Ralf's Daande Lenol A409 Line Amplifiers

Daandé Léñol means "voice of the people" in Pulaar, and is the name of Baaba Maal's band

Daandé Léñol

µ-Follower Line Amp with Pentode-CCS

By accident I came across some curious old triodes: Philips A409 with the very old Telefunken base. The tube is an equivalent to RE074t. The Telefunken socket was used between 1914, when it was introduced, until the end of the 1920ies, when it was succeeded by the Europe 4-pin/5-pin socket. While the old Philips tubes were affordable, there were no Telefunken sockets around. I studied some old pictures and schematics of the socket and told myself: must be possible to "DIY" this socket. Pins of the base are 4 mm, like they are used by the American UX4 base for the filament pins. I found the terminals in standard wall sockets are perfect to hold 4 mm pins. So with a piece of pertinax and a handful of screws the making of 2 Telefunken sockets started:

The socket open: 4 wall socket terminals mounted on a pertinax plate. Also the sockets are isolated from each other with strips of thin pertinax.	The socket closed with a top pertinax plate (top and down side). Ready to be mounted on the underside of a chassis plate.

A409 is a non-spectacular small signal triode from the 20ies. Characteristics are close to American type 30 triode.

	A409 (tfk)
heater voltage Uf	4.0 V
heater current If	0.065 A
plate voltage Ua	150 V
plate current Ia	3.5mA
plate resistance Ri	10.000 Ohm
plate load Ra
grid voltage Ug	-9 V
amplification factor µ	9
socket type	Telefunken 4-pin

In my junk box I had oval MP capacitors left, rated with 20µF for 330VAC. Usual problem is to mount those caps. I solved this issue by glueing M6 screws with flat heads on the underside of the caps (used Uhu Endfest glue), and filled the whole plane of the terminal side with Epoxyd resin. The chassis plate became 2 holes for the screws and 1 oval hole for the terminals. Then the caps were mounted through the chassis plate with large washers and nuts.

Screws glued and cap's terminal plane filled with resin	Cap mounted through the chassis plate

The volume control: having experience with 24 step attenuators, I decided to go for a 12 step type. I choosed a modular switching system from NSF Controls (via Farnell-in-One) to build 2 * 12 step switches, one switch for each channel. Having the choice between different circuit types for shunt attenuators I went for the ladder type. The ladder type ensures the same input impedance regardless of the choosen position. For 12 steps it needs 24 resistors. For better adjustement I calculated a "3 tier sloop" with different attenuation values between the steps. Result of the calculation is this sheet (.gif): 12 step 30dB calculation

It has total attenuation of 31dB with an impedance of 10k. Resistors are 0.25W 1% metall film resistors. This volume control is in the line out of the line amp. To adjust total attenuation a T-pad voltage divider is used. For 10k total impedance a 8.182R and 2.020R combination gives 20dB attenuation, and a 9.387R and 633R combination gives 30dB.

This is the generic schematic of the T-pad and stepped ladder construction (.gif): Ladder

A 2 pole 12 throw rotary wafer switch made by NSF Controls Ltd. / UK ...	... equipped with 1% metal film resistors

Talking about the circuit. To lower the impedance of the line amp's output it is a good idea to use a µ-follower circuit. I found Gary Pimm's developments ( Gary Pimm's DIY audio pages ) very interesting, and had a deeper look on his Pentode-CCS version (.gif) Gary Pimm's 39-44 Plate Load. 39-44 is an old American remote cut-off pentode with an UX5-base. I had a chat with Gary to learn about his calculation, then decided to use a similar version. So the A409 got a Constant Current Source by a 39-44, with a KA431 voltage reference to provide the grid voltage for the pentode. Nevertheless it is a simple circuit: a triode with a CCS and cap coupled output.

This is the schematic of the amp (.gif): Daande Lenol Schematic

For the power supply first I planned to use rectifiers type 80 already in my stock, but I discovered a full-wave rectifier with 5V half-indirect heated developed by Philips, named 1883. This is a funny tube: from the data it is an equivalent to 5Y3GB, the only half-indirect heated version of 5Y3, but it has an European Aussenkontakt 8-pin base. Developed by Philips at the end of the 1930ies for the French market, where 5V was a common heater voltage for rectifiers. From the outline 1883 looks like an AZ1, but it has a nice softstart: needs at least 15 seconds to have full emission.

Playing a little bit with Duncan Munro's PSU Designer 2 my considerations for the power supply finished in a C-R-C-L-C filter combination, with a large currrent sink of about 15mA after the 2nd cap (by a bleeder resistor). To improve the filtering of 100Hz AC noise I've calculated a bypass cap for the choke, value of the cap is 0.168µF. This gives me a filter factor of 13.716 compared to 1.303 without the bypass cap.

Being a battery heated tube A409 needs a filament supply with smooth and well filtered DC. Due to the small current of 65mA I was able to use small 1Hy chokes rated for 120mA. I've build a filament DC-supply with Schottky diodes, this 1 Hy choke, and a low-drop voltage regulator made by Sharp: PQ30RV11.