Cayin Jazz 80, PP Ultralinear valve amp with KT88 or EL34
Verified performance, evidence limits and measurements still pending
What it is: integrated valve amplifier in push-pull class AB1 configuration (one valve pair per channel, four total). It carries four KT88 or four EL34 valves, an ESS ES9018K2M DAC and high-quality Bluetooth 5.0 (LDAC, aptX HD) that only works wirelessly.
Specified by Cayin: 2×40 W KT88 UL / 2×20 W TR; 2×36 W EL34 UL / 2×18 W TR; THD 0.2 % at 1 W/1 kHz; S/N 90 dB (A); 8 Hz–60 kHz (−3 dB); 100 kΩ input; 300 mV sensitivity; 4 Ω and 8 Ω taps; 16.8 kg; 280 W declared maximum consumption.
Measured by AUDIO/stereoplay (11/2023, Lothar Brandt; reproduced by 7Review; KT88 unit): 26 W/8 Ω and 34 W/4 Ω at 3 % THD with sine wave, from the 4 Ω tap; 33 W/8 Ω and 46 W of complex music power with a complex music signal; THD+N 0.18 % at 1 W; 100 dBA at 10 V; consumption 148 W at rest and 231 W at maximum.
P@1 % THD: not available for either KT88 or EL34. No estimated range is included because, although mathematically constructible, it would lead the reader to remember a figure without experimental backing.
Five key magnitudes unmeasured: P@1 % THD, output impedance (Zout), B+ voltage, supply ripple and quiescent current / operating point of the installed valves. Without these five measurements, the technical comparison with the PrimaLuna EVO 300 and the xDuoo A35 remains open.
The essence of this audit: 26 W and 40 W differ mathematically by 1.87 dB, but that comparison is not homogeneous (3 % THD vs. nominal specification). The useful question is not "does it deliver 26 or 40 W?", but "where does THD start to rise, what is its Zout and how much clean SPL does it deliver over the real load?".
GETTING STARTED — WITH NO PRIOR KNOWLEDGE (Basic guide G1–G6)
GETTING STARTED — WITH NO PRIOR KNOWLEDGE
If you have never read a valve amplifier audit, start here. Only high-school physics is needed (logarithms, electrical power, impedance) to understand why the figures in this report are expressed with such care. Technical details begin in §0.
Basic guide: reading this report without being a valve expert
Central idea: the Jazz 80 is well specified and has a sufficiently informative set of external measurements, but it lacks five numbers (P@1 % THD, Zout, B+, supply ripple and quiescent current / operating point of the installed valves) that decide its real behaviour with speakers. This report describes what is known and clearly marks what remains to be measured.
- G1 · how a KT88/EL34 valve push-pull works.
- G2 · which data is reliable and which is not.
- G3 · watts, decibels and SPL: why 26 W is not "little" compared to 40 W.
- G4 · supply, choke filter and ripple.
- G5 · output impedance, damping factor and why they are the absolute priority.
- G6 · conclusions and exercises with solutions.
G1. What a KT88 or EL34 push-pull valve amplifier is
An amplifier takes the weak signal from the player (on the order of 300 mV to 1 V) and converts it into a signal with the voltage and current needed to drive the speaker. In the Jazz 80 this is done by thermionic valves arranged in two blocks: a preamp (two 12AX7/ECC83) that increases the voltage, and a power stage (four KT88 or four EL34, two per channel) that supplies the current. Everything is powered by a supply that produces a high DC voltage called B+.
Player → Preamp (12AX7) → Drivers / Phase splitter (12AU7)
→ Power stage (4× KT88 or EL34)
→ Output transformer (EI, 4 Ω / 8 Ω taps)
→ Speaker
Supply: toroidal + rectification + capacitor bank + choke filter (1 H / 180 mA)The four power valves work in pairs in push-pull: while one conducts more current, the other conducts less, and their currents combine in the output transformer. This has two physical advantages: the DC component cancels in the transformer core (it does not permanently magnetise it) and many even harmonics are cancelled, which reduces distortion. The Jazz 80 operates in class AB1 according to the manufacturer's documentation.
The output transformer matches the impedance: a power valve needs to see thousands of ohms and a speaker has 4–8 Ω. With a turns ratio n, the impedance is multiplied by n². Illustrative example (not an exact Jazz 80 data point): with n = 25, an 8 Ω speaker is seen from the valves as 25² · 8 = 5000 Ω.
The Jazz 80 offers two modes: ultralinear (UL), where the screen grids are connected to intermediate taps of the transformer, with more power and dynamics; and triode (TR), with less power but a smoother, more textured curve. Never change mode with the amplifier on: it can destroy the output valves.
G2. Which data is reliable and which is not
This report is not a measurement of its own: it combines manufacturer specifications with third-party measurements (AUDIO/stereoplay, reproduced by 7Review) and derived calculations. Every statement carries an evidence label (the full legend is in the margin, and the method in §2):
| Label | What it means | Example in this report |
|---|---|---|
| CONFIRMED | Manufacturer specification or documentation. | Nominal power 2×40 W KT88 UL. |
| MEASURED-EXT | Measurement by an identifiable third party with laboratory instrumentation. | 26 W/8 Ω at 3 % THD (AUDIO/stereoplay). |
| CALCULATED | Mathematical result derived from defined data. | 14.42 V RMS corresponding to 26 W over 8 Ω. |
| CONDITIONAL CALCULATION | Calculation depending on an explicit, unverified hypothesis. | SNR at 1 W from the 100 dBA at 10 V reference. |
| ND | Not available in the consulted sources. | Jazz 80 Zout. |
G3. How many watts are needed: decibels and SPL
To know whether an amplifier "has enough", you must start with the speaker and the listening distance. The decibel is a logarithmic scale: +10 dB multiply power by 10 and +3 dB multiply it by 2 (100.3 ≈ 1.995). It is calculated as follows, with a generic reference system:
- Sensitivity. The reference speaker produces 89 dB with 1 W at 1 m. It is a HYPOTHESIS of calculation, not a measurement.
- Distance. The level falls by 20·log₁₀(d) dB with distance. At 3 m: SPL(1 W, 3 m) = 89 − 20·log₁₀(3) = 79.46 dB.
- Power. To reach a level L: P = 10^((L − 79.46)/10).
- The key question. Between 26 W and 40 W there is only 10·log₁₀(40/26) ≈ 1.87 dB of mathematical difference. Between 34 W and 40 W, only 0.71 dB. However: the 26 W come from a measurement at 3 % THD and the 40 W from a nominal specification with no defined condition. It is not a homogeneous experimental comparison.
- Peaks. Music has peaks above its average level. A crest factor of 15 dB multiplies the instantaneous power by ~31.6. That is why you must size for the peak, not for the average.
G4. The supply: choke filter, capacitor bank and ripple
Valves need a high DC voltage (B+) obtained by rectifying the mains. In Spain the mains is 50 Hz, so the ripple appears at 100 Hz. The capacitors charge at the peaks and discharge between them. Since Q = C·ΔV and the time between peaks is 1/f (with full-wave rectification, f = 100 Hz):
ΔV ≈ I / (f · C)
The Jazz 80 includes a choke filter of 1 H / 180 mA labelled on the component itself. That data point is OBSERVED. A coil opposes rapid current variations with a reactance X_L = 2π·f·L: at 100 Hz, 1 H presents 628 Ω, which effectively reduces the ripple passing through the filter.
G5. Output impedance (Zout) and damping factor
A "6 Ω" speaker does not always measure 6 Ω: its impedance changes with frequency. A typical speaker can have a peak of 20–30 Ω in the bass and drop to 4–5 Ω in the midrange. The amplifier's output impedance (Zout) acts as a series resistance with the speaker.
The damping factor is DF = Z_load / Z_out. The higher the DF, the less the varying impedance of the speaker influences the final response. In valve amplifiers without feedback, a DF of 1–3 is usual; with moderate feedback, a DF of 5–10.
Measuring Zout is straightforward: inject a known signal with the amplifier at the output and measure the voltage drop when connecting a known load (for example, 8 Ω) versus open circuit. The difference allows Zout to be calculated precisely. It is a bench measurement that any technician with an oscilloscope and a multimeter can do in half an hour.
