"WZMAK" - That is, an amplifier on 6P6S, ECC82 and EF86

Pltin74 10 Apr 2025 22:46 39 4992 Cool? (+24)

📢 Listen (AI):

TL;DR

A custom tube amplifier built around 6P6S, ECC82 and EF86 evolved from an intended Ampex 2012 clone into a fully original design.
Star grounding, a capacitance multiplier instead of a choke, and a corrected negative-feedback resistor solved hum and restored lost output power.
The project took almost 4 years, and it adds UL/Pentode switching, an input selector, and GZ34-or-silicon rectification.
Pentode mode sounds slightly better than UL, delivers more power, and adds a tube-like compression when listening in full-tube mode.
A slight 50 Hz hum remains in tube-rectifier mode, likely from a tiring GZ34 after one sparking incident.

Generated by the language model.

Hello everyone!

I come to you with a project that dragged on for me for almost 4 years - a tube amplifier, which was my school of tube electronics. During this time I made a lot of mistakes, learned a lot and today I can say with a clear conscience that "I already know something". 😊

Initially, this was to be a clone of the schematic-based Ampex 2012, but it soon became apparent that I did not yet know at that time how the input pentode behaved. Its gain was too high in this circuit, which caused problems - the original design was designed for low input amplitudes.

The tubes, transformers and components were already bought, so something of my own had to be designed.

There was also hum and not much power per channel. I dealt with the former by:
- ground in a star topology, very carefully done,
- and using a capacitance multiplier instead of a heavy and expensive choke.

As for the power - the problem was trivial: a badly chosen resistance of the negative feedback loop. After correction - the power was back in place.

⸻

What did I add?
- UL/Pentode mode switch - gives you the opportunity to experiment with the character of the sound.
- Input selector
- Switch between tube rectifier (GZ34) and silicon bridge - the circuit completely disconnects the GZ34, so it doesn't run and wear out.

⸻

Technical extras:

I also used quite a few parts from the 3D printer - the largest being the transformer covers.

I've tried to keep the interior aesthetically pleasing in a spatial style - I'm a big fan of spatial assembly and I think it looks and works great.

⸻

Impressions
- Looks? In my opinion - excellent.
- Sound? Even better.

In my opinion, the pentode mode beats the UL by a whisker - it gives more power and a characteristic "tube hoarseness". The sound is further 'compressed' by voltage drops on the GZ34, which adds to the atmosphere, but only when listening in 'full tube' mode.

⸻

Concluding remark

I've recently noticed a slight 50 Hz hum in tube rectifier mode - I suspect my GZ34 is running out (I've had an incident of sparking in the bulb), although it may also be a matter of some sort of ground loop - but not likely, as the sound is typical of mains frequencies.

So on a day-to-day basis I listen on silicon, but I still plan to play around with a "pseudo rectifier tube" - i.e. two diodes + a power resistor.

⸻

Photos:

📸

Tube audio amplifier on a workbench in a workshop.

Dark photograph of a tube amplifier with glowing tubes visible.

Tube amplifier with visible tubes and transformers on a workbench.

Tube amplifier with various tubes and transformers on a wooden and metal enclosure.

Rear panel of an amplifier with various connectors and a power socket.

Interior of a tube amplifier in a wooden case with red and blue wires.

About Author

Pltin74 wrote 143 posts with rating 132 . Live in city Tęgoborze. Been with us since 2019 year.

Comments

Add a comment

marweg1967 11 Apr 2025 01:05

Congratulations on a successful build! Four years maybe, but you're happy with the sound effect too, and you've learned a lot in the process. And that's what it's all about. I'm guessing that the output... [Read more]

Pltin74 11 Apr 2025 12:38

yes it is, these are transformers from Mr Ogonowski [Read more]

acctr 11 Apr 2025 13:31

. It was probably the gain that was too low and not the power? [Read more]

rosomak19 11 Apr 2025 16:31

I too have been lurking lately , to make something more than my test 2xECL86 SE . From my resources I have dug out 4 PCC88 tubes and 4 PCL805 . Too bad I'm missing one PL500 for the PP. On the PCC88 maybe... [Read more]

szeryf3 11 Apr 2025 17:42

Congratulations on the design and construction. I like the underside very much. Don't give up on the rectifier tube. Were the transformers specifically chosen for this project? The night photo knocks... [Read more]

dudik56 11 Apr 2025 20:01

Nice design and nice assembly. The overvoltage appearing in the rectifier lamp could be from too high a voltage but it could also be an individual characteristic of the lamp. I once had such adventures... [Read more]

cirrostrato 12 Apr 2025 10:53

These lamps on Allegro now almost at the price of a bottle of natural sourdough for Christmas borscht, I have a lot, zero interest although prices tiny. [Read more]

rosomak19 12 Apr 2025 15:59

So I might ask , is it worth putting together a clone of such a Meloman 25 ? How did the original compare in terms of sound ? [Read more]

Pltin74 12 Apr 2025 21:57

exactly, too little gain to drive a load with 12 watts from the datasheet ;) . bb2168d2d Added after 6 [minutes]: it's a pity that PCC88 is not the same as ECC88 just for a different voltage, but... [Read more]

kris8888 12 Apr 2025 22:27

Why the input selector on the back? Could you throw in a diagram of this amp as I can see quite a few differences from the original. [Read more]

Pltin74 13 Apr 2025 11:09

this is entirely my design, in the sense that it was originally going to be Ampex but however I redesigned it my way :p [Read more]

rosomak19 13 Apr 2025 13:30

. Apart from the series glow , what is the difference between PCC and ECC ? The same as the PCL86 from the ECL 86 . Maybe in the tide of time I will assemble some preamp with tone control on these PCC88s... [Read more]

żarówka rtęciowa 13 Apr 2025 13:46

Hello For dual triodes, the differences are only in the glow parameters. The ECL86 and PCL86 tubes also have different characteristics due to the supply voltages of the circuit for which they were... [Read more]

kris8888 13 Apr 2025 20:17

. This is where you are wrong, these are identical tubes in terms of characteristics, they only differ in their glow parameters. Perhaps you are confusing this with a pair of PL84 and EL84? Because... [Read more]

gulson 13 Apr 2025 20:34

Very nice construction, thanks for the presentations! How long did it take to build? Write me a PW with the Parcel Post and I'll send a small gift. [Read more]

Pltin74 13 Apr 2025 22:38

Thanks a lot! After redesigning without looking for any more shortcomings I think with all parts about a month. A bit long, but that's because I'm actually doing everything myself, apart from the case,... [Read more]

Rocky Horror 14 Apr 2025 04:38

Strange this resting current control. The potentiometer sliders are connected in parallel. No matter how you spin them, the voltage on the first grids of both tubes - more or less - will still be the same.... [Read more]

kris8888 14 Apr 2025 09:58

In my opinion, the diagram probably contains errors. I don't know what the component values are, but according to this diagram, the V2A lamp of the phase inverter should not work at all. I also don't... [Read more]

Pltin74 14 Apr 2025 13:06

fact, I had forgotten about this capacitor, but it seems to me that the cause of the 50hz is different purely visual, I wanted the front to be minimalist ;) . Added after 6 [minutes]:3298872ee... [Read more]

FAQ

TL;DR: This DIY tube amp took 4 years to develop, then about 1 month to build after redesign, and its creator says "the sound? Even better". This FAQ helps builders of 6P6S/ECC82/EF86 push-pull amplifiers solve hum, bias, feedback, rectifier, and transformer-choice questions fast. [#21514922]

Why it matters: The thread shows how small choices in grounding, feedback, rectifier mode, and bias filtering can decide whether a tube amplifier sounds powerful and quiet or hums at 50 Hz.

Topic	Option 1	Option 2	Practical effect
Output stage mode	UL	Pentode	Pentode gave more power and more "tube hoarseness" to the builder.
Rectification	GZ34 tube	Silicon bridge	Silicon was quieter in daily use; GZ34 added voltage drop and character.
PSU smoothing	Choke	Capacitance multiplier	The builder chose the multiplier to avoid a heavy, expensive choke.
Input selector location	Front panel	Rear panel	Rear placement kept the front visually minimalist.

