Saturday, July 25, 2026

Update / 6B4G/302B Single Ended Amplifier.

 I'd like to thank all the people who have sent me good wishes the last year ,  its been very appreciated .

My lastest scans have shown the cancer is stable , and while I have mostly good days , its nice to hear from such nice people on the other days.

It's also been very humbling to hear how so many people are enjoying the Supratek products. I've had messages from across the globe, from New Zealand to Russia , from people who say the Suprateks have enriched their audio lives , and take great pride in owning a Supratek. 

One recent conversation was with a chap who has owned his Supratek since 2007 and is just now replacing a set of tubes - that's pretty good  going in 20 years.

He says he's tried many preamps costing many multiples the price and yet to find something as enjoyable.

That's a common theme- long term ownership that continues to deliver enjoyment from music. 

You can tell I'm quite proud , but it's also been an honour to be able to do this for so long , and to meet so many good people along the way. 

I've just started another preamp build , and while its a long process now , I hope I can make a few more people owners of Supratek preamps.






While I'm still enjoying the Lii Audio F-15/JBL2226/Raal 20-70XR ribbon OB speakers, I did have to spend a fair bit of time finding an amplifier to suit the Lii Audio F-15 driver.

This is what I finally decided on: a 6B4G Single Ended, with options to use the AVVT 302B tube.

The 6B4G tube is not the typical double plate usually used in SE amps, but the Sovtek single plate that was available back in 1990, or even earlier.

It's not a true 6B4G, it has a much bigger plate structure , possibly the 300B Sovtek plate.

It has a plate dissapation of 40 watts, much more then the double plate 15 watt maximum.

It can take a plate voltage of over 400 volts, and current of 100ma (not at the same time!)

 It has a clear, slightly warm tone , but without the tubby , somewhat congested tone that 300B tubes can suffer from.

The driver tube is also Russian, the 6E6P , a high transconductance, high gain tube in triode connection.

I like the single driver configuration when using a preamp, just have to be sure the driver tube can swing enough voltage to get full power from the output tube.

I've also set up this amp so it can use a pair of veteran AVVT 302B tubes. These are "super" 300B tubes, capable of much higher voltage and current than normal 300B tubes

Also higher power, and a cleaner, more engaging sound than typical 300B tubes, which frankly I find quite boring. ( I sold a big stash of original WE300B tubes about 30 years ago - no regrets)

To use both 6B4G and 302B I had to install a switch to change filament voltage, and adjustable fixed bias supply to suit either tubes bias requirements.

The bias supply is regulated with octal gas regulators- 75 volt OA2 for the 6B4G and 105 volt OA3 for the 302B. A simple pull and insert , adjust bias and a different amplifier!

As the 6B4G is an octal tube and the 302B is a 4 pin, I had to make a 4 pin to octal adaptor to use the 6B4G in the amp.

The driver tube has its own shunt regulater , using 6L6 tubes. 

I like regulated power supplies - they dont have the warm tubby sound of ancient tube designs you see everywhere these days , but a solid, detailed, yet natural tone that has significant depth and width .

The Lii Audio F-15 is a very fussy driver that needs a very good, very capable amp. 

LATER: 

After a couple of months of fine tuning and some minor design changes I settled on using the AVVT 302B tube exclusively in this amp. 

The extra power is handy, producing about 1.5 times the power of a 300B , but it is better tube for the sound I want from the F-15.  The 302B also lasts forever as it is just idling in this design. 
















Saturday, June 27, 2026

Taming the Lii Sound F-15 Coaxial.

 



The Lii Sound F-15 Coaxial driver must be the most perplexing driver I've ever come across.

It's sold as a full range driver, and the ears and measurement confirm it's doing a reasonable attempt at that, but its seriously rolled off both in the lower bass and upper treble.

But despite that it has a extraordinary ability to sound quite interesting with exciting dynamics that makes the sound literally jump out of the drivers. 

It sounds much more efficient than its rated dB . Its just an excited driver, extracting a lot more information than you'd expect from a 15" driver.

