.
Early in my career I had the task of trouble shooting a rubber injection molding machine that shot 44 lb. slugs of rubber to make truck tire retread rings. There were no temperature sensors in the rubber flow because "it can't be done". I fabbed some thermocouples on my kitchen table held together with JB weld and used them. They held at 25,000 psi and gave me the data I needed to solve the production problem. Eventually the carbon black loaded rubber flow eroded the thermocouple junction wires away, but the JB Weld never failed. ?????? Bob Benedict?? KD8CGH |
|
Re: #ExhalationValve
#ExhalationValve
My butyl tube valve clocked 350k today.? I pulled it apart to inspect it and there is not much noticeable damage.? Very similar to 200k inspection.??
I plan to test this to failure but I think the key thing I've learned is reducing the amount of lateral and circumferential stretch seemed to help the most.? I did this by adding a grate or cap at the end of the 1" PVC tube.? See 3rd picture.? This was 3d printed but could be manufactured with a drill.? It was just faster for me to print it.? I could still easily breath this the system with this installed.? I hope this helps. AJ |
|
Re: FINISHING: Capturing Baselines, and Adjusting Alarms
I agree that beginners may get intimidated to add features in the codebase without watching videos and asking question.
However, any intermediate embedded SW developer would understand and be able to make changes easily. About having many source files, that is the way professional SW code is designed and organized. Modular, well defined interface and abstraction. Yet that multi-file implementation uses less memory that the "simple" code. Its main loop takes 180 microseconds or less to be processed where the other takes almost a second. As the number of features increases (alarms, sensors etc)? the other code will become unreadable/unmanageable and unsafe. There will be "if... then... else.." hacks everywhere. it is already unsafe... press and hold a key and ventilation is gone, for example. Collaborating with a open ventilator project is not a competition for me. My intention is to help any team in need of a good code for running a Ventilator. I do not really care about competing with Ford, GM, or any other bad code written by hobbyists. Do whatever suits you. |
|
Re: FDA Approval
FIRST TRY at a complete physical implementation of the controller.
1.? I don't have wires yet for controlling the valves brought out -- CB mic connector may work and not expose users to any voltage at all. 2.? I don't have the air pressure lines brought out. 3.? I don't have a connection yet for the MEDICALLY QUALIFIED power supplies I have!? ?So I'm running off battery in this demo 4.? I glued on the nipple on a bmp280....hope to heck that I didn't get any glue on the bmp!!!!? ?The glue took wAY longer to solidify thaN I was Planning 5.? Hole is cut to allow us to switch to 20x4, but at this instant a bezel for 16x2 6.? Positioning is set to allow us to switch to Asher board if we get some made (don't have any assembled yet) Pushbuttons turn out to be more proof against waater than I expected because of their design and I added some dielectric grease to further reduce entrance. The electrical box is THICK and that turns out to cause some problems....but at least for the first time I don't have wires MOVING aall the time.. This will get better, but it is a start. Now i need the softare to get close to fnished! THANKS, folks!! Gordon |
|
Re: Transducer board // power supply
Gordon: Did you test the device listening address? There are a number of programs out there that can determine it for you. This is the first one I found:
On Saturday, April 11, 2020, 9:38:54 AM EDT, Gordon Gibby <docvacuumtubes@...> wrote:
For the first time, i have a working "transducer board" / power supply that can handle both AC and battery inputs. It isn't perfect, it is just all I have at the moment and a lot of blood sweat and tears went into getting even that. the photo is attached (you can see that I had some "issues" such as limited parts availability, unwillingness to risk parts, goofs, etc) The Schematic is attached -- and yes, you will have to pick the values off the photo of the board (will do better later) In the built board, I have only so far put in ONE of the two trickle charging parallel current sources.? ?No matter.?? TWO ISSUES: 1.? The Adafruit board works? (0x77) but I could not get the KeeYee BMP280 board to work.? Why?? I dunno.? ?Wrong bus voltage?? Dunnno.? SEND 3.3 VOLTS TO THE I2C BUS please -- the Ashar 0.1schematic/Stapleton 1.0 Board use 3.3 volt bus.? ? 2.? Dumb me, the AC Power green led stays dimly lit even when the battery is having to run the thing.? ?May have to put a resistor in parallel with the Green LED to get it to quit.? ?Will work on that later.?? This is of course optional.? ?I was instructed to add battery backup (I'm a volunteer, remember?) by the UF Team -- so I did.? ?Now it turns out tht the FDA will accept merely having a commercial UPS running the thing.? Duh!? ?We didn't know that until a day ago. No warrntees; no guarantees; use any of this information at your own risk. |
|
Re: FINISHING: Capturing Baselines, and Adjusting Alarms
All: I worried about the ability of people in lesser-developed countries having C programming expertise to take a project with 25 source code files and being able to make the changes necessary to get it working with the resource pool at their disposal. Who knows what will be used for the Rainbird valves in Kenya or what their parameters will be? As a result, I spent most of my time working on the code from Ashhar, mainly because of that concern. Also, Marcelo's code was very robust from the outset. I was even more hampered by not having any kind of test bed to exercise my code, instead relying on feeding the code to Ashhar and Gordon for testing. I used Marcelo's UI as a baseline, but worked with it on a 4x20 display. Long story short, I got all of the options Marcelo was using in his UI working with Ashhar's baseline code. After several long days of coding, I sent copies to Gordon and Ashhar for testing. Neither even acknowledged receiving it, let alone anything about exercising it. So, I've been following the posts here, and see both tacts making progress, so I'm assuming that my code has either been discarded or absorbed. Either way, it's not a problem, as all I've wanted the entire time was to have a body of code that any C programmer with a year or so experience could take it and get a vent working for them in short order. You guys have done that and I'm really pleased to see the project getting very close to FDA submission. I'll continue to watch the progress as the project marches through the FDA approval process. Gordon: kudos to your Herculean effort on keeping this project focused and on target. I only wish there was more that I could have done to ease your burden along the way. Jcak, W8TEE
On Saturday, April 11, 2020, 7:48:11 AM EDT, Marcelo Varanda via groups.io <mv_email@...> wrote:
What is the reason that Software Developers collaborating with this project are not working on the baseline in CSSALT GitHub? Please be free to openly criticize if there is something wrong there the same way I did so about the other Software. If we had all collaborated with that baseline, for example, had integrated the BMP280 sensor, we would have all these requirements being easily implemented. Thanks, MV |
Bob Benedict, KD8CGH
Marcelo Varanda
jjpurdum
