I went through the KiCad files carefully. Thanks for posting the actual project rather than just screenshots — that makes a proper review possible.
Overall, the LTZ1000A core closely follows the standard datasheet circuit and looks sensible. The LT1013, 120 Ω resistor, 70 kΩ pair, 13.5 kΩ/1 kΩ temperature divider, and external heater transistor are all consistent with the intended topology. The main challenges are likely to be the resistor technology, thermal design, power architecture, and measurement setup rather than the basic LTZ circuit itself.
1. Heaters and fuse. Four 9.6 Ω heaters connected to a fully charged 16.8 V battery could draw about 1.75 A each, or roughly 7 A total, while dissipating close to 118 W. That is not compatible with a 2 A fuse.
I would first estimate the actual heat loss of the enclosure and determine how much heater power is really required. Firmware can limit the number of active zones, but it should not be the only protection against overcurrent.
The heater-current return paths should also be kept completely separate from Reference Low and the precision analog ground.
2. Precision resistors. The output stability will depend heavily on the resistor ratios, especially in the 10 V gain stage and the 1 V divider. Ordinary 100 ppm/°C resistors could easily dominate the drift and largely cancel the benefit of using an LTZ1000A.
Foil resistors or matched low-drift networks would be appropriate for the critical ratios. The LTZ loop resistors should also be selected for low temperature coefficient, long-term drift, voltage coefficient, and power coefficient.
3. Trim range and trimmer influence
The RV1 adjustment range may be adequate only if the 7.138 V value shown in the schematic is the measured output of the actual LTZ device. It will not cover the full device-to-device output spread of an LTZ1000A.
A better approach is to measure the actual LTZ output first, choose fixed resistors for the coarse gain, and leave the trimmer with only a small calibration range.
The same applies to RV2. Its contribution to the 1 V divider is relatively large, so cermet wiper drift will directly affect the output. Most of the ratio should come from fixed precision resistors, with the trimmer used only for fine adjustment.
4. 100 Ω output resistors All three 100 Ω output resistors are outside the feedback loops, so they introduce load-dependent error.
With a 10 MΩ DMM input, the loading error is about 10 ppm. With a 1 MΩ load, it becomes about 100 ppm.
Possible solutions are to sense the voltage after the resistor, provide separate Force and Sense terminals, reduce the resistor value, or explicitly specify that the outputs are intended only for high-impedance loads.
5. Buck converter and LT3045
The buck-plus-LT3045 architecture can work, but it needs to be checked over the full battery-voltage range.
The LMR33630 is a step-down converter only. As the battery discharges, the buck stage will eventually lose regulation, followed by the LT3045. There may be a range where the circuit is still enabled but the nominal 12.4 V analog rail is no longer regulated.
The battery UVLO threshold, buck output voltage, ripple margin, and LT3045 dropout should therefore be calculated together.
The LT3045 can strongly attenuate conducted ripple, but it cannot remove magnetic coupling from the inductor, switching-node electric fields, shared-ground impedance, or thermal gradients. Making the power section detachable is therefore a good design choice.
6. Checks before layout Before starting the PCB, verify every imported or custom symbol against the manufacturer’s datasheet:
* symbol pin number;
* pin function;
* footprint pad number;
* physical package pin.
Then run ERC and resolve all genuine warnings.
The heater MOSFETs should also have gate resistors and gate-to-source pull-downs so that they remain off during Pico startup and reset. Ideally, use MOSFETs with RDS(on) guaranteed at 2.5–3.3 V gate drive.
The Pico is well suited to heater control, temperature logging, battery monitoring, and display functions. Its internal ADC, however, is not suitable for ppm-level characterization of an LTZ1000A. That will require an external precision ADC, a null-comparison method, or a stable 6½- to 8½-digit DMM.
Overall: the project has a solid foundation and looks achievable. Before layout, I would recalculate the heater and power sections, define the precision resistor technology, reduce the influence of the trimmers, and specify the intended output loading. With an LTZ1000A design, the PCB layout, thermal symmetry, and resistor ratios will determine whether the result is merely functional or genuinely stable.
Home
Help
Search
Login
Register
