Thermaltake Toughpower Grand RGB 750W PSU Review @ [H]

[FrgMstr]

Thermaltake Toughpower Grand RGB 750W PSU Review - Thermaltake over the years has built some extremely good computer power supply units and along with that, it has also charged a premium for those which sometimes put it at a severe disadvantage in the value department. That changes today, in a very good way. It has flashy spinny LEDs, which you can turn off easily. And a 10 year warranty.

 

[Twisted Kidney]

OK, I hate frag-harder lights in general. But that's a pretty good PSU for the dollar.

The [H] does excellent PSU tests.

 

[pfc_m_drake]

Yep - another excellent review.
And that DC output quality (which is what ultimately counts in the end) is stellar!

One generic question - it seems like it's been a while since we've seen a unit end up with the red 'fail'. Just wondering if you're only publishing 'passing' reviews these days OR (hopefully!) it's an indication of a general increase in the overall production quality of power supplies. I remember, back in the day, there were a fair number of failues

Thanks again for the great review.

 

[[Spectre]]

One generic question - it seems like it's been a while since we've seen a unit end up with the red 'fail'. Just wondering if you're only publishing 'passing' reviews these days OR (hopefully!) it's an indication of a general increase in the overall production quality of power supplies. I remember, back in the day, there were a fair number of failues

While there are many reasons, perhaps the biggest is logistical. Units that fail take longer to do so they hold up production and that makes things tough when trying to have enough time to cover cool passing stuff.

 

[Bees]

Since working towards my engineering tech degree, I've come more to appreciate the finer details in these reviews. This is no exception, and want to say thanks to Paul_Johnson for testing out and writing this up. Going to delve into the testing methodology later to grasp a better understanding of techniques applied.

A question I have in regards to the ripple and noise testing: As there were only 20mV and 10mV peak-to-peak changes, was the option to perform an additional round of testing available to verify this was an accurate test? To clarify, this is not a knock on the methodology, but rather questioning the (very, very good) quality of the load.

 

[[Spectre]]

A question I have in regards to the ripple and noise testing: As there were only 20mV and 10mV peak-to-peak changes, was the option to perform an additional round of testing available to verify this was an accurate test? To clarify, this is not a knock on the methodology, but rather questioning the (very, very good) quality of the load.

Well, there is always the option do more and more rounds at the expense of time and money. At some point, you have to figure out where the balance is between number of tests and practicality (for which there are various models that you will lean about in school). In the case of reviews like this, there isn't really the time or resources to do full fledged R&D testing program like that. We would need a minimum of 3, but more appropriately 10 units, for each test (and as we do them now each review takes about 18 hours of testing time plus the time to write and make figures). As for accuracy, another test with the current set up would not necessarily net anything about accuracy. Rather, it would speak to precision of the measurements (how reproducible they are). More in depth accuracy determination would be more of looking at the results with another instrument a second time. We did this years ago when setting up the protocol and found that the setup used and a second 20MHz oscilloscope produced the same amplitudes. One of the things you will notice, though, is I liberally use the word approximately or the sign ~ when denoting those values. This is for two reasons. The first is that I am showing you a representative example of the reading that occurs during the 30 minutes (or 8 hours) each test is run rather than a 30 minute time lapse (I do my best to use the worst case in those images). The second is that as the values get towards the edges of any instruments functional range the accuracy is going to wander a bit more than if you are within their operating range.

So, one of the things I usually recommend to people is, due to all of the constraints, to consult multiple reviews on a product like this to observe the performance trends in different testing environments and on different samples.

 

[Zarathustra[H]]

OK, I hate frag-harder lights in general. But that's a pretty good PSU for the dollar.

The [H] does excellent PSU tests.

This. The lighting I'm not much of a fan of, but if it's an otherwise decent PSU, the lights can always be disabled one way or another.

I almost guffawed out loud when I saw the "VR Ready" logo though.

I mean, I was considering getting into VR, but it's too bad. My 1200W Silverstone PSU doesn't have a "VR Ready" sticker