Tube Amps / Music Electronics
|For current discussions, please visit Music Electronics Forum.||The sunn still shines online!|
|10/10/1999 1:33 PM|
||OT protection tests|
I'm writing from my secret lab where I have just been testing some of the means of OT protection I was talking about.
I used the rig I'm working on which has 2 EL34s in push-pull running off 480V, with no feedback. It's biased for 2 x 30mA plate current at idle. The OT is an old hi-fi one and I'd expect it to be of quite high quality.
I installed 2 air gaps 0.3mm wide, one between each plate and the centre tap. I also fitted a 8 ohm resistor in series with a 75V MOV across the secondary, and put a 260 ohm resistor across the secondary too.
Then I crossed my fingers, applied 80 Hz from a signal generator, and turned it up as far as it could possibly go, driving the output stage right into hard clipping.
To my surprise nothing happened. The spark gaps didn't break down, in fact the MOV didn't even conduct. The output was a square wave of 130Vp-p (65V rms) and the power dissipation in the 260 ohm resistor was thus 16 watts.
I tried removing the 260 ohm resistor, leaving only the spark gaps and the MOV/resistor combo. The waveform grew some ringing (to 200vp-p) and the VDR conducted slightly. The gaps still didn't break down!
In all cases, the supply current to the output stage decreased when drivng into clipping, and the plates cooled down.
My conclusions: for my rig at least, a hefty 200-odd ohm resistor across the output gives all the protection I need. No fancy automatic shutdown gizmos required. However, the story might be different for another OT.
I apologise to you all, especially Bruce, for complicating things unnecessarily.
|10/10/1999 5:16 PM|
So you tested the amp with the 260 ohm resistor in parallel to the 8 ohm load, and with the 260 ohm resistor removed, and nothing seemed in imminent danger of self-destruction? Did you also test it with the 260 ohm resistor in place and the 8 ohm load removed (like if the cable into the speaker was unplugged when the amp was cranked up all the way)?
So with both the 8 ohm load and 260 ohm resistor in place you measured 16 watts being dissipated by the 260 ohm resistor? I would have thought that it would be much less than that... Or did you mean that with only the 260 resistor connected to the OT it was dissipating 16 watts? If that was the case then I guess we better use 20+ watt resistors for a 40-50 watt amp.
P.S. BTW did you try testing it with both the 8 ohm load and the 260 resistor disconnected to see if the other components would offer any protection to the tranny?
|10/10/1999 5:24 PM|
If you have the opportunity, it would be interesting to see the same test repeated with a typical guitar amp OT. That hi-fi OT is likely highly interleaved for best high frequency response and very low leakage inductance. The leakage inductance contributes most to the ringing.
|10/11/1999 12:50 PM|
I did all the tests with *no* speaker load connected. That's why I was worried. The voltages, and dissipation figure quoted, were for the 260-ohm resistor on its own. The OT I used had a 16 ohm secondary, not 8.
I think you are right Randall, the results might be a bit more scary with a guitar amp OT. However, I don't have one handy... maybe someone else could try it! If you add a set of air gaps to the primary, they will probably save the insulation on your tranny. An 0.3mm air gap will break over at about 1,000V.
One thing though. While the resistor protected the OT under no-speaker conditions, I did notice that my output tubes drew far too much screen current (150mA peak) If this was allowed to continue for more than a few seconds, the screens could melt.
This suggests that screen current might be a good indicator of a no-speaker fault. Therefore, propeller heads like me and R.G. might want to sense screen current, and shut the HT down if it got excessive. The K.I.S.S. method would be a small fuse in the screen/preamp supply.
It would be nice if someone could repeat this screen current observation with a guitar amp OT.
Keep on tweaking,
|10/11/1999 1:14 PM|
The reason Ig2 goes through the roof is easy to comprehend. If the voltage of the plate is swinging wildly about with great changes in potential, then the period of time that the plate will be much LOWER in voltage than the screen increases. Therefore the screens draw mucho current.
On the load line, this is the case where the line has been rotated counter-clockwise (a much higher impedance) and intersects the plate curves WAY below the "knee." Whenever you do this, it's a guarantee that Ig2 will skyrocket.
You can also look at it this way: in a pentode (or BT for that matter) the cathode current largely stays the same throughout the waveform. That means as plate current falls, screen current is rising. When the OPT represnts such a high impedance, the plate doesn't have the ability to pump a lot of current out, instead generating huge voltage swings. That cathode current still has to go somwhere, and it goes right to the only other positive electrode, the screen.
Perhaps a simple way to prevent screen burnout is merely to limit screen current with an active device? By doing so, as the screens drew more current their voltage would quickly fall, effectively clamping off the tube.
Just some thoughts...
|10/11/1999 2:01 PM|
Thanks for your explanation of increased Ig2. That's exactly what it said in my 'RSGB Radio Communications Handbook', where it recommends tuning tetrode- and pentode- type RF amps by metering the screen current.
Because my amp project has a solid-state HT regulator it will be easy for me to wire a 300 volt Zener from the screens, to the regulator circuit. If the screen current goes too high, the screen voltage dips, current through the zener stops, and I can have the HT locked off using a 2-transistor flip-flop or whatever. I was intending to do this for the bias supply anyway, to shut down in case of bias failure.
Anyone with less of an electronics fetish would probably stick a 50mA quick-blow fuse in the screen supply. If the speaker came unplugged the first good Kerrang would probably blow it and shut down the amp completely.
|10/11/1999 1:11 PM|
Let me clarify the setup I used a little more. The 8 ohm resistor wasn't a load. It was connected in series with a MOV rated at 75V AC, across the speaker terminals. Therefore, the amp was effectively running flat-out with no speaker.
With the resistor/MOV combo in place, and no other load, I observed some ringing, but it wasn't that impressive and the MOV never really conducted much. With the 260 ohm resistor, the ringing was damped completely, even with no other load. (A 510 ohm resistor worked too and would only dissipate 4 watts)
I wasn't able to observe the waveform on the OT primary, because the voltage (even with no ringing) would have destroyed my scope probes.
I'd be very interested to hear other people's results (but please read my post on excessive screen current before you start)
|Page 1 of 4||Next>||Last Page>>|