From diode to tube rectification in my Aikido

[cbutterworth]

Well I did it! This morning, with a feeling of hope in my heart, I pulled apart the PSU of my Aikido. Actually, I didn't need to do too much. I removed the 125-0-125 trafo, omron relay, diode bridge and snubber network. In place, I added a 275-0-275.

Once the universal primaries had been set for 120VAC, I was going good. The octal socket for the Amperex delay tube was re-wired for a tube rectifier, and it was ready to fire up.

Being cautious, I pulled the B+ and PSU GND from the Broskie PCB, plugged in and measured 370VDC on the PSU output - Phew! Then I re-wired the PCB, added my cheap Russian nasties and fired-up the Aikido again. It was great to watch the voltage rise to around 370VDC, then after around 15 seconds (my guess as the 6SN7s warmed-up), the B+ dropped to around 312VDC - PHEW and more PHEW!

Initial playing with my good Sylvania and RCA in place of the nasties revealed a couple of things:

1. My low level, undescernable hum was replaced with a very, very quiet but clearly 120Hz hum. My ear really needs to be close to the speaker to hear it.

2. The buzzing B+ trafo has stopped!

3. The music sounded much improved over the diode bridge. How different? Well, it is hard to say, but it seemed smoother and warmer, and Kara reckoned that it sounded much clearer.

It's early days and the new PSU components need to settle, but I love it!

As for the hum, well I can live with it, as it is very difficult to hear. BUT, being 120Hz, I feel comfortable about it being the result of poor filtering. My tubes pull around 24mA which for a B+ of 312 and chokes of 10H, in a LC filter means that I am not pulling enough current to allow the chokes to function properly. It could be that, but I am using a 0.68uF "tuning" cap so maybe the PSU is more like a CL filter. I can replace the 0.68uF with a 1uF or 1.5uF, which will bring the PSU into much clearer CL filter. I could also add a current sink across my final cap to pull a little more current, maybe up to 34mA.

Anyway, I feel pleased at knowing that easy "tweaks" will likely bring the PSU to a state in which hum is a thing of the past.

Charlie

[ray_moth]

I doubt if your 0.68uF tuning cap will change the electrical characteristics to a noticeable degree. You could try a bleeder, as you say, to increase the choke current.

Since you have 120Hz ripple, it should be possible to reduce it simply by adding more filtering. Increasing the capacitor after the choke is the first obious thing to try but, if that's the final stage in the filter, you should make sure it's a good quality cap.

The ultimate answer would be a regulator = no hum.

[jayme]

Good to hear!

BTW: what value are you using for the "C" in the LC filter?

[cbutterworth]

Ray-moth and Jayme,

I could go with regulation using tube rectifiers, I think I would need three of them.

For my caps, I am using:

0.68uF - "tuning" cap
60uF motor cap
60uF motor cap

Yes....I could use a larger value for the final cap as it is always better to have the final cap slightly larger than the others.

The bleeder resistors should make a difference. The trafo is only 50mA DC rated, so I want to be very sure about not exceeding that value.

Are you aware of Broskie's tables for tube and resistor selection? If you are, then for example if I read for a 6SN7 300V at B+ and a cathode resistor of 680R, then does his column labelled lk refer to current drawn by the tube? If it does, does it mean each section or the entire tube?

I have always taken it to refer to current drawn by the entire tube. In which case, my Aikido draws around 7.2mA for each output tube and 4.5mA for each input tube. This gives a total of 23.4mA.

If I am reading the table correctly, I can add sufficient bleeding to bring current draw up to around 34mA. Alternatively, I could use current-drop with and without tubes to estimate current draw for the entire circuit, then add a little for draw across my safety resistors on the caps.

Charlie

[cbutterworth]

Hey....

Where and how would I install an arc snubber in my new PSU? My old snubber was between trafo and bridge.

Thanks,

Charlie

[jayme]

What did that snubber circuit do in the old PSU?

[cbutterworth]

Jayme,

It was supposed to prevent kick-back from the diode bridge at switch-off. I cannot even remember where I got the circuit details.

Charlie

[jayme]

If all it did was protect the SS diodes from kickback, you don't need it with a tube rectifier.

[Stixx]

Are you aware of Broskie's tables for tube and resistor selection? If you are, then for example if I read for a 6SN7 300V at B+ and a cathode resistor of 680R, then does his column labelled lk refer to current drawn by the tube? If it does, does it mean each section or the entire tube?

Charlie,

as I read the aforementioned table the current draw of the tubes depends on B+ and the value of the cathode resistor. With a recommended value of 470R for the first stage the tube draws 7.2mA EACH half as listed in the table. With rk = 240R for the output stage the tube draws 10mA (each side) for a grandtotal of 34.4mA. I assumed that you are using 6SN7 in both positions.

Plugging that into PSUDII and assuming/vaguely remembering that you are using two 10H chokes I am nowhere near 300V at the psu output with a .68uF first cap. But it is easy to play with this value to arrive at the desired voltage. But then, I am just guesstimating your actual psu...

...using tube rectifiers, I think I would need three of them.

You won't need three rectifiers for a regulated psu, just build a psu that goes like: trans-rectifier-filter section-regulator. But your current psu has the potential to be quiet, I think.

Hope this helps just a little...

[cbutterworth]

Stixx,

I always assumed that Broskie's table listed values for the entire tube (as far as lK was concerned). Thus for a stereo board I estimated:

B+ 300V with 470R and 1000R for rK, I would get 4.5mA x 2 for the 470R tubes and 7.2mA x 2 for the 1000R tubes totalling 23.4mA.

By your reckoning, I should be doubling this number because the lk's listed by Broskie are per side of tube not the whole tube. This would mean that the actual tube current draw would be more like 46.8mA.

Given that I am unsure of this, how would I measure it, or use tube datasheets to get an idea of the total current drawn by four tubes. Sy did once try to explain it, but it seemed like something was missing. The 300V gets halved so that each side of the tube sees more like 150V for a B+ of 300V.

My new PSU consists of: 550VAC; 5V4GA; 0.68uF; 10H; 60uf; 10H; 60uF. When I model this in PSUDII for 23mA, I get a voltage of 317VDC. If I model it for 46mA, I get more like 260VDC. In reality, I see something like 311VDC on the output.

If you have any firther enlightenment, please let me know.

Charlie