G6. Conclusions and exercises
- The Jazz 80 has manufacturer specifications, a sufficiently informative set of external measurements (AUDIO/stereoplay, reproduced by 7Review) and a reasonably documented functional architecture (push-pull AB1, 1 H/180 mA choke, EI transformers, LDAC Bluetooth + ES9018K2M DAC).
- The measured power at 3 % THD (26 W/8 Ω and 34 W/4 Ω from the 4 Ω tap) is below the nominal 40 W, but the mathematical difference is only 1.87 dB, and the comparison is not homogeneous.
- The five magnitudes that would decide its real behaviour — P@1 % THD, Zout, B+, supply ripple and quiescent current — remain ND. Without them, the comparison with PrimaLuna and xDuoo remains open.
- Neither a table nor a review decides whether an amplifier "is better": that requires comparable measurements (THD+N vs power, Zout, response with real load, noise).
Exercises to practise:
- A speaker has 92 dB/W/m. What level does it produce with 1 W at 2 m?
- With the 89 dB/W/m speaker from this guide, what power is needed for 95 dB at 3 m with a single speaker?
- How many dB of difference are there between 26 W and 40 W?
- A speaker has a 24 Ω peak in the bass and a 4.8 Ω minimum in the midrange. With Zout = 1.5 Ω, what level variation does that produce between those two points?
See solutions
- 92 − 20·log₁₀(2) = 92 − 6.02 ≈ 86 dB.
- 10^((95 − 79.46)/10) ≈ 35.8 W.
- 10·log₁₀(40/26) ≈ 1.87 dB.
- Δ ≈ 20·log₁₀[(24·(4.8+1.5)) / (4.8·(24+1.5))] ≈ 20·log₁₀(1.2353) ≈ 1.84 dB.
0. Executive summary and thesis
20-second read
What it is: integrated push-pull AB1 valve amplifier (one valve pair per channel, four total). Four KT88 or four EL34, with ESS Sabre ES9018K2M DAC and Bluetooth 5.0 (LDAC, aptX HD) that only works wirelessly. CONFIRMED
What has been measured (KT88 unit): 26 W/8 Ω and 34 W/4 Ω at 3 % THD (sine wave, from the 4 Ω tap); 33 W/8 Ω and 46 W of complex music power (discrepant impedance: 3 Ω in the lab box / 2 Ω in the body text); THD+N 0.18 % at 1 W; 100 dBA at 10 V; consumption 148 W at rest and 231 W at maximum. MEASURED-EXT
What is NOT demonstrated: five magnitudes — P@1 % THD, Zout (output impedance), real B+, supply ripple and quiescent current / operating point of the installed valves. All are ND.
The essence: 26 W and 40 W differ mathematically by 1.87 dB, but it is not a homogeneous comparison. The useful question is not "does it give 26 or 40 W?", but "where does THD start to rise, what is its Zout and how much clean SPL does it deliver over the real load?". New terms? Start with the basic guide.
Audit thesis
The Jazz 80 is sufficiently measured to characterise part of what it does, but still insufficiently measured to explain quantitatively why it behaves that way under different loads. This sentence summarises the real state of knowledge about the amplifier and defines the objective of the document.
The objective is not to fill the gaps with assumptions, but to make clear what the unknowns are, what consequences they have and what specific measurements would allow them to be resolved.
Synthesis in three layers
- Layer 1 — Specified (CONFIRMED). Cayin publishes the architecture (including AB1-Push-Pull-Verstärkung in the official manual), the valves, the nominal powers, the THD at 1 W, the S/N ratio, the bandwidth, the sensitivity, the input and output impedances, the dimensions, the weight and the maximum consumption. This whole layer can be verified against the manufacturer's documentation.
- Layer 2 — Measured by a third party (MEASURED-EXT, KT88 unit). The AUDIO/stereoplay review (11/2023, Lothar Brandt), reproduced by 7Review, provides the sine power at 3 % THD from the 4 Ω tap (26 W/8 Ω, 34 W/4 Ω), the complex music power (33 W/8 Ω, 46 W with discrepant 2/3 Ω impedance), the THD+N at 1 W (0.18 %), the S/N at 10 V (100 dBA) and the consumptions.
- Layer 3 — Calculated, conditional, hypothesis or ND. The conversions to other magnitudes (SNR at other powers, theoretical SPL) and the explanations of causes (why the PrimaLuna seems more authoritative, why the response depends on the load) belong to this layer and are explicitly labelled.
The essence of this audit
The 26 W power and the nominal 40 W are not acoustically different in a proportional way: the mathematical difference is 1.87 dB, but the comparison is not homogeneous. The useful question is not "does it give 26 W or 40 W?", but "where does THD start to rise, what is the Zout and how much clean SPL does it deliver over a real load?".
The same applies to the PrimaLuna EVO 300: the comparison is not "one sounds better than the other", but "one has the measured Zout and the other does not, so the rigorous comparison remains pending".
1. Object and scope of the audit
What is audited
The Cayin Jazz 80, integrated valve amplifier in its KT88 and EL34 versions, as marketed in October 2026.
What is analysed
- Electrical architecture and topology (push-pull, preamp, drivers, transformers, supply).
- Verifiable performance: power, THD, S/N, frequency response, consumption.
- Behaviour with the load and real meaning of the measured power.
- Relationship between power, SPL and dynamic margin in a specific listening system.
- Compatible valves and their operating requirements.
- Proposed modifications — as experimental protocols, not as demonstrated improvements —.
- Comparison with three references: CS-55A (predecessor), PrimaLuna EVO 300 (higher range) and xDuoo A35 (direct competitor).
What is out of scope
- Subjective listening assessments not correlated with electrical magnitudes.
- Modifications without prior measurement ("premium" capacitors, silver wiring, etc.).
- Unverifiable commercial claims.
- Estimates that substitute measured data without being labelled as such.
Guiding principle
Every statement in the report must be traceable to one of the categories in §2. When not possible, it is marked ND or DO NOT DECIDE YET. When there is no evidence, it is not filled in with plausibility.
2. Evidence criteria
Classification used in this report
| Label | Definition | Implication for the reader |
|---|---|---|
| CONFIRMED | Manufacturer specification or documentation. | Official data, subject to the accuracy with which the manufacturer publishes it. |
| MEASURED-EXT | Measurement carried out by an identifiable third party with laboratory instrumentation. | External data, not reproduced by the author of this audit. |
| CALCULATED | Mathematical result derived from sufficiently defined data. | Valid under the stated hypotheses; it is not a measurement. |
| CONDITIONAL CALCULATION | Calculation depending on an explicit, unverified hypothesis. | Only valid if the hypothesis is accepted; the hypothesis must be cited. |
| OBSERVED | Element visible or physically documented. | Photographic or documentary evidence, not electrical. |
| HYPOTHESIS | Possible causal explanation, pending verification. | Must not be cited as a demonstrated cause. |
| ND | Magnitude not available in the consulted sources. | Requires direct measurement or access to the service schematic. |
| DO NOT DECIDE YET | Question on which current evidence does not allow a conclusion. | Must not be resolved by analogy or commercial claim. |
| AUDIOPHILE PROPOSAL — S3 | Experimental modification without prior A/B measurement. | May be proposed as a trial, not as a demonstrated improvement. |
Note on the method itself
This audit explicitly separates what is specified, what is measured by third parties, what is calculated, what is conditional and what is pending measurement. The reader can use the above classification as a tool to read the rest of the report with appropriate scepticism.
Application rule: no statement in the report should be raised to a category higher than that of its source evidence. In particular, a CONDITIONAL CALCULATION data point must not be read as CONFIRMED, and a HYPOTHESIS must not be read as a demonstrated cause.
3. Introduction and positioning
Origin and philosophy of the product
The Cayin Jazz 80 is an integrated valve amplifier from the Chinese brand Cayin (Zhuhai Spark Electronic Equipment Co., Ltd.), presented in September 2023 as an evolution of the successful CS-55A. The magazine fairaudio describes it as "one of the bestsellers in the Cayin programme", highlighting that its reliable Bluetooth function and sensitive remote control make it attractive also to younger listeners.
Cayin's own CEO, Thomas Deyerling, explained the origin of the name: "We chose the Jazz designation because there is already a Soul amplifier in the Cayin line. So the designation by a musical genre suggested itself".
Technical proposition
The Jazz 80 combines a push-pull class AB1 power stage with KT88 or EL34, switchable operation between ultralinear and triode, individual bias adjustment via VU meters, and Bluetooth 5.0 connectivity with an integrated ESS ES9018K2M DAC (accessible only wirelessly). The construction mixes point-to-point wiring in the main analogue section with printed circuit boards, and uses EI-core output transformers.
Positioning in the Cayin family
The Jazz 80 should no longer be presented simply as "the model that succeeds the CS-55A". Cayin has since added the Jazz 70 and Jazz 90 models to the family, and fairaudio tested the Jazz 90 in 2026. The Jazz 80 is the model that consolidated the brand's Jazz branch and is still marketed in both KT88 and EL34 configurations.
Reference gallery
4. Technical specifications
4.1 Manufacturer specifications (CONFIRMED)
| Parameter | KT88 | EL34 | Evidence |
|---|---|---|---|
| Product type | Stereo integrated valve amplifier | CONFIRMED | |
| Configuration | Push-pull, class AB1 («AB1-Push-Pull-Verstärkung» in the manufacturer's manual) | CONFIRMED | |
| Power valves | 4× KT88 | 4× EL34 | CONFIRMED |
| Preamp valves | 2× 12AX7 (ECC83) | CONFIRMED | |
| Phase splitters / drivers | 2× 12AU7 (ECC82) | CONFIRMED | |
| Power (ultralinear, 4/8 Ω) | 2 × 40 W | 2 × 36 W | CONFIRMED |
| Power (triode, 4/8 Ω) | 2 × 20 W | 2 × 18 W | CONFIRMED |
| Frequency response | 8 Hz – 60 kHz (−3 dB) | CONFIRMED | |
| Declared THD (1 kHz, at 1 W) | 0.2 % (THD; not including noise) | CONFIRMED | |
| S/N ratio | 90 dB (A) (measurement condition unspecified) | CONFIRMED | |
| Input impedance | 100 kΩ | CONFIRMED | |
| Input sensitivity | 300 mV | CONFIRMED | |
| Output taps | 4 Ω and 8 Ω | CONFIRMED | |
| Dimensions (W × D × H) | 300 × 385 × 190 mm | CONFIRMED | |
| Weight | 16.8 kg | CONFIRMED | |
| Declared maximum consumption | 280 W | CONFIRMED | |
| Colours available | Black or Silver | CONFIRMED | |
4.2 Connectivity
| Type | Detail |
|---|---|
| Analogue inputs | 3 × RCA Line (stereo) |
| Headphone output | 1 × 6.35 mm jack (360 Ω output impedance) |
| Output voltage (headphones) | 1.5 V at 32 Ω / 8.4 V at 300 Ω |
| Speaker outputs | Screw terminals (4 Ω and 8 Ω) |
| Bluetooth | Version 5.0, Qualcomm QCC5125 chipset |
| Bluetooth codecs | SBC, AAC, aptX, aptX HD, aptX Low Latency, LDAC (990 kbit/s) |
| Integrated DAC | ESS Sabre 32 (ES9018K2M), accessible only via Bluetooth |
| Wired digital inputs | None (no optical, coaxial or USB) |
5. Internal architecture and topology
5.1 Circuit topology (CONFIRMED)
The Jazz 80 uses a push-pull class AB1 topology with four power valves. The preamp section uses two 12AX7/ECC83 valves to amplify voltage, while two 12AU7/ECC82 act as phase splitters and drivers for the power valves.
RCA → Selector → 12AX7 (preamp) → 2× 12AU7 (phase splitter / driver)
↓
4× KT88 / EL34 push-pull class AB1
↓
EI output transformers
↓
4 Ω / 8 Ω5.2 Transformers (OBSERVED)
- Power transformer: toroidal type according to the manufacturer.
- Two output transformers: EI (laminated core) type, with taps for 4 Ω and 8 Ω.
5.3 Power supply
The power supply includes a choke filter with a confirmed value of 1 H and 180 mA, according to the component label visible in the interior photographs. This data point is OBSERVED in the photographic documentation.
The supply, however, is better documented by architecture than by electrical magnitudes. The B+, the capacity of the main capacitors, the real HT current, the screen current, the choke DCR, the resulting ripple and the dynamic impedance of the supply are unknown.
5.4 Bias adjustment
The Jazz 80 allows individual bias adjustment for each of the four power valves via potentiometers accessible from the top. The left VU meter can be converted into a quiescent current monitor to facilitate adjustment. A central switch changes the left VU meter to bias mode, and switches A1/A2 and A3/A4 select the valve to adjust.
The manufacturer's documented procedure includes: initial stabilisation of approximately 10 minutes; switching to BIAS mode with the volume at minimum; individual adjustment of A1, A2, A3 and A4 bringing the needle to the centre of the scale; cross-check of the four positions (the valves influence each other); re-check after another 20 minutes; and consideration that the bias point depends on the mains voltage.
5.5 Construction and materials
- Chassis: aluminium, black or silver finish.
- Valve cage: removable metal cage.
- Feet: four vibration-absorbing feet.
- Wiring: combination of point-to-point wiring in key sections and printed circuit boards (PCB).
- Accessories: cotton gloves, cleaning cloth, spare fuses, standard power cable and manual.
5.6 Critical reading of the architecture
What is confirmed: the Jazz 80 is a push-pull class AB1 with four output valves, 12AX7/12AU7 in small signal, EI output transformers with 4/8 Ω taps, and a choke identified as 1 H / 180 mA in the photographic documentation.
What should not be over-interpreted: knowing the choke does not by itself allow reconstructing the complete filtering capacity, the B+ ripple, the source impedance or the peak margin. Knowing that the output transformer is EI does not allow deducing its core section, primary inductance, copper, air gap or thermal power.
Bias: the VU meter system allows adjusting the four valves individually according to the device's procedure, but it does not turn the VU meter indication into an absolute measurement of plate or cathode current. To characterise the operating point, a direct electrical measurement is needed.
6. Functional characteristics
6.1 Operating modes: triode vs. ultralinear
| Mode | Power (KT88) | Power (EL34) |
|---|---|---|
| Ultralinear (UL) | 2 × 40 W | 2 × 36 W |
| Triode (TR) | 2 × 20 W | 2 × 18 W |
Listening description from reviews — NOT OBJECTIVE
| Mode | Subjective description |
|---|---|
| Ultralinear (UL) | More energetic, dynamic and controlled. |
| Triode (TR) | Warmer, more fluid and textured. |
6.2 High-resolution Bluetooth
The Bluetooth module is encapsulated in its own metal sub-compartment to mitigate interference towards the analogue circuitry. It uses the Qualcomm QCC5125 chipset with Bluetooth 5.0 and supports LDAC (up to 990 kbit/s, 24-bit/96 kHz), aptX HD, aptX Low Latency, aptX, AAC and SBC. The antenna is screw-on at the rear.
6.3 Integrated DAC
The ESS ES9018K2M Sabre32 is a 32-bit DAC. It is only accessible through the Bluetooth input; there are no wired digital inputs (optical, coaxial or USB). fairaudio explicitly regrets it: "Unfortunately, the integrated D/A converter is only accessible via this route; there is no way to connect it by cable. A pity".
6.4 Dual-function VU meters
- Sound level indicator during listening.
- Bias monitor: the left meter allows checking and adjusting the power valve bias.
6.5 Remote control and volume control
The remote is metal, solid and with fine, precise volume control. Its functions are limited to Input, Mute and Volume. Volume control is via an Alps potentiometer.
6.6 Headphone output
Includes a 6.35 mm output with an output impedance of 360 Ω. Output voltage: 1.5 V at 32 Ω and 8.4 V at 300 Ω. According to fairaudio, it sounds "fine, but not exceptional"; it cannot replace a dedicated headphone amplifier.
6.7 Burn-in and warm-up
fairaudio reports a recommendation from Cayin's CEO of 50 h of burn-in, preferably 100 h, for a new unit. The amplifier takes 20–30 seconds after power-on to be ready.
7. Laboratory measurements
7.1 Frequency response
AUDIO/stereoplay describes the response as "ausgeprägte Lastabhängigkeit", literally, an "accentuated load dependence". The observation appears associated with both ultralinear (UL) and triode (TR) operation, and the review presents 8 Ω / 4 Ω / 2 Ω curves.
In a valve amplifier with appreciable Zout, the amplifier response and the speaker impedance form a system. A flat response over a laboratory resistor does not guarantee the same response over a real cabinet. This reinforces the importance of measuring the Jazz 80's Zout.
7.2 Output power (MEASURED-EXT)
| Condition | Power (8 Ω) | Power (4 Ω / 3 Ω / 2 Ω) | Evidence |
|---|---|---|---|
| Nominal power (KT88, UL) | 40 W | 40 W | CONFIRMED |
| Power at 3 % THD (sine; measured from the 4 Ω tap) | 26 W | 34 W (at 4 Ω) | MEASURED-EXT |
| Complex music power | 33 W | 46 W — discrepant impedance: 3 Ω (box) / 2 Ω (text) | MEASURED-EXT discrepancy |
7.3 Harmonic distortion
The AUDIO/stereoplay measurements show that H2 (second harmonic) is the dominant distortion product, with H2 clearly above H3 in the published graph.
7.4 Signal-to-noise ratio
The manufacturer declares 90 dB(A) without specifying the measurement condition (power, load, weighting). AUDIO/stereoplay publishes a single external reference: 100 dBA at 10 V. With that figure, the equivalent noise is 100 µV RMS, whatever the load over which those 10 V are applied.
To translate that reference to other powers, it is assumed that the noise level is constant with the signal, that is, that the S/N scales with the applied output voltage. Under that hypothesis:
| Magnitude | Data | Status | Condition |
|---|---|---|---|
| Manufacturer S/N | 90 dBA | CONFIRMED | Measurement condition unspecified |
| S/N AUDIO/stereoplay | 100 dBA | MEASURED-EXT | 10 V |
| S/N equivalent to 1 W / 4 Ω | ≈ 86 dBA | CONDITIONAL CALCULATION | Constant noise; 2 V |
| S/N equivalent to 1 W / 8 Ω | ≈ 89 dBA | CONDITIONAL CALCULATION | Constant noise; 2.83 V |
| S/N equivalent to 10 W / 4 Ω | ≈ 96 dBA | CONDITIONAL CALCULATION | Constant noise; 6.32 V |
| S/N equivalent to 34 W / 4 Ω | ≈ 101.3 dBA | CONDITIONAL CALCULATION | Constant noise; 11.66 V |
7.5 Other measurements
- THD+N at 1 W: 0.18 % MEASURED-EXT.
- Consumption at rest: 148 W MEASURED-EXT.
- Measured maximum consumption: 231 W MEASURED-EXT — recorded by AUDIO/stereoplay under the conditions of their review.
- Declared maximum consumption: 280 W CONFIRMED — declared by Cayin.
- Complex music power: 46 W with discrepant 2/3 Ω impedance MEASURED-EXT.
7.6 What we can really conclude from the external review
The combination of 26 W/8 Ω and 34 W/4 Ω at 3 % THD (from 4 Ω tap) with 0.18 % THD+N at 1 W provides useful information, but does not allow reconstructing a complete P–THD curve. The music power of 46 W also must not be converted into a continuous power specification, and even less so while the associated impedance remains documentarily discrepant (3 Ω box / 2 Ω text).
The SNR measurement at 10 V is an external reference point. The values at other powers are mathematical transformations under an explicit hypothesis and do not constitute new measurements of the amplifier.
High-value pending data point: a THD+N versus power curve, at 4 Ω and 8 Ω, would allow directly determining P@1 %, P@3 % and the progression towards clipping. That test is more informative than inferring a single figure from a single power at 3 % THD.
7b. Power, voltage, SPL and dynamic margin
7b.1 Power → voltage → current conversion (CALCULATED / CONDITIONAL CALCULATION)
From P = V²/R and from P = I²·R, the electrical magnitudes corresponding to each condition are derived. The rows corresponding to defined data are CALCULATED; the row corresponding to 46 W has the status CONDITIONAL CALCULATION because the source impedance is documentarily ambiguous.
| Condition | P | V RMS | I RMS | Evidence |
|---|---|---|---|---|
| 26 W over 8 Ω (from 4 Ω tap) | 26 W | 14.42 V | 1.80 A | CALCULATED |
| 34 W over 4 Ω (from 4 Ω tap) | 34 W | 11.66 V | 2.92 A | CALCULATED |
| 33 W over 8 Ω (music power) | 33 W | 16.25 V | 2.03 A | CALCULATED |
| 46 W / 2 Ω (text) | 46 W | 9.59 V | 4.80 A | CONDITIONAL CALCULATION |
| 46 W / 3 Ω (box) | 46 W | 11.75 V | 3.92 A | CONDITIONAL CALCULATION |
| 40 W over 8 Ω (nominal) | 40 W | 17.89 V | 2.24 A | CALCULATED |
| 36 W over 8 Ω (nominal EL34) | 36 W | 16.97 V | 2.12 A | CALCULATED |
- The 26 W, 34 W, 33 W, 40 W and 36 W rows are conversions of defined data (measured or specified). Their status is CALCULATED.
- The two 46 W rows are conditional conversions: the source impedance is not resolved in the primary source. Their status is CONDITIONAL CALCULATION.
- The value of 4.80 A (46 W/2 Ω) or 3.92 A (46 W/3 Ω) does not imply that the output transformer or the power stage have been measured physically delivering that RMS current.
- In the case of the 46 W, the source defines the data point as complex music power; therefore, the calculated RMS values must not be interpreted as measurements of continuous sine delivery at 2 Ω nor at 3 Ω.
7b.2 Power → SPL in a generic reference system (CONDITIONAL CALCULATION)
For a generic speaker with a sensitivity of 89 dB/W/m and a listening distance of 3 m, the free-field geometric loss is:
20·log₁₀(3) = 9.54 dB
Level with 1 W at 3 m:
89 − 9.54 = 79.46 dB
| Idealised power | SPL at 3 m (free field, 89 dB/W/m) |
|---|---|
| 1 W | 79.5 dB |
| 2 W | 82.5 dB |
| 5 W | 86.4 dB |
| 7 W | 87.9 dB |
| 10 W | 89.5 dB |
| 18 W | 92.0 dB |
| 20 W | 92.5 dB |
| 26 W | 93.6 dB |
| 34 W | 94.8 dB |
| 40 W | 95.5 dB |
| 46 W | 96.1 dB |
7b.3 Specific application to Crossover A v2.4 MR-A-01 (CONDITIONAL CALCULATION)
For Crossover A v2.4 MR-A-01 we start from a sensitivity of 89 dB / 2.83 V / 1 m over a reference nominal impedance of 5.3 Ω. The conversion to sensitivity per watt is:
P = 2.83² / 5.3 = 1.511 W
S(1 W) = 89 − 10·log₁₀(1.511) = 87.21 dB/W/m
At 3 m in free field:
S(1 W, 3 m) = 87.21 − 20·log₁₀(3) = 77.67 dB
| Idealised power | SPL at 3 m (free field, 87.21 dB/W/m) |
|---|---|
| 26 W | 91.8 dB |
| 34 W | 93.0 dB |
| 36 W | 93.2 dB |
| 40 W | 93.7 dB |
| 46 W | 94.3 dB |
| 54.1 W | 95.0 dB |
7b.4 What 26 W vs 40 W really means (CALCULATED)
| Power | Δ with respect to 40 W |
|---|---|
| 20 W | −3.01 dB |
| 26 W | −1.87 dB |
| 30 W | −1.25 dB |
| 34 W | −0.71 dB |
| 40 W | 0 dB |
| 46 W | +0.61 dB |
The mathematical difference between 26 W measured at 3 % THD and 40 W nominal without THD criterion is 1.87 dB. Between 34 W and 40 W it is only 0.71 dB. Although in percentage 26/40 = 65 %, it is not a homogeneous experimental comparison.
7b.5 Joint interpretation
The data in this section allow the question about power to be reformulated in a technically sound way:
- The Jazz 80 delivers 26 W/8 Ω and 34 W/4 Ω at 3 % THD (from 4 Ω tap, AUDIO/stereoplay).
- It delivers 33 W/8 Ω and 46 W of complex music power (discrepant 2/3 Ω impedance, AUDIO/stereoplay).
- Its P@1 % THD is ND.
- Its Zout is ND.
- Its B+ and supply ripple are ND.
In a system with Crossover A v2.4 MR-A-01 (87.21 dB/W/m, historical reference), that power is compatible with usual listening levels and with peaks of approximately 92–94 dB SPL in a free-field approximation. The real margin must be determined through the available power at low distortion and the behaviour against the real load, not through the nominal number of watts.
Note on tap selection: Crossover A does not present an impedance that corresponds exactly to either 4 Ω or 8 Ω. The tap selection (4 Ω or 8 Ω) constitutes a variable of the test, not a trivial data point. Ideally, the S0 test should measure both.
INTERPRETATION — §8 to §10
8. Sonic performance
This section explicitly separates three levels: third-party listening descriptions, available electrical data and what cannot be related to them.
8.1 Listening description — fairaudio
- Warm, textured and with great presence, without falling into excess coloration.
- Balanced and musical, with silky highs, midrange with "smoothness" and neutral bass.
- Three-dimensional, with excellent localisation precision and scene depth.
fairaudio summarises the character with the expression «goldener Mittelweg» («golden middle way»): it does not seek extremes in any parameter. It describes it as "more a means of enjoyment than a studio tool".
Bass: neutral and well integrated, without extreme extension. Macro-dynamics: solid middle ground. Stage width slightly narrower than depth.
8.2 Available electrical data
- THD+N 0.18 % at 1 W (AUDIO/stereoplay).
- S/N 100 dBA at 10 V (AUDIO/stereoplay).
- Response with "ausgeprägte Lastabhängigkeit" in UL and TR (AUDIO/stereoplay).
- H2 dominant over H3 (AUDIO/stereoplay).
- Measured power 26 W/8 Ω and 34 W/4 Ω at 3 % THD (from 4 Ω tap).
- Zout: ND.
- B+ and ripple: ND.
8.3 What cannot be related
No cause-effect relationship can be established between the listening descriptions and specific electrical magnitudes when those magnitudes are not measured. In particular:
- The difference in bass with the PrimaLuna cannot be attributed to Zout, because the Jazz 80's Zout is ND.
- The smoothness of the highs cannot be attributed to H2 dominance without measuring the full spectrum.
- The dynamics cannot be attributed to the supply or the choke without measuring B+ and ripple.
9. Comparison: KT88 vs. EL34 vs. KT77
9.1 Valve technical data (valve manufacturers)
| Aspect | KT88 | EL34 | KT77 (Genalex Gold Lion) |
|---|---|---|---|
| Type | Beam tetrode | Pentode | Beam tetrode |
| Max plate dissipation | 42 W | 25 W | 32 W (source-dependent; see note) |
| Max plate voltage | 800 V | 800 V | ≤ 800 V datasheet; 600 V according to some sources |
| Power (UL, 8 Ω, in the Jazz 80) | 2 × 40 W | 2 × 36 W | Not published for the Jazz 80 |
| Power (TR, 8 Ω, in the Jazz 80) | 2 × 20 W | 2 × 18 W | Not published for the Jazz 80 |
| Matched quartet price | ~€200–400 | ~€120–250 | ~€200–250 |
9.2 On the sonic description of the valves
The usual descriptions (KT88 "powerful", EL34 "warm", KT77 "middle ground") are kept in this report as review literature, not as a technical criterion. There is no Jazz 80 measurement with each valve set that allows these descriptions to be verified.
For an audit report, datasheet data are more useful: maximum dissipation, transconductance, maximum voltage, current, bias requirements and plate power.
10. Market and alternatives
10.1 Prices observed in Spain (October 2026)
| Version | Official RRP | Promotional price observed |
|---|---|---|
| Jazz 80 EL34 | €1,890 – €1,990 | €1,323 (Novomusica, −30 %) |
| Jazz 80 KT88 | €1,990 – €2,198 | €1,390 (Novomusica, ≈ −30 %) |
10.2 Market alternatives
| Model | Type | Power | Bluetooth | Approx. price |
|---|---|---|---|---|
| Cayin Jazz 80 (KT88) | Push-pull valves | 2×40 W | LDAC + aptX HD | €1,990–2,198 |
| Cayin CS-55A | Push-pull valves | 2×40 W | No | ~€1,500 |
| xDuoo A35 | Push-pull valves | 2×45 W | aptX HD | ~€1,339 |
| Denon PMA-1700NE | Transistors | 2×70 W | No | ~€1,500 |
| PrimaLuna EVO 300 (valves) | Push-pull valves | 2×42 W | No | ~€4,500–4,900 |
| PrimaLuna EVO 300 Hybrid | Hybrid (MOSFET) | 2×100 W | No | ~€6,300–7,500 |
| VTL IT-85 | Valves | 2×60 W | No | €8,690 |
EXPERIMENTAL PROTOCOLS — §11 to §15
11. DIY cable — experimental protocol
11.1 Description of the proposed cable
| Component | Specification | Evidence |
|---|---|---|
| Mains filter | Schaffner FN9244B-3-06 (IEC C14, 3 A, 250 V AC, 13.45 mH, Cx 0.1 µF) | CONFIRMED |
| Ferrites | 2× Würth 742721 (impedance ~59 Ω at 100 MHz) | CONFIRMED |
| Cable | 2.5 mm², copper strands | OBSERVED |
| IEC connector | Schurter C13 (female) at the amplifier end | OBSERVED |
11.2 Proposed mechanism (HYPOTHESIS)
RF/EMI noise travelling over the mains could degrade the SNR through non-linear mechanisms in the amplifier's active stages (intermodulation in valves, demodulation in semiconductor junctions, parasitic oscillations). The ferrites would act as lossy inductances for high-frequency common-mode currents; the Schaffner filter would block both common mode and differential mode.
The magnitude of the effect depends on the noise source, the couplings, the earthing, the layout and the susceptibility of the device. Effect on SNR: ND.
11.3 Verification protocol
- First measure the factory amplifier's noise and spectrum under controlled conditions: RCA input shorted, FFT 20 Hz–20 kHz, FFT up to 200 kHz if the analyser allows it, A-weighted noise, components at 50/100 Hz and harmonics, digital spuriae.
- Record the results as a baseline.
- Install the cable and repeat the measurement under identical conditions.
- Compare the results. The intervention is only justified if the measurement demonstrates a reproducible reduction of RF/EMI components at the output or in the supply.
- Evaluate whether the change translates into a variation of integrated noise, spuriae, IMD or THD+N. If those magnitudes do not change significantly, there is no technical basis for attributing an objective improvement.
12. Bluetooth module disconnection — experimental protocol
12.1 Approach (HYPOTHESIS)
The Bluetooth module is made up of the Qualcomm QCC5125 chipset and the ESS Sabre ES9018K2M DAC, housed in a shielded metal sub-compartment. The shielding mitigates much of the radiated interference, but the module remains connected to the amplifier's power supply. The digital circuits generate high-frequency noise that, in theory, could couple to the analogue stages. Effect on SNR: ND.
12.2 Verification protocol
- RCA input shorted (baseline): FFT 20 Hz–20 kHz, FFT up to at least 200 kHz, A-weighted noise, components at 50/100 Hz, mains harmonics, digital spuriae.
- Bluetooth connected but not playing: repeat measurements.
- Bluetooth playing digital silence: repeat measurements.
- Bluetooth electrically disconnected (according to the manufacturer's service procedure): repeat measurements.
- Repeat with analogue source to rule out the effect coming from the signal source.
