Mine has no power switch, just like many other high-end amps.
My view is that if the amp requires to be on all the time in order to provide good sound quality, there's an issue with the design. Of course, this doesn't prevent any manufacturer from removing the power switch (saving cost) and marketing that as a high end feature.
Tom
In my experience as an IC circuit designer for 10 years and electronic hobbyist for 30+ years, if the circuit performance varies significantly between hot and cold, it's not production worthy.
I think the whole "amp warmup" issue is left from the days of vacuum tubes. In tube amplifiers, the performance does improve as the tube warms up. There's also significant burn-in that happens. In my tube amps, I re-bias new tubes after a couple of hours of operation because the operating point has drifted significantly. I re-bias again after another couple of hours. After a good 8-10 hours, the tube characteristics have finally settled.
With semiconductor amps, there are no such issues. The device chemistry does not change as long as you operate the devices within the specs of the part. You'll need to heat the parts up to several hundred ºC to cause the diffusions to shift. There's also the little detail called "loop gain". Solid state amps rely on loop gain to linearize the circuits. The loop gain does depend on temperature (less loop gain at hot), but the variation is generally not enough to cause any meaningful variation in the closed loop performance.
I maintain that if the performance of a solid state amp depends on the operating temperature, there's something wrong with the design. I'd also claim that any perceived difference in sound quality between hot and cold is caused by expectation bias or confirmation bias; unless there is a design flaw in the amp, that is.
And yes. It's an interesting quirk of the human psyche to cherry-pick data points and perceive things that back up preconceived notions.
Tom
I think the whole "amp warmup" issue is left from the days of vacuum tubes. In tube amplifiers, the performance does improve as the tube warms up. There's also significant burn-in that happens. In my tube amps, I re-bias new tubes after a couple of hours of operation because the operating point has drifted significantly. I re-bias again after another couple of hours. After a good 8-10 hours, the tube characteristics have finally settled.
With semiconductor amps, there are no such issues. The device chemistry does not change as long as you operate the devices within the specs of the part. You'll need to heat the parts up to several hundred ºC to cause the diffusions to shift. There's also the little detail called "loop gain". Solid state amps rely on loop gain to linearize the circuits. The loop gain does depend on temperature (less loop gain at hot), but the variation is generally not enough to cause any meaningful variation in the closed loop performance.
I maintain that if the performance of a solid state amp depends on the operating temperature, there's something wrong with the design. I'd also claim that any perceived difference in sound quality between hot and cold is caused by expectation bias or confirmation bias; unless there is a design flaw in the amp, that is.
And yes. It's an interesting quirk of the human psyche to cherry-pick data points and perceive things that back up preconceived notions.
Tom
Last edited:
When you first bring up a Pass F5 or F5T, the performance of the semiconductors *will* vary as their temperatures change. You can see the variance of operation on the graphs supplied from the manufacturer. Rebiasing is necessary after an hour, then a few hours, then a day then a week or so, until the unit stabilizes.
Does this mean it is poorly designed?
Does this mean it is poorly designed?
Not knowing how significant loop gain may vary with temperature, nor have I seen any data as to how it may vary across the spectrum with temperature, but I have just dinished fine tuning equalization about 1db variation. This makes a difference when varies over some frequency ranges. Then the phase also changes, where causing a system to come closer to linear phase does make a difference as well. So to say warming up does not make a difference in solide state electronics just means one has not experienced it.
Lots of people experience such that warm up takes significantly longer in solid state electronics due to the lowere power. Is a warmed up device always the best with solid state electronics? I would say not necessarily, but you always want the thing on a few days before you decide.
Lots of people experience such that warm up takes significantly longer in solid state electronics due to the lowere power. Is a warmed up device always the best with solid state electronics? I would say not necessarily, but you always want the thing on a few days before you decide.
My personal experience is that sometimes it is obvious sometimes indistinguishable. When there is a difference, you hear some rough edges when cold, after about 30 minutes or more, it will become smooth and open. The longest demonstrated to me was about two hours, which was while I was in Texas. The machine was turned on, we listened, went out for dinner, then came back to listen. No wine was involved.
I would argue that it is poorly designed. The circuit should be designed to be largely independent on the performance of anything but the passive components (which can be manufactured to be near-ideal). If this requires some sort of servo to handle the biasing, build that into the circuit. No semiconductor circuit should require periodic tweaking. I'm willing to accept an initial bias adjustment, but that's it.
If the Pass F5/F5T requires you to tweak, re-tweak, and re-re-re-tweak, then it's not production worthy in my book. It may be a good DIY circuit, as many DIYers enjoy tweaking. I much prefer a set&forget approach. No offense to Mr. Pass. I'm sure he reserves his best circuits for the jobs that pay. That seems like a reasonable optimization to me.
Of course the performance of the semiconductors vary with temperature. I've never argued against that. My main argument is that the closed-loop performance of the amp should be mostly independent of the performance of the semiconductors. Loop gain is one way to accomplish this. In a "no global negative feedback" design, the loop gain is the gain in the many local loops present in such amplifiers.
You could take any schematic for a discrete amp and plop it into a circuit simulator. Simulate the loop gain vs temperature. That happens to be part of my day job. I get pretty disappointed if I can't get the loop gain to vary less than ±5 dB across -40 C to +135 C (including power supply voltage variation as well). ±5 dB may sound like a lot, but when you start at 80-120 dB of loop gain, ±5 dB results in a negligible difference in the closed-loop performance.
If you don't "believe" in simulation, you can build the circuit and put it in a thermal chamber. Measure the loop gain with a network analyzer. You'll see the same thing.
Tom
Last edited:
Curious: Is the amp always on? What happens when the amp is turned off, is allowed to cool down, and then turned back on? Does it go through another week-long warm-up?
Tom
Glad to see my "switch from hell (x4)" has perpetuated its evil legacy by igniting a spirited discussion. Happy to share gentlemen.
I defer to others as to electronics, but I am very comfortable in many areas of chemistry and physics and as such I cannot comprehend the thermodynamics of a device "warming up" beyond an equilibrium temperature that is likely reached in 30-90 minutes. Help me understand please, beyond perceived SQ is there an explaination that lines with established laws of physics?
I defer to others as to electronics, but I am very comfortable in many areas of chemistry and physics and as such I cannot comprehend the thermodynamics of a device "warming up" beyond an equilibrium temperature that is likely reached in 30-90 minutes. Help me understand please, beyond perceived SQ is there an explaination that lines with established laws of physics?
Member
Joined 2006
A decaying surface charge redistribution phenomenon perhaps...happening at the interfaces of plastic and metal ?..Similar effects like in cables?.It takes time for electronic charges to "leak out" at the plastic interface to meet up with the conducting metal to reach a.steady state. So its not a warming up on the active devices side issue per se I think... Tube amps need warming up as well in this way..Just a thought... :
Member
Joined 2006
A bit off topic...some like old western electric cable as IC, why,?perhaps because the jacket is made out of .....cloth!, storing no charge at all at the interface....maybe its as simple as that..haha..and old copper surface already passivated...:and skin effect.then..etc etc..but thats another.issue...
Anyway, once electricity is off..charges and traps get redistributed again, only when you turn the power on...they will come out again and haunt your signal..and until.....the steady state arrives..haha
An analogy would be like clean river streams in a somewhat muddy river bed.....when the dirts get settled..you got the point...
Anyway, once electricity is off..charges and traps get redistributed again, only when you turn the power on...they will come out again and haunt your signal..and until.....the steady state arrives..haha
An analogy would be like clean river streams in a somewhat muddy river bed.....when the dirts get settled..you got the point...
Last edited: