I wasn't saying it would die right there on the spot.
I was trying to point out that repeatedly disconnecting it from AC (either by switch or physically unplugging it) would shorten it's lifetime. This means that somewhere down the road those repeated surges could take their toll and it would quit before it's time... It might be a year instead of 5 but each major surge does bring you one step closer.
Especially in laptop chargers, feeding a device with a large current inrush, like these mini-amps can play havoc because they may not be designed to cope with it. Perfectly reliable with a battery to buffer the surge... but somewhat unpredictable otherwise. I've had a couple that died when retasked to mini-amps.
External SMPS bricks, which are now readily available, are assumedly designed as you described... with better cold start protections, but we can't always count on that.
The PC reference was because people were shutting down, then turning off the baby sitter switch with the BIOS set for "On after power loss" so that the AC switch effectively booted up the system. I had several supplies that failed because of this, blown rectifier diodes. The failure was not immediate, in one case it took a couple of years... but it does significantly shorten the power supply's lifetime. Most PC supplies last well beyond the computer's useful life, except for this.
You need to see and hear the spark when you plug them in to understand just what kind of surge I'm talking about.
This is why I advised the OP to plug it in and leave it connected... to avoid repeated surges.
I don't know how this got so confused and I certainly don't know how "shortened lifetime" got translated into "catching fire"...
Last edited:
Yes, I get what you are saying, I didn't understand why you said there was no inrush protection. Also, as Kay says that will also protect it independent of load. A few people have mentioned catching fire, so it is an issue that failure could also cause in a cheap, poorly protected smps.
There sometimes isn't. Especially in the cheaper ones.
No argument. It's the right thing to do.
First I've heard of that ... but duly noted.
It looks terrible. 1. No manufacturer name. 2. no independent laboratory mark (UL, Tuv...) 3. if you remove the plastic feet, and you see screws inderneath, this means that they did not have a ultrasonic welder and they bought a generic plastic shell or that they don't want a permanent seal because they may need to do a few tweaks on the PCB before selling it.
As I said on a previous post, long term reliability of this kind of products cannot be determined. I would avoid plugging/unplugging it from the mains too often.
This behaviour is undesirable also because when the power supply is off, the PC clock will drain the 3V motherboard lithium battery that also keep the Bios/Uefi configuration parameters. Off-brand manufacturers don't test for leakage on filter capacitors on this supply line, and I've seen batteries totally drained in a month. This may be a big issue for non technical people that are unable to replace the battery and select the correct parameters. Some boards don't even have a proper quartz oscillator and the clock will drift when running from the slightly lower battery voltage.
I feel the same way - lots of powering on and off is hard on modern electronics.
Thermal cycling (going from hot to cold and back repeatedly every time the device is powered off and then back on) is yet another mechanism that can cause failure over time, if power to the device is frequently cycled.
Even when we were all using rotating mechanical computer hard drives, the recommendation was to leave computers on all day, and maybe continuously for days, weeks, or months on end, rather than turning them on and off. Otherwise several of the failure mechanisms mentioned shorten the life of motherboards, disc drives, CPUs, and power supplies.
It seems counter-intuitive that a constantly spinning platter would last longer than one that spent lots of time at rest, but that is in fact the case.
-Gnobuddy
I've rarely heard of PC power supplies going bad, let alone the PC's themselves.
But I'm sure an occasional one does take a crap.
I surmise that manufacturing build quality is to blame there.
I've had computers since 1990, I started out with Windows 3.1.
Since then, and many computer "upgrade" since, I've only had ONE serious malfunction - a hard drive going bad.
In another, the memory stick crapped out.
But I've never, ever, had an issue in 29 years with shutting OFF my computers, or laptop - completely, when I'm through using them.
Cheap PC supplies go bad all the time. Often taking other components with them.
Well made, properly designed switching PSUs are indeed very reliable.
It's when you get into no brand or bottom rung stuff you have problems. Poor design, little or no protective circuitry, and corners cut on components all play a part.
