Your original question was too vague to address. If you have already written code, and have specific questions about it, then please follow the forum guidelines and post it, using code tags.
1 Like
Hye there, i really like with what s are going on with your project. Can you share what could be with your update there?
1 Like
So here is my update. I went through research papers to understand how pulse oximeter works and what parameters are taken into account as i am using a pulse oximetry sensor. Then i have compared the formula to what a Near infrared spectroscopy uses, bear in mind that i am no student in medical science of some sort but i have used what i could understand . In the end of comparing formulas i went into the pulse oximeter codes and found out that is takes samples of infrared reflected by the blood. It takes a minimum and a maximum and averages the data, a bit similar to how Near infrared spectroscopy used. Finally i took the formula from the code took the average and parsed it through another formular called beer lambert's law which resulted in a somehow working stuff and on top it gave me pretty accurate reading for being an amateur invention. To sum it all i am not some expert in medical science, to some this may seem useless stuff but for me it was a challenge which i took and as i mentioned is it reliable : no , does it work : yes . Can this be improved : certainly. If one day i get more support i shall work with other pros to make a legit reliable non invasive glucometer.
If you believe it works, then apply for the nobel prize.
When you are at revolutionizing science you could go and patent the one-lead-ecg - also nobel prize worthy.
could you share the coding? i didnt get it? how do you insert the formula?
Indeed. It should be noted that Apple and others are trying to measure blood glucose from their watches. They've been working on this for the better part of a decade, and have probably spent hundreds of millions of dollars on the efforts of hundreds of engineers, so far without success. So if you can make this work, you might want to hire an agent to negotiate for you.
Can please share the code of glucometer with max30102. I need it badly
Hi, your project seems so interesting!, I'm working with something similar but as of being new here in the forum as well I don't seem to find a way we can communicate directly to share ideas better and clearly
Hello noobieG69, could you please email me whenever you can
nadiazoubir10@gmail.com
1 Like
hello I am very interested in the project you are developing, and this also happens to be my thesis project at university.
Would you mind sharing the steps of how you worked on the project, creating formulas, coding the program, etc.?
I would really appreciate it if you could share.
Thank you, success always 
1 Like
Hello falah27,
I don't think that you will receive an answer from noobieG69.
If you look at his profile at forum.arduino.cc/u/noobieg69/summary, then you will see that he hasn't been on the forum since December 2023.
1 Like
okee thank you for your information
1 Like
Hello all . Really sorry the delay . Was really busy. I'll try to answer all your questions at once and I'll freely share the code in the upcoming weeks .
So to make it short what i did was learn the algorithm for the Pulse oximeter calculation. What is does and how it is being calculated . As i said I'm no doctor but in thinking if the pulse oximeter is being able to interact with blood and make calculation so by adapting the algorithm i can get li to estimate blood sugar as blood sugar in itself causes some variation in blood cells . Worked my way around the code ,fitted it the beer Lambert spectroscopy calculations alongside some miscellaneous bits and voila the stuff kinda work . So for all needing this project and the documentation just send me an email list if possible I'll forward all of you the project. As i said I'm no engineer nor doctor . I tried and worked for me . The only thing i ask is for a credit in your thesis and projects would be nice .
And if in the near future i have enough funds I'll carry forward in making a sensor powerful enough to be able to have accurate reading.
1 Like
The problem of Glucometer with NIR is serious. The water part of blood itself has many spectraö maximums in NIR.
So you have to find an area where detection of glucose make sense. This are around only 4 spectral areas from which 2 only are actual usable because of availability of cheap emitter and detector. This is the area around 1605 nm and the area around 1706 nm.
You need a minimum of two spectral maxima here at least for calculating concentration.
Here you can try - first with in vitro solution and than in reality. But you have to see that skin components and other parts in blood are complicating the problem. So you will face serial problems here.
I had also make some experiments here in that field with a raspberry pi pico and got some positive results which need by far a further development.
Lambert Beers law you discussed here is so far the right solution but parameter determination works not perfect actually.
With test solution in water and with light through absorption measure it seems to work in acceptable limits.
1 Like
The problem is common. You are not using a single wavelength with max30102 but a wide area - a spectra instead.
The lambert beer law works only perfect with a emitter of a single wavelength. You need a source of small wavelength - for instance a laser or a led with small emission wide and in a wavelength frame where water has low absorption but glucose has a high one.
The NIR spectra of water and the spectra of glucose have a lot of maxima - near together. I actually make a similar project. If you want to measure glucose concentration you have to concentrate here on special wavelength where water absorption is low and glucose absorption is high.
The max30102 here is not so good for that task, because the single receiver diode in that sensor has a wide sensitivity from 660nm for the installed red led to 880 nm for the installed infrared led - used and optimized so far for blood oxygen detection.
A much better area and choice here is a special sensor - one with 1605 nm and one with 1706 nm. Every one with small emitter wide and alsio a optimized photo receiver for that range.
Therefor I have developed in my project a sensor pcb which I use for 1706 nm and for 1605 nm as well. There are other values here over 2000nm but the receiver and transmitter diodes are better available in the 1600 - 1700 nm range.
There is a project also where people get acceptual measurment accuracy with the max30100 - but they resolved this task with a special software.
I also use the max30100 for puls and oxygen alone.
1 Like
HI Nikolai, can you pls send me the code as I would like to understand were you are at with your latest version ? Thank you.
Can you share your mail ?
//jf
Hello guys sorry for the delay. I've uploaded the code to git repo .
All the best with your endeavors and i hope this will help you. i know this project is out of my knowledge bound but it somehow work enough to present for a live demo. I'll be happy to hear from you if you've made any progress in this project.
Thank you.
This quackery is based on IRvalue being greater than 50000 or less than 50000 (while ignoring 50000) and a button?
Yes and the Github has no hardware listing.
No schematic.
The max30105 looks in the wrong part of the IR spectrum to do a proper analysis.
Who wrote it then?
Tom.... 
