The World's first all valve 5.1 surround sound amplifier!

300 Watt RMS 6 channels, mercury rectifiers and triode regulated HT supplies.

Featuring microprocessor control and monitoring with LCD display and full remote control!


Why?

Why did I spend nearly a year of my life designing and building this gargantuan amplifier?

Simple really. I wanted a 5.1 surround amplifier that was all valve and could supply copious amounts of power, high quality 'valve' sound and look nice as well. I searched the Internet, Google, eBay and everywhere else and it appeared that no such amplifier existed. So I decided to build the worlds first all valve 5.1 surround amplifier, and incorporate some nice modern touches like remote control and intelligent monitoring and control. This site showcases the result!

Was it worth it? I believe it was. I love this amp' and hope that others can appreciate it even if just from the pictures on the site!

Power supply.

The starting point for the power supply was a redundant 3KVA 110v site transformer. It was an old wooden boxed one dating from about 1960 and was begging to be recycled rather than thrown away! Once removed from its box it was obvious that this was a high quality varnish dipped transformer. It was carefully dismantled and once all the laminations had been removed it was unwound to reveal the turns ratio the designers had used. This turned out to be 1.8 TPV. The original wire was very heavy gauge and is now waiting to be taken to the scrap dealer.

A new former was made using the plastic bits remaining from the original one, and new cheeks were made from hardboard. The primary was wound first using the original turns ratio. This was then insulated and a copper screen added. Then the secondaries went on. The HT first, 500-0-500v tapped at 300 and 250v 1.5A. This was followed by the bias winding 88v @ 20ma. Then the 12v 3A and 5v 10A windings for the control circuits. Finally the heater windings went on; four lots of 6.3v @ 25A and two 2.5v @ 10A for the rectifier heaters.

The transformer was then assembled with the original laminations and tested. It proved to be very quiet. Silent with no load, and only the slightest mechanical hum at full load. It was tested at full load for six hours and the core temperature rose to 37 degrees Centigrade during that time. I was more than happy with this performance and it was put to one side ready for use in the new amplifier.s PSU.

The chassis for the PSU came next. This was made from 15mm steel angle welded together to form the rather odd shape required to accommodate the transformer, chokes and the myriad big electrolytics. Finally it was given a few coats of blue Hammerite paint.

The 560v choke was made next, and this along with an old 8H one acquired from eBay were mounted onto the chassis. The capacitors followed, 16 high voltage electrolytics and one paper in oil were fixed to the chassis. This was followed by adding the control board. Originally this was made from veroboard and used 800v triacs for controlling the various outputs. This proved to be a bad move though. The veroboard had a habit of flashing over and after repairing it three times I redesigned it using big relays and point to point wiring for everything except the bias supply and the low voltage (+5v and +12v) regulated supplies. Finally a large heatsink was added to the top of the control box and the regulators were bolted to the bottom of it through the lid of the enclosure.

The power supply was then bench tested for a day and found to be reliable. It was sprayed with protective lacquer and put aside until needed.

Chassis.

The chassis was designed using the trial download of the rhinoceros V4.0 cad program from www.rhino3d.com this program is very versatile and allowed me to play around with the layout until I had something that looked balanced and .right.. Once I had the layout sorted I printed out the top and front panels full size and used these as templates to make the real thing from 1.5mm aluminium sheet. The main chassis was made from 16mm square steel tube welded together and painted with the same blue Hammerite used for the PSU chassis. The top, bottom and sides are bolted to the steel chassis with countersunk 3mm screws for which the chassis was drilled and tapped every 100mm around each perimeter.

HT Rectification and Regulation.

The rectifiers are 866a mercury rectifiers. I spent a long time researching these as there are all sorts of alleged dangers with using them, especially in an exposed design such as this where they form a sort of focal point for the whole look of the amp. My main concerns were X-radiation and UV radiation. I wasn.t particularly worried about the fact that they contain mercury as in my house this doesn.t present any risk. There are no children or pets around and so they lead a quiet unmolested life, the mercury content may present a risk if you have small people or animals likely to cause a breakage risk though. However in this case you can substitute 3B28.s which are Xenon gas rectifiers which are clearly safer as they contain no mercury at all. They don.t look quite as nice in operation as the 866a in my opinion though as they have a dimmer purple glow and I like the bright blue of the 866.s In this circuit the 866.s are operating at around 300v 500ma. From my research it appears they are unlikely to produce any X-radiation at this voltage. They are rated up to 10KV and the radiation increases with voltage. The danger of UV radiation at this voltage is also minimal, to quote one source .you get 100 times more UV radiation from watching a single lightning strike than from an evening staring at an 866a.. This coupled with the fact that my UV exposure meter wouldn.t even register when put right up against their envelope convinced me that the risk was practically non-existent.

The main HT regulation circuits were inspired by the .Optimised valve voltage regulator. published by Morgan Jones in his excellent book entitled simply .Valve Amplifiers.. Apart from a couple of component changes the circuit is almost a direct copy of his. The major change was to use 6c33c triodes as the main series regulators rather than the 6080.s specified in Morgan.s design. These valves are actually designed to be used as voltage regulators although if you look on eBay there are a lot of people selling them as audio triodes! But crucially they are BIG and can handle a lot of power. Also they are quite cheap, I bought mine from eastern European sellers on eBay, strangely they seem almost extinct in the western world. The circuit is basically a comparator and uses 85a2 neon regulator tubes as the reference.