I'd be pretty comfortable (and indeed I do) leaving power bricks running if they're made by a reputable company, ie laptop bricks which came with dell, HP and the likes. They still get bad ones sometimes, but usually they're subject to recalls. And they will at least have been be UL listed, safety tested, etc.
Some no brand $20 thing from AliExpress? You takes your chances. Is the secondary properly isolated? Did they use new components or salvage? Are the components rated correctly for the application? Does it have proper safety cutouts in case things go wrong, like it overheats or shorts out? Did they use a fire retardant plastic? You have no idea. It's a gamble.
Well made, properly designed switching PSUs are indeed very reliable.
It's when you get into no brand or bottom rung stuff you have problems. Poor design, little or no protective circuitry, and corners cut on components all play a part.
I'd be pretty comfortable (and indeed I do) leaving power bricks running if they're made by a reputable company, ie laptop bricks which came with dell, HP and the likes. They still get bad ones sometimes, but usually they're subject to recalls. And they will at least have been be UL listed, safety tested, etc.
Some no brand $20 thing from AliExpress? You takes your chances. Is the secondary properly isolated? Did they use new components or salvage? Are the components rated correctly for the application? Does it have proper safety cutouts in case things go wrong, like it overheats or shorts out? Did they use a fire retardant plastic? You have no idea. It's a gamble.
Do you have any recommendation for 32V SMPSs that can supply at least 5A?
For my TPA3255 amp, I may be able to buy Mean Well LRS-350-36.
For my TPA3255 amp, I may be able to buy Mean Well LRS-350-36.
Last edited:
Both should be happy at 36v. You could even go higher for the TPA3255. This assumes the rest of the board components are up to it.
Here are a couple of suggestions, there's plenty of options.
Mean Well GST120A24-R7B
Mean Well GST220A36-R7B
You've been luckier than I have. I've replaced several dead power supplies in my own PCs during roughly the last 20 years.
Quite recently, a friend complained to me about the stench in his spare room after the power supply in his old PC fried, taking the motherboard with it, and leaving behind the acrid stench of burned electronics. He was quite pleased to discover that (a) I had a spare power supply I gave him (it won't fit any of my current PCs), and (b) that I had a commercial-grade ozone generator, which made short work of the stench.
Over the last 20 years, I've assembled almost all of my PCs myself, rather than buying Dell or similar large-volume manufacturers products. Make what you will of that - I was using aftermarket power supplies in all my PCs.
Statistics have been kind to you. 🙂
I've had to replace more than one dead hard drive over the years. Fortunately I had a one-day-old backup each time, and nothing crucially important had been done in the last 24 hours.
Once a day, more or less?
I switched entirely to Linux at home starting in 2001. Linux boxes inherited some characteristics from older Unix boxes that ran for months or years on end; so Linux boxes traditionally wake up in the wee hours of the morning and do various bits of house-cleaning, so it is best to leave them on, so that they can run these various cron jobs.
The world is full of server farms that run 24/7 nowadays, and the hardware is not always different from your generic consumer PC. I remember reading a story about a journalist's visit to Google in their early days - their server room was filled with vast numbers of wire shelving racks, with metal trays placed on them, each tray with a motherboard and hard drive sitting in it, humming away quietly.
They didn't bother to put a box around the things - all an enclosure does is take up room and impede cooling, right? So the motherboards sat out in the open, roughly 100,000 of them at the time. If one unit died, you just unplugged power and networking, slid out the tray, and replaced it with another one.
(I assume there were stand-offs or an insulating layer keeping the motherboards from shorting out on the metal tray it sat in.)
-Gnobuddy
Pure speculation...
It might ... if the bulk capacitors in the high voltage primary are slowed in their initial charge, they may transition through a longer stage of instability that the switcher is not designed to cope with causing spurious startups on the outputs... which could be even more damaging than frying the input rectifier (which is what usually happens).
