Amp's tone changes when hot

[daz]

I'm just looking for ideas as to what the most common reasons for this are. I notice that after about 15-20 minutes my amp will start changing, and the difference i believe is enough so that i'm not imagining things. I had a marshall that did this too. I'm thinking tubes and that a fan may be in order. But before i try that i'd like to hear any other suggestions. The chassis is good size and not at all cramped, so i don't think thats an issue. the circuit is similar to a JCM800 2204. Outputs are SED el34's and the circuit from the PI on back is almost exactly 2204. Tubes are biased so that it tops when warm at about 38Ma. In fact i checked them when it changed last nite and with 424 VDC on the plates they were 37 ma after the change occurred.

Should i just put a fan in or does anyone have other ideas? Also, if i do put a fan in it what is the most efficient way...blowing at the tubes or away? Before you answer, note that the combo is very open in the back and doesn't even have a panel across the tubes like most combos have. (i made the cab and just never made one for it)

[km6xz]

The first thing is to take baseline measurements of basic operating condition: AC mains current draw and set a standardized mains voltage with the Variac, set a standard input amplitude, measure noise floor, frequency response and harmonic content with a spectrum analyzer/Tracking generator, and log the information. If the amp has a power amp in or effects return run it again with power amp only. Log this information, it comes in very handy if you get into the habit of doing with everything that in on your work bench.
It might take 10 seconds to establish these specs. If you do not have a spectrum analyzer, use your sound card to generate the sweep tones and view the spectral display on your computer. There are a number of free software packages that so what you want.
Here is a good one. Visual Analyser details

Then play. If you hear a change, run the same tests to determine if there was a change in the amplifier or room acoustics or ear fatigue. Don't discount the last one, it really changes our perception of sound throughout a session.

There are many parts that are expected to be heat dependent for parameters, and many parts that are not normally temperature dependent but a particular part might be and thus out of tolerance. It could be tubes, 1ma drift should not be a matter for concern however. Even speaker voice coils change with heat.

[overtone]

Daz, most of my old seventies English amps, like your Marshall mentioned, do this.
I put it down to the polyester caps and resistor drift as the amps heat up.
Have never taken any data, just accepted it and dealt with it.
Some have commented that the amps actually sound better after an hour.
"They hit you in the solar plexus" was one comment I particularly remember.
Does your's sound worse?

[R.G.]

Good one, overtone. Resistors and caps change value with temperature. The manufacturers list a temperature coefficient in the datasheet. Electrolytic caps *especially* do this, as do older carbon comps, perhaps others. Often the cathode caps on tubes are electros.

[Bruce / Mission Amps]

Good one, overtone. Resistors and caps change value with temperature. The manufacturers list a temperature coefficient in the datasheet. Electrolytic caps *especially* do this, as do older carbon comps, perhaps others. Often the cathode caps on tubes are electros.

Also, within a short amount of time at medium to loud volume levels.... your frickin' ears can start to shut down, become slightly fatigued and change your perception of tone.

[daz]

Of course i thought of component drift, but at a bit more than bedroom level for 15 min it seems like that might affect tube mounted components like screen or grid resistors on the power tubes, but it doesn't seem like such low levels in such a short time should affect the board. Certainly not ear fatigue at that level and amount of time i would think.

[exclamationmark]

+1 on the loud volume thing. My hearing seems to be really sensitive to big SPL's - If I don't play gigs/practice with proper plugs, within a minute of playing my ears will go from being able to identify all instruments clearly, to some horrible smeared mess where I can't even tell what I'm playing (nor anyone else for that matter!). I'm fairly sure this is one of those perception things, where no physical change actually happens, rather my brain is just trying to cope with horribly loud noises. The problem with comparing a hot amp to a cold one, is that there is easily a large time gap for you get 'ear fatigue' (even at low volumes) as well as the fact we really aren't very good at remembering what something sounds like 15 minutes later. Not to say that the tone doesn't change while hot, I just think we need actual measurements.

[pdf64]

My suspicion with this type of thing is the speakers, as mentioned earlier; especially if the power amp is run open loop, or there's a lot of presence boost. It's surprising how hot speaker magnets can get, particularly unvented low or medium weight types, after playing loud for a while. Celestion's site advises to leave the amp on for a few minutes before playing, to warm the voice coil up; just normal background hiss they reckon will do it.
And the hearing level shift / compression effect; we can kid ourselves that we aren't playing that loud but a dB meter generally indicates otherwise. Guitar speakers are so efficient that levels in the room, especially with distortion, are easily in the high 90s and as that's sustained / continuous / so much louder than normal ambient levels, our hearing system will likely react to it.
Pete

[Steve Conner]

I've definitely felt my speaker magnets get pretty hot. Sometimes I've even smelt toasty voice coils. :O However, I've never noticed the sound changing enough to be annoying. My ears certainly change as they get blasted, but I've had this happen often enough that I know what to expect (and use earplugs if the session turns out exceptionally loud)

I use polyester film caps similar to the Vishay 368 series (another "orange drop" lookalike) For small resistors I use 1% metal film, and for big ones, either metal oxide, carbon film or wirewound, 5%.

[km6xz]

I would agree with the amp-on for a few minutes but not for warming up the voice coil. That does not make sense, the total power dissipated in the VC on a typical amp while idling with audible but normal low hiss would be in the 2mw or less range. I am testing a 40 watt amp on the bench right now. Home brew 4xEL84. The residual hum and noise 150mv and the owner thinks that is too much. Sitting on the cabinet for a few minutes would transfer more heat than that to the VC.

To determine if the perceived change in tone is human or electronic in origin, getting the cold baseline measurements would be the first step otherwise it is all guessing. Our perception will change the most over a period of time of loud playing. Walking a few feet while playing changes the sound even more and any thermal drift of a functioning amp.
Every part changes in value with temperature but some have positive and others have negative temp coefficients, some parts drift in the opposite direction and negative feedback can't be ignored either in mitigating the effects of drift. When designing a circuit we take thermal change into consideration which is why there are some parts designed to have the opposite tempco. as conventional parts of the same type. We can get an instrument amp or oscillator to be less influenced by thermal drift by adding parts that cancel the drift of others.

Remember that we are also have changing sensitivities to different frequencies based on loudness. The "Loudness" compensation control on stereos alters the eq curve with volume to try to comply with our changing sensitivity by frequency versus loudness.

You have likely heard of the equal-loudness contours originally created by Fletcher and Munson which still bear their names 80 years later. The sound really IS different to our brains at different sound pressure levels.

[daz]

Guys, as i said i played it 15 min at home levels, about a bit louder than TV level. The speaker isn't going to over heat at those levels, especially a EV12L. As to hearing changes, I have no doubt that happens to me, as i note it every time i play for anything more than 5 minutes. Problem is, this is a much more obvious change and also affects the dynamics in that the clean tones i get by rolling back the guitar were staying considerably more dirty than at first.

[R.G.]

Problem is, this is a much more obvious change and also affects the dynamics in that the clean tones i get by rolling back the guitar were staying considerably more dirty than at first.

Hmmm. That's a useful bit of additional information. If the amp gets dirtier as it gets hot, that indicates to me that either the gain is going up or the bias is shifting on one or more tubes.

The tubes themselves do heat up, but the cathodes get heated up pretty quickly to nearly their final value, as the vacuum around them is a pretty good insulator. They can cool only by radiation, and that stabilizes quickly. The plates heat, but they don't have much say in tube gain, and I speculate that what hotter plates would do to a triode is reduce gain as secondary emission goes up from the hot plate surfaces.

Electrolytics do get lower ESR as they heat, and there is an electrolytic on the cathode of many preamp tubes. That cap directly affects the gain of the tube stage. It would be an interesting test (to me at least ) to temporarily replace the cathode caps with film instead of electrolytic and see if the heating-up change still happens.

Bias shift can also make an amp sound dirtier, especially as two stages one after the other tend to clip opposite peaks of signal if they drift the same direction because of the inversion of the signal in the first one. I don't think film caps drift much thermally, though.

Oh. I just thought of another test. Doh! It would also be interesting to set up an experiment where the amp got into its hot-mode tone shift, then cool specific parts with a can of freeze spray and a long nozzle. That was one of our favorite tricks to find components with thermal sensitivities back in the power supply lab. It doesn't help a huge amount if the thermal effect is distributed over many parts, but if one part is going hinkey, it's deadly.