BUT, part of that excitement comes from a far from flat frequency response. It's nothing like the response graph provided by the factory, which is pretty much flat.

The F-15 has a huge boost and plateau from 6-8Khz , and a dip around 2Khz. 



This is hardly what you'd call hi-fi, although I suspect it might be something approaching the inverse of what a lot of older guys with limited hearing loss have.

Maybe it is a quite clever marketing ploy by Lii Song - give all the old hi fi guys a driver that restores their old ears , a sort of giant hearing aid?

Is that too skeptical ? 🤫

Nevertheless, despite it being quite interesting for a while, the brain/ear interface is too clever to be fooled for long , and the F-15 soon starts to sound a bit too detailed and eventually quite shrill.

It's too inaccurate to be called even remotely "hi-fi" , and I'd call it a novelty driver not to be taken that seriously.

Still I did recognise something in it that deserved a bit of work to see if it could be brought to the fore.

I was using a JBL 2226 15" bass driver under the F-15 for bass enforcement , but the problem of the brash upper mid/treble had to be solved.

First thing I tried was a notch filter , but that was mostly ineffective- the plateau is just too broad and high to be tamed by a notch filter , and I've never liked using them, as they can often do more harm than good .

Then I tried a low pass filter around 6Khz to see if would generate a reasonably flat response - it worked better than the notch filter, but still not the smooth response I was after.

Looking at the raw response, it really is quite hopeless and it was then I decided I had to totally remove everything above 2Khz , and use another driver to fill in the 2Khz-20Khz response.

I eventually settled on a 2Khz 2nd order low pass filter that worked satisfactorily , if not perhaps perfectly.

There's still a bit of a boost at 6Khz - it really needs a 24dB 4th order crossover, which I will build. 


Interestingly the values I used - 0.82mH and 9.4 uF are nowhere near what the crossover calculators suggest, but measurement confirmed they are reasonably close in achieving the desired roll-off.


Now I had the problem of what to use for the 2Khz to 20Khz range, and more importantly would it work with the difficult F-15.

I had some horn tweeters I could try but decided to try a pair of RAAL 70-20XR ribbon tweeters.



I was worried that the directivity of the ribbons would clash with the F-15 , especially as I was using the F-15 off axis , to avoid some undesirable beaming. 



The RAAL 70-20XR has a very good reputation , and I've used it very successfully in other speaker designs. 

Its expensive at over $1000 each , but I believe it lives up to the manufacturer's sales blurb:

This model is somewhat different from other models. The ribbon of 70-20XR is wider in comparison to the others, which is indicated by its label. Ribbon is 70mm long and 19mm wide. Housing is made out of 8mm thick steel. At the bottom of the chamber there is an adapter transformer, which is completely covered with resin. Net volume of the chamber in between the transformer and the magnet is reduced with wooden tiles to the required level and that space is very precisely filled with sound absorbing material. In this ribbon, the filling is made of finest medical cotton wool from pure cotton. The strong neodymium magnets are the same as those installed in all other models except for the 64-10.

The transformer in this model is special, also wound by hand and completely handmade. The primary is wound with a copper wire with a round section and high purity, while the secondary is being wound with copper foil of a 0,1mm thickness in three layers. The transformer is special because it has 3 outputs: 6,3 ohms, 9 ohms, and 12,6 ohms. In this way, the user can choose which sensitivity fits the best. The outputs of the transformer can have other values which can be chosen by the user himself on special request.

The next feature that sets this ribbon apart from the others is that the ribbon is made like a sandwich. It’s made out of two layers of aluminum foil with a thickness of 8 microns, which is ironed and stiffened by a very demanding process. In between the two layers of aluminum foil a special grease is extremely precisely measured and applied, which has the role of soothing the ribbon in an even shorter time after the impulse stops. This ribbon doesn’t have carrier foil made out of MYLAR, it’s in direct touch with air. This provides the smallest possible moving mass, several hundreds of times smaller than any classic DOME tweeter and far smaller than RIBBONS with a MYLAR carrier foil.

