Megatron Electrostatic Headphone Amplifier

[JoaMat]

Here is my solution for filament transformer… 

 …mounted on bottom case plate, belonged to a now retired Blue Hawaii. I also found a piece of Bergquist Gap Pad for the transformer to rest on.

It seems that cathode resistor voltage is much lower for Emission Labs tube than for EL34, so I decreased the negative high voltage from 460 V to 400 V.

 The lowered voltage called for change of Cathode resistor and trimmer, now 402R resistor and 100R trimmer. A few things had to be removed before component changes.

[JoaMat]

This Emission Labs amplifier wasn’t planned, it just happened along the way. Although it seems to work, I have no idea how to best use the direct heated triodes in an electrostatic headphone amplifier. 

Here are maximum conditions stated in datasheet from Emission Labs with note 6.

 
Note6) Maximum conditions do not apply simultaniously. This is not a working point.

I think I’m within most of the maximum limits, but what how about Power Output in Class A: 8W and Grid resistor: max 470K? 

Is there anything to consider in my case? Now I’m running at 400V/27mA roughly 11W and grid resistors are 680K as in the original Megatron. Any advice and comments are appreciated.

Edited December 20, 2021 by JoaMat

[JoaMat]

i'm done

[Kerry]

That’s awesome!  Wonderful work Kevin  

[JoaMat]

On 11/26/2021 at 5:29 PM, audiostar said:

JoaMat,

question on umbilicals (or better the connections though both) - are those absolutely identical for both channels?
Probably shouldn't, as Safety Earth (that must connect both Amp/PSU cases to Earth in the PSU case and to circuit/star ground as well) should be carried through one only, right, in order to avoid loops.

What about circuit ground, as lately in the Stax connector both left and right channel ground connect together, so if ground is carried through both Umbilicals, this would introduce a loop as well? 

 

Sorry for a really slow response.

Yes, the umbilicals are identical.

Some background – my first electrostatic build was a KGSShv with a PSU with on board heat sinks in its own house. Second build was the DIY T2 and I liked the chassis design. Thereafter I have only built things (almost) with off board heat sinks. With T2 I built an original T2 PSU in a Modushop case. Later I built another T2 PSU with my own designed PCBs. That are the only PSUs I’ve built, along with the first KGSShc supply. But I’ve built a lot of Kevin designed amplifiers, and all have been power by those two T2 PSUs. I found the T2 PSU with its two identical umbilicals quite convenient. With the good quality Amphenol connectors it’s easy to connect/disconnect as wanted when I want to try/test one channel at the time.

What about ground loops? I’ve read/heard about it, but so far, I haven’t, or I don’t know I have had ground loops. For me it’s more important I always have a safety ground connected. Therefor I’ll continue to use two identical umbilicals …

P.S.
I’ve had some issues with mini T2 that was solved by removing one 580V bias wire, so I probably have had a loop. problem. But safety ground wires will stay!

[audiostar]

On 1/7/2022 at 2:15 AM, JoaMat said:

Sorry for a really slow response.

Yes, the umbilicals are identical.

Some background – my first electrostatic build was a KGSShv with a PSU with on board heat sinks in its own house. Second build was the DIY T2 and I liked the chassis design. Thereafter I have only built things (almost) with off board heat sinks. With T2 I built an original T2 PSU in a Modushop case. Later I built another T2 PSU with my own designed PCBs. That are the only PSUs I’ve built, along with the first KGSShc supply. But I’ve built a lot of Kevin designed amplifiers, and all have been power by those two T2 PSUs. I found the T2 PSU with its two identical umbilicals quite convenient. With the good quality Amphenol connectors it’s easy to connect/disconnect as wanted when I want to try/test one channel at the time.

What about ground loops? I’ve read/heard about it, but so far, I haven’t, or I don’t know I have had ground loops. For me it’s more important I always have a safety ground connected. Therefor I’ll continue to use two identical umbilicals …

P.S.
I’ve had some issues with mini T2 that was solved by removing one 580V bias wire, so I probably have had a loop. problem. But safety ground wires will stay!

Thanks, JoeMat and a Happy New Year as well! 

So, you carry circuit ground through both umbilicals which you connect separately to both channels in the amp's chassis).
Safety Earth, you carry through both umbilicals as well and connect both to a single point in the amp's chassis?
Bias line you split in the PSU chassis and carry through both umbilicals as well, but connect in the amp's chassis only one of them to the actual Bias output?

In the power supply chassis you tie ground of all PSUs together and connect through a circuit breaker (or not, or simple 10R, etc) to the chassis itself & Safety Earth coming off 1) the IEC mains connector and 2) both umbilicals off the amp's chassis.

[kevin gilmore]

IT FUCKING WORKS. makes music and everything.

transformers stone cold after 1 hour.

 

[Kerry]

That's awesome!

Congrats  

[Beefy]

Holy shit, it looks glorious!