Key insight: The biggest lesson is that hum and lost output power did not come from one dramatic flaw. They came from small implementation details: feedback resistor value, star-ground execution, and bias-supply filtering.

Quick Facts

The amplifier evolved from an Ampex 2012 idea into a custom push-pull design using 6P6S, ECC82, and EF86, after the original input pentode gain proved too high for the intended source levels. [#21514922]
Output transformers are commercial units from Mr. Ogonowski, confirmed by the builder when asked whether the amplifier is push-pull. [#21515365]
The reported lost power was traced to a wrongly chosen negative-feedback resistor; after correction, the builder says output returned to the expected level, citing 12 W from the datasheet discussion. [#21517044]
The remaining hum is 50 Hz and is mainly heard in tube-rectifier mode; after adding an extra capacitor and returning that ground to the star point, the hum was reduced but still audible very close to the speakers. [#21520337]
The 3D-printed transformer covers use PET-G, with 3 mm walls and quoted softening around 80°C; the builder reports no visible deformation in use. [#21520337]

How did the design evolve from an Ampex 2012 clone into a custom 6P6S, ECC82 and EF86 tube amplifier?

It started as an Ampex 2012 clone, then became a custom design when the builder found the input pentode gain too high for the intended signal levels. He had already bought the tubes, transformers, and parts, so he redesigned the amplifier around those components instead of forcing the original schematic. He later confirmed that the final circuit was entirely his own design rather than a close Ampex copy. [#21517470]

What causes 50 Hz hum in a tube amplifier with a GZ34 rectifier, and how do star grounding and power-supply filtering help reduce it?

50 Hz hum here came from power-supply grounding and filtering details, not from the audio path alone. The builder first reduced hum with careful star grounding and a capacitance multiplier. Later, he identified another issue: the multisection capacitor shared ground with the amplifier section on the PCB before returning to the star point. After adding an extra capacitor with its return taken to the star, hum dropped, but remained faintly audible in GZ34 mode when standing very close to the speakers. [#21520337]

What's the difference between UL and pentode mode in a 6P6S push-pull amplifier, and how does each mode affect sound and power?

In this build, pentode mode gave more output power and a stronger tube character than UL mode. The builder described pentode mode as having more power and a slight "tube hoarseness," while UL sounded a little more restrained. He also noted that the effect becomes more pronounced in full-tube operation, because the GZ34 adds voltage drop that further compresses the sound. [#21514922]

How do you choose the negative feedback resistor value in a tube amplifier when the gain is too low to reach the expected output power?

Choose the feedback resistor so the amplifier keeps enough closed-loop gain to reach the target output. In this project, the builder lost power because the negative-feedback resistor value was wrong. After changing it, the amplifier recovered the expected performance. He later clarified that the real issue was insufficient gain to drive the stage to about 12 W from the datasheet. That makes the feedback network a power-limiting element when set too aggressively. [#21517044]

What is a star ground in tube amplifier construction, and why is it used to fight hum?

"Star ground" is a grounding layout that brings separate return currents to one common point, reducing shared impedance and hum coupling. The builder used a very carefully executed star topology to fight mains-frequency noise. That matters because shared ground paths can let power-supply ripple enter sensitive input or bias circuits. In this thread, moving returns toward the star point was part of the successful hum-reduction strategy. [#21514922]

What is a capacitance multiplier in a tube amp power supply, and when is it used instead of a choke?

"Capacitance multiplier" is a power-supply filter stage that uses an active device to imitate a much larger capacitor, lowering ripple without a large inductor. The builder chose it instead of a choke because a choke would have been heavy and expensive. In this amplifier, that substitution was part of the solution for hum, alongside careful star grounding. [#21514922]

Which output transformers were used in this build, and how suitable are Mr. Ogonowski transformers for a 6P6S push-pull amplifier?