The result is shockingly rich, detailed, the sound image is unencumbered by any colorations.

The ribbon itself is fixated on the ends to the supporting frame via tensioners. After installation, the ribbon is being precisely tensioned in order to achieve optimal performance. In this model, the ribbon is the same length as in the 70-10 model. In all models, the first number indicates the effective length of the ribbon, and the second number indicates the width of the ribbon. Model 70-20 is user serviceable, which means that the user can replace the ribbon himself if necessary.

Like every other model, this one is characterized by exceptional openness, detail, spaciousness and depth of the sound image. The sound is fluffy, colorless, it will not hide or mask even the finest information read from the source.

However, I didn't really expect it to work optimally with the F-15, as they are such different drivers.

To say I was pleasantly surprised is an understatement; I was shocked at how well they integrated .

To be honest this is the most natural, immediate, and vivid vocals I've ever heard, and I've had half a century playing around with electrostatic, horns ,  planar  and just about every speaker you can imagine. 

It has the positive attributes of all these speaker types , without any negative attributes .

Excellent imaging and soundstaging , utilising the best of well mastered recordings. 

I would never have thought that the combination of a Chinese 15" co-axial driver and a Serbian ribbon tweeter would make such a grand musical sound. 

The F-15 does a superb job restricted to upper bass/midrange , and the Raal 70-20 XR fills out the treble in a splendid fashion. 

I do think that the Loftin White 6B4G/2A3 SE amp driving the Raal 70-20XR would be contributing to the beautiful, smooth sound , as it is a gorgeous sounding amp. 

Technical description: a Supratek Grange preamplifier with dual adjustable level outputs - one output for low bass and the other for upper bass, mid and treble. 

JBL 2226 15" bass driver used in open baffle configuration. Graham Maynard T-Bass OB EQ. Minidsp digital crossover 10hz 48dB high pass, 180Hz 24dB low pass. Carver solid state amplifier.

Lii Song F-15 15" co-axial driver. Passive low pass passive crossover 12dB 2Khz , reversed polarity. 300B/6B4G SE amplifier.

Raal ribbon tweeter. 10uF capacitor in conjunction with inductance of internal transformer forms a 12dB 2Khz high pass crossover , with 24dB acoustical crossover- see Raal website for more info. 6B4G Loftin White SE amplifier.

The dual outputs of preamplifier allows me to make fine adjustment of bass to mid/treble balance of each recording. 

There is no perfect setting of this balance as every recording is influenced by the quality of the recording , the mastering engineers perception , and various other influences.

I like being able to make fine and subtle adjustments to suit my system, room  and ears, rather than being stuck with one "suits all" option.

It's amazing how much difference just a slight change to the bass balance can make to the overall sound. 

So a very surprising and satisfactory result, that I'm enjoying very much. 

Always something to learn in this peculiar hobby!




                                                              JBL 2226 Bass response



                                                  JBL 2226 Bass, F-15, Raal 70-20XR








Tuesday, March 17, 2026

New Email Address - Updates

 micksupratek@gmail.com


After 25+ years of using supra@supratek.com.au as my email address , it has become cluttered with spam , which my current email provider is useless at removing.

This has led to missed and confusing emails , and I apologise if anyone has been disadvantaged by this.

I'm now using micksupratek@gmail.com as my primary audio email address.

Please take note and advise anyone you know who is in the Supratek circle.

On other news , I'm slowly building a few preamps - just finished a Dharma , starting another and then will do a couple of Granges, and a pair of Malbec power amps. This will likely take a long time, so no promises on orders. 



Dharma with NOS 6J5 tubes  and Power supply with NOS 6F6G, 5V4G tubes.

All are using genuine early 20th century NOS tubes from my tube stash , apart from rectifier tubes , which are becoming harder and harder to find.



I've also treating myself to some new power amp builds .  Designs which I've wanted to try for decades, but have been constrained by time and the demands of Supratek. 

These are the power amps I'm using in my system at the moment.

