logo elektroda
logo elektroda
X
logo elektroda
Advanced Search
logo elektroda

Class A amplifier, full Dual Mono

Maciek_C 11286 34

TL;DR

  • A custom full Dual Mono Class A integrated amplifier combines a phono preamp, Bluetooth, DAC, control logic, and speaker protection in a Modushop 1NPD02300B enclosure.
  • The phono stage uses a FET-input circuit inspired by Pioneer and Denon designs, and the power amplifiers are redesigned from the Pioneer A-09 with current, inexpensive transistors.
  • The build took seven months, and the DAC uses Cirrus Logic CS4334 converters with DIR9001 SPDIF decoders and a 4052 source selector.
  • Startup was mostly straightforward: the analog section needed only offset and bias adjustment, and the finished amplifier sounded clearly better than the earlier GainClone and Vincent.
  • The 30 cm-deep case left very little space, the heatsinks run around 75 degrees, the subtitles are too dim, and the DIR9001 needed delayed Reset timing to start correctly.
Generated by the language model.
ADVERTISEMENT
Treść została przetłumaczona polish » english Zobacz oryginalną wersję tematu
Report a violation of the law
Reply Cool? Ranking DIY | New topic
Notify about new articles
📢 Listen (AI):
  • #31 20687839
    Maciek_C
    Level 15  
    Posts: 249
    Help: 14
    Rate: 254
    Hi,
    ad. 1. Potentiometer R9 on the power amplifier board,
    ad. 2. I set the offset so that "cold" has about -100mV, after heating it has +20mV,
    ad. 3. Loaded with 6 ohm speakers, after an hour of moderate play,
    ad. 4 and 5. The given values were calculated (with a margin for drops in the bridge, pulsations, etc.), then checked in practice (several watts more),
    ad. 6 and 7. I was playing it safe - the transformers are very close to the transistors of the voltage amplifier, and since JFETs work there, I was afraid of hum. Steel sheet, because ferromagnetic, brass on top to improve the visual experience (flatter the ego),
    ad. 8. About 35dB - the value of the feedback resistors of the preamplifier and power stage was changed relative to the diagram,
    ad. 9 and 10. The amplifier with the connected generator was loaded with resistors made of resistance wire (4 and 8 ohms), the voltage value was measured at the output (power control), and the shape of the voltage was checked on the emitter resistors (indirectly, the current). I assumed the shape of the emitter current as the limit value, which, at the peak of the sinusoid, did not drop to zero yet, but was already visibly deformed (cross distortion) - this occurred when the emitter current decreased to about 40mA.

    Ps. Thanks again my friend aaantek for helping me half a year ago (of course thanks also to my friend jozgo) :)
  • ADVERTISEMENT
  • #32 20687870
    aaanteka
    Level 42  
    Posts: 7683
    Help: 728
    Rate: 1186
    Maciek_C wrote:
    ad. 3. Loaded with 6 ohm speakers, after an hour of moderate play,

    I was also concerned about the ambient temperature. And since it's class A, there should be no difference with and without load. As one of my colleagues mentioned earlier.

    Maciek_C wrote:
    ad. 4 and 5. The given values were calculated (with a margin for drops in the post, pulsations, etc.), then checked in practice (several watts more),

    Practically it's like: - for the whole band, for 1kHz on what load(s)?

    Maciek_C wrote:
    the transformers are very close to the transistors of the voltage amplifier, and since JFETs work there, I was afraid of hum. Steel sheet, because ferromagnetic

    Have you thought about the magnetic permeability of steel at 50 Hz?

    Maciek_C wrote:
    and the shape of the voltage (indirectly the current) was checked on the emitter resistors. I assumed the shape of the emitter current as the limit value, which, at the peak of the sinusoid, did not drop to zero yet, but was already visibly deformed (cross distortion) - this occurred when the emitter current decreased to about 40mA.

