Ayon Spirit V EN
Ayon Spirit V — Architecture, circuit, measurements and technical assessment
Analysis of a KT150/KT170 push-pull integrated amplifier based on an internal inspection, a power test and external documentation. Special attention to DMP, the power supply, the transformers and the limits of the evidence.
Sources: supplied technical video, interior photograph and manufacturer documentation · Updated 18 November 2026
What is analysed: Ayon Spirit V, KT150/KT170 architecture, DMP control, power supply and output power.
What we know: ≈60.5 W/8 Ω at 1 kHz near clipping in the KT150 unit inspected.
What we do not know: THD+N versus power, output impedance and conduction class at full power.
Who it is for: readers of technical audio and owners of loudspeakers with variable electrical load.
Conclusion: an interesting design, but without enough data to claim metrological superiority.
0. Executive summary

The Ayon Audio Spirit V is an Austrian valve integrated amplifier with two KT150 per channel, 12AX7/6SN7 drivers, output transformers with ultralinear taps as described in the technical inspection, fixed bias with automatic control and DMP damping adaptation. The technical inspection of a KT150 unit provides internal data that do not appear in the commercial specification sheet.
MEASURED-EXT: the technician reports approximately 22 V RMS into an 8 Ω resistive load at 1 kHz, close to visible clipping: 60.5 W calculated. He does not report THD+N or continuous power with both channels driven.
Relevant finding: the examined unit uses an adjustable global feedback path taken from the 4 Ω output, even with DMP in OFF according to the technician's description. This requires qualifying the commercial «0 dB» specification.
Main uncertainty: measurements are missing for the Spirit V of output impedance, damping factor, THD+N versus power in both modes (up to 40 W in triode and 65 W in pentode mode, per the declared power), IMD and behaviour with reactive loads.
1. Sources, scope and method
1.1. Primary source: inspection and tests
The basis for the internal analysis is the video «Ayon Spirit V — Review, analysis, technical details and measurements», roughly 37 minutes long. The work used its Spanish transcript, which contains automatic-recognition errors (for example «ION» for Ayon, «vías» for bias). Technical denominations are normalised only when the context is unambiguous.
The video shows a repaired KT150 unit; it must not be assumed that its circuit or components are identical to those of every KT170 revision.
1.2. Classification of evidence
The same evidence categories as the Magnat MA 900 report are used. CONFIRMED identifies data appearing in official documentation, but does not mean that its performance has been independently measured. MEASURED-EXT is reserved for third-party instrumental measurements: the video reports a voltage close to 22 V RMS into 8 Ω at 1 kHz, near clipping, without a full distortion curve. REPORTED-EXT identifies information communicated by third parties without a verifiable instrumental measurement. Technical explanations are separated from facts through PROBABLE, HYPOTHESIS and ND.
1.3. Complementary sources
Ayon's commercial specification sheet and reviews from HiFi.nl, Soundrebels, HiFi-IFAs and Alpha Audio. Listening impressions are kept separate from instrumental results. Links are collected in the sources annex.
2. Specifications and revisions
| Parameter | Value | Source / notes |
|---|---|---|
| Topology | Stereo push-pull | CONFIRMED; inspection |
| Output valves | 4 × KT150 or KT170 | CONFIRMED; the video unit uses KT150 |
| Small-signal valves | 2 × 12AX7 + 2 × 6SN7 | OBSERVED in the video |
| Modes | Triode / «pentode» (sheet) · ultralinear taps ≈43 % (video) | CONFIRMED (commercial naming) / REPORTED-EXT (taps); see note |
| Power, general sheet | 65 W pentode mode (commercial naming) / 40 W triode per channel | CONFIRMED; no THD specified |
| Power, 2023 KT150 list | 70 W pentode mode / 45 W triode | Historical commercial document; not standardised without conditions |
| Power, 2023 KT170 list | 75 W pentode mode / 50 W triode | Historical commercial document |
| Nominal output impedance | 4 Ω and 8 Ω taps | CONFIRMED; not equivalent to Zout |
| Input sensitivity | 500 mV RMS | CONFIRMED; sensitivity for maximum power per the sheet |
| Input impedance | 100 kΩ at 1 kHz | CONFIRMED |
| Declared response | 10 Hz–60 kHz (−3 dB) | CONFIRMED; power condition ND |
| Feedback | «0 dB» declared / DMP observable | Documentary discrepancy; see §6 |
| Dimensions and weight | 480 × 370 × 250 mm; ≈33 kg | CONFIRMED |
2.1. What changes with KT170
Ayon described a power-supply and driver upgrade to accommodate KT170. The historical figures of 70 vs. 75 W represent only 0.30 dB of nominal difference; 45 vs. 50 W in triode represent 0.46 dB. On their own they do not justify an audible improvement or a higher price.
2.2. Terminology
The KT150 is a beam power tetrode. «Ultralinear» describes the connection of its screen grid to a tap on the primary; «triode» describes the equivalent triode connection. Neither of those terms on its own determines class A or AB.
3. Audio circuit and transformers

3.1. Signal stages
| Block | Description | Evidence |
|---|---|---|
| First stage | 12AX7 in common cathode with active load; one triode acts as the load of the other | OBSERVED / partial reverse engineering 24:06 |
| Driver / phase splitter | 6SN7 in differential configuration; in-phase and anti-phase signals | OBSERVED 23:13 |
| Coupling | Two capacitors per channel towards the power stage; drive-balance adjustments | OBSERVED 23:45 |
| Output | Two KT150 per channel, push-pull; relays switch triode / UL | OBSERVED 25:56 |
3.2. Output transformer
The technician reports ≈5,300 Ω plate to plate as the reflected primary impedance and 43 % for the ultralinear taps (25:29). This is the nominal load reflected to the primary, not a 5.3 kΩ DC resistance.
The assessment of low leakage inductance and small parasitic capacitance comes from the technician; the transcript does not provide numerical measurements of primary L, leakage inductance, distributed capacitance or phase response.
3.3. Square-wave test
The technician shows a roughly rectangular signal with small transients at the edges (34:35). This is a favourable observation, but the frequency, amplitude, exact load and time scale are not sufficiently documented in the transcript. It does not allow the bandwidth to be calculated or stability with capacitive loads to be guaranteed.
4. Measurements and reproducible calculations
| Parameter | Result | Condition / limitation | Evidence |
|---|---|---|---|
| Output voltage | ≈22 V RMS | 8 Ω, 1 kHz, 65 W mode (ultralinear per the video), near clipping | MEASURED-EXT 30:53 |
| Derived power | 60.5 W | P = 22²/8; THD not specified | CALCULATED |
| Difference vs. 65 W | −0.31 dB | 10 log10(60.5/65) | CALCULATED |
| Primary load | ≈5.3 kΩ | Value reported by the technician | REPORTED-EXT 25:44 |
| Ultralinear tap | 43 % | Value reported, no measurement procedure | REPORTED-EXT |
| THD+N at 1/5/10/20 W | ND | No distortion analyser shown | ND |
| Damping factor | ND | No Zout given per DMP position | ND |
| Response with reactive load | ND | No standardised test available | ND |
P = V² / R = 22² / 8 = 60.5 WThe video describes the visible clipping as level is increased. The figure is not equivalent to 60.5 W at 1 % THD. The test does not document a THD+N versus power sweep, nor the thermal duration of the run.
4.1. Ideal primary excursion
Assuming an ideal transformer, 60.5 W and a primary load of 5,300 Ω:
V plate-to-plate (rms) = √(60.5 × 5300) ≈ 566 V; peak excursion per half-primary ≈ 400 VThis is not a measurement of plate voltage. It ignores losses, magnetising inductance, dynamic valve impedance and load variations.
4.2. Comparison with the declared 65 W
The agreement between power at clipping and the general specification is reasonable, but it does not allow the linearity to be judged. An amplifier can reach its nominal power at a THD higher than is tolerable for a high-fidelity comparison.
5. Power supply, filtering and startup
5.1. Architecture
The inspection distinguishes permanent control supply, main audio supply, high-voltage taps and sequencing relays (14:17). The technician estimates the final B+ at around 470 V and describes an initial startup with a lower-voltage tap, followed by relay switching.
5.2. High-voltage capacitors
8 × 220 µF / 450 V are described (14:57). With two equal capacitors in series, each pair is ideally 110 µF and withstands a combined 900 V rating only if voltage sharing is adequate. If four pairs were in parallel on the same node, the equivalent would be 440 µF; the distribution between nodes is not verified.
5.3. Choke filtering
A main C–L–C filter and a second choke dedicated to the driver are described (17:48). This favours ripple reduction and decoupling, but the L values, winding resistances and residual-ripple measurements are missing.
5.4. Filaments and grounding
According to the technician, the 12AX7 and 6SN7 use regulated 6.3 V DC, while the KT150 use 6.3 V AC with a balanced reference. The reported consumption of ≈2 A per KT150 implies about 50.4 W in the four power filaments alone (6.3 × 8 A). Separate per-channel ground returns and dedicated volume-control connections are visible (19:15).
6. DMP, feedback and output impedance
6.1. Apparent contradiction with «0 dB»
The commercial sheet mentions 0 dB of feedback. In contrast, the technician identifies a path from the 4 Ω secondary to the first voltage stage and explains that the DMP selector modifies the global feedback (5:48). He explicitly states that a small amount of feedback remains in OFF, not a completely open loop.
Without the schematic and measurements, it cannot be determined whether the commercial figure corresponds to a specific configuration, to a naming convention or to a different revision.
6.2. Selector positions
| Position | Supported information | Not demonstrated |
|---|---|---|
| OFF | Minimal feedback per the technician | Exactly 0 dB or open loop |
| 1 | Alternative feedback setting | Exact NFB magnitude and Zout |
| 2 | Alternative feedback setting | Exact NFB magnitude and Zout |
Reviews associate more intense DMP positions with difficult loads. The transcript is not enough to assign a numerical damping value to each position.
6.3. Power does not necessarily decrease
Increasing negative feedback usually lowers the closed-loop gain. With the same 500 mV input the output voltage may fall; that does not prove a decrease in maximum achievable power if the drive is increased. The test must be repeated to the same clipping threshold in each position.
6.4. Load-interaction model
ΔdB = 20 log10[(Zmax/(Zmax+Zout)) / (Zmin/(Zmin+Zout))]| Hypothetical Zout | Speaker Z 5 Ω | Speaker Z 15 Ω | Difference |
|---|---|---|---|
| 0.5 Ω | −0.83 dB | −0.29 dB | 0.54 dB |
| 1 Ω | −1.58 dB | −0.56 dB | 1.02 dB |
| 2 Ω | −2.92 dB | −1.09 dB | 1.84 dB |
This simplified resistive model does not include phase angle, frequency dependence of Zout or transformer non-linearity.
7. Biasing, control and safety
7.1. Auto-Fixed-Bias
The technician describes fixed bias with a negative grid voltage controlled by a Microchip microcontroller, based on current sensing through low-value cathode resistors (16:45). This is not classic cathode self-bias. A front-panel instrument and a selector allow each of the four valves to be checked.
7.2. Startup and shutdown sequences
A countdown close to one minute on power-up and a supervised shutdown sequence are described. The technician warns against cutting power during the sequence (12:11). The exact scope of the configuration loss and its recovery procedure has not been demonstrated by a systematic test.
7.3. SGND and protective earth
The SGND switch changes the relationship between signal ground and earth, according to the inspection (4:31). It can help with ground loops, but the chassis protective earth must never be removed. Without a service schematic, altering internal connections or bypassing protections is not recommended.
8. Class A, AB and interpretative limits
After studying how the biasing is adjusted and how the amplifier is protected, a different question remains: for what portion of each cycle does each valve conduct as power increases? That is the question that separates classes A and AB.
Ayon presents the Spirit V as class A, while some reviews describe triode in class A and ultralinear in class AB. The video does not document quiescent current or conduction angle up to maximum power.
With 5.3 kΩ plate to plate and ≈60.5 W, the ideal voltage excursion is high; that is not enough to demonstrate class AB. Nor is a static bias reading enough to demonstrate class A up to 60 W. Instantaneous currents of both branches, plate and screen voltages, and load conditions are required.
