Aman Singhal
← All work

MedTech · Freelance

Jan Elaaj —
Biosignal Platform

A handheld device that takes photoplethysmography and electrocardiography from the same sensor, on the same board, at the same time — then runs long enough in real clinics to find out whether the readings hold up.

Role
Mixed-signal hardware architecture
Engagement
Freelance
Core silicon
MAX86150
Board revisions
Three
Clinical pilot
10 months, New Delhi

The brief

Most low-cost health screening devices measure one signal well. Jan Elaaj needed two — PPG for the optical pulse waveform and ECG for the electrical activity of the heart — because a number of the health parameters the product wanted to derive only fall out when you have both, time-aligned against each other.

I owned the hardware architecture: choosing the acquisition path, designing the boards, and getting the analog front end quiet enough that the signals were worth processing at all.

Needs your detail

Who was this for, and what was the product actually trying to do for them? Screening camps, clinics, home use? That framing decides how the rest of the page reads.

Why the MAX86150

The MAX86150 integrates the optical PPG front end and an ECG channel in a single package. Putting both on one device removes an entire class of problem: the two signals share a clock and a reference, so their timing relationship is fixed by the silicon instead of having to be reconstructed in software across two separate subsystems.

Needs your detail

What else was on the table before you landed on it — discrete AFE, MAX30102 plus a separate ECG front end, something else? Naming the alternative and why it lost is the part that shows the engineering.

Three boards

The design went through three PCB iterations. Mixed-signal layout is where this kind of product is won or lost — the optical path wants current, the ECG path wants microvolt-level quiet, and they sit centimetres apart.

Needs your detail — this is the most valuable section

What changed between rev 1, rev 2 and rev 3? Every revision exists because something did not work. Grounding, supply noise, electrode placement, mechanical fit, something failing only on real skin? Even one sentence per revision is enough, and it is the part other engineers will actually read.

The electrodes

The device uses custom gold-plated dry electrodes. Dry electrodes avoid the gel that makes clinical ECG impractical outside a clinic, at the cost of a higher and less stable skin-electrode impedance. Gold plating is the usual answer to the corrosion and polarisation that would otherwise show up as baseline drift over a long recording.

Needs your detail

Why custom rather than off-the-shelf? What did the geometry have to solve — contact pressure, hand position, repeatability between users? And did plating thickness or base metal end up mattering?

Ten months in the field

The platform ran a 10-month clinical pilot across NABL-accredited labs and hospitals in New Delhi, deriving 40 health parameters. A pilot that length stops being a demo — it surfaces the drift, the outliers and the handling problems that a bench test never will.

3
PCB revisions
40
health parameters derived
10
months of clinical pilot
Needs your detail

What did the pilot change? Anything that worked on the bench and failed on real patients? Was there a validation reference you measured against, and how close did you get? If any of this is under NDA, say so and I will write around it.

What I would carry forward

Needs your detail

Two or three sentences in your own voice: what this taught you that you now apply elsewhere. This is usually the section people remember, and it cannot be written by anyone but you.