    Wait a minute, but how does this relate to going from A to AB? Cross distortions concern class B - no quiescent current or its insufficient value.
    Did you use the transistors in the diagram, including the TIP3055/TIP2955 power transistors?
  • ADVERTISEMENT
  • #33 20687891
    Maciek_C
    Level 15  
    Posts: 249
    Help: 14
    Rate: 254
    aaanteka wrote:
    Maciek_C wrote:
    ad. 3. Loaded with 6 ohm speakers, after an hour of moderate play,
    I was also concerned about the ambient temperature. And since it's class A, there should be no difference with and without load. As one of my colleagues mentioned earlier.

    Room, approx. 23 degrees.

    aaanteka wrote:
    Maciek_C wrote:
    ad. 4 and 5. The given values were calculated (with a margin for drops in the post, pulsations, etc.), then checked in practice (several watts more),
    Practically it's like: - for the whole band, for 1kHz on what load(s)?

    1kHz, load as in point 9 and 10.

    aaanteka wrote:
    Maciek_C wrote:
    the transformers are very close to the transistors of the voltage amplifier, and since JFETs work there, I was afraid of hum. Steel sheet, because ferromagnetic
    Have you thought about the magnetic permeability of steel at 50Hz?

    NO.

    aaanteka wrote:
    Maciek_C wrote:
    and the shape of the voltage (indirectly the current) was checked on the emitter resistors. I assumed the shape of the emitter current as the limit value, which, at the peak of the sinusoid, did not drop to zero yet, but was already visibly deformed (cross distortion) - this occurred when the emitter current decreased to about 40mA.
    Wait a minute, but how does this relate to going from A to AB? Cross distortions concern class B - no quiescent current or its insufficient value.


    Maybe I named them wrong. My point is that near the transistor plugging area, the voltage across the emitter resistor was starting to deform noticeably.

    aaanteka wrote:
    Did you use the transistors in the diagram, including the TIP3055/TIP2955 power transistors?

    Yes.
  • ADVERTISEMENT
  • #34 20688000
    OTLamp
    Tube devices specialist  
    Posts: 2336
    Help: 334
    Rate: 504
    In general, the moment of transition to class B can be detected by measuring the current drawn by the output stage from the power supply. As long as it works in class A, the current consumed practically does not depend on the drive, after switching to class B it starts to increase with the drive.
  • #35 20688019
    Maciek_C
    Level 15  
    Posts: 249
    Help: 14
    Rate: 254
    Great tip, hadn't thought of that, thank you :)
Reply Cool? Ranking DIY | New topic
Notify about new articles
📢 Listen (AI):
Report a violation of the law

Topic summary

✨ The discussion revolves around the design and construction of a Class A amplifier with a full dual mono configuration. The creator initially used a Koda AV-600X receiver enclosure and aimed for a design similar to the Gamut DI 150 amplifier. Participants praised the aesthetic and technical aspects of the build, discussing various methods for achieving effective backlighting and visibility of inscriptions on the front panel. Concerns were raised about the heat management of the transformers and the quiescent current settings, with suggestions for optimizing performance and reducing heat generation. The creator confirmed the use of custom-designed PCBs and shared details about the amplifier's specifications, including power output and component choices. The conversation also touched on the potential for future improvements and measurements related to distortion and performance.
Generated by the language model.

FAQ

TL;DR: With 2x28V DC rails and a builder-verified "about 32W at 8Ω" in pure Class A, this FAQ helps DIY audio builders understand a full Dual Mono amp with modular boards, relay-switched digital power, and a DAC reset bug that blocked startup until Reset was delayed. [#20673453]

Why it matters: This build shows how to combine Class A sound, upgradeable module design, and practical troubleshooting in one custom integrated amplifier.

Alternative What it achieved Main drawback reported
PCB with black soldermask under glass Labels stayed hidden when off Passed too little light
Etched thin laminate + smoked plexiglass Proven workable by multiple users Needs darker cover material
Tinted glass Most "pro" finish and cleaner surface Can reduce IR range
Laser-printed film on foil Simple and cheap Print could show through

Key insight: The strongest lesson is thermal and mechanical planning. The electronics started easily, but the 30 cm chassis and undersized heatsinks pushed temperatures to about 75°C and made the build much harder than expected.