The company I worked for used to sell a line of PCs that we assembled in the back room. We used quality parts and sub-assemblies throughout and we still had occasional power supply failures... almost always it was the primary rectifiers in the SMPS module that got fried.
It took some time to figure it out but mostly it was due to "night time shutdown" and the inrush of current that followed in the next morning's startups. Contacting the manufacturer (In-Win) gave us the answer that "These are not meant to be turned on and off every day".
Yes ... and it's pretty easy... use the one that came with your amp. There is likely to be a good reason why they chose that particular module.
Lets get something clear ...
These chip amps are inexpensive, not because they're poorly made, but because Class D topology reduces the complexity to the point where the most expensive single part is the enclosure. A 100w amp, 20 years ago would cost upwards of $500... the mind numbing simplicity and effiency of these new amps reduces that price to less than $100, and leaves enough room to give you a supply brick in the process.
Do not mistake "inexpensive" for "cheap" ... things have changed, rather a lot, in the last few years.
Last edited:
Douglas Blake, The chinese generic TPA3251/TPA3255 amplifiers(39~41USD) I bought from aliexpress.com didn't come with a power supply. The seller recommended a 32V 5A power supply for the generic TPA3251 amp, and I ordered one(~18 USD) from aliexpress.com.
The generic TPA3255 amp that will arrive today says
Amplifier chip: TPA3255
Output power: 300W+300W
(Actual THD is ~1% at 4ohm, 225W+225W, and 48V. THD seems to be 0.1% at 4ohm, 200W+200W, and 48V. 200W+200W at 4ohm and 48V is more realistic.)
Number of channels: 2.0 channels/Stereo
Power supply: DC24-48V (outer diameter 5.5mm inner diameter 2.5mm)
Recommend DC48V 5A power adapter
(The seller doesn't know much about amplifier. 48V may be dangerous for some components on the amp board. I will inspect board components.)
Impedance: 4-8 Ohm
I'm considering three options for this TPA3255 amp.
* The chinese 32V 5A adaptor that I have now
* Mean Well LRS-350-36 (36v 350w)
* Mean Well LRS-350-48 (48v 350w)
The generic TPA3255 amp that will arrive today says
Amplifier chip: TPA3255
Output power: 300W+300W
(Actual THD is ~1% at 4ohm, 225W+225W, and 48V. THD seems to be 0.1% at 4ohm, 200W+200W, and 48V. 200W+200W at 4ohm and 48V is more realistic.)
Number of channels: 2.0 channels/Stereo
Power supply: DC24-48V (outer diameter 5.5mm inner diameter 2.5mm)
Recommend DC48V 5A power adapter
(The seller doesn't know much about amplifier. 48V may be dangerous for some components on the amp board. I will inspect board components.)
Impedance: 4-8 Ohm
I'm considering three options for this TPA3255 amp.
* The chinese 32V 5A adaptor that I have now
* Mean Well LRS-350-36 (36v 350w)
* Mean Well LRS-350-48 (48v 350w)
Last edited:
I agree. I remember when I needed a massive finned metal heatsink for a 30 watt, solid-state, discrete, class AB push-pull power amp - a heatsink about the size of a hard-cover novel!
I think one of the reasons why people get the wrong impression about today's class D amps is the manufacturer's insistence on quoting power output with the amplifier deep into clipping, at 10% THD. That's a lot of very nasty-sounding distortion.
But take a closer look at the datasheet, and you'll find the distortion is inaudibly low - below the threshold of human detectability - over most of the output power range. Keep one of these inexpensive little class-D chip amps out of clipping, and it's audibly perfect.
Imagine, lower distortion and bigger output power than the famed Williamson amps of yesteryear, at a tiny fraction of the size, weight and cost. Incredible, but true!
-Gnobuddy
And this needs a three page thread with you reasking the same question over and over because???
You asked if it is safe to leave it on all the time... the answer is YES.
And any one of those power supplies will work just fine.
Last edited: