form a component durability standpoint (especially OPS transistors), when not listing to music, is it better to:
- leave your amp in hot standby (power on no signal applied but idling)
- power it off
i always thought that it was the thermal cycles (heat up then cool down) that wore out components, so i have always left my amps on continuously. but i was talking to a stereo guy i met recently and he advised to always turn amp off because its the heat itself that kills, not the cycling between hot and cold.
which is right?
- leave your amp in hot standby (power on no signal applied but idling)
- power it off
i always thought that it was the thermal cycles (heat up then cool down) that wore out components, so i have always left my amps on continuously. but i was talking to a stereo guy i met recently and he advised to always turn amp off because its the heat itself that kills, not the cycling between hot and cold.
which is right?
hey thanks for the reply. yeah it idles pretty cool when i touch the heatsinks when nothing's been playing for a while i barely feel warmth. just enough to let me know its on. in fact, i use the 3 second rule on all of my amps ... if i can't keep my fingers comfortably on the component for 3 seconds without pain, its too hot.
some may say thats overly conservative but i like my stuff to run cool enough to touch.
some may say thats overly conservative but i like my stuff to run cool enough to touch.
There is a term called "Working hours lifetime" for all components.
It may be better to leave your amp on for a short time (few hours) if you know that you will be listening again very soon, however if the design is made right, the transistors won't suffer from switching the amp On/Off when not listening.
Much worse is it for the electrolytic caps (PSU caps etc.)! These have normally only a working lifetime of 2000 hours, so by leaving your amp On all the time quickly eats up the caps lifetime.
It may be better to leave your amp on for a short time (few hours) if you know that you will be listening again very soon, however if the design is made right, the transistors won't suffer from switching the amp On/Off when not listening.
Much worse is it for the electrolytic caps (PSU caps etc.)! These have normally only a working lifetime of 2000 hours, so by leaving your amp On all the time quickly eats up the caps lifetime.
What about metal fatigue, would it be an issue particularly the bonding wires that connect the pins to the die? Hot/cold/hot/cold ect will take a toll on the wire bond strength, just like the filament in a light bulb. Incandescence bulbs tend to last longer if they are left on continuously compared to intermittent use. I saw a program once about an old firehouse light bulb that has been on for like 75 years. Nobody dares turn it off because the filament may break from temperature change and age and no one wants to be the one responsible for it’s death.
They just don't make 'em like they use to anymore.
They just don't make 'em like they use to anymore.
It depends a bit on the amp.
ClassB-amp: no thermal cycling to speak of when switching on/off; only thermal cycling is caused by listening to music. Leaving it on all the time will hurt your power supply caps (as ACD already wrote).
(By the way there's the old rule for temperature: every 10°C temperature increase cut lifetime of the part in half. So a 85°C cap with 2000 hours lifetime translate into 32000 hours at 45°C - in about 4 years permanent on the caps are outside their specified lifetime.)
ClassA-amp: thermal cycling due to Class A; however there's not much benefit again for leaving it on as the caps have to endure high ripple current and higher ambient temperature. Both reduce lifetime.
In short: switch it off at least for the night.
Have fun, Hannes
ClassB-amp: no thermal cycling to speak of when switching on/off; only thermal cycling is caused by listening to music. Leaving it on all the time will hurt your power supply caps (as ACD already wrote).
(By the way there's the old rule for temperature: every 10°C temperature increase cut lifetime of the part in half. So a 85°C cap with 2000 hours lifetime translate into 32000 hours at 45°C - in about 4 years permanent on the caps are outside their specified lifetime.)
ClassA-amp: thermal cycling due to Class A; however there's not much benefit again for leaving it on as the caps have to endure high ripple current and higher ambient temperature. Both reduce lifetime.
In short: switch it off at least for the night.
Have fun, Hannes
Hi
I guess I go along with ha and ACD.
In practice I have found that the longest use I have had from a piece of equipment was from a Sony professional tv. That lasted 20 years. If you assume 2 hours per day that gives 600 hours per year ( time off for hols ) or 12,000 hours.
I have also assumed that electrolytic caps dry out slowly over the years and this is exascerbated by keeping them warm. ( find thaty electrolytics last a long time if stored in a sealed plastic bag.)
I switch equipment off at nights.
Don
I guess I go along with ha and ACD.
In practice I have found that the longest use I have had from a piece of equipment was from a Sony professional tv. That lasted 20 years. If you assume 2 hours per day that gives 600 hours per year ( time off for hols ) or 12,000 hours.
I have also assumed that electrolytic caps dry out slowly over the years and this is exascerbated by keeping them warm. ( find thaty electrolytics last a long time if stored in a sealed plastic bag.)
I switch equipment off at nights.
Don
some of you all's are forgetting your basic electronic theory who are talking about capacitors and temperature and failure times.
capacitors do not produce heat. voltage and current are 90 degrees out of phase, thus no power is being produced in the device
therefore, caps operate at ambient temp the whole time. sorry as much as i hate correcting people some needed to be reminded of this fact about capacitive reactance.
capacitors do not produce heat. voltage and current are 90 degrees out of phase, thus no power is being produced in the device
therefore, caps operate at ambient temp the whole time. sorry as much as i hate correcting people some needed to be reminded of this fact about capacitive reactance.
PMA said:
then it is coming from some external source. heatsinks perhaps if class A? trust me or not, but nature wont allow a capacitor to generate any amount of heat under any circumstances. this isn't to say you might have a corrosion on a cap terminal that is creating resistance. now that WILL heat up the cap, but again, the key is the source of the heat is external.
so does this mean that continuous operation extends MTBF? thanks very much for your reply btw.
It is not important (external or internal).
I can see you (gain) do not know much about caps. Imagine polyester foil cap, with max allowed dv/dt. In case you exceed dv/dt for longer period of time, the cap self heats (dielectric losses) and may explode. This happened to me when I made longterm test of power amp into dummy load by 10kHz square - the foil polyester cap overheated and exploded.
I can see you (gain) do not know much about caps. Imagine polyester foil cap, with max allowed dv/dt. In case you exceed dv/dt for longer period of time, the cap self heats (dielectric losses) and may explode. This happened to me when I made longterm test of power amp into dummy load by 10kHz square - the foil polyester cap overheated and exploded.
PMA said:
this is not a valid counterpoint imho. you were operating the cap way outside of its specified parameters. of course if you hook up a 50V cap to a 500V rail its going to destroy itself.
if you don't believe me, which obviously you don't, just do the math yourself. take cap equation
q = CV = CVsin(wt)
take derivative with respect to time
dq/dt = I = CVwcos(wt)
so V = Vsin(wt) and I = CVwcos(wt)
hence they are 90 degrees out of phase!!! hence no power (heat) can be produced within a cap!
PMA said:
me neither
gain said:
MTBF extends when the system is not powered
However, MTBF worsens when at turn on turn of conditions. There is a simple light bulb test, leave it off and it will last for an infinite period, zero stress just fatigue. Turn in on and leave it, it will run for its predicted life MTBF. Turn it on and off for a thousand times and the life is reduces exponentially.MTBF reduces by the number of components likely to fail in a system and would normally boil down to the first most likely component failing.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.