HEG_ESP32 icon indicating copy to clipboard operation
HEG_ESP32 copied to clipboard

status of Josh's HEG projects in September 2023

Open wjcroft opened this issue 10 months ago • 10 comments

Josh, hi.

Saw your comment on the BrainsAtPlay Discord, that V2 was a disappointment due to faulty components.

Where does that leave owners of HEGduino V1. Is that still viable? Was rather sad to learn that your WIX website disappeared:

https://hegalpha.com/

Did find this page on Wayback:

https://web.archive.org/web/20210122150409/https://www.hegalpha.com/forum/

But that was years ago.

Regards, William

wjcroft avatar Sep 06 '23 21:09 wjcroft

The site is archived here: https://brewsterjoshua1.wixsite.com/website

Otherwise I am meandering through to a new device that I can cheaply scale to as many 24 bit channels as I want and I have a couple SBIR grants I applied for. The last of the V2 devices went out through crowd supply finally. They'll remain supported in my software going forward but the new projects I am working on are taking a while.

On Wed, Sep 6, 2023 at 2:37 PM William J Croft @.***> wrote:

Josh, hi.

Saw your comment on the BrainsAtPlay Discord, that V2 was a failure due to faulty components.

Where does that leave owners of HEGduino V1. Is that still viable? Was rather disappointed to learn that your WIX website disappeared:

https://hegalpha.com/

Did find this page on Wayback:

https://web.archive.org/web/20210122150409/https://www.hegalpha.com/forum/

But that was years ago.

Regards, William

— Reply to this email directly, view it on GitHub https://github.com/joshbrew/HEG_ESP32/issues/5, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEK2PHY2ENMLXUCCFYOXOZDXZDUJ5ANCNFSM6AAAAAA4N6P2PE . You are receiving this because you are subscribed to this thread.Message ID: @.***>

joshbrew avatar Sep 07 '23 03:09 joshbrew

Josh, thanks much for that link to the archived site.

Not sure I understand WHY the site is not live, so users can support each other via your Forum? This would be a kind thing to offer as the user base are now your current 'experts'. Are the monthly fees for Wix just too high?

Back to the original question at the top of this 'Issue':

Can you point me to HOW-TO documentation links for V1 that describe recommended steps to install latest firmware and use the device for training? Same for V2?

Are those the steps I see in the README?

https://github.com/joshbrew/HEG_ESP32

Since the dates are between 2 and 3 years old at this point, are they still working and current?

So this is V2,

https://github.com/joshbrew/HEG_ESP32_Delobotomizer

Do any of the webapps developed for V2 work for V1?

Thanks,

wjcroft avatar Sep 07 '23 03:09 wjcroft

This repository is for the V1. The Delobotomizer repository is for the V2. The drivers are the same otherwise relative to any software. Everything worked last time I checked.

Easiest way to install drivers (for either device) is:

Download this repository.

Install the Arduino IDE.

Install the Espressif ESP32 libraries through the Arduino IDE library manager.

Download and unzip the dependencies that go with the drivers into the documents/Arduino/libraries folder (make it if not there)

Select the Lolin32 or Huzzah32 via the Tools > Boards Manager in the ESP32 section.

Change the partition to "Large Apps with minimal SPIFFS" (paraphrasing), and the clock to 80MHz (Lolin32 only afaik)

Compile and upload. It is pretty slow.

And yes I mainly don't want or need to be paying the wix fees right now as I don't have time to keep the site up to date and both designs are fairly deprecated. I've not sold more V1 kits myself in years and they're really unreliable, especially compared to what I'm working with now. The remaining V2 kits were back orders. Otherwise I'm gonna be mostly under the radar while I work on my newer projects.

On Wed, Sep 6, 2023 at 8:20 PM William J Croft @.***> wrote:

Josh, thanks much for that link to the archived site.

Not sure I understand WHY the site is not live, so users can support each other via your Forum? This would be a kind thing to offer as the user base are now your current 'experts'. Are the monthly fees for Wix just too high?

Back to the original question at the top of this 'Issue':

Can you point me to HOW-TO documentation links for V1 that describe recommended steps to install latest firmware and use the device for training? Same for V2?

Are those the steps I see in the README?

https://github.com/joshbrew/HEG_ESP32

Since the dates are between 2 and 3 years old at this point, are they still working and current?

