uk_headhunter

HeadHunter

Headphone Tube Amplifier (OTL)

Page 3: April 25, 2007

Building the amp

Building the amplifier after choosing a circuit design always takes me a while. I must confess that I really dislike drilling holes and that's just what needs to be done in this stage of the project.

In order to be able to test the amp before building, I will make a test construction first. This will allow me to test the heat dissipation of the tubes before mounting them and avoiding overheating of chassis and or other components such as caps.

For the prototype I use the same board and setup as ued for the Martian. Normally, if everything sounds nice without hiss and humm then I'm pretty sure that when building the final amp in a shielded enclosure these noise levels will even be lower.

The prototype that is built can be found on the previous page as the prototype circuit. If all successful I will promote that circuit to the final version. However, reality means that I will probably have to change many things in order to get to the final version.

Proto 1 ...

Yesterday (April 25) I finished a first prototype of the amp. Hmmm, or better hummmmm. Although the amp was making music, wich tells me that I probably connected all wires correctly, there is a background humm which kills the absolute listening pleasure at very low listening levels. Not that I cannot predict any source for humm: I'm still waiting for a new Amplimo transformer to arrive, the combination of a 60VA, 12V transformer with a step-up for the plate voltage is far from ideal. So the question is whether the final version will not have humm.

The amp (not the powersupply) has not changd between proto1 and proto2, so the picture above gives an idea of the setup of the amplifier part of headhunter. The two (small) 5670 tubes are mounted horizontally on a L-shaped aluminum profile.

Proto 2 ...

This a picture of the second prototype of the amp. As you can see on the picture below I use a separate transformer for the heaters (front left) with the LT1083 voltage regulator mounted on a large cooler and one transformer for the B+ rail (front right) with the three large capacitors (two stage RC filters). And the back you see the two 6AS7 tubes standing. The amplifier part itself is built on the aluminum L-profile.

But it is clear that the 6AS7 is a terrific show-off tube: Large coke bottle shape and nice glow. Only there seems to be a big variation in performance between samples of the same tube. In proto 1 I clearly heard a difference between both channels so I'll have to find out whether this was because the poor power supply setup or that the tubes indeed differ too much. The solution could be to use one tube for the lower side of the power stage and one tube for the top part of both channels. But probably it is equally simple to use local feedback.

Proto 3 ...

After listening to the amp for a few hours, I was convinced abouts it's quality regarding control over headphones and the ability to reproduce lowest frequencies. I was not entirely sure that the amp did show its full potential with high frequencies. I was suspecting the output capacitors of Nichicon, 470uF/200V and therefore added two parallel MKP capacitors of 1uF. These two small value caps kick in at about 1kHz and will make sure that the high freqs get the best response possible.

After listening again, I think I solved the missing high frequencies. It's time to work on the final version now...

Circuit of the final version

I also made some small changes to the PSU. Mainly I added 2 resistors of 0.39 ohms in series with the heaters with following consequences:.

The initial inrush current to the heaters is much, much more than the 2.5A which it is as the heaters reached their temperature. This means that the voltage regulator has a difficult task of keeping the current to the accepatble 7A level and therefore nitially it takes some time before the voltage level reaches the 12V. The resistors total 2*0.39 is about 0.8 Ohms in series with the heaters which makes sure that the current keeps within acceptable limits.
The voltage regulator needed to adjust the voltage from 18 Vdc (after 15Vac rectification) to 12.6V regulated. Ths meant that a lot of heat was generated by the voltage regulator and it required massive cooling. After adding the two resistors the voltage regulator stayed warm but not hot. Of course the side effect is that both resisors get hot now, but that is not so much of an issue.
I needed to change the resistor close to the regulator from 220 ohms to 180 ohms in order to keep sufficient headroom for regulation.

Assembling the final version

I need to work out a few things before I can build the final version of Headhunter.

For the enclosure I'll use two identical hammond die-cast alu boxes. I used such a box also for headline.
I'll have to find a connector pair for connecting the PSU and the amp. I'll probably settle for Hirschmann as this will make the amp compatible with previous phono amps (as long as I stick to 12Vdc for heaters).
I'll need to find a power cord (4 leads) capable to support about 4A for the heaters and at the same time 220Vdc for the B+ . Though it needs to be a safe solution, I would like to have a flexible unbillical cord.
Heaters and stand-by switches need to be on the amp itself, the PSU needs to have a mains switch to power up/down everything.
Both amp and PSU need bleeder resistors (also for the heater supply!)
When space permits I would like to use a rotary switch for volume control instead of a stereo pot.
I would like to position the input on the front panel, as I found out that in most cases I listen to smaller CD, MD and MP3 devices.

Anyway, I'v decided not to use Boutique parts for the amp. Well that's no surprise if you read my other projects. It is important to get the B+ power supply right, and get the heater supply DC reglated as I do notice when using badly regulated PSU.

Also, I still feel very strong about the use of aluminum enclosures as these enclosures do protect from PSU influences and are more "silent" than the steel ones.

Above you see the two boxes of die-cast aluminum that I bough for headhunter. I painted them black, just so I'll have another color on my desk.

Things to do later

I need to consider crossfeed input for this amp. I did not use crossfeed input in my projects before.
Need to do some objective testing for the amp and based on the outcome decide whether or not to use feedback

Platenspeler.Com