Summary
R1 (1kΩ) limits the TL431 reference current budget to ~2.5mA (as labeled on the schematic). This isn't enough to excite a standard 350Ω full-bridge load cell — the most common bridge resistance for the low-cost cells this board is marketed for.
Problem
With a 350Ω bridge connected, the excitation voltage on E+/E- collapses from the nominal 2.48V down to well under 1V under load, since a 350Ω bridge needs ~7mA at 2.5V while R1=1kΩ only leaves a fraction of that available. Since REFIN shares the same node, this also pulls the CS1237's reference below its datasheet minimum, resulting in a saturated/stuck ADC reading on the analog channel (the on-chip temperature channel, which doesn't depend on this reference, reads fine — confirming the issue is external to the ADC).
Suggested fix
Reducing R1 to roughly 68–100Ω raises the available current enough for a 350Ω bridge while keeping TL431 dissipation negligible — a single-component change.
Might be worth a note in the README about this, since the limitation isn't obvious from the schematic without doing the current-budget math against your specific load cell's resistance, and this reference design seems to be shared across several similar CS1237 boards from different vendors.
Happy to share measurements/scope captures if useful. Thanks for maintaining this repo!
Summary
R1 (1kΩ) limits the TL431 reference current budget to ~2.5mA (as labeled on the schematic). This isn't enough to excite a standard 350Ω full-bridge load cell — the most common bridge resistance for the low-cost cells this board is marketed for.
Problem
With a 350Ω bridge connected, the excitation voltage on E+/E- collapses from the nominal 2.48V down to well under 1V under load, since a 350Ω bridge needs ~7mA at 2.5V while R1=1kΩ only leaves a fraction of that available. Since REFIN shares the same node, this also pulls the CS1237's reference below its datasheet minimum, resulting in a saturated/stuck ADC reading on the analog channel (the on-chip temperature channel, which doesn't depend on this reference, reads fine — confirming the issue is external to the ADC).
Suggested fix
Reducing R1 to roughly 68–100Ω raises the available current enough for a 350Ω bridge while keeping TL431 dissipation negligible — a single-component change.
Might be worth a note in the README about this, since the limitation isn't obvious from the schematic without doing the current-budget math against your specific load cell's resistance, and this reference design seems to be shared across several similar CS1237 boards from different vendors.
Happy to share measurements/scope captures if useful. Thanks for maintaining this repo!