Adaptavist Inc. designs the instruments that keep machines level when the ground will not stay still, working from its instrument hall at 12110 Sunset Hills Rd Ste 600, Reston - 20190-5916, United States (US). Our engineers build gimbal stabilisation platforms, inertial sensor fusion stacks and calibration loops that hold their reference through heat, vibration and long unattended duty. Every programme begins with a measured baseline and ends with a signed alignment report that a client engineer can reproduce. See the full practice scope and how each service is delivered →
Adaptavist Inc. opens its winter alignment cycle, reserving bench time in the Reston instrument hall for platforms that must be re-zeroed before the colder months change their mechanical behaviour.
Most control problems that reach our desk are not software problems. They are reference problems. A platform drifts because two sensors disagree and nobody has taught the system which one to trust under load. A camera mount swims because its gimbal ring is stiff at the wrong point in its travel. A test rig produces beautiful charts that no one can defend in a design review because the rig was never calibrated against a traceable standard.
Adaptavist Inc. works on the reference itself. We start by measuring the mechanical behaviour of the structure, then the thermal behaviour, then the way the sensors behave when both act at once. Only after that do we write control law, and we write it so that a maintenance technician can diagnose it a year later with a handheld tool and a short checklist. The result is a system that stays aligned because the alignment is a property of the design, not a lucky state that someone once observed.
The practice is deliberately narrow. We do not build general purpose machinery, we do not resell hardware catalogues, and we do not take on work where the reference cannot be measured. Within that narrow band we can carry a programme from first sketch to field acceptance. Our six service lines form one continuous chain, and clients often enter at any point along it.
Each disc below represents a discipline we own end to end. They are not packages to be bought separately; they are the stations a programme passes through. Read the plate under each disc for the short version, then follow the link into the detailed practice page.
Two and three axis gimbal assemblies that keep a payload level while the frame beneath them pitches and rolls. We size the bearing stack, the ring stiffness and the drive torque together so that the platform has no dead band at the end of its travel.
Gimbal platforms →Firmware that combines gyroscope, accelerometer, magnetometer and external aiding into one attitude estimate. We tune the filter against recorded field data rather than a clean bench, so the output survives vibration and magnetic disturbance.
Sensor fusion →Closed loop procedures that bring a platform back to a known reference and prove it with residuals. Each loop is documented with tolerances, drift budgets and a repeatability figure so that two different technicians get the same answer.
Calibration loops →Actuators, drives, safety interlocks and supervisory software joined into one coherent machine. We take responsibility for the interface between vendors, which is usually where integrated systems quietly fail after commissioning.
System integration →Purpose built rigs that reproduce the motion, temperature and load a platform will meet in service. We design the rig, the fixture and the data acquisition together so that a test tells you something about the product rather than about the rig.
Test rigs →Engineers who travel to the installed platform, re-establish its reference on site and leave a signed record behind. The service covers transport induced drift, foundation settling and the slow changes that a laboratory can never reveal.
Field alignment →When a client brings us a platform, we ask the same questions in the same order. What is the reference, how is it traced, what moves it, and what would a technician find at the end of the warranty period. The answers shape the whole programme, from bearing choice to the wording of the maintenance card.
A reference that cannot be written down cannot be defended. Before any design work begins we identify the physical standard the platform will be measured against, confirm its traceability, and state the uncertainty it carries. Platforms built without this step tend to pass acceptance and then drift away from each other on site, because each one was aligned to a different idea of level.
Gimbal rings fail at the joint, not in the middle of the tube. We map where compliance lives in a structure and add material only at those points, which keeps mass and inertia low. A platform that is stiff everywhere is usually heavy everywhere, and heavy rings respond badly to the small, quick disturbances a vehicle produces.
The best calibration loop is one a tired technician can perform correctly at the end of a long shift. We write the loop, the fixture and the software together, reduce the number of live decisions to a minimum, and make every step produce a visible residual. Boring calibration is repeatable calibration.
A platform is only aligned once it has been aligned where it lives. Transport, foundations and climate all move the reference, and no amount of laboratory work can predict the exact amount. Our field alignment service exists because the laboratory answers a different question than the installation does.
Adaptavist Inc. operates in computer systems design and related services, and most of our clients arrive from adjacent technical fields. The common thread is a machine whose behaviour matters more than its appearance, and a team that needs the behaviour to be provable.
Survey crews rely on platforms that were aligned separately and then compared on a hillside. We design the reference structure, the re-levelling procedure and the record keeping that lets two instruments reach the same answer on different days. The work overlaps with field alignment and with precision calibration loops.
Laboratories need benches that stay quiet over hours of exposure. Motion at the micron level is enough to spoil a long integration, so the design conversation is about damping, thermal expansion and the way a sensor is mounted, not about the sensor itself. Sensor fusion helps here when multiple instruments must share one attitude frame.
Industrial robots and automated cells need a stable base frame and a control loop that tolerates the compliance of a real floor. We integrate drives, safety interlocks and supervisory logic, and we test the cell under the load profile it will actually see. Control system integration is the usual entry point for these clients.
Some engagements are reviews rather than builds. A client has a design that should work and does not, and the team needs an independent read on where the reference budget was overspent. We take these reviews on a short cycle and deliver them as written findings that can be circulated inside the client organisation.
A typical engagement starts with a one day site or bench survey, where we record the structure, the existing sensors and the way the machine is used. Within two weeks the client receives a short written baseline with the measured behaviour, the reference options and the budget in mass, power and cost. From there the programme moves through design, build, calibration and acceptance, with a written decision at each gate.
We keep the client engineering team inside the loop rather than behind it. Reviews are working sessions with drawings and data on the table, not presentations with conclusions already fixed. That habit is the reason our acceptance records survive audits years after the project has closed, and it is the reason clients return with the second and third platform rather than rebuilding the relationship from the start.
The same discipline applies to the paperwork. Every alignment report carries the date, the instrument used, the reference standard, the ambient conditions and the residual figure. When asked why a platform behaves as it does, we can point to a page rather than to a memory.