Quick Facts

  • Power rails were 2x28V DC, with 375 mA per branch and 1.5 A total quiescent current per channel. The builder calculated 2x35 W into 8Ω and 2x70 W into 4Ω, then checked that practice gave a few watts more. [#20662842]
  • Pure Class A operation was not full-scale. The builder later estimated about 32 W into 8Ω and 18 W into 4Ω before the stage moved into AB operation. [#20673453]
  • Thermal behavior was the main hardware limit. The chassis reached about 75°C after 1 hour of moderate playback into 6Ω speakers in a 23°C room. [#20687891]
  • The front-panel insert measured 74 x 154 mm. Suggested label solutions included etched 0.3-0.6 mm laminate, smoked plexiglass, tinted glass, and window-tint film. [#20663210]
  • The control section used an Atmega16, and the volume control was an Alps RK271 2x50k potentiometer. Spare JLCPCB boards, Gerbers, Eagle files, and MCU code were offered privately. [#20661024]

What does a full Dual Mono amplifier design mean, and what advantages does it bring in a Class A build?

A full Dual Mono design means the left and right channels are built as largely separate signal and power paths. In this build, each channel has its own mirrored audio supply board, and even the DAC works in Dual Mono. That separation reduces shared supply interaction, simplifies channel-by-channel servicing, and fits the builder’s modular upgrade goal in a hot Class A amplifier. [#20661024]

How was this Class A amplifier divided into separate modules, and why does that make future upgrades easier?

It was split into separate boards for input selector and phono stage, preamp, power amp, DAC, control, standby and soft-start, audio power supply, and speaker protection. The builder explicitly said the design uses separate module plates so parts can be upgraded or replaced easily. That modular layout also let him mirror one channel and keep digital, control, and audio sections physically distinct. [#20661024]

Why did the DIR9001 DAC chip fail to start correctly, and how do you delay the Reset pin relative to the power supply?

The DIR9001 failed because Reset came too early relative to the supply voltage. The builder said he underestimated the need to delay the chip’s Reset pin, which cost him several hours of debugging. He fixed it by adding an inertial delay circuit on Reset, so power stabilized first and the SPDIF receiver then started correctly. [#20661024]

How do you determine when a Class A amplifier based on the Pioneer A-09 topology leaves pure Class A and starts working in AB?

You determine it by watching output-stage current behavior as drive rises. 1. Load the amp with 4Ω or 8Ω resistors and feed a sine wave. 2. Measure voltage across the emitter resistors while increasing output. 3. Mark the point where emitter current at the sine peak no longer stays clear of zero and begins to deform. The builder used that method and estimated the transition near 32 W at 8Ω and 18 W at 4Ω. [#20687839]

What output power can this 2x28V DC amplifier deliver into 8Ω and 4Ω, and how were those values calculated and checked?

It can deliver about 2x35 W into 8Ω and 2x70 W into 4Ω. The builder said those values were first calculated with margin for bridge drops and ripple, then verified practically and came out a few watts higher. He also stated the amplifier runs from 2x28V DC rails with 375 mA quiescent current per branch. [#20687839]

How do you adjust DC offset in this power amplifier, and why does the offset drift with temperature until the amp is fully warmed up?

You adjust DC offset with potentiometer R9 on the power-amplifier board. The builder set the amplifier to about -100 mV when cold, which moved to about +20 mV after warm-up. The drift matters because this Class A design runs very hot, so transistor operating points shift with temperature until the chassis and output stage stabilize. [#20687839]

What is DC offset in a power amplifier, and why is offset drift a concern in a hot-running Class A design?