I've had a half dozen CV1076 (DA41) for at least 30 years, but never had the opportunity to do something with them.

https://www.r-type.org/exhib/aaa1214.htm

They are somewhat between a 300B and a 211 triode .  From the transmitting triode family with a thoriated tungsten filament , they have the same "lighthouse" radiation of the 211, 845 etc tubes.  They get quite hot, but look very cool! They were designed as Class B tubes for high power PA audio , but we want to use them for Class A high fidelity! 

The circuit was originally designed by a Japanese guy back in the fifties , and although not popular , quite a few Japanese and some Western tube diy guys have built modified versions of it. 

It uses a small power tube like a 6V6, 6K6, 6F6 etc as a cathode follower, with the cathode tied directly to the grid of the CV1076. 

This way it can provide a POSITIVE grid bias and deliver grid current to the CV1076 so that it can run in Class A2 . 6-8 watts .

It sounds glorious , makes 300B SE amps sound like mud. Has a real 3D soundstage, deep imaging and detailed but still very natural and organic sound. 

One of the best SE amps I've built , very happy with it. 






Of the "conventional" SE tubes, the 2A3 is probably the most renowned.  As Joe Roberts said, back in the 90's - I've never met a 2A3 amp I didn't like" .

My Loftin White 2A3 SE amp is another beautiful sounding amp- just the bees knees for female vocals and music that is not too complex and dense. 

I'll do a write up of that later, but here's a description of my 6B4G parallel push pull amp .

The 6B4G is the 6.3V version of the 2A3 - the 6B4G has 6.3V filament v 2.5V  of the 2A3. 

This amp has seen many variations as I've used it as a test mule for various driver stage designs. 

This version I've made it as simple as possible - one driver tube , an interstage transformer and the output tubes and transformer. 

The driver tube is the Russian 6E6P - a high conductance/high gain tube with low impedance that can swing a lot of drive volts . The Inter-stage transformer is 1:1+1 and provides more voltage than the 6B4G will ever require, and will easily drive the 6B4G stage into Class AB2. 

The 2A3/6B4G tubes have very high transconductance and require separate biasing for both phases of the push pull output stage.

In addition, the output transformer is a toroid design , and they dont like any imbalance between the two primary windings . With the separate cathode bias configuration, balance between the windings can be adjusted as low as 1 mA .

This amp starts to clip at 20 watts , which is far more than my high efficiency speakers need, but it has good solid bass , and the typical 2A3/6B4G sweet sound. 

Such a simple circuit gives an unencumbered , pure sound . The interstage does produce a bit of ringing , but the typical loading to reduce it does more harm than good , its minimal, and virtually inaudible. 


6B4G PPP amp. Monoblock - one channel. A bit untidy but it has been many amps over the years! 





I've had some rare and very expensive Tango X-10S output transformers for at least 40 years.  

These were most often used for 211 tube SE amps and sometimes 845's, although the 10K impedance is a bit high for 845 tubes.

I made a number of 211 amps with them,  and then they sat on the shelf for a good number of years. 

I found some very ancient VT25 tubes in my stash, that were used in aircraft radios post WW1 .

I realised that the VT25 tubes and the Tango OPT's could be used to make a very nice Class A2 6 watt SE amp , with a 3A/167 (WE437) tube driving an interstage transformer.

This build is underway, but will be a long-term project - I'll do another write up when it's finished.


                                                   Initial Circuit design VT25 SE Amp




The Loftin White circuit goes back to the 1930's , nearly a hundred years old! It was pretty much forgotten till the Japanese DIY guys started using it again, around 1970.

It's an all direct coupled circuit, no capacitors between stages. It's tricky to set up , but sounds great. 

It really works well with the 2A3 tube , highlighting the purity and beautiful tone the 2A3 is capable of. 

It needs a single high gain driver tube, the 12AX7 is often used , but the octal 6SL7 is a true audio triode and sounds much better .  The ancient 6SF5 is also a good choice.

Not much power, you're lucky to get 3 watts from it, but you'd be hard pressed to find a more intoxicating and beguiling sound.

Simple 6SL7/6B4G Loftin White SE amp . Power supply not shown.