Valve buffered IGC - volume pot causes DC offset, crackling

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've just finished building a valve buffered IGC based very closely on this design here http://www.customanalogue.com/diytubegainclone/index.htm .

It is all up and running, but unlike my previous GainClone (unbuffered , non-inverting), when I turn the volume pot, I get crackling and, according to my multimeter, DC offset of up to about 200 mV, which then quickly goes away (presumably cancelled out by the amp's DC blocking capacitors).

Is this normal? Or is something wrong? Is this safe for my speakers?

Also just wondering, how much DC offset is it safe for an amp to generally put out? One channel of mine constantly puts out 90 mV. Is this okay?

Thanks in advance!
 
It is all up and running, but unlike my previous GainClone (unbuffered , non-inverting), when I turn the volume pot,

There must be some DC on the pot. Could be, the +/- voltages for the tube are not symetrically?

Is the tubes filament voltage correct, e.g. does the tube work?

Please measure the DC voltage at the potis center pin (full volume and turned downwards). It should be always 0V.

It makes sense to do the measurements with and without attached input device.

Franz

P.S.
200mV bias is on the very high side, for some speakers a problem. 90mV for a non inverted gainclone: lets say in the max range. You could adjust by inserting a trimmer between +input pin and ground.

/Edit
Be carefull with your speakers during startup of the tube: you will have full DC (high current) voltage on the outputs during some seconds! This can overheat the voice coil of the speaker! I recommend to switch on the chip by a delay circuit some 20 seconds after powering up the filament or by grounding the chips input pin.
 
Yikes - I measured the voltage on the middle pins of the pot - i.e. the input pins. When the pot is all the way down, voltage is effectively zero, but when the pot is all the way up, the voltage on the middle pins is around 8 V! Lucky that didn't fry my source. Lesson learned.

The voltage rails for each side of the valve are balanced to within 0.1 V (+-35.1 V), so I don't think that's the cause of the problem.

The problem is the same for both channels, so I don't think it's likely to be a stray wire or anything like that.

One thing I did slightly against recommendation was to use a single transformer for both the valve and chip amp power supplies (+-25 V ac), but I do have a separate rectifier bridge for each section, and for each channel (four rectifier bridges total). Grounding was done according to the diagrams on this page http://www.customanalogue.com/diytubegainclone/ps.htm

I even tried two different valves. And yup, the amp works.

Thanks so much for your help, I really appreciate it! And thanks also for the warning about powering on the amp. Fortunately I did already know this, but thanks for taking the time to make sure :)

Any ideas about what could be causing this problem?
 
Any ideas about what could be causing this problem?

Must be some problem with this valve.

A "proper" circuit would consist of the poti, followed by a coupling cap. From the other side of the coupling cap to ground a so called grid resistor. This circuit is also a high pass filter.

This would also protect the input device in case of a valve fault.

But it is not very attractive, when your input device already has some output caps.

Franz
 
So there should be a coupling capacitor/high pass filter between the pot and the valve

Indeed, but maybe the sound quality is better without. Soundwise it is still true: the best cap is no cap...

Do you think that in addition to this I also have a faulty valve?

When I understand you right: one is working, another not, then I guess one has a fault.

What tube type do you use?

Franz
 
There is absolutely no need for additional caps. You have wired something wrongly and that's why you get the strange behaviour. What is the voltage at the cathode? If the tube is correctly biased the grid voltage is definitely zero, irrespective of pot position. Tube failure at such low voltage is extremely unlikely.
 
Cathode voltage (with valve plugged in and everything warmed up) is -6.6 V on one channel and -4.6 V on the other. Grid voltage is -2.7 V and -2.1 V. That is with the volume pot all the way down.

With the volume pot all the way up, cathode voltage is -10.7 V and -12.4 V, and grid voltage is -9.0 V and -10.8 V.

This is all with no source plugged into the input.

I may not have been clear about the valves - I tried two, and both exhibited the same behaviour. I assume neither one is faulty.

The valves I am using are 6922's made by Shuguang.

Thanks again!
 
Problem solved - partly. Due to a problem with my DC heater filament supply, the valve was only receiving about 3.5 V. The main problem was that 5R6 resistor (in the design at http://www.customanalogue.com/diytubegainclone/images/ps-2.gif), it was using up a lot of the voltage.

With that issue fixed, depending on what valve I use, grid and cathode voltages can be as moderate as -0.3 V with volume pot all the way down and -3 V with volume pot all the way up.

Is this still unusual? I still get the DC offset and crackling when I turn the pot, but much less so.

Should I try increasing the heater voltage from 6.0 V to 6.3 V? Or is there likely to still be some grounding problem (I haven't found one yet)?

Or, my valves are not actually clearly labelled. I bought them on eBay from a random trader in Taiwan, although they did have 100% positive feedback. Perhaps I am using a valve which actually wants a higher heater voltage?

Really appreciate all this help :)
 
Yes, you solved your problem partially.

Do you remember: my first question was the filaments voltage! I know, this circuit is working (with high distortion), even when there is no filament voltage. And in this case, you have this DC on the potentiometer.

6V is within the +/- 10% tolerance and O.K.

But you have another problem, as the voltage on the grid must be zero Volts.

Did you ground the center of the tubes power supply?

Franz

/Edit
Did you ground the filaments voltage?
 
Gaplessophile


As Franz mentioned the heaters voltage is not particularly critical. There is something seriously wrong somewhere as you do get grid current. I just simulated your cathode follower and this is the operating point irrespective of the pot position.

A picture of the valve won't be a bad idea if there seems to be any doubt whether it is indeed an ECC88. A picture of the entire build will also help: there may be something simple you are missing.
 

Attachments

  • ecc88 follower.pdf
    35.4 KB · Views: 101
So sorry for not discovering the heater supply problem earlier. I checked it when the valve was not plugged in, and of course it showed the correct voltage. My silly mistake.

The heater supply and the centre of the valve power supply are both grounded to the centre tap of the transformer powering the valves.

To try to further isolate the problem, I disconnected the valve circuit from the chip amp circuit - cutting off from the 3u3 capacitor onward. I also cut off everything from the grid pin of the valve. Under these conditions, cathode voltage was approximately -25 V, and grid voltage was approximately -12 V.

I then connected a 30k resistor from the grid to ground. Grid and cathode voltages were now between -2 V and -3 V.

analog_sa, my setup looks exactly like that in the file you attached, except that I assume I am right that plate should be pin 1, grid should be pin 2, and cathode should be pin 3 (or pins 6, 7 and 8 respectively for the other channel), and the heater supply should go to pins 4 and 5.

I am really so grateful for all this help!

Attached are pictures of my build and of the valve (of which there is no DC resistance between pin 9 and 4 or 5). The two Rectifier bridges and clusters of capacitors on the left of the picture are the valve supplies, and the other two on the right are the chip amp supplies.

The writing on the valve has partly worn off. On one line I can make out the characters "6N", and on the next line, "722". I also have two other valves which were also sold to me as 6922's and look virtually identical to this one, but with no marking, and they perform with the same problems in this amp.
 

Attachments

  • 00.jpg
    00.jpg
    86.4 KB · Views: 251
Well, the centre of the valve supplies and the negative rail of the heater filament supply is connected to the central ground, along with the centre of the chip amp supplies, and this central ground is then connected to the centre tap of the transformer (only one transformer for both valve and chips), and to the chassis ground and to the earth wire in the power cord.

Is this okay?

Thanks for the heads up about the heatsinks!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.