I use two hifi amps, one SS and one tube amp single ended 6V6 and have a speaker switch. Stupid me forgot to turn the tube amp off when I switched to SS amp. Has been running for about 12 hours without load. But nothing happened to it as far as I can hear. I'm just a little concerned about tubes, what are the consequences of running them without load?
But I'm extremely happy that output transformers survived it.
But I'm extremely happy that output transformers survived it.
without any input signal, that amp is expected to live...
i do that to my amp builds to see if anything will burn out...when first powered up..
i do that to my amp builds to see if anything will burn out...when first powered up..
Dave Gillespie over on AK runs his amps with no load and drives them with input signals all the time for stability testing. So it all depends on what you do to it while you have no speakers attatched. The constantly repeated warnings about never running a tube amp with no load is just a half truth that continues to be told by folks who don't really understand the condition and just want to give their best advice.
Last edited:
I think Dave always uses a load for stability testing. Either a simple resistor or an RC in parallel. There isn't anything to stability test without a load.
He very explicitly details his tests into, no load, capacitive only and resistive. His last testing was posted about some Heathkit amps he was tweeking within the last 3 weeks. In each case he tries to shock the amps or sends a sweep into them to try to create oscillation.
Last edited:
Maybe "no load" means: without external load.
I am quite confident of my builds, but I would really never leave the secondary of the OPT completely unloaded. I generally solder a 100R resistor at the speaker terminal, so that it can indeed run without an external load.
When i was new to testing anything with a signal generator, I burnt out a Dynaco Mk 4 mono output trans once driving it with a square wave. No load except the scope.
The nice square corners suddenly started to fade and round and then the smell!
Real dumb!
The nice square corners suddenly started to fade and round and then the smell!
Real dumb!
As Erik said.
In the heyday of tubes it was common practice to wire 1K.ohm over the output terminals. For 8 ohm output load I wire 680 ohm across the output terminals. I don't want to sound boastful, but I always had doubts about amplifiers which go unstable with no (external) load. It is simply easy enough to get NFB to be stable with some resistance over the output, when the rest of the design is proper.
In the heyday of tubes it was common practice to wire 1K.ohm over the output terminals. For 8 ohm output load I wire 680 ohm across the output terminals. I don't want to sound boastful, but I always had doubts about amplifiers which go unstable with no (external) load. It is simply easy enough to get NFB to be stable with some resistance over the output, when the rest of the design is proper.
Someone deliberately running a valve amp with no load (even with no input signal) is either foolish or sufficiently wealthy to afford new OPTs from time to time. This is the 'NASA' approach to safety: "it hasn't happened yet, so it won't happen"; then one day it did happen.
30 yrs ago (hly sht, already 30) we had a bass head in our band, anything else was SS. That seemed like the someone's homebrewed, AFAIR made on a quad of 829B-likes amp. Worked fine, but once on rehearsal been turned on with no load: all 4 burnt with sparkles in few minutes. The trafo survived, but it was pain in the arthe to find the replacements.
Consider a pentode wired amp that has negative feedback.
Suppose it has a 6k Ohm plate to plate primary impedance when there is an 8 Ohm load on the 8 Ohm tap.
Now, instead, put a 80 Ohm load on the 8 Ohm tap. Now the primary impedance could be as high as 60k Ohms plate to plate.
The open loop to closed loop gain ratio has increased by 10 times (20 dB!).
If the negative feed back was 20 dB with an 8 Ohm load, it is now effectively 40 dB with the 80 Ohm load.
You ask yourself, how much phase stability does the amp have in this condition?
Typically, since you did not hear any sound, you turned the volume control to full.
If the amp is stable under that condition, you are lucky.
And if it does not arc either the output transformer, the output tubes, or both, you are lucky.
If not, you are not so lucky.
Good luck!
Now, consider a push pull triode (or triode wired pentode) amp with no negative feedback, and a damping factor of 3.
Load the 8 Ohm tap with 8 Ohms. The gain of the output stage is 3/4 times the output tube mu (u).
Next, do not connect any load on the 8 Ohm tap. The gain of the output stage goes up by 1.33 times (about 2.5 dB).
You have a better chance that it will not be damaged.
But again, you turn the volume up Full, hoping to hear something.
But, do not leave ANY amp unloaded, because "Your Mileage May Vary".
Good Luck!
Suppose it has a 6k Ohm plate to plate primary impedance when there is an 8 Ohm load on the 8 Ohm tap.
Now, instead, put a 80 Ohm load on the 8 Ohm tap. Now the primary impedance could be as high as 60k Ohms plate to plate.
The open loop to closed loop gain ratio has increased by 10 times (20 dB!).
If the negative feed back was 20 dB with an 8 Ohm load, it is now effectively 40 dB with the 80 Ohm load.
You ask yourself, how much phase stability does the amp have in this condition?
Typically, since you did not hear any sound, you turned the volume control to full.
