HUUUUUMMY new build

[dehughes]

A cap to ground from the center tap wouldn't do it. The whole point of the exercise is to make the heaters sit at some DC level higher than the preamp tube cathodes. The output tube cathode is a "free" source of such a voltage. You don;t need to make a DC supply to provide the voltage then.

The voltage itself doesn;t matter much, as long as it is more positive than the cathode voltage in the preamp tube, it will prevent electrons from flowing heater to cathode in that tube.

Makes sense. I'll have to do that first, then. Thanks. If it doesn't do the trick, then I'll look to a DC heater system...

[dehughes]

Soooo, I hooked up a 6v lantern battery to the 6SN7 heaters, and voila! Dead quiet! AC heater hum was indeed my problem. So, I've gone and installed this DC supply:



I'm reading about 7VAC on each side of the filament transformer's secondary. I'm running two 10000uf caps in series. One MUR410G diode is connected on each side of the tranny, with the cathodes connected to the + of the caps. The - of the caps is connected to the terminal strip lug on the left.

HOWEVER, I'm only reading .002vDC after the two diodes. Shouldn't I be reading something closer to 6.5vDC or so? What might I have done wrong?

[WholeToneMusic]

Filament wiring

It is hard to tell from the pics, however, Could it be preamp tube 1 may not be wired hum cancelling, If I make the pic out right. I like the zip ties! I use those also.

[dehughes]

It is hard to tell from the pics, however, Could it be preamp tube 1 may not be wired hum cancelling, If I make the pic out right. I like the zip ties! I use those also.

The tubes aren't wired in yet, as I wanted to check the DC voltage first in case it was too high (over 6.5v). Do I need to wire in the tubes to do this? I'd think one could just check the voltage on the top of the caps, without the tubes in the circuit.....or am I totally wrong...

[bnwitt]

If I'm seeing your layout correctly in the picture you need to turn one of the diodes around. You need positive and negative DC for the heater. Frankly you should use 4 diodes in a bridge rectifier layout or a modular bridge rectifier rather than a full wave layout. But, you can try the full wave. Just turn one diode around and make sure you have a positive and negative output by separating the downstream side of the two diodes to two different terminals.

Look at the attached schematic from a Marshall JCM2000. Notice the 4 filter caps. Two are between the plus and negative outputs of the rectifier and two are between the output legs and ground(W1 Chassis) in opposite polarity. Works for Marshall on two 12AX7's

[WholeToneMusic]

The tubes aren't wired in yet, as I wanted to check the DC voltage first in case it was too high (over 6.5v). Do I need to wire in the tubes to do this? I'd think one could just check the voltage on the top of the caps, without the tubes in the circuit.....or am I totally wrong...

I was referring to a possible cross wiring of the filament circuit which could cause hum if the output tubes aren't wired in parallel correctly. Testing the voltages with no tubes is valid.

[dehughes]

Cool. Thanks. I'm trying to emulate the DC heater system on the Komet Constellation: Filament transformer, two diodes into a 10000uf cap, then off to a 6SN7 and 12AX7.

I assume in that situation, one would ground the AC side of the filament tranny (CT on the secondary) and then let the DC side just be grounded at the - on the cap and then at the tube sockets as well?

[Robert M. Martinelli]

If I'm seeing your layout correctly in the picture you need to turn one of the diodes around. You need positive and negative DC for the heater. Frankly you should use 4 diodes in a bridge rectifier layout or a modular bridge rectifier rather than a full wave layout.

Hi,
the problem with your layout is not related to the diodes' polarity, which is correct, as, in a 2-diodes FW rectifier, each anode goes to one side of the 6,3VAC PT winding, and the cathodes go together to the smoothing cap(s) positive terminal. The actual problem is that, when using a 2-diodes rectifier, you need to have a center tap to connect to the negative terminal of the smoothing cap(s) and to the load. If I read your pic correctly, you seem to miss the PT center tap on the 6,3 VAC winding, so the only solution is to use a 4-diodes Graetz ( Bridge ) rectifier, which does not need a center tap.