So this is V2,

Do any of the webapps developed for V2 work for V1?

Thanks,

— Reply to this email directly, view it on GitHub https://github.com/joshbrew/HEG_ESP32/issues/5#issuecomment-1709418088, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEK2PHY2GGCGCRVTJMON333XZE4RVANCNFSM6AAAAAA4N6P2PE . You are receiving this because you commented.Message ID: @.***>

joshbrew avatar Sep 07 '23 05:09 joshbrew

Josh, thanks. Replied to your brewster.joshua1 email address. I assume you got that?

I opened up the CrowdSupply HEGduino V1 bag (preassembled) that I received in March 2020, and it works fine with the https://hegalomania.netlify.app/ webapp over the usb serial port on my Mac here.

Regards, William

wjcroft avatar Sep 08 '23 17:09 wjcroft

Josh,

[oops, ok realizing now I should have just used the software/DataCharter present on the Github, nicely done! Still would be helpful to see trends across CSVs.]

Assuming I save the CSV recording file for each session, is there any recommended way to chart / track the session to session progress over say days or weeks? I guess the obvious thing would be to read the CSV into a spreadsheet, then average the ratio column and see how grows over time.

Here is the 30 minute session graph of ratio. It gently rises over the first 20 minutes or so, then gently declines, I imagine from 'fatigue' of the biofeedback process. So 20 minute sessions may be preferable to 30 min.

Also @joshbrew wondering if you might comment on the 'noise' level present in the trend line. I suspect this is normal, but kind of surprising that it bounces around this much. Is this reflecting actual physiology, or is it inherent in the signal to noise ratio achievable with this ADC and sensor? I'm fine with the noise, I can visibly see the trend increasing and will be able to track the centerline of that trend session to session.

ratio1

Best regards, William

wjcroft avatar Sep 08 '23 18:09 wjcroft

Here is my second session, making good improvement, 40 minutes long. Not much fatigue, even at end. I have the impression now that the live webapp is smoothing the data somewhat for presentation of the ratio etc. That smoothing is not present here in the raw CSV.

ratio2

wjcroft avatar Sep 08 '23 23:09 wjcroft

Also @joshbrew wondering if you might comment on the 'noise' level present in the trend line. I suspect this is normal, but kind of surprising that it bounces around this much. Is this reflecting actual physiology, or is it inherent in the signal to noise ratio achievable with this ADC and sensor? I'm fine with the noise, I can visibly see the trend increasing and will be able to track the centerline of that trend session to session.

OK, I think I'm now getting the impression that the variations of the 'ratio' signal are naturally caused by the blood volume / pulse variations, the raw PPG photoplethysmogram signal, which is also the basis for optical heart rate and SPo2 devices.

wjcroft avatar Sep 12 '23 19:09 wjcroft

@wjcroft a lot of this is being addressed in my new work as this was pretty much all a bandaid software solution based on my inexperience. Have had a lot of hitches in my own life and so far finding experienced collaborators has been impossible and/or I can't afford salaries so it's just slow going. It'll probably be another year at this rate till I can publish a proper app for all this since it'll all be my own work. I am currently learning wavelet filtering and classification to turn this up another notch on top of my new 24 bit + multimodal hardware, oh and MBLL proper as the ratio is a relative metric of total blood O2 volume (as opposed to SpO2). The hardware will cost about the same to produce as the originals too, ironically.

The SNR of the HEGduino V1 is also very bad. You should be able to improve it significantly by adding a 200pf capacitor between V and GND on the photodiode as a simple lowpass filter. It is enough to elucidate a trend and that is it, but yes the CSVs in this version are not capturing the smoothed line as I put it before the filter pass for some reason.

joshbrew avatar Sep 12 '23 19:09 joshbrew

Josh, hey thanks. I left a comment for you over on Discord / Brains at Play.

Here is another question, is it possible to save an OFFLINE version of the V1 webapp? https://hegalomania.netlify.app/ Tried doing that by just telling Chrome to just Save As, "Webpage Complete", but that is not functional. The online version of the webapp works fine. I'm using the usb serial port interface to keep things simple. Thanks,

wjcroft avatar Sep 12 '23 20:09 wjcroft

@wjcroft it should be a PWA

Capture

joshbrew avatar Sep 12 '23 20:09 joshbrew