• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

EL802 tube giving strange noise.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hign transconductance valves were often used on PCB mount sockets for exactly that reason - the transmission line properties of the connecting traces can be controlled far better than the properties of flying leads. Although there's no need to go as far as to connect an SMT resistor to a clipped socket lug (in the 60's they were not available, yet the TV sets worked), small size of the suppression component is a bonus. 1/4 to 1/8 W resistors are better in this role than higher rated ones, although they may look a bit out of place in a tube amplifier, where all components are supposed to be a bit bulky. Best to just get over that feeling...

But on the other hand, just look at this picture:

http://www.philipstv.org.uk/blog/wp-content/uploads/2012/03/img042.jpg

B401 is the PL802 (the 300 mA series string heater version of the EL802). The g1 connection is kept short, but the rest of the connections (including g2) are routed more liberally. The anode lead flies off to the little PCB at the CRT's base. Now, a CRT is a very different load than an interstage transformer. Also, as can be seen here, there's a suppression resistor in the anode load circuit as well, but it's not on the board:

http://www.philipstv.org.uk/blog/wp-content/uploads/2012/03/img047.jpg

So it might be a good idea to solder a grid stopper to g1 and g2, and then to steer our gaze away from the grids, and to start looking at the anode and cathode leads. The cathode signal input looks OK to me for now (bias circuit might be reviewed later), but the anode lead could benefit from a 220 Ohms series resistor (soldered directly to the socket) in order to tame the oscillation. Could you try that?
 
The cathode signal input looks OK to me
I take that back, it's not OK. The signal routing is OK, but the power supply routing is not. This is one of the many cases where star grounding simply doesn't work.

You need a supply decoupling cap between the EL802's cathode and the supply side of the interstage transformer. The cap should be soldered directly to the tube's cathode pin, and the other side should go to the transformer's supply connection. Keep the loop as small as you can make it. If you also include the 220 Ohm anode resistor (it doesn't hurt anything here, the tube's Ri is about 1.7 k Ohms), and use the right grid stopper (no wirewounds please on g1 and g2, just use cheap, ordinary 1/4W carbon or metal film) you'll likely get rid of your oscillation issue.

At the moment, you are relying on the HF properties of relatively long flying leads (not good) and on the properties of a very old electrolytic can capacitor. The Siemens cap must be at least 18 years old, because in 1999 Siemens became Epcos and now it is TDK...
 
Last edited:
Sadly, I cannot change the electrolytic cap's position at the moment. I can only add a bypass capacitor.

My first try would be changing the g1 resistor with a carbon comp and higher value one, as well as putting an anode resistor in series with the IT, as you recommended. The IT itself as enough primary inductance and can be driven even by 3k Rp, so 220R won't hurt a bit.

What if I also put a small unbypassed cathode resistor, say 10-20 ohms? It will give local NFB, lower gain and increase plate resistance slightly, but could it help suppress oscillations? It will also include a slight degree of auto bias.
 
I have been chasing similar problems for years in my D3A/GM70 fully transformer coupled amps. I always suspected the tube (D3A) and the problem generally occurred only in one of my two amps, sounds exactly as you described and also confusingly followed at least some tubes. There was more than one tube that would do it, in fact many more than one tube, always a tube that measured amongst the best. Some tubes would be fine for the first hours of use and develop symptoms after a few hours, others did it immediately, and a dwindling number in my stock didn't do it.

Years of pulling out my hair I identified that in my design tubes above a certain transconductance threshold oscillated at a frequency beyond the range of either of my scopes, consequently could not see it.

Recently I fixed it, I replaced the grid stoppers with two in series, halving capacitance, doubling resistance and inductance. (The same resistance value by itself did not solve the problem)

I also increased the value of the screen stopper resistor.

I've tried beads and they didn't work, apparently not lossy enough at the frequencies where instability was occurring. Rearranged grounds to reduce inductance, etc.

Now the amplifiers show no sensitivity at all to the tube plugged into the driver stage socket.
Great post, Thanks for sharing.
 
Hey hey,

I powered the amplifiers an hour ago and I'm currently listening to music. I made three changes to the amplifier:
-the series to g1 resistor of 270R MF with 1k CC
-the g2 resistor from 470R WW to 470R CC
-loaded the secondary of the Interstage transformer with 50k.

Now, the left channel, contrary to the right started to exhibit the same noise. The good news is, the noise has audibly a much lesser amplitude and if I tap the tube gently, it almost disappears. Then, the psshhsh is dimly apparent in both channels, which means I have to stick my ear to the speaker in order to hear it.

I prefer letting the amplifier play a bit more for a more precise conclusion. If there is persistent oscillation, I'll also add a resistor series to the anode.
 
Last edited:
However, how I am supposed to make its effect useful, while the leads to the interstage are 30cm long each and I can't shorten them? The ground can be made short, directly to the pin, yes. But the B+ lead?

I mean, is it effective to add a bypass capacitor in this case, when the lead's inductance will spoil its benefits?
Also, the interstage transformer itself is a big series inductance, together with a Cp and Ls
 
I'm sorry folks. It turned out the noisy tube had a faulty heater joint. I realised it when it went noisy again and I gently moved it with my finger, I saw the filament end changing intensity, flashing and dimming and going off at some point. I had to tap it again to make it glow.

I don't think I'm abusing the filaments. The voltage is about 6.45VAC at night, when the mains voltage has reached maximum value. The amplifier startup also doesn't give filament flash to any tube, like it happens in some ECC series.

Probably bad luck with a bad tube? I'm out of EL802, just installed my last pair :D
 
I think you have nothing to lose. So let's try to weld the broken joint: TZake a 10mF electrolytic, charge it to about 12V and connect the heater to it. As soo as it gets contact, a rather high current will flow, which might fix the Joint.


Or it might damage the heater definitvely ...


Best regards!
 
I was waiting to hear when the solder joints were checked, given that it could be mechanically induced! Great job! As an electronics guy, fixing things is one of the greatest joys I get out of my job.
I see many attacking problems from a view that they have an idea what the problem is. I wasn't a world class troubleshooter until I threw all that out the window, and only looked at what info was obvious by eliminating variables.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.