This protocol would allow checking whether the QCC5125/ESS module is injecting noise and to what magnitude. Since Cayin already uses a metal compartment, the hypothesis of a large improvement by disconnection must be considered low priority until it is first demonstrated that a problem exists.
12.3 Warning
13. Combined improvement — protocol
13.1 Protocol
- First measure the factory amplifier's SNR under controlled conditions.
- Modify one variable at a time (cable first, then Bluetooth) and measure again.
- Compare the results under identical load, bandwidth and level conditions.
- Do not assume any improvement based solely on theoretical plausibility.
The combined value is not an arithmetic sum nor a simple logarithmic correction. The real improvement will depend on the respective noise powers and their correlation.
13.2 Indicative comparison (not measured)
| Aspect | Factory Jazz 80 | Modified Jazz 80 |
|---|---|---|
| SNR at 1 W (4 Ω) | ≈ 86 dBA (conditional) | ND (not measured) |
| RF/EMI filtering | Metal chassis + BT module shielding | Configuration with Schaffner filter + ferrites (effect not measured) |
| Internal digital noise | Present (BT module connected) | Potentially reduced if disconnected (effect not measured) |
14. KT77 and B+ — state of the evidence
14.1 Bias adjustment with KT77
The Jazz 80 has a manual bias adjustment system assisted by VU meter. The general procedure, documented by the manufacturer for the factory valves, consists of switching the central switch to bias mode, selecting the valve with switches A1/A2 (left channel) or A3/A4 (right channel), turning the corresponding trimmer until the needle aligns with the centre mark, and repeating for the four valves.
This system allows equalising the four valves to the meter's reference point. However, the VU meter is calibrated for the factory valves and does not indicate the real current nor the correct bias value for valves of another type. To know the real cathode current, it would be necessary to measure the voltage across the cathode resistors (if they exist and have a known value) with a multimeter.
14.2 Plate voltage (B+): not published
| Parameter | Status | Evidence |
|---|---|---|
| Jazz 80 B+ | Not published in the consulted sources | ND |
| Consumption at rest | 148 W | MEASURED-EXT |
| Measured maximum consumption | 231 W | MEASURED-EXT |
| Declared maximum consumption | 280 W | CONFIRMED |
14.3 Bias for KT77: no documented value
Cayin does not publish a specific bias value for KT77 in the Jazz 80. The values recommended for KT77 in technical forums vary widely (30–45 mA) depending on the amplifier, plate voltage and operating mode. There is no specific recommendation for the Jazz 80.
The official manual lists as replacement valves KT88 → KT88 / 6550 / 6550A / KT90 and EL34 → EL34 / 6CA7 / 6L6 / 5881. The KT77 does not appear in that list. That does not prove electrical incompatibility, but it confirms that there is no manufacturer recommendation for KT77 in this amplifier.
I_K = V_K / R. Note that the cathode current is not identical to the plate current: it includes the screen grid current.14.4 Safety
The interior of the amplifier contains potentially lethal voltages (on the order of hundreds of volts in the power supply) even after disconnecting it from the mains. The electrolytic capacitors can remain charged for an indefinite time.
14.5 Protocol before installing KT77
Before installing KT77s, at least the following should be known:
- Real B+ (plate voltage).
- Screen voltage.
- Grid voltage.
- Cathode current.
- Cathode resistor, if present.
- Real trimmer range.
- Plate dissipation.
- Thermal stability.
- THD before and after.
- Absence of red-plating.
Solid protocol: measure the device with the original valves → document the operating point → install the new set → measure again → check current, dissipation, stability and distortion. Without that data, the correct recommendation is DO NOT DECIDE YET.
15. Other modifications
15.1 Methodological classification
| Type | Modifications | Status |
|---|---|---|
| A. Verifiable | Replacement of out-of-spec components; documented bias adjustment; EMI reduction when a measurement demonstrates it exists. | Verifiable with measurement |
| B. Potentially audible but not quantified | Tube rolling; mechanical isolation. | Not quantifiable without specific measurement |
| C. Without demonstrable benefit without prior measurement | "Premium" coupling capacitors; silver wiring; supply electrolytics without measuring prior ripple; feet with attributed dB values. | AUDIOPHILE PROPOSAL — S3 |
15.2 Type B: potentially audible (no dB value assigned)
Tube rolling. Changing the 12AX7 for a low-noise, low-microphony one (Tung-Sol, Gold Lion, Mullard NOS) could reduce background noise. For the 12AU7s (driver/phase splitter), a matched pair improves push-pull symmetry. The effect is potentially audible but not quantifiable in dB without specific measurement. Remember: with new valves, trimmers A1–A4 must be brought to minimum before the first power-up.
Mechanical isolation. Valves are microphonic. The Jazz 80's factory feet already include damping. The improvement with alternative feet is potential but not quantifiable without specific measurement.
15.3 Type C: audiophile proposals without demonstrated benefit
The following modifications are usually cited in forums, but there is no measurement demonstrating that they produce an audible improvement in the Jazz 80:
- Replacement of coupling capacitors with "premium" models (Mundorf, Jupiter, Jantzen, ClarityCap). The electrical improvement in ESR/DF is measurable on the bench, but has not been shown to be audible in this amplifier.
- Silver or OFC copper internal wiring. The electrical improvement is marginal and not quantified in audio.
- Replacement of supply electrolytics with higher-capacity versions without first measuring the real ripple.
COMPARISONS — §16 to §18
16. Cayin Jazz 80 vs PrimaLuna EVO 300
16.1 Comparison (OBSERVED + MEASURED-EXT, not a like-for-like comparison)
fairaudio describes the PrimaLuna EVO 300 (valve version, EL34) as an amplifier with more bass presence and greater dynamic authority than the Cayin Jazz 80. The same review describes the Jazz 80 as neutral in bass, with correct integration but without extreme extension, and with solid mid-level macro-dynamics.
| Aspect | Cayin Jazz 80 | PrimaLuna EVO 300 (valves) |
|---|---|---|
| Price (EU) | €1,990–2,198 (KT88 RRP) | ~€4,500–4,900 |
| Total amplifier weight | 16.8 kg | 31 kg |
| Power (EL34, UL) | 2 × 36 W | 2 × 42 W |
| Power (EL34, TR) | 2 × 18 W | 2 × 24 W |
| Output transformers | EI | Wideband, toroidal core, resin-encapsulated |
| Power supply | One toroidal + 1H/180mA choke filter | Dual monophonic (two toroidals) |
| Auto-bias | Manual (with VU meters) | Adaptive AutoBias (manufacturer terminology) |
16.2 Technical factors: strict separation of data / hypothesis / ND
| Factor | PrimaLuna EVO 300 | Cayin Jazz 80 |
|---|---|---|
| Output impedance (Zout) | MEASURED-EXT 2.86–3.65 Ω in UL; 2.28–2.68 Ω in TR (HiFi News). | ND — not declared. |
| Damping factor (DF) | CALCULATED ≈ 2.2–2.8 at 8 Ω. | ND — without declared Zout, it cannot be calculated. |
| B+ (anode/screen voltage) | MEASURED-EXT 417 V (HiFi News). | ND. |
| Power at <2 % THD | MEASURED-EXT 55 W (HiFi News). | ND. |
| Output transformers | Toroidal, resin-encapsulated. | EI. |
| Power supply | Dual monophonic (two toroidals, one per channel). | One toroidal + choke filter. |
| Total weight | 31 kg. | 16.8 kg. |
16.3 Why the PrimaLuna can sound more authoritative without a single identifiable cause
The EVO 300 has more laboratory data available that allows it to be characterised precisely. But none of that data by itself demonstrates the observed listening difference. The possible causes include:
- Measured Zout (2.86–3.65 Ω) interacting with the varying load impedance.
- Dual monophonic supply (two toroidals, one per channel).
- Greater weight and iron volume, whose specific attribution is ND.
- Bias point and measured B+ (417 V).
- Behaviour near clipping.
- A combination of all of them, or others not considered.
16.4 Comparison conclusion
The PrimaLuna EVO 300 (valve version) is an amplifier of a higher range than the Cayin Jazz 80 in price, weight, power and construction complexity. The reviews describe a sound with greater authority in bass and dynamics. The Jazz 80 offers, at less than half the price, a different sonic balance, with integrated Bluetooth and DAC connectivity that the PrimaLuna lacks.
17. Cayin Jazz 80 vs Cayin CS-55A (predecessor)
17.1 Context
The Jazz 80 is presented as an evolution of the CS-55A, the brand's best-selling amplifier. The Son-Vidéo review states that the Jazz 80 "takes the key elements of the best-selling CS-55A in an even more seductive new version" and that it "uses the same amplification" as its predecessor. The specifications and laboratory data reveal differences that go beyond aesthetics and connectivity.
17.2 Specification comparison
| Parameter | Cayin CS-55A (KT88) | Cayin Jazz 80 (KT88) |
|---|---|---|
| Nominal power (UL, 8 Ω) | 2 × 40 W | 2 × 40 W |
| Power (Triode, 8 Ω) | 2 × 22 W | 2 × 20 W |
| THD (1 kHz) | 1 % (power unspecified) | 0.2 % (at 1 W) |
| S/N ratio | 92 dB | 90 dB (A) (condition unspecified) |
| Frequency response | 5 Hz – 48 kHz (−3 dB) | 8 Hz – 60 kHz (−3 dB) |
| Weight | 17 kg | 16.8 kg |
| Dimensions (W × D × H) | 360 × 334 × 180 mm | 300 × 385 × 190 mm |
| Declared maximum consumption | 270 W | 280 W |
| EL34/KT88 valve change | Rear panel switch | Not documented |
| Wired digital inputs | USB-B (32-bit / 384 kHz) | None |
| Bluetooth | No | Yes (LDAC, aptX HD) |
| MM phono | Yes | No |
| VU meters | Single central instrument (bias) | Two VU meters (level + bias) |
17.3 What the spec sheets publish about power
The CS-55A publishes 40 W in ultralinear and, on another line, THD 1 % (1 kHz) without indicating at what power. The Jazz 80 publishes 40 W in ultralinear and 0.2 % at 1 W. AUDIO/stereoplay measured 26 W/8 Ω and 34 W/4 Ω at 3 % THD on the Jazz 80 (from 4 Ω tap); there is no equivalent measurement for the CS-55A.
17.4 Weight, transformers and harmonic signature
The almost identical weight (17 kg and 16.8 kg) does not prove that the transformers are equivalent, but nor is there data to attribute differences to NFB, bias or iron size. The declared consumptions (270 W and 280 W) are similar. To explain any difference, the P–THD curve and the Zout of both would be needed.
The AUDIO/stereoplay measurements show H2 dominant in the Jazz 80. There is no harmonic spectrum for the CS-55A, so its signature cannot be compared. The descriptions "warm" or "analytical" are kept as listening language from the reviews, not as a diagnosis of unmeasured harmonics.
17.5 Comparison conclusion
Jazz 80 and CS-55A share architecture (push-pull, valves, similar weight and declared consumption). The Jazz 80 adds LDAC Bluetooth, DAC (Bluetooth only) and dual-function VU meters; the CS-55A includes USB, MM phono and an EL34/KT88 switch. The tuning differences (NFB, bias, Zout, P@1 %) are not measured in either of the two.
18. Cayin Jazz 80 vs xDuoo A35
18.1 Context: direct price competitor
The xDuoo A35 is an integrated valve amplifier that competes directly with the Cayin Jazz 80 in the ~€1,300–1,400 price range. It shares a similar proposition: an EL34 push-pull stage, an integrated ESS ES9018K2M DAC and high-resolution Bluetooth.
18.2 Specification comparison
| Parameter | Cayin Jazz 80 (EL34) | xDuoo A35 |
|---|---|---|
| Power valves | 4× EL34 | 4× EL34 |
| Preamp valves | 2× 12AX7 + 2× 12AU7 | 1× 12AX7 + 2× 12AU7 |
| Nominal power (UL, 8 Ω) | 2 × 36 W | 2 × 45 W |
| Specified THD | 0.2 % (at 1 W) | THD+N ≤ 1 % (power unspecified) |
| S/N ratio | 90 dB (A) (condition unspecified) | 93 dBA |
| Frequency response | 8 Hz – 60 kHz (−3 dB) | 20 Hz – 60 kHz (±1 dB) |
| Declared Zout | ND | ND / secondary data pending confirmation |
| Total weight | 16.8 kg | 20 kg |
| Maximum consumption | 231 W measured (KT88) / 280 W declared | ~232 W declared |
| Bluetooth | 5.0 (LDAC, aptX HD) | 5.1 (aptX HD, SBC, AAC) |
| DAC | ESS ES9018K2M | ESS ES9018K2M |
| Price (EU) | €1,890–2,198 (RRP; EL34 from €1,890, KT88 from €1,990) | ~€1,339 |
18.3 On the A35's declared Zout
In the sources consulted for this audit, a published Zout figure for the xDuoo A35 cannot be confirmed. Therefore, the A35's Zout is treated as ND / secondary data pending confirmation.
If a low Zout were confirmed, it would be one of the most interesting data points of the comparison: a low Zout in a valve push-pull is usually associated with a significant amount of global negative feedback. But that inference cannot be used while the data point is not verified.
18.4 Weight and preamp valve
The A35 weighs 3.2 kg more than the Jazz 80. That difference is physically real, but does not allow identifying its origin: it could come from transformers, chassis, heatsinks, boards, components, shields or mechanics. The weight is compatible with a heavier construction, but does not allow quantifying the size or capacity of its transformers.
The A35 uses 1× 12AX7 for preamp, while the Jazz 80 uses 2× 12AX7. Since the 12AX7 contains two triodes, this difference suggests different philosophies (simpler preamp stage in the A35 vs. possibly one valve per channel in the Jazz 80), but does not by itself establish a quality hierarchy.
18.5 Comparison conclusion
xDuoo A35 and Jazz 80 compete on price and share a proposition: EL34 push-pull, ES9018K2M DAC and Bluetooth. The A35 declares 45 W/channel versus the 36 W/channel of the Jazz 80 EL34, a nominal difference of 25 %. Without a comparable P–THD curve, we do not know what difference exists at 1 % THD. The A35's Zout is pending confirmation. Any difference in "muscle" remains a HYPOTHESIS until comparable measurements are available.
CLOSING — §19 to §21
19. Limits of inference — what this audit does NOT claim
This audit does NOT allow claiming that:
- the Jazz 80 delivers 40 W at 1 % THD;
- it delivers 26/34 W at 1 % THD;
- its Zout is high or low;
- its damping factor can be determined without measuring Zout;
- the 1 H / 180 mA choke is responsible for the 46 W of complex music power;
- the 46 W of complex music power demonstrate the amplifier's current capability: the impedance associated with the data point is documentarily discrepant (3 Ω box / 2 Ω text) and has not been resolved;
- the Jazz 80's B+ is equivalent to that of another amplifier using KT88;
- the KT77 are electrically interchangeable without verifying the operating point;
- the EMI filter produces an audible improvement;
- disconnecting Bluetooth necessarily produces a noise improvement;
- the weight or size of the transformer by itself allows inferring greater energy capacity;
- a subjective difference with respect to the PrimaLuna can be attributed to a single electrical variable;
- a particular brand of capacitor, resistor or cable produces a sonic improvement without a controlled test;
- the manufacturing plant or its commercial relationship with another brand determines the product's quality.
Specific limit of the comparisons
This audit does not compare the Jazz 80 with the PrimaLuna EVO 300, the CS-55A or the xDuoo A35 under equal experimental conditions. The comparisons are, in most cases, between a measured data point and an unavailable data point, not between two comparable measurements.
Any claim about "better" or "worse" performance between these amplifiers must be understood as a description from reviews or as a difference in published specifications, not as the result of a direct bench comparison.
20. Conclusions
What has been measured, what has been calculated and what remains to be measured
The report relies on three clearly differentiated layers of evidence: Cayin's factory specifications, the independent measurements by AUDIO/stereoplay (11/2023, Lothar Brandt) and HiFi News (the latter on the PrimaLuna, not on the Jazz 80), and the author's own estimates, always labelled as CALCULATED or CONDITIONAL CALCULATION.
