On this page: Architecture · Modules · Licensing · Deployment · Jobs and exports · Instrument interfaces · Calculation engine · Security and compliance · Sizing and network · FAQ · Check your instruments
System architecture and platform
Deployment model
| Architecture | Client-server, zero-footprint deployment |
| Workstation install | None. Executables run from a shared folder through a standard LAN shortcut — no local registry entries, no DLL installation |
| Network protocol | TCP/IP over LAN or WAN |
| Server environment | Windows Server: on-premises, virtualised or private cloud |
| Remote access | Citrix and Microsoft RDS supported |
Database tier
| Engine | Microsoft SQL Server 2016 to 2022, Standard or Enterprise |
| Connectivity | ODBC |
| Data integrity | Relational integrity enforced at engine level; transaction logging configured for point-in-time recovery |
| Audit trail | Field-level and immutable: record inserted or modified, user, timestamp, previous value, new value |
| Multi-database | Several databases can be selected at login |
Execution engine
| Core runtime | Native C++ core with a C#/.NET application layer (DevExpress) |
| Processing model | Multi-threaded, deterministic execution |
| Calculations | Synchronous background processing for heavy analytical mathematics |
Modules
All modules share one Microsoft SQL Server database and the same zero-install deployment. Management modules are licensed per site with unlimited users; instrument modules are licensed per instrument.
| OnLAB | Central data backbone: reception and registration, barcode generation, work orders, client metadata, preparation control (pulverising, crushing, split weights, sieve tests), result entry with range checks and detection limits, multi-level sign-off and certificate distribution in PDF, CSV or Excel |
| OnWSH | Online Worksheet — the electronic laboratory notebook and Level 0 instrument capture. Layout templates group samples with QA/QC standards, blanks and duplicates in sequence, and instrument output maps straight onto worksheet positions |
| OnQC | Statistical quality control (formerly OnLQC): real-time analysis of certified reference materials, Shewhart, X-Control, CUSUM and Thompson-Howarth control charts, bias against certified value, and alarms (2SD warning, 3SD rejection) that block sign-off when QA/QC fails |
| LabData | Read-only portal for clients and stakeholders: results, crosstab views, export and analysis requests, without giving access to the laboratory applications |
| OnConfig | The control centre: system configuration, reference tables, test libraries, price lists, roles and audit logs — so the laboratory configures the LIMS without external support |
| OnQMS | Accreditation evidence: ISO/IEC 17025 clause-by-clause coverage with drill-down to the real records, proficiency-testing control charts, country packs, and an optional local AI assistant whose data never leaves your network. Fully bilingual EN/ES |
Instrument interfaces (Level 0)
Physical connectivity
| Serial | RS-232 and RS-485, directly or through hardware converters and Moxa device servers |
| Network | TCP/IP and UDP sockets |
| File-based | Watched folders (polling) for CSV, TXT, XML and proprietary instrument export formats |
| Coverage | 150+ instruments integrated over serial, plus 25+ configurable file parsers. New instruments are added by configuration, not by recompiling |
| Bidirectional operation | Supported on many instruments: OnLIMS writes a text file with the sample IDs for the run — and, where the instrument needs it, the sample weights, as with ICP — the instrument software reads it, and OnLIMS then captures the instrument's own output file. Which instruments accept a sequence file depends on their control software — send us your instrument list and we will confirm yours |
What we can connect
The rule is simple: if the instrument can send data through a serial or USB port, or write a file, OnLIMS can capture it. Thirty years of connecting laboratory instruments means most of what a mining, smelter or environmental laboratory runs is already in the driver library — and an instrument released next year is added by configuration, not by recompiling. If a driver does not exist yet, we write it as part of the project.
| Already integrated | 150+ instruments over serial, plus 25+ configurable file parsers |
| Instrument classes in production | Atomic absorption, ICP-OES and ICP-MS, XRF and XRD, carbon/sulfur and mercury analysers, titrators, UV-Vis, fire assay balances with auto-tare, and analytical balances |
| Examples of makes in service | Agilent, Shimadzu, Thermo, PerkinElmer, Spectro, Malvern Panalytical, Bruker, Rigaku, LECO, Mettler Toledo, Sartorius and Ohaus, among others — the list is not a limit |
| Not on the list? | Send us your instrument list and we will tell you what we read today and what would need a parser |
All trademarks are the property of their respective owners.
Licensing model
| Management applications | OnLAB, LabData and OnConfig are licensed per site, with no restriction on the number of users |
| Instrument applications | OnWSH and OnQC are licensed per instrument, whatever the instrument is. A laboratory with two balances, one LECO and two AA needs five OnWSH licences; if the LECO is later replaced by an XRF it is still five. Chemists may run OnWSH and OnQC on any machine in the laboratory without additional licences |
| Support and maintenance | The first year after installation is included. After that, an annual agreement at 18% of the total licence cost covers software updates, instrument drivers and parsers, and small fixes. Enhancements are quoted before any work starts |
Deployment footprint
A single shared folder holds the applications and their components; it can sit anywhere on the LAN and access is normally controlled through Active Directory.
