Tube Phono Equaliser

“Tweety”

Three in a row ...

Last modified: 31-Oct-2006

Introduction

Some time ago I received the hint from a colleque on the hifi-forum to simulate my own designs with E-CAD software. I liked that idea and since I have done simulations on scratch, hounddog and some commerically available projects on the internet.

In case of the hounddog, I discovered that hounddog will probably not follow the RIAA curve the way it is designed, scratch and later projects are within 0.3dB of the ideal curve. On the other hand, we're talking simulation here, which is different from measuring with a scope and a signalgenerator in the lab. Slight deviations are inevitable as it is difficult to model all parameters of a tube in an electronics package. I know TriodeDick has done some measurements on his scratch and he found that the finished amplifier is even slightly better than the simulations had indicated (end these were already very good).

For a better understanding of the background of RIAA filtering and simulation I studied the books of Rainer zur Linde and Menno van der Veen on this subject. I also wondered whether I would be able to design a phono equaliser based on the SRPP principle myself. This should not be too difficult, especially when the simulation tools help you to design RIAA circuits that are close to the curve. The design principles of the SRPP are more or less standard, it's the design and use of the RIAA reproduction network (one stage, two stage, passive, active etc.) that make the difference.

I was lucky enough to have two small enclosures available that I originally bought for scratch but that were too small to accomodate the tube rectifier circuit and larger transformer. So I convinced myself I had most of the components already in-house available and it would be easy and cheap to build another phono pre. Consider it an exercise: I am very satisfied with the sound of scratch but I would like to try out several things myself.

My goals were as follows:

1. Although Hounddog with its active filter and feedback has proven to have a low noise level, I decided to build the next project with a passive filter again.
2. But I want a low signal/noise ratio, I'm considering to use an external power supply design to keep transformer noise etc. away from the sensitive phono circuitry.
3. I want to build a RIAA filter with just two capacitors and a resistor, all balues to be standard values. In theory this is possible according to initial simulations.
4. I will use components of high quality, but not the high-end exclusive rip-you-off stuff.
5. The two enclusores made of aluminium are small, so I will need to work on the square mm. Even so I want to be able to change the filter setup or exchange otehr components at a later moment. For scratch, this is no problem at all, but Hounddog is hardwired with point-to-point cabling and is a dog to repair.
6. Apart from simulation and listening tests, I will do measurements at a later stage to check theory against real-life.
7. I want more gain than hounddog, preferrable the same as scratch. My UL-40 amplifier does not react too well on low-level inputs.
8. Tweety will have a nice yellow lab-connector for mass connnections.

Why Tweety?

Why Tweety? I am an absolute fan of cartoon network, and Tweety is the singing bird. We have a canary at home which we call tweety too and it seemed like a good idea to name the amplifier after our own singing bird. An Tweety isn't whistling, it's singing (and can talk too) and that makes all the difference. Making an amplifier which is humming or whistling is not that difficult, but making it sing is something special.

In both cases it's no problem calling it Tweety, but it's just to give you my background for naming this amp Tweety.

I had to start with making the amplifier itself. Because I can connect this amplifier to the powersupply of Scratch this gave me the opportunity to build and test Tweety and then start working on the power supply. It was necessary too by the way, as the transformer for the power supply took several weeks to arrive. I will increase the B+ to 300V to keep both parts of the SRPP stage happy.

Here is the description of Tweety and the power supply.

Building Tweety

As shown on the top of the page, the cabinet for Tweety had the claustrofobic dimensions of 16*10*10 cm. This width of just 10 cm really was not sufficient to mount the three noval sockets next to each other with sufficient space between them. By each turning the sockets a few degrees the mounting holes of two adjacent sockets were just far enough apart.

In front of the tube sockets I placed a 8 cm strip of soldering lugs and spaced two cm apart another strip of soldering lugs. This way I could solder caps and resistors between these two strips allowing me to easily upgrade th amp at a later stage. Also, it provides a better overview in the case.

Ik heb vlak voor de buizen eerst een rij van soldeersteuntjes neergelegd, en twee cm. verder nog zo'n rijtje. M'n idee is om onderdelen tussen deze twee rijtjes in te solderen zodat ik gemakkelijk van waardes kan wisselen.

As said before, initially I use simulation software to design and build this amplifier. If at a later stage, during final testing and measurements it should appear that I have to change values, at least I can do so with reasonable effort. Something which would be impossible in such a small room if not specifically taken into account during the component layout phase.

After all, I have to house two Jensen audio caps of 0.36 uF, and these babies are laaaarge. I will fasten the Jensens to the side of the box with a tie-wrap and a cablesupport (plastic) glued to the box. Unfortunately these Jensens are so large that should I need to swap components of the filter of cathode circuit I have to remove the Jensens first.

Tweety is a Songbird...

After two days of work, Tweety was finished. It is nice and small and built with good components. On a temporary basis I had to borrow the powersupply of the Hounddog or Scratch since Tweety's powersupply could not be finished because of the late arrival of the power transformer.

At a later`stageI will add some more pictures of Tweety, but the fever of completing Tweety was so high that I did not take the time to shoot some pictures. You know the feeling? On this page therefore only photos of the first days of Tweety's existence.

Tweety now has to break in, because I'm not really impressed with it's performance the first time. Initially I only had two ECC83 tubes for the second stage and I used a EE81 as the first amp stage. This was because I was worried about having too much gain but it turned out that a third ECC83 tube added just a little more gain and more "umpf" to the sound. I decided to go for a setup with three ECC83 tubes.

After some more experiments with grounding of the filaments Tweety became less nervous and there was more ease in the sound. It became time to sit back and enjoy.

After two days of continuous operation Tweety has gained a lot in musicality. Something that cought my eye (uh ear) was the good signal to noise ratio of Tweety. If I'm able to keep the current sound quality and SR values I'm happy. I have hardwired Tweety with solid 0.8mm fine silver wire which is shielded with a copper braid and then isolated with red crimp sleeve. Clearly shown in the picture are the two big Jensen audio capacitors (older type, blue color).

Todo

The following issues needed to be solved later in the project:

• Power Supply for Tweety (275-300 Volt). See next pages
• A nice LED on the front plate of Tweety to show that the amplifier is in action (done, see next pages).
• measurements of Tweety against the ideal RIAA curve.
• Maybe change the capacitor of 8.2 nF with one of 6.8nF and change the resistor value from 10 Kohm to 11 KOhm. The should result in slightly better high frequencies without changing much in the accuracy of the filter setup.
• A better flexibele power interconnect between Tweety and the Power Supply

References and Links

• Tube formulas and spreadsheet; calculate output impedance, miller capacity, gain of tubes
• RIAA cookbook; how to build your RIAA filter even without SPICE tooling, including an analysis of Tweety's filter

© Maarten&Annemarie, March 2002, 2003, 2006