"DC offset" is a power-amplifier output condition that leaves a steady DC voltage at the speaker terminals, instead of a zero-centered audio-only signal, and its key characteristic is that heat can shift this voltage as transistor conditions change. In this build, offset moved from about -100 mV cold to +20 mV hot, so the builder had to set it on a fully warmed amplifier. [#20687839]

Which transistors were used instead of the original Pioneer A-09 parts, and how well do TIP3055 and TIP2955 fit this redesign?

The builder confirmed he used the transistors shown in his redesign, including TIP3055 and TIP2955 as the output devices. He also stated the Pioneer A-09-based stage was redrawn around cheap, currently available parts. In practice, the analog section started immediately and only needed current and offset adjustment, which suggests the TIP3055/TIP2955 pair fit the redesign functionally. [#20687891]

What is SPDIF, and how was it routed here between the Bluetooth module, 4052 multiplexer, DIR9001 receivers, and CS4334 DACs?

"SPDIF" is a digital audio interface that carries stereo PCM data as a serial bitstream, and its key characteristic is that one source can be selected, decoded, and then converted to analog by a separate DAC stage. Here, SPDIF from Bluetooth, coaxial, or optical first goes into a 4052 multiplexer, then through optocouplers to separate DIR9001 decoders for each channel, and finally into CS4334 converters. [#20661024]

Why was the DAC and Bluetooth supply switched by relay only for digital inputs, and what noise or power benefits does that give?

The DAC and Bluetooth supply was relay-switched so those circuits only receive power when a digital input is selected. The builder said the digital section’s secondary side is powered from the audio supply through a relay, and when analog sources play, no voltage is applied to DAC or Bluetooth. That reduces unnecessary powered digital activity inside the chassis and avoids keeping extra sections energized all the time. [#20661024]

What are the best ways to make illuminated front-panel inscriptions visible only when lit, using PCB soldermask, etched laminate, smoked plexiglass, tinted glass, or laser-printed film?

The best thread-backed options were etched thin laminate behind dark cover material or darker tinted glass. The builder’s PCB with black soldermask hid labels well when off, but blocked too much light. Other users recommended 0.3-0.6 mm etched laminate with smoked plexiglass, laser-negative print on film, or darker glass such as Antisol, because those methods preserve the hidden-until-lit effect better than weak tint. [#20663469]

Tinted glass vs smoked plexiglass vs car-window film: which works best for hiding IR receiver cutouts and front-panel labels in an amplifier?

Tinted glass gave the most professional finish in this discussion. One experienced user said foil on plexiglass kept producing bubbles, scratches, and artifacts, then switched to cut and polished tinted glass with better visual results. The main limit was IR: with 6 mm black Antisol glass, the remote range dropped to about 3 m, so darker hiding can cost remote performance. [#20663469]

How hot is too hot for a Class A amplifier heatsink, and what case size or radiator changes would help if the chassis reaches about 75°C?

About 75°C was high enough here that the builder regretted the chassis choice and wanted larger heatsinks. He used a 30 cm deep case and later said he wished he had chosen a 3U enclosure with bigger radiators. In practice, that means the thermal design was workable but undersized, especially for a 1.5 A per-channel quiescent current Class A build. [#20661024]

How do soft-start and speaker protection work in an Atmega16-controlled amplifier with standby power and relay-switched transformers?

The Atmega16 control board runs from a standby supply and manages transformer switching, soft-start, and speaker protection relays. The builder said the control section handles backlight, buttons, and power switching, while the speaker-protection board is supervised by the microcontroller and mounts directly at the speaker terminals. That setup lets the amp power up in a controlled sequence and disconnect speakers under fault conditions. [#20661024]

What model was the volume potentiometer in this build, and where can I get the preamplifier PCB files, Gerbers, or spare JLCPCB boards?

The volume potentiometer was an Alps RK271 2x50k. The builder said he still had extra JLCPCB boards left, including 3 power-amplifier and selector boards and 4 of the other boards, because the minimum order was 5 pieces. He also offered to share Gerbers, Eagle files, and microcontroller code, asking interested people to contact him by private message. [#20672089]
Generated by the language model.
ADVERTISEMENT