Bin (read) | Applications, DLLs and INI files |
Help | Help files and general documentation |
Inst | Instrument data: drivers and instrument files |
RepoForm | Configuration scripts for Certificate of Analysis forms |
Sql | Optional, for the daily database backup |
Sent | Default LIMS export folder |
Temp | Temporary data written by the applications |
Job types and exports
| Batch jobs | A fixed number of samples, the usual shape of a geology job with drill core |
| Routine jobs | An endless job created once: new samples are added by date or another unique identifier. Designed for fixed sampling points feeding a historical series, with Min, Max and Std limits per point for QC checks on production data |
| Sample login | OnWSH templates, job sample masks and macros register tens or hundreds of samples in one command — no scripting required |
| Export formats | Generic: TXT, CSV, XLSX, database views and stored procedures. Custom exports are built to the client’s specification, up to database-to-database transfers |
| Labels and certificates | Labels print from OnWSH, OnLAB and LabData; label layout and Certificate of Analysis formats are configured through INI files |
Data processing and calculation engine
| Engine | Proprietary mathematical interpreter built natively in C# — no Excel runtime dependency |
| Formulas | User-configurable algebraic equations referencing sample, dilution, moisture, tare, gross and CRM values |
| Unit conversion | Automatic (for example ppm to g/t, % to ppm) |
| Detection limits | Dynamic lower and upper limit management (for example <0.01 or >10000) |
| Significant figures | Global and element-specific rounding rules |
| Dilution factors | Automated back-calculation of the original concentration from dilution volume and aliquot weight |
Security, access control and compliance
| Authentication | Native database credentials or Windows Active Directory single sign-on |
| Authorisation | Role-based access control with fine-grained permissions: read-only, result entry, worksheet creation, QC validation, final release |
| Result lifecycle | Enforced Pending → Ready → Approved → Published. On publication results are password-sealed under supervisor control and every later change is captured in the audit trail |
| ISO/IEC 17025 | Built-in controls that support ISO/IEC 17025 workflows: document control, traceability, reagent tracking, instrument calibration schedules and audit logging. Accreditation is granted to the laboratory, not to the software |
| 21 CFR Part 11 | Controls that support Part 11 workflows: user-attributable electronic sign-off, system logs and audit records. Part 11 compliance is assessed on the validated system as deployed, not on the software alone |
Sizing, network and environment
| Database growth | The unit that drives growth is the result, not the sample: about 0.5 MB per 1,000 results including the full audit trail. Measured across production databases with 13 and 14 years of history, one reporting 26 results per sample and the other 187 |
| Growth per year | Depends on how many results the laboratory produces, not on how many samples it receives. Measured: 1.5 GB per year in a laboratory producing about 3.2 million results a year, and 4.1 GB per year in one producing about 7.3 million |
| Transaction log | Sized separately from the data file and driven by the recovery model and backup schedule, not by laboratory volume |
| Latency — direct client | Under 10 ms recommended |
| Latency — remote | Up to approximately 150 ms via Citrix or Microsoft RDS |
| Bandwidth | Low footprint during active sessions thanks to optimised SQL execution |
| Instrument workstations | One PC per connected instrument, each with an available COM port for serial instruments |
Frequently asked questions
Which versions of Microsoft SQL Server does OnLIMS support?
SQL Server 2016 through 2022, Standard or Enterprise edition, on-premises, virtualised or on a private cloud Windows Server.
Does OnLIMS need to be installed on each workstation?
No. Workstations run the application from a shared folder through a standard LAN shortcut, with no local registry entries and no DLL installation.
How much database storage does a laboratory need?
Plan on about 0.5 MB per 1,000 results including the full audit trail, which came to 1.5 GB a year in one production laboratory we measured and 4.1 GB in another, depending on how many results each one produces. The unit that drives growth is the result rather than the sample: two production databases we measured hold a similar volume of data, even though one reports 26 results per sample and the other 187. The transaction log is separate and depends on your backup policy.
How much network latency does OnLIMS tolerate?
Under 10 ms is recommended for a direct client connection. For remote sites the architecture supports Citrix and Microsoft RDS with latency up to around 150 ms.
How does OnLIMS connect to laboratory instruments?
Through serial RS-232 and RS-485 (directly or via Moxa device servers), TCP/IP and UDP sockets, and watched folders that pick up CSV, TXT, XML and proprietary exports. New instruments are added by configuration, not by recompiling.