Endurance Test
What was tested, for how long, under which load, with which instruments, and what happened.
A VENDOR.Max laboratory prototype completed a continuous 532-hour run with an electronic load set to approximately 4.0 kW and kept continuously connected. Over that interval the operating regime did not collapse, no hardware failure was recorded, and the customer port was not disconnected by the output protection. The run was ended by the operators. What follows is the test record.
One uninterrupted interval under a constant connected load.
Laboratory and engineering operation across recorded sessions.
Electronic load kept continuously connected throughout the run.
The 16 A output protection did not disconnect the customer port at any point in the run.
Approximately 4.0 kW describes the electronic-load condition used for the endurance run. It is not a product rating, and it is not published here as an independently verified active-power value at the customer port.
Status: company-recorded operating observation. The complete Frame 0 balance has not yet been measured by an independent laboratory.
What this record establishes
This page records an engineering result, not an energy-source attribution. A physical prototype was operated under controlled laboratory conditions, challenged through several independent forms of loading and observation, and repeatedly demonstrated in operation. The complete energy-source attribution at the device boundary is a separate measurement task.
An explanation of the observation must account for the recorded facts together: the absence of a complete Frame 0 balance does not erase the observation, and the observation does not by itself identify the source.
What these measurements can and cannot answer
Engineering measurements answer whether the operating state exists, how it responds to adjustments and loads, and whether it remains stable. A complete energy-source attribution requires a different experiment: synchronized accounting of all flows across one chosen boundary over one common window.
Electrical voltage and current can be combined into a power or energy result only when their measurement passports are compatible: the same defined electrical port, synchronized in time, over the same observation window and with compatible quantity classes. Other channels are measured in their own natural variables. The values on this page come from different boundaries, different windows and different classes.
| Value | What it is | Boundary and time class | What it is not |
|---|---|---|---|
| 9 V | Voltage during the startup event | Startup port · event of about 10 s | An energy value; the continuing input of the 532-hour run |
| 24 V, 24 A | Internal quantities at one measurement point | Internal plane · separately reported mean values | A Frame 0 input power; a published synchronized same-port active-power measurement |
| up to 30 kV, up to 150 A | Amplitudes of the resonant state | Internal resonant plane · amplitude | Active power delivered to the customer |
| 450 V | State variable of a storage node | Post-bridge capacitor · DC level | A power value |
| approximately 4.0 kW | Electronic-load condition of the endurance run | Customer side · long duration | A product rating; a value derived from the startup voltage |
| 532 h | Elapsed operating interval | Endurance run · duration | An energy value by itself |
No denominator for an energy ratio is published on this page. The startup event, the internal operating quantities, the load condition and the duration belong to different boundaries and different observation windows, so they do not form the numerator and the denominator of one ratio. No energy-gain figure and no end-to-end efficiency can be calculated from the published values.
A complete calculation would require one common Frame 0 observation window: all incoming and outgoing boundary flows, including the startup-port integral if startup falls within that window, together with the change of all internal stored energy. Customer-port power, thermal flow and electromagnetic flow are each measured in the natural variables of their own channel. This record does not contain such a window.
What was controlled
Six controls of the test programme, stated before any result. This section reports the configuration.
A historical VENDOR.Max laboratory prototype. Not a production unit and not a pre-sale configuration.
This page describes the prototype on which the tests were performed and for which this record was captured. It does not describe the present state of the system. What VENDOR.Max is and where the programme stands now is set out in VENDOR System.
1. Startup-source boundary
The prototype existed in two startup configurations.
Revision A. An external battery is connected to the startup port for initiation only and is physically disconnected once the operating regime is established. In individual demonstrations this happened after roughly 10 to 15 seconds. During subsequent operation the source is physically absent from the circuit, not merely de-energised by logic.
Revision B. Startup is performed by the internal system; the battery remains in the loop as a bidirectional buffer. Its role is to close the temporary gap when the load changes faster than the system can move between operating points, not to serve as the source of steady-state customer power.
The revision is stated wherever the test record establishes it. The 532-hour run was performed in the Revision A configuration, with startup at 9 V for about 10 seconds.
2. Load-circuit configuration
The load was connected to the customer output as a consumer. On the load side there was no separate connection to the grid, to a battery, to a generator or to any other independent power source capable of continuously returning energy into the prototype.
A reactive or rotating load can briefly return part of the energy it has previously received. Such a return flow is not an independent energy source: it originates from energy previously delivered to that same load.
Single connection to the test load.
No grid connection. No external battery. No external generator.
If present, not an independent source.
3. Output path and customer-port limit
The historical prototype used a household AC output path. The customer output was implemented as a protected 220 V / 16 A interface. That applied to the output port and its electrical implementation, not to the internal class of the installation.
