Eaton Powerware 3105 UPS, some observations

[Graham.]

I install a lot of these small Powerware 3105, but I noticed that the ones being supplied for the last 6 months or so are a fraction of the weight of the ones we used to get I've also noticed we get the occasional one with something rattling about inside.

I decided to investigate.

The full model number of both samples is identical PW3105 5001 There is a P/W number that differs slightly

103004250-5591 for the heavy one and 103004250-5592 for the new light one Both have 500VA outputs

Here is the inside of the old one

here is the new one I kind of expected that the laminated iron cored transformer will have been replaced with some kind of torroidial thing, but I was totally shocked that the same or similar performance could be obtained with such a tiny torroid.

Has some major advance in ferrite or other technology been made in the last few years that has totally passed me by.

Why couldn't this lighter and no doubt cheaper design have been used all along?

Oh and note the piece of ferrite ring that has come adrift, it's quite a common occurrence. Its not broken, rather the adhesive between the two halves has come unstuck.

[Mike Tomlinson]

In article , Graham. writes

I think if you put a 'scope on the outputs while running on battery you would find that UPS1 puts out an approximation of a sine wave (it's a switching power conversion circuit), while UPS2 puts out a proper sine wave due to use of a transformer.

Both units look like cheap Chinese s**te built down to a price. No way would I use them for my own or customer's kit.

[Dave Liquorice]

There has also been a major (aka complete) redesign of the main board. The two devices may have similar ratings but are producing that by different means. Looking at it a bit more closely I think the small torroid and C's on the little board are just a mains filter rather than anything to do with the power conversion.

[Andrew Gabriel]

being supplied

been replaced with some kind of torroidial thing, but I

with such a tiny torroid.

I suspect the design has changed to be a Class D amplifier generating a

50Hz sine wave. If so, the large transformer is replaced by the transformer on the rear PCB, and the small todoid you're looking at is just filtering out the switching frequency from either the supply or the output.

[Graham.]

being supplied

been replaced with some kind of torroidial thing, but

with such a tiny torroid.

Would you expect the new one to be more efficient in terms of how long it will run on a given load on its 5A/h (20hr) battery?

[Andrew Gabriel]

being supplied

been replaced with some kind of torroidial thing, but

with such a tiny torroid.

Can't really guess. I wouldn't expect it to be worse, and they could have made it more efficient. If you have one of each, test them (probably have to use same battery to ensure that's not a factor). The other thing would be if there's a noticable difference between the two in heat given off when running on battery, but it probably doesn't run for long enough to get a good feel (even if that large transformer generates 10 times more losses, it probably won't heat up enough during the run time to notice the extra heat).

If I wanted to know this for sure, I would use a fixed load, and measure the battery current in each case when driving it.

The output waveform may be different too. In principle, a Class D output can generate a perfect sine wave very efficiently, which can't be done with older designs. That's not to say they have done that in this case (and it's only a guess that they've changed to a Class D output stage - I can't see enough detail in the picture and I don't have hands-on with any recent UPS's to know if the industry is moving that way). If it is generating a better sine wave, some loads may be more efficient with it, even if the UPS itself isn't.