For this reason, the rigorous classification of the Spirit V must remain ND as regards the maximum power sustained in class A. Triode/ultralinear and class A/AB describe different properties.
9. Listening, speakers and applications
9.1. What the reviews say
HiFi.nl describes a powerful and relatively neutral sound; Soundrebels points to resolution capability and differences between modes; Alpha Audio provides a comparative listening. These are testimonials in different systems, not measurements or blind tests of superiority.
9.2. Power and dynamics
Illustrative power-demand scenario, not a measurement of the Ayon–MoFi combination. As an example, a pair of MoFi SourcePoint 10 Master Edition coaxial high-fidelity speakers is taken, listened to at 3 m. The measured sensitivity is 89.2 dB/2.83 V/1 m, averaged between 300 Hz and 3 kHz, according to Erin’s Audio Corner on Spinorama. For the calculation, 89.2 dB/W/m is adopted as an approximation, assuming a resistive 8 Ω reference impedance; the actual impedance is not assumed to be constant. The manufacturer declares 91 dB/2.83 V/1 m. These are assumptions, not measurements: −9.5 dB from 1 m to 3 m, +4 dB room and +3 dB for the sum of two channels.
Under those assumptions, peaks of 95/100/102 dB require approximately 6.8/21.6/34.2 W per channel. The results change with distance, room acoustics and music content; they do not predict the maximum SPL of the Spirit V in any room.
9.3. MR-A-01
What MR-A-01 is. It is a passive three-way speaker project, with a crossover designed for a relatively uniform response and a predicted minimum impedance close to 4.9 Ω. It is used here as an example of a real load, not as a speaker tested with the Ayon.
For the MR-A-01 project (predicted impedance minimum ≈4.9 Ω), it is particularly relevant to measure Zout and response with a real load in the three DMP positions. A change in damping may alter a carefully tuned axial response. It is incorrect to select the DMP position exclusively by nominal impedance.
9.4. High-output sources
The 500 mV nominal sensitivity indicates considerable overall gain. With 2 V RCA or 4 V XLR sources, it is advisable to check input headroom and effective gain, especially when using the direct input. According to the technician, the XLR connection is internally converted into an unbalanced signal (21:12); that alone does not demonstrate the common-mode rejection quality of its input interface.
10. Comparison, price and maintenance
10.1. Technical comparison
The amplifiers with a dedicated Nauscopio report are linked in the table; the remaining analyses are in the technical reports index.
The power values are each manufacturer's specifications unless an external measurement is indicated. ND means that no independent, comparable measurement with documented conditions is available here. The absence of a value does not prove worse performance.
| Amplifier | Declared power | Comparable THD+N | Comparable Zout | Comparable DF | Note |
|---|---|---|---|---|---|
| Ayon Spirit V | 65 W pentode / 40 W triode | ND | ND | ND | Video: ≈60.5 W/8 Ω near clipping, THD not quantified |
| Fezz Titania MK2 | 60 W | ND | ND | ND | Comparable DF: ND. Sheet DF: >40, non-standardised conditions |
| Cayin Jazz 90 | 50 W UL / 28 W triode | ND | ND | ND | Sheet THD data without a homogeneous independent test |
| Cayin CS-55A | 40 W UL | ND | ND | ND | Power quoted under its own criterion |
| Acronos RS-1 | 25 W | ND | ND | ND | Single-ended topology: different from push-pull |
To compare THD+N, Zout or DF, identical load, frequency, level, output tap and operating configuration would be required. A catalogue DF is not a comparable measurement if those conditions are missing.
10.2. Price and second-hand purchase
Verifiable current price: ND. Without a dated and traceable offer, no current price or price range is quoted here. For a used unit, it is advisable to check the KT150/KT170 revision, serial number, repair history, valve condition, Auto-Fixed-Bias operation and parts availability.