There are two regulators, one for the 250v line and one for the 300v, these are used to power the input and drive circuits, along with the channel amplifier output stages. The 560v HT is unregulated and only powers the output stage of the bass amp.

The only other point worthy of note about the regulators is that they are very noisy! The regulator part of the chassis is completely screened from the rest of the amplifier. Until I installed these screens there was lots of switching noise being picked up by the main amplifiers, but simple earthed copper clad board enclosing the entire underside of the chassis containing the regulators did the trick.

Channel amplifiers.

The channel amplifiers, of which there are five use two cathode biased EL34.s in ultralinear push pull. I played about with many different designs for this and eventually settled on a simple classic. The output transformer is an off the shelf V78A01F. HT is only 300v but as I don.t require a lot of power from these amplifiers this works fine. The HT is regulated and each amplifier produces around 30W RMS. This is plenty for the purposes of its intended use. The drive and phase splitter is an ECC83 connected as an AC long tailed pair. The input stage is one half of another ECC83 connected as a classic common cathode amplification stage, the other half being used as the drive for the EM84 VU indicator.

Bass amplifier.

The bass amplifier uses four KT88.s in parallel ultralinear push pull. Developing around 150W RMS. This amp. was interesting to design as the frequency response is only required to extend to 500Hz. This meant that I could play around with the design of the output transformer a bit .. I used a stack of grain oriented silicon steel E I cores from my local transformer supplier and wound it without any interleaving of the windings as I wasn.t worried about HF response. I used the highest number of turns I could fit in the window with wire of an adequate gauge to carry the quiescent current of 120ma per side. Add 40% UL taps and job done. For a quick and dirty the resulting transformer has surprisingly good response with the lower -3dB point sitting at around 16Hz. The upper one was around 3KHz.

The driver stage is an ECC81 biased hot and connected as a cathode follower, so no gain, buplenty of current to drive those greedy KT88.s.

The phase splitter is an ECC83, exactly the same as the channel amps, likewise the input stage and EM84 driver.

Microprocessor monitoring and control.

The heart of the control system is a PIC16F877 microcontroller. This was chosen for its abundance of i/o pins. The code was all written in assembly and compiled using the MPLAB IDE software; I used the Picstart plus programmer. The main criterion for the controller was to:-

  • Monitor the power switch and instigate turning the thing on and off.
  • Provide an intelligent HT delay to the rectifiers, in other words the longer the amp is powered down the longer the delay next time it is powered up, this is to protect the 866a's. One thing you really must do is ensure that the vapour pressure is up before you apply the HT!
  • Provide a soft start to allow the reservoir caps to charge slowly!
  • Provide constant monitoring of all the output valves and shutdown the offending amp if any go out of limits.
  • Provide constant monitoring of the HT rails and shutdown the whole amp if they go out of limits
  • Provide a visual indication of the bias conditions of the KT88.s allowing for manual biasing.
  • Provide an interface for the remote control receiver.
  • Drive an LCD display to indicate the status to the user.

This was all accomplished over many nights using a test board I made for the purpose with the LCD connected and LED.s, pots and switches representing the various inputs and outputs. The program runs to some 2900 lines of code! A pair of 4067 analogue switches were used to multiplex the A to D input, giving 32 analogue inputs.

A second microcontroller PIC16F84 was used for the remote receiver. I wanted the receiver to be a .learning. receiver so that I could use any remote control with it. I found an interesting article http://www.pjrc.com/tech/mp3/ircontroller/index.html and used this as the starting point for the receiver. I had to heavily modify the code to do what we want, but my thanks go to Todd Elliot for his article which provided all the building blocks! There is a 9 pin D-SUB socket on the front of the amplifiers control .pod. which connects to the PC.s serial port allowing me to program the remote receiver as necessary.

There are only four remote instructions; Power, mute, Volume up and Volume down. That seems adequate though and if I wanted to extend it there are a further 28 outputs available from the receiver.

Control Pod.

Mounted underneath the main amplifier is the control pod. This houses individual level controls for each input as well as a switch to select .maintenance mode., six gold phono input sockets and the four recessed multi-turn trimmers for setting the bias on the KT88.s. There is also a headphone socket and the 9 pin serial connector mentioned above. The maintenance mode switch, when operated, re-powers any failed channels and displays the bias level for each KT88.

Valve Line-up.

  • Two 866a Rectifiers
  • Two 6c33c Power Triodes
  • Four KT88 Output Pentodes
  • Ten EL34 Output Pentodes
  • Fourteen ECC83 Double Triodes
  • One ECC81 Double Triode
  • Four EF91 Signal Pentodes
  • Six EM84 Indicator Tubes
  • Two 85a2 Voltage Reference Tubes
  • Two EL84 Output Pentodes (Headphone Amplifiers)

Interesting facts.

  • The amp uses a total of 47 valves.
  • It draws a cool 760W from the mains whilst at low volume settings.
  • At full power it pulls just over 1300W input power!
  • Structural damage is a distinct possibility when used at high volume!
  • It sounds absolutely awesome when listening to action films with Dolby Digital.
  • It produces an incredible bass experience when listening to modern dance music
  • It is able to delicately and faithfully reproduce each and every instrument when listening to classical music.
  • It has a lovely 'warm' sound when listening to traditional Jazz
  • It makes an extremely efficient room heater for use in winter time!
  • Bearing in mind all the above it does everything a good amp should do and I am satisfied that my decision to design and build the worlds first 5.1 all valve amplifier was well worth all the work that went into it .
Return To HOME page.
Copyright © Bob Howlett 2009