Be aware that the "Graetz" arrangement will bring the DC voltage to its peak input value, so a 6,3 VAC ( RMS ) at the input will result in : 6,3*1,4142= some 8,9 VDC at the output.

HTH

Best regards

Bob

[conisat]

get rid of VERY strong AC hum

First I would like to complement you on a nice build, at first thought it sounds very like an earth fault, check with a continuity meter all points on the earth, I cant tell by the photo's but have you used the chassi as the only earth path this could cause some RF issues, use solid conductors to link all earth busses starting at the power side then PI then preamp the ground chassi as close to the input jacks as possible. The other main cause of AC hum is the heater ellements acting as diodes to solve this try injecting a DC voltage on the centre tap of the 6.3V winding (make shure you remove this from the earth path first) the DC voltage shoulde be slightly greater than your bias voltage on the PI (check data sheets for maximun heater possitive to ground voltage) Hope this solves it for ya.

[bnwitt]

Hi,
the problem with your layout is not related to the diodes' polarity, which is correct, as, in a 2-diodes FW rectifier, each anode goes to one side of the 6,3VAC PT winding, and the cathodes go together to the smoothing cap(s) positive terminal. The actual problem is that, when using a 2-diodes rectifier, you need to have a center tap to connect to the negative terminal of the smoothing cap(s) and to the load. If I read your pic correctly, you seem to miss the PT center tap on the 6,3 VAC winding, so the only solution is to use a 4-diodes Graetz ( Bridge ) rectifier, which does not need a center tap.

Be aware that the "Graetz" arrangement will bring the DC voltage to its peak input value, so a 6,3 VAC ( RMS ) at the input will result in : 6,3*1,4142= some 8,9 VDC at the output.

HTH

Best regards

Bob

Well maybe I didn't explain it fully but you've given him the same direction. A bridge rectifier is needed without a center tapped filament winding. With the full wave set up and no center tap he would have to make a posititve negative source by inverting one diode as I described. The other option is an artificial center tap with two 100 ohm resistors as the negative reference. All of that being said, however, the Marshall layout I attached is going to give him a much smoother DC filament supply than what he is trying to set up.

[dehughes]

So yeah....you can tell the (very short) limit of my knowledge about diodes and rectification....

I do have a CT on the filament transformer's secondary, but I mis-understood the goal and left it outside of the chassis. I've brought it in the chassis now, and will connect it to the - terminal of the two 10000uf caps. I didn't quite get the concepts around the two types of rectifiers created with diodes, but I get it now. Thanks!

So, I'll go wire in the CT and see what voltages I get. Any recommendations as to how best I should reduce the voltage, if I find it to be a bit high? Resistor in series....add another diode?

[dehughes]

So, now that I wired things up correctly (thanks Paul) I'm getting 9.56vDC after rectification....waaaaay to much for my poor little tubes.

How best should I reduce that?

[bnwitt]

Don't forget, that voltage is without a load. When you put the tubes on the supply the votage will come down but probably not to the typical 6.3V. To determine the size of a Dropping resistor you need to determine the current the tube or tubes will be pulling so you can calculate the voltage drop. How many 6SN7's are going to be on this supply?

You now the fact that you don't have your filament center tap wire grounded to your chassis may have something to do with your heater hum.

edit: Ok, I see in your early photos that you do have the CT of the filament winding chassis grounded. Did you scrape away the chassis coating at all of your grounds?

[loudthud]

The 9.56V will drop down with a load. A resistor tends to even out line voltage fluxuations. Diodes tend to send all the variation to the load. If you are running a 6SN7 and 12AX7, that's 900mA. The math says 3.62 ohms but with the voltage drop, start with 3 ohms at 5W. If it's too much drop, parallel 10 or 22 ohm 2W to bring it up.

[dehughes]

Right on. Thanks. I didn't think about the load...

I did ground the filament CT...just not in that picture.

So, loudthud, I'm not understanding what you're saying...are you saying I should put in a 3 ohm resistor to even out the flow? Or is that for the purposes of a voltage drop? Or...?

Thanks.