diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   c3g self oscillation problem - a newbie needs help! (http://www.diyaudio.com/forums/tubes-valves/221092-c3g-self-oscillation-problem-newbie-needs-help.html)

nlinus 7th October 2012 02:09 PM

c3g self oscillation problem - a newbie needs help!
 
4 Attachment(s)
Hi all.
I am a complete newbie at dealing with vacuum tubes, and my plan is to build a class A PP amp with c3g tubes for phase inverter and driver stage.

I have wired up three c3g tubes and been doing some tests. It seems to me like they are self oscillating. The oscilloscope screen shots shows the plate to GND voltage (oscilloscope in AC mode) of the PI stage. As you can see, there is an oscillation at about 400kHz, and I can not get rid of it.

The setup looks pretty much like the drawings attached (sorry for the quality, I don't have a scanner), but I have tried some different grid leak resistors, grid series resistor etc. Also, the 100R resistor is not used.

In the c3g datasheet, there is no indication that I should use grid resistors for g2 and g3. Maybe I should? Could the lack of those be the cause of the oscillations?

I'm very thankful for all input.
Please feel free to ask for more info if needed.

tinitus 7th October 2012 03:19 PM

shouldnt there be a grid resistor, for one

but C3g as phase splitter ?

MarcelvdG 7th October 2012 03:26 PM

Whether you need grid stopper resistors and if so, what value is difficult to predict as it depends on the wiring. The valve can for example act as a parasitic Colpitts oscillator, the tank being formed by the inductance of the grid wire, the grid to cathode capacitance of the valve and the capacitive load on the cathode side. If the wire connected to grid 2 is long, you might need a grid stopper on grid 2.

Is the phase inverter stage loaded by the output stages when you measure this oscillation? If so, is it stable when you disconnect the output stages? Do you use any overall feedback?

You wrote that you have tried grid series resistors on grid 1, but were they placed close to the grid? They get less effective when they are far away.

wknight 7th October 2012 03:27 PM

Hi
Maybe your input resistor 470k is working like feedback resistor.
And your input capacitor and input resistor 470k are rotating the phase of the signal.
It is possible to check my idea.
Please, change the value of the capacitor or of the input resistor and you will get oscillation on another frequency.

nlinus 7th October 2012 03:59 PM

Let's see...

c3g as phase inverter... As I said, I'm a newbie. These seems like good tubes. Maybe they are not a good choice in a PI?

I have tried grid resistors of different values, but the oscillations stay. What's strange is that now when I started it all up after a trip to the supermarket, the oscillations are absent.

The PI stage was not loaded when I made the measurements. There is no global or local feedback. At least not intentional. I guess the quite high resistance (about 95 ohms) in the 10H coil could result in some kind of feedback?

The grid resistor is connected directly to the grid tab of the socket.

Actually, I have found one error. This issue with bad measurements has puzzled me for quite a while. The 400kHz signal comes from my secondary TFT monitor. My primary does not seem to be causing similar disturbances. What's strange about this is that hen the Oscilloscope (Rigol DS1052E) is just started and has not warmed up fully, these disturbances are not visible. :cuss:

I am still having some stability issues though. Have to check further into this. I guess I'll come back with an update pretty soon...

MarcelvdG 7th October 2012 04:04 PM

What kind of stability problem have you got now that you've identified the 400 kHz as interference from a monitor? By the way, are any of the resistors connected straight to the valve wirewound?

tinitus 7th October 2012 04:22 PM

Quote:

Originally Posted by nlinus (Post 3192833)
What's strange is that now when I started it all up after a trip to the supermarket, the oscillations are absent.

have heard its a bit 'nervous' tube
I guess it means microphonic

but I also have some, so Im interested in how best to use them

nlinus 7th October 2012 04:30 PM

1 Attachment(s)
The stability issues I'm experiencing right now are like this:

The B+ voltage swings about +- a couple of volts (max) at a frequency of 3 Hz. The swing does not really form a sinus curve, but rather a quite straight (horizontal) line with dips every 330ms.

This swing is clearly caused by the circuit (forget everything I said before about mains voltage - I must have been tired when I came to that conclusion...).

kevinkr 7th October 2012 04:49 PM

I don't see any grid stopper resistors in your schematic, with the C3G and other high transconductance pentodes it is a very good idea to install a series resistor at the screen (G2) and control grid (G1) of the tube. (Not needed on G3.) I'd recommend something like 1K, and install them right on the socket.

I know you have eliminated the 400kHz issue as being the result of a noisy monitor, but you may have HF/VHF oscillations that you cannot see with moderate bandwidth scopes, sometimes there will be other manifestations though. Snooping with an AM/FM portable radio may help to identify if the driver stages are oscillating.

I don't see a problem with using the C3G as a phase inverter in triode connection, or as part of a QUAD II style amplifier implementation if you have an appropriate OPT.. One thing I would say though is that your plate and cathode resistors should be roughly twice the values shown minimum for good linearity with the C3G.. Interesting adjustable bias scheme, I would however restrict the bias range so that you always have at least -1V of bias no matter what the pot setting.

The C3G has a reputation for filament fragility so I would let them cool for a couple of minutes before removal. I'm not sure they are any more microphonic than other similar types. I have a few and almost used some in my latest dac project, but they take up too much space.

nlinus 7th October 2012 05:56 PM

Wouldn't the 470k resistor be the grid stopper, or have I misunderstood something?
I have tried different resistors in series with g1 but they are not in the drawing. Maybe a 1k resistor at g2 would be a good idea.

Regarding the load resistors, I kind of went with the recommended values from the Siemens c3g data sheet. Maybe higher values would be a good idea. I though the voltage drop over the resistors would be a bit low with the current values.

I'm not sure I understand that about "filament fragility" and "cool for a couple of minutes before removal".

Regarding the bias setup; I just thought I'd try something I had not seen yet. Don't know if this is better than any other solution, but LED biasing seems quite popular and the pot gives me the ability to adjust the bias voltage.


All times are GMT. The time now is 12:37 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2