This amplifier uses push-pull output transformers from Mr. Ogonowski. The builder confirmed both the push-pull topology and the transformer source directly. The thread does not give primary impedance, bandwidth, or power rating, so it does not prove exact electrical matching. It does show practical suitability: the amplifier was completed, ran, and delivered the expected output after feedback correction. [#21515365]

Why would someone place the input selector on the back panel of an amplifier instead of the front?

The builder put the input selector on the back purely for appearance. He wanted a minimalist front panel and said the rear selector does not get in the way during use. That choice trades some convenience and labeling clarity for a cleaner faceplate. Other users questioned the ergonomics, but the builder kept it because the visual result matched his design goal. [#21518736]

How long does it typically take to build a DIY tube amplifier like this one once the design is finalized and all parts are ready?

In this case, the finished redesign took about 1 month to build once the design was settled and all parts were available. The builder also said the whole project stretched across almost 4 years because he used it as a learning process and repeatedly corrected mistakes. That gives a realistic split: weeks for assembly, years if the design is still evolving. [#21518155]

How should the bias supply for Grid 1 be filtered in a fixed-bias tube amplifier to avoid injecting AC ripple and hum?

Filter the Grid 1 bias supply with a capacitor to ground and return that ground cleanly to the star point. One reviewer noted missing filtering after diode D2 or near resistors R19 and R20 would leave large AC voltage on the bias network. The builder later confirmed the working design already had those capacitors and further reduced hum by adding another capacitor with a direct return to the star. 1. Add the filter capacitor. 2. Keep the bias return separate. 3. Tie it at the star ground. [#21520337]

What is a phase inverter in a push-pull tube amplifier, and why does its biasing method matter for long-term stability?

"Phase inverter" is a driver stage that creates two equal, opposite-phase signals for a push-pull output pair, so both output tubes share the waveform symmetrically. In this thread, one user accepted the stage would work with the shown values but warned that fixed phase-inverter bias is less forgiving than autobias. He pointed to aging high-value resistors and tube wear as long-term risks, because drift can push one half toward cutoff and raise distortion. [#21518834]

How do PCC88 and ECC88 differ, and how is that comparison different from ECL86 versus PCL86 or EL84 versus PL84?

PCC88 and ECC88 differ only in heater parameters, not in the triode system itself. One poster stated that for dual triodes, the differences are only in the glow parameters. Another user added that ECL86 and PCL86 are also identical in characteristics and differ only in heater requirements, while EL84 and PL84 are a genuinely different pair electrically. So PCC/ECC and ECL/PCL are heater variants, but EL84/PL84 is not the same kind of substitution. [#21518055]

What should you check before using 3D-printed PET-G transformer covers inside a hot tube amplifier chassis?

Check the material, its softening temperature, wall thickness, and whether the part visibly deforms in service. The builder used PET-G, not PLA, and stated that PLA softens around 50°C while PET-G softens around 80°C. He also used 3 mm walls and reported no distortion after use. That makes PET-G more appropriate here, but only if the cover stays below its thermal limit in the real chassis. [#21520337]

How do you calculate the right output transformer impedance for PCL805 tubes in push-pull class AB, and when is 5 kOhm suitable?

Use the tube curves and the class AB operating point, then check whether the load line reaches the intended maximum power. The builder answered that directly when asked about a 5 kOhm / 15 W speaker transformer for PCL805 in push-pull. He did not give a blanket yes. He said the correct method is to plot the curves, choose the operating point, and verify that the chosen impedance lets the pentode deliver the target power without violating the selected conditions. [#21521271]

What is a 'pseudo rectifier tube' made from diodes plus a power resistor, and how does it compare with a real GZ34 and a silicon bridge?

A pseudo rectifier tube is a diode rectifier with a series power resistor added to imitate some voltage drop of a real tube rectifier. The builder mentioned it as a future experiment: "two diodes + a power resistor." In this amplifier, the real GZ34 added sonic character through voltage sag, while the silicon bridge was quieter in daily use and did not wear like a tube. A pseudo rectifier aims for a middle ground between those two behaviors. [#21514922]

Generated by the language model.