Meanwell (and other) SMPS module reliablity?

Hello,

I'm looking at moving from LM317 based powersupplies to a premade SMPS module, most likely the Meanwell PD-2515.

http://www.meanwell.com/webapp/product/search.aspx?prod=PD-25

Has anyone any experience of these sort of things? I'm working on the premise that these modules will be inherently less reliably than a linear supply however, is that actually the case anymore?

Meanwell's test report indicates a lifetime of around 92k hours which would be fine but that assumes no defects, which would be the main problem. (As a side note, PD-2515 has 'high reliability' listed as a feature and they offer a 2 year warranty, other models don't list high 'reliability' but have a 3 year warranty...bit odd).

So yeah...any real life experience of Meanwell and other manufacturer's SMPS modules? Maybe you know about large audio companies using these?

Many thanks.

I would be reluctant to drop a switching supply retroactively into the middle of sensitive audio circuits, while scratch designs can incorporate  appropriate filtering.  In theory the lower heat should benefit everything inside the box, but PS is an unknown. I hope some here have experience they can share (I don't).

FWIW one of my first technician gigs was working on a DC to DC switching supply back in the 60s. Modern switching technology can be suitably quiet when done properly. I wouldn't expect a generic off the shelf switcher to be the quietest, but hopefully they publish a spec sheet.

JR

Noise spec on that meanwell is 50mV max whilst their test report showed 6mV at full load IIRC.  I'm not too worried about noise at the moment, I'll try it and I'll be able to see instantly if it's fine or not, long term reliability however is a little harder to ascertain.

I have used several of their 48v supplies, both for phantom power in mic preamps and for powering Yamaha PM-1000 modules.  Some of those are at least three years in the field with no issues.

They are solid.  There's at least one mastering compressor that uses an outboard Meanwell and achieves lower noise than with the onboard linear they previously used.  I've got a bunch of onboards in various projects with good noise specs and reliable so far. 

FETlife said:

Hello,

I'm looking at moving from LM317 based powersupplies to a premade SMPS module, most likely the Meanwell PD-2515.

http://www.meanwell.com/webapp/product/search.aspx?prod=PD-25

Has anyone any experience of these sort of things? I'm working on the premise that these modules will be inherently less reliably than a linear supply however, is that actually the case anymore?

Meanwell's test report indicates a lifetime of around 92k hours which would be fine but that assumes no defects, which would be the main problem. (As a side note, PD-2515 has 'high reliability' listed as a feature and they offer a 2 year warranty, other models don't list high 'reliability' but have a 3 year warranty...bit odd).

So yeah...any real life experience of Meanwell and other manufacturer's SMPS modules? Maybe you know about large audio companies using these?

Many thanks.

Click to expand...

An SMPS has more parts for sure. But good manufacturers seem to make them pretty damn reliable. It's just not the weakest link anymore.

There are so many other advantages with SMPS I don't know why anyone would use linear anymore.

Noise is not an issue if you have a good common mode choke for each voltage. That in itself is probably quiet enough for most things but I also regulate down to the desired voltage. I actully use two stacked +30V SMPS and then regulate down to +-15 in the target gear.

Note that your +-15 supply after regulation is going to yield ~+-12V. So SMPS are not like transformers where you get a transformer that is +-15 RMS and then regulate down to +-15 DC. You need an SMPS that is ~+3V higher than the final regulated voltage. I'm not sure if they have +-18V supplies. You might have to just stack two +24V supplies. I used +30 V because they're constant current [1].

Otherwise the only real negative about SMPS is that it has to be external. Or at least far enough away from any high impedance high gain circuitry that might pick up on the super fast high current switching going on.

[1] My supply is completely ripped off of Sound Skuptor's PS2 which uses LPC constant current supplies. Not sure why they used constant current but the Sound Skulptor stuff is clearly well designed so I assume they know something I do not. Unless they deliberately lied in their assembly document to throw off copycats like me!

That particular product is VERY mature. I used it in the 1980s. I think it is widely used from computer/network gear to medical scanners. While the details must have changed, I'm sure they know to keep the reliability up.

What "defects"? Any defects not found by final inspection will count in the life-testing. If a seal has a 1000-hour leak, failures rise in the second month.

The number I see is "MTBF  507.9Khrs min. MIL-HDBK-217F (25deg)" The average death is a half-million hours or 57 years down the road. Ignoring infant mortality and wear-out, there's a 10% chance of failure before 5.7 Years. Which may seem short. OTOH they didn't test for 57+ years yet. What typically happens is they put a rack on long-term test and don't get enough failures to give a valid number. The longer the test runs and the more failures, the better the number. I know in hard-drive tests the product goes out of production and warranty before they have solid numbers. The test is just to see if they have some unexpected time-failures.

I would not bet on every linear supply hitting 0.5MHr MTBF. That's an exceptional standard of soldering and cleaning, beyond part-stress.

IIRC, these things are cheap enuff you can buy one and a spare. OTOH in 5 or 10 years you may be able to buy a replacement cheaper.

The MIL- ...217.. is a calculation method based on assigning a reliability factor to each and every part and  process in an item of equipment.  It is a method of comparing different equipments using a common 'ruler'.  It is often used as a guide for anticipated failure rates based on the parts and their environment, assuming that any part failure would cause an equipment performance problem.

I've been using outboard Meanwell supplies and like them. An EE friend of mine dropped a inboard module into a case that had high gain phono preamps without any additional filtering or shielding and said it was dead quiet.

Excellent, thanks chaps. Just what I wanted to hear.
Now, if only I could find a module with 5V and decent current on the -15v rail...

I put PD-2515 in my recent builds. Even without an additional LC low pass filter on each rail, it's quiet.
If I'm not mistaken, Apogee also use a customized SMPS made by MW.