At those nominal values, the interface corresponds to an apparent-power scale of 3.52 kVA.
16 A is the nominal current of the protected interface, not a hard instantaneous trip threshold. Protection behaviour depends on the characteristics of the specific protective device and its operating conditions.
Delivering 5 kW as a nominal rating through one such household interface was not part of the design.
5 kW described the development class of the historical prototype. 220 V / 16 A described the specific household AC implementation of its customer port. These are different characteristics.
The approximately 4.0 kW electronic-load setting was selected as a deliberately demanding long-duration condition for this particular AC implementation, while higher loads approaching the 5 kW development class were used only in separate short-duration peak tests. The continuous run therefore examined both long-term retention of the operating regime and the behaviour of the inverter and 16 A output protection under sustained loading.
The historical test configuration is recorded as 220 V / 16 A. This describes the nominal convention and the interface used at that test location; it is not a statement that the instantaneous port voltage remained exactly 220 V throughout operation.
The physical core, the specific component implementation and the customer interface are three different levels of description, and a characteristic of one is not a characteristic of another.
4. Calorimetric cross-check of the load
Electrical observations were not the only way the load response was assessed. A separate series used water calorimetry: water quantity, temperature rise and elapsed time were measured directly.
The figures are given in the results section.
5. Operating-state monitoring
Throughout the test programme the operating state was monitored instrumentally at several measurement planes.
Observations at different planes belong to different physical quantities, different charge ensembles and different temporal classes. Key points of the operating state are recorded in the results section, each with the class of quantity it belongs to. Their interpretation and the map of local increments belong to Where Is the Plus.
6. Internal electromagnetic screening
The test programme included internal observations of the electromagnetic environment, among them monitoring of the radiated high-frequency field with a HyperLOG 7040 measurement antenna, and point field measurements.
This is engineering screening, not a certification test for electromagnetic compatibility. Numerical readings are not published: individual point readings are not used to claim conformity with standards.
Instruments
| Equipment | Role |
|---|---|
| HyperLOG 7040 measurement antenna | Observation of the radiated high-frequency field |
| SOEKS Quantum dosimeter | Point check of the environment |
| MEGEON electromagnetic field meter | Point check of the environment |
| ITECH IT8626 electronic load | Controlled long-duration loading for the 532-hour endurance run |
| Scales | Mass of the complete prototype |
| Measuring vessel, thermometer, stopwatch | Calorimetry |
| Resistive heater 2.4 kW; inductive load 800 W | Real load appliances |
The test programme
Seven records. Each answers a different question.
Continuous endurance run
Revision A · 532 h · electronic load set to approximately 4.0 kW and kept continuously connected. Long-term retention of the operating regime and the behaviour of the output path under that load condition.
Cumulative engineering sessions
All configurations · 1,000+ h. Repeated starts and stops, varied and peak load levels, protective disconnections, instrumental working sessions.
Calorimetry of the load
Revision A · five repeats. Water-heating result obtained by a method that does not require high-frequency wattmetry.
Combined dynamic load
Revision B · 20 minutes. Behaviour with an inductive component and repeated inrush peaks.
Destructive testing
Early builds. Deliberate loading of the power path to failure in order to locate the limiting element.
Filmed working demonstration under load
Revision B · video record. Operation with real household AC appliances.
Off-site demonstrations
Revision A and Revision B · several locations outside the laboratory. Repeated starts and operation under load in both startup configurations.
Results
The test programme used several independent forms of loading and observation. They answer different questions and are not combined into one calculated energy balance. The continuous endurance run used an electronic load at approximately the 4.0 kW test level; separate calorimetric tests established a thermal result on a resistive appliance; dynamic and peak-load sessions tested transient behaviour; off-site demonstrations tested repeatability of the observable operating sequence outside the laboratory.
Continuous 532-hour run
| Parameter | Record |
|---|---|
| Continuous interval | 532 h |
| Configuration | Revision A |
| Load condition | Electronic load, approximately 4.0 kW, continuously connected |
| Startup | 9 V for about 10 s, then the startup source physically disconnected |
| Collapse of the operating regime | None |
| Hardware failure | None |
| Disconnection of the customer port by the 16 A output protection | Not recorded across the interval |
| Termination | Controlled, by decision of the operators |
The load configuration remained constantly connected. There were no peak regimes within the run.
Recorded operating time
| Reading | Meaning in the test record |
|---|---|
| 532 h 23 min 11 s | Reading associated by the test record with the continuous run |
| 1,000 h 00 min 06 s | Total operating time, as labelled by the interface |
The figures 532 h and 1,000+ h used on this page are roundings of these readings.
Recorded key points of the operating state
| Point | Value | Quantity passport |
|---|---|---|
| Startup | 9 V, about 10 s | Event value; the startup source is then physically disconnected |
| Input of the resonant part | 24 V, 24 A | Separately reported mean internal values at this measurement point; not published as a synchronized same-port active-power measurement and not a Frame 0 input power |
| Resonant part | up to 30 kV, up to 150 A | Amplitude values of the resonant state; they are not a synchronized active-power measurement, and their numerical product is not customer power |
| Capacitor after the bridge | 450 V | DC voltage level |
| Customer load | approximately 4.0 kW | Electronic-load condition of this run; not a product rating and not derived from the nominal 220 V / 16 A interface values |
| Duration | 532 h continuous; 1,000+ h cumulative | Prototype runtime counter |
These values belong to different boundaries, different charge ensembles and different temporal classes. They are not multiplied by one another, not divided by one another, and do not form an end-to-end energy result. Internal voltages and currents describe circulation, not active power at the customer port.
The 532-hour run was performed with an electronic load set to approximately 4.0 kW. The duration is not multiplied by that load level on this page to derive delivered electrical energy: the figure is published as the documented load condition of the test, not as a measured continuous active-power value at the customer port.
Regime and control frequencies, window and burst durations, pulse counts, quality factors, coupling coefficients, ratios between the points and the distribution of the bus are not published.
Calorimetry
| Parameter | Value |
|---|---|
| Load | Resistive heater, nameplate 2.4 kW |
| Water quantity | 1.0 L |
| Temperature | 20 to 100 °C |
| Elapsed time | 240 s |
| Repeats | 5 |
| Water-heating energy per run | 334.9 kJ · calculated from measured inputs |
| Total across five repeats | 1.67 MJ · calculated |
| Mean water-heating power | 1.40 kW · calculated |
For the energy calculation 1.0 L of water is taken as approximately 1.0 kg.
The calculation uses the water quantity, the temperature rise, the elapsed time and the standard specific heat capacity of water. It does not use the phase of the electrical signal, RMS values, duty cycle or power factor.
The value of 1.40 kW is a water-only thermal result. Heating of the vessel, heat transfer to the surroundings and the energy of evaporation are not included in the calculation. It is therefore not an estimate of the full electrical power of the heater.
Combined load
The operating regime remained stable for 20 minutes under combined resistive and inductive loads with repeated peaks. There was no collapse of the regime.
More than 1,000 hours does not represent a single operating regime
532 hours is a continuous endurance run. 1,000+ hours is cumulative operation across recorded engineering sessions.
Those sessions include repeated starts and stops, varied and peak load levels, combined loads, protective disconnections and instrumental working observations.
The filmed working demonstration and the off-site demonstrations form a separate, broader body of evidence. They are not counted separately towards the 1,000+ hour figure on this page.
A protective trip and a system failure are different events
| Term | Meaning |
|---|---|
| Protective trip | Intended disconnection of the customer port on reaching a protective condition |
| System failure | Unintended loss of the operating state, or a hardware failure |
Output-protection trips occurred during peak tests in other sessions and were not treated as failures of the installation. In the 532-hour run the customer port was not disconnected by the output protection.
Separate peak-load sessions reached load levels approaching the 5 kW development class. Short-duration operation at such levels did not invariably cause an immediate protection trip; protective trips were also recorded in other peak-test conditions. These sessions were separate from the 532-hour continuous run.
Filmed working demonstration under load
The record shows the sequence: startup, establishment of the regime, connection of household loads, operation, shutdown.
The recording shows operation with a resistive load of 2.4 kW nameplate and an inductive load of 800 W nameplate. Nameplate values of load appliances are not measured transferred energy.
Off-site demonstrations
The prototype was repeatedly started and demonstrated outside the laboratory in both configurations.
Revision A. An external battery was connected to the startup port for the start. Once the regime was established — after roughly 10 to 15 seconds — the battery was physically disconnected from the startup port. The installation then ran in demonstration mode with a load connected.
Revision B. Startup was performed by the internal system. During off-site demonstrations the installation was repeatedly switched off and restarted from the main switch, after which the same observable sequence followed: startup, establishment of the regime, connection of load, operation, shutdown.
Across different locations the observable startup and operating sequence was reproduced repeatedly.
Boundary of this record
Frame 0 is the product and test control surface drawn along the device enclosure. Frame 1 is the engineering control boundary of the working core. These are different control volumes with different flows crossing them, so an internal Frame 1 flow cannot be labelled as the external input of Frame 0.
In the Revision A configuration the external startup source crosses Frame 0 only during startup and is then physically disconnected. After that there is no architecturally provided continuing external power port. This is a topological statement, not a statement that every possible external energy channel is absent: a complete Frame 0 balance requires those channels to be inventoried and measured together with the change in internal stores.
Internal voltages and currents belong to different temporal classes and different charge ensembles. Multiplying them arithmetically, or comparing them in sequence, does not form an end-to-end energy balance. Without a correct distinction between planes, temporal classes and the external boundary, identical voltage and current values admit physically different interpretations.
The site therefore keeps two languages apart: the charge model is used for the local mechanism and the organisation of motion, while the energy balance is used for quantitative accounting at an explicitly chosen boundary.
This page publishes the conditions and results actually observed. Working planes and local changes are covered by Where Is the Plus. The rules of quantitative energy accounting are covered by Energy Model. The complete Frame 0 balance — inventory of all channels crossing the product control surface, measurement of the customer output, and tracking of the change in internal stores — belongs to Technology Validation.
How the 532-hour run proceeded
The causal sequence of the run, from a de-energised state to a controlled shutdown.
- T0
- Installation de-energised.
- T1
- External startup source connected: 9 V for about 10 s.
- T2
- Operating regime established.
- T3
- External startup source physically disconnected.
- T4
- Regime maintained.
- T5
- Electronic load set to approximately 4.0 kW connected.
- T6
- Continuous operation with that electronic-load condition — 532 hours.
- T7
- Load removed, run ended by the operators.
The run was ended by the operators after 532 hours. It was stopped because it required continuous human supervision of an operating experimental prototype, not because of a failure of the installation or a protective trip.
Peak regimes and protective trips are not part of this run and belong to other sessions of the programme.
Physical record
What can be seen directly.
The tested AC implementation and the failures that occurred
Which part of the configuration the run actually stressed, and which limiting elements were identified in earlier work.
The 532-hour run related to one specific historical AC implementation of the output path: an inverter and a protected 220 V / 16 A customer port. Over that interval the inverter did not fail and the protection did not disconnect the port.
The result relates only to the tested historical AC implementation and does not automatically transfer to other output configurations. How the functional route is arranged, and which output configurations exist, is set out in How It Works.
The operating state was developed experimentally. Component values, switching conditions and control settings were iterated against measured behaviour until a repeatable operating regime was obtained. This record describes what that physical implementation then did; it does not reconstruct the result from a prior end-to-end energy model.
During development, early builds were deliberately driven to destruction. Three limiting failures were recorded in sequence: a wound assembly, a ferrite magnetic core and output-stage power transistors.
After each failure was addressed, the next element of the power path became the limiting one. These results were used in the subsequent revision of the prototype.
What is recorded and what is not established
The record supports statements about system operability. It does not substitute for independent verification.
Recorded
- the prototype was built and operated, and the record of that operation was captured;
- the operating regime was retained over a long interval with a constantly connected load configuration, and was not a short-lived effect;
- over the 532-hour run no hardware failure of the AC output path was recorded, and the customer port was not disconnected by the protection;
- the operating regime was reproduced in repeated sessions;
- outside the laboratory, repeated start and operating cycles were demonstrated in both revisions of the prototype: in Revision A the external startup battery was physically disconnected once the regime was established, and Revision B was restarted from the main switch using the internal startup system;
- in peak tests of other sessions, trips of the output protection were observed; in the continuous 532-hour run with a constantly connected load configuration the protection did not disconnect the port;
- three limiting failures were established in real destructive tests;
- in the calorimetric series the input quantities were measured and the water-heating thermal result was calculated; the method does not require high-frequency wattmetry.
Not established
- the energy balance at the complete device boundary;
- end-to-end efficiency of the complete device;
- compliance with safety and electromagnetic-compatibility requirements;
- formal environmental qualification across defined ranges of temperature, humidity, vibration, dust, ingress protection and electromagnetic environment;
- characteristics of a production unit.
The next layer of measurement
| Recorded | Next layer of measurement |
|---|---|
| Revision and unit | Synchronous complete Frame 0 balance |
| Startup procedure; disconnection of the startup source | Integral battery balance, where a battery is present in the revision |
| Connected load and the way its power is stated | Complete thermal flow |
| Duration of operation | Electromagnetic flow across the boundary |
| Retention of the described regime; key points of the operating state | Independent laboratory and an uncertainty budget |
An unresolved energy balance does not invalidate an observed operating result; it limits what may be inferred from that result. The engineering observation is not a theory of the energy source.
The record was made by the company; the complete-boundary balance has not yet been measured by an independent laboratory — that is the next layer.
This is an internal laboratory record, not independent certification. Independent verification is the next evidence layer: Technology Validation.