The five unknowns that remain explicitly open are: P@1 % THD, output impedance (Zout), B+ voltage, supply ripple and quiescent current / operating point of the installed valves. None can be resolved without specific instrumentation or access to the service schematic.
Product profile and positioning
The Jazz 80 is an integrated valve amplifier with reasonably documented functional architecture, a sufficiently informative set of external measurements (although corresponding to the KT88 unit) and a feature set (LDAC Bluetooth, integrated DAC, triode/ultralinear modes, bias adjustment with VU meters) that is unusual in its price range. The manufacturer declares THD of 0.2 % at 1 W/1 kHz. The measured power at 3 % THD (26 W/8 Ω and 34 W/4 Ω from 4 Ω tap) is below the nominal 40 W, but the mathematical difference is 1.87 dB and does not constitute a homogeneous experimental comparison.
Compared to the CS-55A, the Jazz 80 gains LDAC Bluetooth and dual-function VU meters, and loses USB, MM phono and an EL34/KT88 switch. The spec sheets give 40 W in ultralinear for both and similar declared consumptions.
Compared to the PrimaLuna EVO 300, the reviews describe more authority in bass and dynamics, and the Jazz 80 costs less than half. The PrimaLuna measures Zout 2.86–3.65 Ω and weighs 31 kg versus 16.8 kg; the Jazz 80's Zout is ND, so the cause of the difference cannot yet be attributed.
Compared to the xDuoo A35 (45 W nominal, 20 kg, similar price), the Jazz 80 offers LDAC and dual-function VU meters. The A35's Zout is pending confirmation, and any dynamic advantage is a hypothesis until comparable measurements are available.
Priority of the pending measurements
| Priority | Magnitude | Utility |
|---|---|---|
| S0 | Zout (20 Hz–20 kHz, 4 Ω and 8 Ω outputs, UL and TR) | Allows DF, speaker interaction and direct comparison with PrimaLuna. |
| S0 | P–THD curve (4 Ω and 8 Ω; 20 Hz, 1 kHz, 10 kHz) | Determines P@1 %, P@3 % and the progression towards clipping. |
| S0 | B+ and ripple | Allows the choke and the supply to be genuinely evaluated. |
| S0 | Response over a representative load of Crossover A v2.4 MR-A-01 — Z(f) + phase, measured at 4 Ω tap and at 8 Ω tap | Directly connects Zout → impedance interaction → frequency response. The phase of Z(f) is as critical as its magnitude: an amplifier with feedback can behave very differently on 5 Ω with a −60° phase than on purely resistive 5 Ω. |
| S0 — prior | Z(f) + phase(f) of Crossover A v2.4 MR-A-01 | Without this curve, a representative load cannot be built for the S0 test. Sequence: measure the Z(f)/phase of Crossover A → build representative load → measure Jazz 80 on 4 Ω and 8 Ω → compare response → evaluate deviation. |
| S1 | Real bias current; operating point of the installed valves | Needed to characterise the operating point and for tube rolling. |
| S2 | Noise FFT; THD vs frequency | Better characterises dynamic and load behaviour. |
| S3 | EMI cable; ferrites; capacitors; tube rolling; wiring; feet | Only makes sense after identifying a measurable problem. |
Tap selection as an experimental variable: Crossover A does not present an impedance that corresponds exactly to either 4 Ω or 8 Ω (historical reference: 5.3 Ω nominal, minimum ~4.9 Ω). The tap selection is not a trivial data point: ideally, the S0 test should measure both taps and compare results. If only one had to be started with, the 4 Ω tap is the pragmatic choice for proximity to the real impedance, but that does not make it "the correct one" in a technical sense.
Method conclusion: the next quality leap for the report is not adding more "premium" components, but obtaining three numbers from the real device: Zout, P–THD and B+ with ripple, plus Crossover A's Z(f)/phase curve. With them, the comparison can be closed rigorously.
Final synthesis
The improvement in this audit does not consist of adding data where it is missing, but of making explicit what we know, how we know it and what we cannot yet know.
There is no lack of general information about the Jazz 80. What is lacking is specifically decisive information to determine its electrical interaction with Crossover A v2.4 MR-A-01.
The available external measurements show 26 W/8 Ω and 34 W/4 Ω at 3 % THD (from 4 Ω tap), plus 33 W/8 Ω and 46 W of complex music power, but this last figure shows an internal inconsistency in the original source regarding the load impedance: 3 Ω in the laboratory box and 2 Ω in the body of the article. The discrepancy is preserved unresolved.
For the decision about Crossover A, P@1 % THD, Zout and response over complex load Z(f)/phase remain critical. Without them it is not possible to determine whether the Jazz 80's output impedance introduces a relevant deviation nor whether its dynamic reserve allows the established listening objective to be satisfied.
DO NOT DECIDE YET. Not for lack of general information about the Jazz 80, but because the three experimental magnitudes that really resolve the question are missing: Zout → P@1 % THD → response with realistic load Z(f)/phase.
DO NOT DECIDE YET21. Primary sources and documentation
| Source | Type | Link |
|---|---|---|
| AUDIO/stereoplay, 11/2023 Lothar Brandt | Primary source for the laboratory measurements. Jazz 80 review and laboratory box. Contains the internal 3 Ω / 2 Ω inconsistency in the 46 W. PDF hosted on cayin.com. | en.cayin.cn (Jazz 80, AUDIO review) |
| Cayin Jazz 80 official manual | Primary source for architecture and operation. Documents AB1-Push-Pull-Verstärkung, valves, bias procedure, safety instruction of trimmers at minimum with new valves, and replacement list (KT77 not included). | en.cayin.cn (Jazz 80 manual) |
| 7Review.com | Secondary reproduction of the AUDIO/stereoplay review. Does not constitute a second independent measurement. | 7review.com / Cayin Jazz 80 |
| Cayin (manufacturer) | Official Jazz 80 specifications | en.cayin.cn (Jazz 80) |
| fairaudio.de | Review with sonic description and comparison (no measurement bench) | fairaudio.de / Cayin Jazz 80 |
| HiFi News | PrimaLuna EVO 300 Lab Report (Zout, B+, power) | hifinews.com / PrimaLuna EVO 300 Lab Report |
| Son-Vidéo.com | Review and prices (consulted on homepage) | son-video.com (homepage) |
| Novomusica | Spanish distributor of the Jazz 80 (consulted on homepage) | novomusica.com (homepage) |
| PrimaLuna (manufacturer) | EVO 300 specifications | primaluna.nl |
| ShenzhenAudio | xDuoo A35 spec sheet (consulted on homepage) | shenzhenaudio.com (homepage) |
| Audiophonics | xDuoo A35 spec sheet (consulted on homepage) | audiophonics.fr (homepage) |
| Genalex / New Sensor | KT77 Gold Lion datasheet | genalexaudio.com |
| TE Connectivity (Schaffner) | FN9244B-3-06 datasheet | te.com |
| Würth Elektronik | 742721 datasheet | we-online.com |
Note on the method of this audit itself
This document separates what is specified, what is measured by third parties, what is calculated, what is conditional, what is hypothetical and what is pending measurement. The reader can use the evidence criteria of §2 as a tool to read the rest of the report with appropriate scepticism.
No statement in the report should be read as a category higher than that of its source evidence. In particular, a CONDITIONAL CALCULATION data point is not a CONFIRMED data point, and a HYPOTHESIS is not a demonstrated cause.
Discrepancies preserved explicitly: the 2 Ω / 3 Ω inconsistency in the 46 W of complex music power is preserved as an internal discrepancy of the primary source (AUDIO/stereoplay), without being resolved by editorial inference. The discrepancies between manufacturer specification and external measurement (S/N 90 vs 100 dBA, consumption 280 vs 231 W) are presented as magnitudes of different origin, not as contradictions.
Calculators — SPL and Zout variation
SPL and required power calculator — Jazz 80
Formulas: SPL_1W@d = sens − 20·log10(d) and P = 10^((target SPL − SPL_1W@d − factor)/10). Results per channel.