If the amp is stable under that condition, you are lucky.
And if it does not arc either the output transformer, the output tubes, or both, you are lucky.
If not, you are not so lucky.
Good luck!
Now, consider a push pull triode (or triode wired pentode) amp with no negative feedback, and a damping factor of 3.
Load the 8 Ohm tap with 8 Ohms. The gain of the output stage is 3/4 times the output tube mu (u).
Next, do not connect any load on the 8 Ohm tap. The gain of the output stage goes up by 1.33 times (about 2.5 dB).
You have a better chance that it will not be damaged.
But again, you turn the volume up Full, hoping to hear something.
But, do not leave ANY amp unloaded, because "Your Mileage May Vary".
Good Luck!
Last edited:
sings like a turn table pick up,
I think the output tubes takes extra heat dissipation and output transformer acts as a speaker coil, if it overheats or get too high voltage it is going to be good for garbage.
I think the output tubes takes extra heat dissipation and output transformer acts as a speaker coil, if it overheats or get too high voltage it is going to be good for garbage.
6A3sUMMER,
You are assuming that tube gain is directly related to Rl - it is not. (Otherwise you will end up with a stage gain of > tube µ, which is clearly impossible.) Also, the stability is not directly related to the gain - with a higher load e.g. leakage reactance influence diminishes.
Yes, with straight pentode output stages the problem becomes more likely, far less so with triode or UL feed. Otherwise you will have to explain why stability is not inversely proportional to the load. See my previous post: There are many designs which are quite happy without an external load, having been properly designed to begin with. It is not a simple case of stability vs. load - you have to bring Nyquist criteria in. Not to boast, but I have never put out a tube design which was not stable witout external load (and I have been at it for a fair time ....) I did mention earlier that mostly a nominal internal load of some 600E - 1K resistance is required. Some designs feature a Zobel network at the output.
This does not infer that loads can be removed willy-nilly. - I am not saying that. I have also seen amplifiers go berserk wihout a load, or enough capacitance across the load. I am simply stating that a tube design can be designed to be stable without an external load.
You are assuming that tube gain is directly related to Rl - it is not. (Otherwise you will end up with a stage gain of > tube µ, which is clearly impossible.) Also, the stability is not directly related to the gain - with a higher load e.g. leakage reactance influence diminishes.
Yes, with straight pentode output stages the problem becomes more likely, far less so with triode or UL feed. Otherwise you will have to explain why stability is not inversely proportional to the load. See my previous post: There are many designs which are quite happy without an external load, having been properly designed to begin with. It is not a simple case of stability vs. load - you have to bring Nyquist criteria in. Not to boast, but I have never put out a tube design which was not stable witout external load (and I have been at it for a fair time ....) I did mention earlier that mostly a nominal internal load of some 600E - 1K resistance is required. Some designs feature a Zobel network at the output.
This does not infer that loads can be removed willy-nilly. - I am not saying that. I have also seen amplifiers go berserk wihout a load, or enough capacitance across the load. I am simply stating that a tube design can be designed to be stable without an external load.
I thought bypassed common cathode triodes had gains of (u * RL)/(RL + rp).
Yes, agreed, that means gain is not Directly proportional to RL. It has to have the rp term accounted for too.
Yes, agreed, that means gain is not Directly proportional to RL. It has to have the rp term accounted for too.
No it wasn't deliberate. I am well aware the OT can be damaged without load. Happened to me when I was a kid being amazed that I could hear sound from transformers if I disconnect the speaker. It eventually got hot and died, hehehe.
But I've build this amp and made sure I have headroom for errors.
But I was mainly asking what happens to output tubes running with no load (speaker). The dissipation must go somewhere but plates weren't glowing or anything.
But I've build this amp and made sure I have headroom for errors.
But I was mainly asking what happens to output tubes running with no load (speaker). The dissipation must go somewhere but plates weren't glowing or anything.
If the amp is stable, nothing special will happen with the output tubes. If they are in Class A they will dissipate V*I, so max power (which they should be able to handle) as no power is being delivered to the speakers. In Class AB also the V*I, which is probably lower than the max anode dissipation anyway.
so nothing to worry about!
so nothing to worry about!
For a pentode gain = gm x Rl, to a good approximation. Pentodes have very high mu, so high that it is rarely mentioned as it is not usually a constraint.Johan Potgieter said:
Of course, but usually this involves some form of internal load which is always connected. In most amplifiers this protection is omitted so most amplifiers should not be run with no external load.
There are two dangers with no load: feedback loop oscillation, parasitic oscillation within the output stage. Both depend on parasitic reactances, so both are difficult to design out without careful testing. Necessarily, such testing will put the OPT at risk so it is rarely done. How many people can say that they have tested all possible (or all likely) output loads and so found that their amp is stable under all possible (or plausible) load conditions?
- Status
- Not open for further replies.