Embedded Files
TVA-1 from Michaelson & Austin
Tim De Paravicini designed the TVA-1 ~
Tim's Valve Amplifier ~ The first ~ 1 hence , TVA-1 .
Tim De Paravicini worked for Michaelson and Austin from 1978 to 1983.
Kevin Austin and Anthony Michaelson started assembling the first amplifiers in the shop premises at this very fine address in London back in 1978.
140 New Bond St. London, UK
Mr.Anthony Michaelson, and mr. Kevin Austin owned the Michaelson & Austin company at the time where they produced valve different valve amps including
the TVA1 and TVA10.
On December 21, 2009, the author purchased the Michaelson & Austin TVA-1 shown below.
I bought this amplifier from Berlin via an auction on Ebay.de for 578 Euro and subsequently had to have transported the 40 kg of iron to Denmark.
The woman who sold the amplifier was not aware of whether the amplifier worked and she could not lift the 40 kg. I was lucky.... I persuaded UPS to ship the TVA-1 to Copenhagen.
Fortunately, the super good Partridge output transformers were intact and the amplifier played but with a lot of hum.
I didn't care... the icon was at home. From here began the work of understanding the background of Tim De Paravicini's circuit and why people talked about the TVA-1 being an unstable amplifier, but the explanation about instability turned out to be very simple.
First some photos from 2009 and then some drawings of the vital prints and a component list which took several months to collect and never existed, as well as a revised circuit that changed over time.
This is the author´s TVA-1 serial no. 553 as the amplifier arrived
Received with a soldering tower and a "bird's nest" of wires
Below is one of the circuits that started it all
TVA-1
The original TVA-1 circuit shown in Jean Hiraga's book "Rörförstärkere" 1986
A lot of information about valve amplifiers and history about the developement and all the guys who invented the circuits like Harry F. Olsson, Frank Mac Intosh, David Hafler, Harold J. Leak, Saul B. Maranz, Peter Walker, D. T. Williamson a.o.
Note the huge mains transformer
Firstly, Paravicini used the famous Partridge output transformers, and without too much knowledge of the well-kept winding technique, it looks as if Partridge has wound the coils "trifilart", with three copper wires at a time.
This way of winding gives better control of the capacity and that it increases the frequency range considerably. The size of an output transformer gives an indication of whether the transformer can reproduce deep low-frequency tones, and the TVA-1's output transformers are huge.
The TVA-1 can reproduce very low frequencies below 10 Hz! Note that the size of the mains transformer was increased during production time, as there was no negative bias tap on the first mains transformers.
Because of this missing outlet we have developed a low-impedance negative stable
power supply
But first the original data sheet from UK.
As can be seen on this erly TVA-1 this amplifier has a small mains transformer without a winding for the negativ bias. But overall the transformer was sufficiant for the heaters and the high voltages.
On the original TVA-1 schematic a 100VAC winding can be seen, but we don't know from which serial nr. this outlet was added?
The following circuit shows safe and stable low impedance negativ supply that could be used in the TVA-1...taken from a high voltage wire.....
later on a fixed negative bias with a capacitor from high voltage
The TVA-1 driver board from the solder side
The driver board from the component side
For the first time a drawing of the driverboard with wire indicators
The first TVA-1 component list ever
..if seen before it's mine....
The TVA-1 bias board
Note the original NCT in-rush limiter.....
Note ! when changed there will be 100K ~ 39K ~ and approx. (1/2 20K pot) 10K from G1 to ground!
Total within 150K from KT88 grid to ground....explanation later, see below!
The first drawing of Soundlease Services FB1 bias board from TVA-1 serial nr. 553
Note the missing bias tap on this TVA-1 serial nr. 16
For TVA-1 owners this overall wiring diagram could be useful
The picture is a bit blurry but it shows a negative bias supply on a small circuit board next to the mains transformer...the yellow wire is the output! The supply is taken from the existing 370VAC outlet on the mains transformer.
The negative bias supply has low impedance and the bias is very stable.
TVA-1 serial nr. 16 before repair
560VDC very close to the chassis!!
The driver board is hanging from some screw posts and the noval sockets are made of plastic, where I would have preferred ceramic sockets. But the sockets seems to be ok!
After I found 479VDC across the treminal of capacitor C5 , I had to use (2 x 350V in series) a 700VDC capacitor..!
Find the plate voltage for the ECC81 driver tubes on the diagram below.....
Note the C5 of 18uF reaches at least 480VDC by start-up, and there are not much space below the driver-board....so two axial 350VDC in series was used.
Note C20 in this serial nr. 16.... 470nF/630V goes to 1N5408 and delivers -224VDC for the negative bias. Note the inrush limiter after the 1N5408 bridge.
But most important...Note the bias configuration and that there are approx 150K from the KT88 grid G1 to ground.....!!
The same amplifier but TVA-1 serial nr. 553 runs on lower voltage of approx. 530VDC ...almost same performance inspite the lower voltage....C5 could be a 500VDC type!
Instability
Regarding instability in the TVA-1, there may be some talk if signal capacitors with too low voltage values have been used. But you can reduce the risk of a capacitor being exposed to grid rectification by trying to use signal capacitors with voltage values up to 1000V in the amplifier between the tube stages as a preventive measure. If a capacitor is exposed to grid rectification, this will mean that increased voltage values on the signal capacitor of up to 1000V do not move the tube's operating point to the same extent, and that the extra distortion does not occur in the amplifier stage....a distortion that makes the TVA-1 unstable. So the idea of using 1000V capacitors in an amplifier with +B of 550VDC – 600VDC may be one way forward that can provide greater stability.
Grid to ground resistor
But most importantly ...The distance of the grid resistor to ground: Another and more significant cause of instability in the TVA-1 is the grid resistors in the bias circuit! My danish hifi and PA friend (also my retired colleague) who has worked with 6550A and KT88 since the sixties, has repaired many tube-based hifi and guitar amplifiers and he has uncovered the following:
In the TVA-1 original bias circuit it is necessary to reduce the shown 330 K resistors in the bias circuit to max 100K... even down to 50K, if you have enough voltage swing from the driver tube!
The path from the KT88 grid (G1) to ground goes through the revised diagram (see jpg) of the TVA-1: 1. The grid stop resistor of 2K2 and 10K (there must be 2 resistors in series)......2. then a bias circuit resistor 100K...... 3. into the voltage divider trimpot middle leg 20K (approx 10K) ......4. 39K to ground! The Ohmic resistance from the KT88's G1 grid gives approx. 10K + 100K + 10K + 39K in total 150-160 K measured from grid to ground....try as far as possible to keep the measured value of 150K between G1 and ground (cathode)
Stability and Zeners
Stability is improved by the tube having difficulty drawing grid current when the G1 grid is closer to ground (cathode).
The high G2 voltage is a bit critical for some tube types/makes. (newer types of KT88 output tubes) Therefore, you can lower the voltage for G2 by using 60V/5W Zener diodes. Zener diodes are very low-impedance and can provide a fairly stable voltage. To show what my colleague has done to stabilize his own TVA-1, I have drawn on an overview of my TVA-1 double 5W zener diodes of 60V facing the transformer's UL outlet (opposite end to G2 (grid 2 on KT88)). Two must be used in parallel per tube as the KT 88 can draw up to 6-7 Watts. (According to the data sheet on connection).
You will have to reset the bias, as the total current draw across the cathode drops significantly. ( The screen grid voltage is a powerful booster of the tube's performance ).
Trying a zener to lower the G2 voltage is a matter of taste.... but listen and replace if the sound doesn't suit you...... that's the method! The screen grid is very uncritical as far as voltage is concerned. You can possibly shunt a capacitor across the zener. Try without it first. If you like the sound, you can continue with the shunt.
The problem...KT88
But the whole explanation for our current TVA-1 instability problems comes from one component and that is...KT88!
When Tim De Paravicini designed the TVA-1 in the seventies, the development was based on the supplier of KT88 and that was the MOV / GEC /Golden Lion KT88.
The KT88 was introduced by GEC in 1956 as a larger variant of the KT66. It was manufactured in the U.K. by the MOV (Marconi-Osram Valve) subsidiary of G.E.C, also labelled as IEC/Mullard, and, in the U.S., Genalex Gold Lion.
Here it is....
No, and it bears repeating...no modern KT88 tube contains the same high quality materials as the original MOV GEC KT88 or Genalex Gold Lion, and no KT88 tube produced today can withstand the same, and is as long-lasting as the original MOV KT88!
The old MOV KT88 lasted much longer than today's tubes, and did not "run" with bias when their lifespan was coming to an end. The problem was not a known issue when the TVA-1 was designed.
Therefore, the mentioned changes to the bias circuit in the TVA-1 are very crucial in the effort to achieve a stable amplifier. There have been quite a few comments from many tube people and others who have been interested in talking about the TVA-1, either because it sounds fantastic (because it does) or because it was important to talk about tubes that "runs" hot, such as this conversation:
Posted by J.. . May 14, 2001:
I have a Michaelson and Austin TVA-1 amp, which I have some doubts about. Anyone know much about these amps? J...
Answer: In reply to: Michaelson and Austin TVA-1 posted by J.. .....on May 14, 2001 at 06:33:30:
Dear J..
The M&A TVA-1 was built in New Bond Street London, very prestigious address! Between 1980 and 1983, it uses KT88 in push-pull and literally eats valves.
It was designed by Tim de Paravicini. It is not a bad sounding amplifier for a KT88 unit, but beware its propensity for eating valves, it becomes expensive pretty quickly.
Sincerely, Pet.. Q.......
As an end of my TVA-1 story I have to say:
First of all the TVA-1 was produced between 1978 and 1982 , it sounds absolutely fantastic with punch in the bass and fine details in the mid- and high frequence region ...
I can just say that the answer above shows that you either haven't studied or understood Tim De Paravicini's TVA-1 circuit, or that one perhaps would like to offer another amplifier instead of the TVA-1?
But no matter what... before one speaks it is necessary to investigate things .....
Page updated
Google Sites
Report abuse