10.3. Maintenance and cost of ownership
| Item | Documentary status | Recommended check |
|---|---|---|
| Power valve set | 4 × KT150 in the version analysed; KT170 revisions exist. Replacement cost: ND | Request a quote for a matched set and a bias check |
| Signal valves | 12AX7 and 6SN7 in the described circuit. Total replacement cost: ND | Verify the number and exact references of the unit |
| Valve lifetime | ND for this unit and usage conditions | Assess hours, electrical condition and settings; do not fix a universal period |
| Warranty and service in Spain | ND for a specific date, distributor and serial number | Confirm coverage, service centre and parts supply in writing |
| Idle consumption and heat | ND: no reproducible measurement of mains input or thermal dissipation is available | Measure absorbed power at thermal steady state; ensure ventilation |
| Residual noise | ND under comparable laboratory conditions | Measure with terminated inputs and a known level reference |
The circuit includes a microcontroller, relays and bias control. The possibility of obtaining repair and spare parts is therefore an important consideration for a long-term purchase.
11. Conclusions and verdict
Architecture: technically ambitious construction, with an elaborate power supply, differential drivers, ultralinear transformers and bias control.
Power: for the KT150 unit in the video, approximately 60.5 W/8 Ω is calculated from 22 V RMS into 8 Ω near clipping, in line with the general specification of 65 W, but without a quantified THD criterion.
DMP: a feature of great interest for speakers with variable impedance; its effect has not been instrumentally characterised in the examined documentation.
Caveat: electrical superiority cannot be claimed over Fezz Titania MK2, Cayin Jazz 90 or other candidates without homogeneous tests of THD+N, Zout, noise and reactive load.
ANNEX A — SOURCES AND TRACEABILITY
Annex A. Sources and traceability
A.1. Primary sources
Technical video: Ayon Spirit V — review, inspection and measurements. Video transcript and interior photograph provided for the report.
Manufacturer: Ayon Audio — Spirit V specifications · Ayon price list dated 1 January 2023 (copy hosted on Scribd). Commercial specification sheets may change; record the revision of each unit.
A.2. Reviews and cross-checking
HiFi.nl — René van Es · Soundrebels — Spirit V · HiFi-IFAs — Spirit V · Alpha Audio — amplifier comparison.
Not all external publication and access dates could be verified. The report's update date should not be interpreted as the publication date of those references. The historical commercial price list quoted corresponds to 2023 and does not certify current prices.
Example in §9.2. MoFi SourcePoint 10 Master Edition measurement data published by Erin’s Audio Corner on Spinorama. Manufacturer specification (91 dB) according to Audio Advisor.
A.3. Editorial record
Authorship: maty @nauscopio + IA (research, calculations, writing and layout). Version: V1.8.5.2. Date: 18 November 2026.
Scope: technical synthesis based on the supplied inspection and the identified sources; no in-house laboratory test was performed.
A.4. Change history
V1.8.5.2 (EN): English translation of V1.8.5.2 (ES); Spanish version at ayon-spirit-v.html. V1.8.5.2: pentode naming in tables and summary, source for the §9.2 example and A.2 wording. V1.8.5: editorial cleanup, note on operating mode, internal links and print compatibility. V1.8.4: editorial close: unified version, corrected SVG attributes, verified links and precise evidence labels. V1.8.3: references to other reports added. V1.8.2: accessibility fixes and new power-supply and power figures. V1.8: ND-homogeneous comparison, reworked §10 and functional diagram. V1.7: simplified CSS and JavaScript. V1.6: JSON-LD and dedicated script for annexes. V1.5: collapsible annexes.
ANNEX B — PENDING MEASUREMENT PROTOCOL
Annex B. Pending measurement protocol
| Test | Proposed conditions | Purpose |
|---|---|---|
| THD+N versus power | 1 kHz; 0.1–65 W; 4/8 Ω; both channels | Clean power and curve shape |
| Multifrequency THD+N | 20 Hz / 1 kHz / 10 kHz at 1/5/10/20 W | Linearity and transformers |
| Output impedance | 20 Hz–20 kHz; DMP OFF/1/2; 4/8 Ω taps | Damping factor and response |
| Square wave | 100 Hz / 1 kHz / 10 kHz; resistive/reactive load | Stability and transients |
| Noise | Terminated inputs; 1 W reference | SNR and hum |
| Conduction current | Per branch and per mode; safe measurement | Actual class A / AB |
All tests require documentation of mains voltage, warm-up, valve condition, DMP configuration, output tap, power and distortion threshold.