[jamesmking]

3 hours ago, kevin gilmore said:

IT FUCKING WORKS. makes music and everything.

transformers stone cold after 1 hour.

 

 

 

nicely megafixed (tm) 🙂

 

[kevin gilmore]

done and ready to ship

 

[Pars]

Very nice Kevin! I read your comments on HF as well. What does the plaque on top read?

[kevin gilmore]

you can expand the picture and read it

NOT Felista Audio

OTL Push-Pull Electronic

Headphone Amplier

 

For the record this is not a push pull amplifier.

[mwl168]

There seems to be no trim pots nor servo circuits. How is the offset and balanced adjusted? 

Edited February 2, 2022 by mwl168

[kevin gilmore]

by matching the tubes. its +/-5 volts typically.

[ibuski]

Kevin, I remember your old post telling that the best design CCS is a high voltage tube with a cathode resistor, i.e. Megatron.  It should be better theoretically than any cascoded FET or T2 style BJT CCS, I read.   Is that still true?  I'm tempted to think Megatron as my next project because it doesn't require many rare semiconductors and the PCB is available from Aliexpress while I haven't seen other Gilmore design PCB on sell.   Finding reasonable Tubes shouldn't be problem and anyway I have some already.  The amount of heat from 8 EL34 makes me a bit nervous though.

  BTW following above statement about CCS comparison, putting Megatron's CCS into T2 seems to be the ultimate amplifier, but I haven't seen anybody do so.  just because it's too hot?

[kevin gilmore]

you are going to need a pile of floating filament transformers. although now that SumR has finally produced transformers that are the right voltage and run cool this is easy. this one has regulated 300v and RCLCC high voltage supplies each with tube diodes. that is 2 more transformers. This one is 211 watts.

A standard T2 is about 250 watts. replacing the current sources with tubes will add another 80 watts. and lots more physical space. and more filament transformers

This modified to fit amplifier board will not be published and would be useless to everyone. The standard amp board should be used. The 300v regulated tube board with the tube hv diodes would be publishable once i put all the tube sockets back on standard distances. The hv power supply board would be publishable if people are interested.

both of these would only be released to people who promise not to let them fall into the hands of FUCKING INCOMPETENT MORONS.

Edited February 2, 2022 by kevin gilmore

[ibuski]

Thank you Kevin, I haven't thought about transformers.   Tube diode wasn't in my focus.   For any design of eStat amp (KGSSHV, BH, MT, GG, etc..), I think most of people now select GRHV type PSU as it's known as the best performance.  Or is there special synergy when tube PSU is paired with full tube amplifier like Megatron?   From appearance point, yes...

[kevin gilmore]

A tube diode CRC power supply driving GRHV would have slightly lower noise.

[ibuski]

3 hours ago, kevin gilmore said:

A tube diode CRC power supply driving GRHV would have slightly lower noise.

Oh really?  Then ultimate amplifier would be combination of

Tube diode CRC driving GRHV + T2 + Megatron style CCS ?

full of tubes...

[kevin gilmore]

posted for the owner who will comment much further. (for some reason he cannot publish high res pictures)

[orientexpedite]

The sound: it’s spectacular!  Compared to a Stax SRM-007tii with 60’s Mazda Brimar 6CG7’s or 50’s RCA long black plates with D getters, the sound is far more relaxed, analog, detailed, expansive, dynamic and musically involving.  I can’t wait to get some NOS EL34’s in here.  The Berkeley Reference DAC shows off the Megatron’s speed and transparency; the DAC and Megatron are both detail monsters.  The Megatron is a far more capable amp when connected to the Berkeley Reference DAC than a current low-noise, low-distortion reference amp, such as the Topping A90, which I also have.

The balanced inputs: dead quiet - perfect with the Wireworld Platinum 8 series XLR interconnects.

The voltage gain: I can’t imagine needing any more gain than I’m getting with the 1958 Holland D foil getter E180CC’s (thanks to the community for the tip on these) and the 1948 RCA 5751 triple mica black plates.  I will say that this amp takes a solid hour to warm up, however.

Temperature: just touched 126F on the exterior of the right transformer case after a couple of hours - the internal temperature was typically within a couple of degrees when previously measured, likely due to the machined vents around the tubes.  SMALL UPDATE: Kevin suggested some whisper fans to deal with the heat buildup; this is an excellent idea.

Hum: there is none whatsoever, there is only dead silence, irrespective of where the pot is positioned - it’s absolutely impossible to tell whether the Megatron is on, when powered from a current-gen PS Audio regenerator set to 120V.

Definitely T2 level stuff!

I CANNOT THANK KEVIN ENOUGH

[mwl168]

@JoaMat: I see you have been using Mundorf Silver/Gold/Oil capacitors in the coupling positions of your Megatron. These caps are rated at 600vac/1000VDC and being an oil-impregnated caps they are susceptible to high temperature (factory rating is up to 85C/185F). 

Also, the Megatron PCB shows the coupling caps between 12au7 and 12ax7 rated at 600v and 1000v for the caps between the 12ax7 and EL34. 

Have you experienced any issues with using these Mundorf caps in Megatron over time? 

Edited May 7, 2022 by mwl168

[JoaMat]

No, I haven’t noticed any issues over time. But they haven’t been used that much so I have no idea how they change at higher temperature over a longer period of time.

[dingding123]

Built a Megatron. Seems to be the perfect amp to build during the seemingly unending shortage of 10m90s.

I used up my final few pieces of 10m90s and C2M1000170D for this simple build, and it gave me some mild anxiety and regret for not hoarding up those sands.

However, the build experience and result are worth it.

 

Build process and after thoughts

I started off with a pair of generic chassis from Taobao. 32cm wide, 8cm tall, 35cm deep. Silver and black anodised aluminium. They weren't available in other finishes but this looks clean enough to me. They do have some minor scratches and dents if you look hard enough though.

The casework was done by a laser etching service provider I found on Taobao.

The end result was a clean and modern appearance.

Inside the PSU are a pair of +/-450V GRHV, +/- 15V GRLV, delay board and a 270VA toroidal transformer. The idea was to create a PSU that can be used by both the Megatron and the Grounded Grid amp units.

These components fit the chassis nicely with little room for a third GRHV board for the 300v rail for Megatron.

Outputs are split into AC and DC outputs.

Transformer specs:

Spoiler

Pri: 115x2; Sec: 363v x 0.2a x 2; 15v-0-15v x 0.2a x 1; 15v x 0.1a x 1; 12.3v x 1.5a x 1; 6v x 4a x 4

The heatsinks for the amp unit were probably redundant since most of the heat comes off from the top. The amp board could probably fit into a smaller and lighter chassis, but I chose the same chassis as the PSU just so that they can stack.

The amp build was straight forward and relatively easy compared to a Carbon build or any of the KG dynamic amp builds, due to the low component count.

I did not use any boutique components in this build. The coupling caps are Wima, and the voltage rail decoupling caps are Siemens. 

Volume pot is an EIZZ-style stepped attenuator, internal signal wirings are pure copper.

Special care was taken when routing the heater filament wires to avoid hum.

Amp testing and adjustment was simple. Just ensure the all the rails are as spec'ed and that's about it. The voltage offset settles automatically after power on. No need to fiddle with any potentiometer when the amp is powered on, so it is much safer to test than the Carbon or the T2 family.

 

Sound

I used the Megatron with ES-1A, and here is my impression of the Megatron after comparing it with GG (on the same PSU), Carbon (450V GRHV 15V GRLV) and Mini T2 (triple GRHV).

Currently I am using the Megatron with a quad of Mullard XF2 and PSVane EL34PH. The front end tubes are all Valvos.

The sound of the Megatron is warm, slightly thick in the mids, and non fatiguing and airy in the treble. It sounds very natural and the imaging is holographic with good recordings. While it is very detailed, it does not ruin joy if the recordings are of poorer quality. Older or poorly mixed recordings have a smooth timbre and organic vibe when heard from the Megatron. While the Megatron's treble is not as extended as the Carbon's, it complements the overall signature to create a very complete soundstage and convincing sense of separation and space.

What makes the Megatron standout from crowd lies in its low end - it is an amp for bassheads. It has by far the most THICC bass compared to my other amps, even more than the Carbon. The bass extends as deep as the Carbon, but the Megatron has a nice lift in the mid bass that makes my ES 1a slam extra hard. Listening to EDMs and fast pop tracks on E-Stats is finally satisfying, and I can finally stash away my planars.

Carbon is like an antithesis to the Megatron. Carbon is like a clean cutting razor with little tolerance for inaccuracies where as Megatron adds a lot of joy and flavour into the listening experience. Carbon also sounds a little linear. Although the stage is wide, it is not as deep as the Megatron's.

GG sounds more organic than Carbon in general. The GG's signature is quite dependent on the tubes. With the XF2s, GG is warm and clean, but doesn't deviate far from the Carbon sound. The bass isn't as tight and fast, and treble could sound stiff and slightly shrill with the wrong tubes - such as the re-issue Mullard EL34.

Mini T2 sounds slightly leaner and cleaner than the Megatron, and has a much smaller bass. It is as enjoyable a listening experience for me, but the Mini T2 was much harder to build - by sheer component count.

 

Tube placement and choices

Placing the EL34s in this manner shown below allows one to use two matched quads of EL34 if matched octets are unavailable.

Initially I used cheap Linlai tubes for the front end and had issues with sound imbalance and hum - turned out that the tubes were the culprits. Switched to some cheap NOS and the issues were gone. So my advice would be to use well built tubes from reputable makers.

The 12AU7s affect the sound significantly. The option to roll tubes makes Megatron extra fun to use.

 

More photos of the internals and the back of the amp coming up soon, when I am more free.

I would say that the Megatron is my favourite amp - until I build the T2.