[daz]

I can put a fan in back as a test. I will have to wait tho, as i can't play it more than a few minutes here and there more than about once a week. (complaints) I think however i may try a different set of tubes too because these are pretty well used. Or monitor these tubes better. I left the meter attached while playing and I have dual bias and a switch that goes from one tube to the other so i can check each one seperately. But i didn't switch it when i looked at the bias so the other tube may have been responsible. I'll check that next time to make sure it's not one tube's bias, then swap tubes if thats not it, them try a house fan.

But while on the subject, whats the best was to install a regular cooling fan in it? I may just do that anyways just for the peace of mind that the temps will not be as likely to get too bad. It's a big combo with no panel across the tubes. So it's very open therefore i figured blowing ON the tubes might be the better way. Maybe between the front of the amp and the power tubes so it's both blowing air out of the amp and over the tubes too? Thoughts?

Hmmm. That's a useful bit of additional information. If the amp gets dirtier as it gets hot, that indicates to me that either the gain is going up or the bias is shifting on one or more tubes.

The tubes themselves do heat up, but the cathodes get heated up pretty quickly to nearly their final value, as the vacuum around them is a pretty good insulator. They can cool only by radiation, and that stabilizes quickly. The plates heat, but they don't have much say in tube gain, and I speculate that what hotter plates would do to a triode is reduce gain as secondary emission goes up from the hot plate surfaces.

Electrolytics do get lower ESR as they heat, and there is an electrolytic on the cathode of many preamp tubes. That cap directly affects the gain of the tube stage. It would be an interesting test (to me at least ) to temporarily replace the cathode caps with film instead of electrolytic and see if the heating-up change still happens.

Bias shift can also make an amp sound dirtier, especially as two stages one after the other tend to clip opposite peaks of signal if they drift the same direction because of the inversion of the signal in the first one. I don't think film caps drift much thermally, though.

Oh. I just thought of another test. Doh! It would also be interesting to set up an experiment where the amp got into its hot-mode tone shift, then cool specific parts with a can of freeze spray and a long nozzle. That was one of our favorite tricks to find components with thermal sensitivities back in the power supply lab. It doesn't help a huge amount if the thermal effect is distributed over many parts, but if one part is going hinkey, it's deadly.

[R.G.]

But while on the subject, whats the best was to install a regular cooling fan in it? I may just do that anyways just for the peace of mind that the temps will not be as likely to get too bad. It's a big combo with no panel across the tubes. So it's very open therefore i figured blowing ON the tubes might be the better way. Maybe between the front of the amp and the power tubes so it's both blowing air out of the amp and over the tubes too?

Blowing air on the tubes is better than exhausting. And it also gives you the chance to ameliorate one of the issues with fans - dust buildup. The computer world's use of fans has made filters that fit right over small box fans really common. If you put a filter over the outside of the fan and aim the air from the fan into the amp, the filter grabs the dust first and minimizes what gets into the amp.

Also, the filter is outside and easy to see when it gets dirty. I would put the fan on the back with a filter on the fan-intake side, and aim the stream of air over the tubes, letting the air get out of the back any way it can. The filter will also help damp down the acoustic noise of the fan.

12Vdc fans for computers can be bought now that are very, very quiet. Another approach to quiet fans is to get a small AC powered box fan rated for 220/240vac and run it on the 120Vac in the amp. This makes it turn slowly and quietly.

You don't need a howling storm of air. Even a modest movement of air over power tubes will keep them quite cool.

[daz]

Thanks, thats what i'll do then. The euro voltage fan idea is a good idea and i will go that route if i have to buy one. I have a lot of them sitting around tho and i will check to see how many are A/C. I know at least one is but it's ancient.

Blowing air on the tubes is better than exhausting. And it also gives you the chance to ameliorate one of the issues with fans - dust buildup. The computer world's use of fans has made filters that fit right over small box fans really common. If you put a filter over the outside of the fan and aim the air from the fan into the amp, the filter grabs the dust first and minimizes what gets into the amp.

Also, the filter is outside and easy to see when it gets dirty. I would put the fan on the back with a filter on the fan-intake side, and aim the stream of air over the tubes, letting the air get out of the back any way it can. The filter will also help damp down the acoustic noise of the fan.

12Vdc fans for computers can be bought now that are very, very quiet. Another approach to quiet fans is to get a small AC powered box fan rated for 220/240vac and run it on the 120Vac in the amp. This makes it turn slowly and quietly.

You don't need a howling storm of air. Even a modest movement of air over power tubes will keep them quite cool.