[pu82]

I finally got a chance to squeeze in testing of new Hitachi 400GB PATA 7200RPM drives in RTV4000, 4500, 5000, and 5500 units. So far, here are the results: the 400GB drive has successfully functioned in my test 4040 and 4504 and provides 407 hours in standard quality. Divide by 2 for medium and 3 for high. I will follow up with the results of my thermal and stress testing. My other 400GB drive has failed to boot up in either the 5040 and 5504 test machines. It seems to be a similar incompatibility that the Samsung Spinpoint SP drives have with all of the newer ReplayTV models. After I complete the 4000 and 4500 testing, I will reformat both drives and swap machines to see if the failure is to the drive itself or to the drive model. After that, I expect to establish a new ReplayTV record of 825 hours of capacity. Actually, I will do that in a couple of days. I have a pair of 400GB Seagate PATA drives en route for more fun testing.


The moral of the story: don't buy these drives until I complete the testing......

 

[Crunchy Doodle]

While our culture has shown that you actually can be too skinny and too rich, it will never be shown that you have too much space on your ReplayTV. Keep up the good work.


Bye.

 

[Amazingly Smooth]

Actually, I think too much space on an RTV is possible, but too many RTV's isn't.

 

[BaysideBas]

Yeah, really, haven't you heard the one about eggs and one basket?

 

[rm -rf *.*]

In an effort to avoid the "The drive is loud." and "No it's not, you cant even hear it over the fan..." curd, any possibility of adding taking a few SPL readings to your test procedure?


Just a thought... if you can great, if not oh well... Thanks either way!

 

[pu82]

Sorry, I don't have an SPL meter to do your sound measurements in my nonexistent anaechoic chamber. I can give you a qualitative evaluation. These 2 Hitachi drives are quieter than the Maxtor Maxline II 300GB Quickview drives. You can't hear either model unless you are really close to the drive anyway.


I can measure temperature with my laser thermometer. The temperature during the stress tests at the top and sides of the Hitachis are within a few degrees F higher than the Maxtor Maxline II 5400s. A little bit warmer but that is expected for a 7200RPM drive.


Back to the testing progress. I reformatted the drive that would not function in the 5040 and the 5504. It is now happily functioning in a 4040. It does seem to be a model incompatibility. Should I dare to say "The Hitachi 400GB drive is not compatible with the RTV5500 and RTV5000 series."?


I'll try the 400GB that is now in the 4504 to the 5040 and 5504 to get an additional data point.

 

[l8er]

Has anyone ever come up with an explanation on why certain brands of drives will not work in 50XX/55XX ReplayTVs?

 

[pu82]

I just got my first pair of the new Maxtor 7200RPM, 16MB cache, 250GB hard drives. I installed the first one on my 4504 and it works like a champ. No specs posted on Maxtor's web site on this drive, 6B250R0, but it has a 3 year warranty and fluid dynamic bearings. The manufacture date on it is 11OCT04 so I may have one of the first to hit our shores.


I'm not a big proponent of the advantages of the 8MB cache vs the standard 2MB, but 16MB may actually be noticable. It is large enough to cache the HiRes mode for smoother skipbacks, at least for several in a row.


I will take temperature measurements after the drive has been stressed for a few hours and report back. The other 6B250R0 is formatted for my test 5504 and I will let you know how that goes.


For now, the Hitachi 400GB drive has not successfully functioned in a 5000 or 5500. It is strictly a 4000/4500 compatible drive.


I will have screen shots of the killer 800 hour RTV40800 real soon, after I test the 6B250R0. A pair of 6B300R0s are on their way to me as well.

 

[icecow]

Quote:

Originally posted by pu82 I'm not a big proponent of the advantages of the 8MB cache vs the standard 2MB, but 16MB may actually be noticable. It is large enough to cache the HiRes mode for smoother skipbacks, at least for several in a row.

The cache does not work like that. The cache is not a buffer.



Quote:

Originally posted by pu82 For now, the Hitachi 400GB drive has not successfully functioned in a 5000 or 5500. It is strictly a 4000/4500 compatible drive.

Is that a fact? Are you sure? I did not know that. I've heard lower sized Hitachi drives work fine. Do you know which size/model Hitatchis work with 5000s?

 

[pu82]

My bad if I am mistaken. My understanding is that the cache keeps the last xx MB of data read from the drive in the cache so that future calls for the same data would not have to take place as a read from the disk which is slower. Even as a read ahead cache, it would improve performance. 16MB may be large enough to make a difference in the DVR world.


As for the Hitachis, I last tested the 180GXP and it worked on all of the 4000/4500/5000/5500 test machines. The only other compatibility issues I have come across is the Samsung 7200 series and some Seagate BIOS issues where some BIOS versions worked and some did not.

 

[sfhub]

Quote:

Originally posted by icecow The cache does not work like that. The cache is not a buffer.

Unless you know which caching algorithms the particular drive is using

I'm not sure how you can say this so definitively.

 

[icecow]

I haven't seen any stickers on harddrive boxes saying 'With Ubercache! Optimized for Video file playback and Media needs!' so i'm left to believe they doing the cache the same-til I hear different-old way .. predictive of repeating info.


you know different or just bustin my gyros?

 

[sfhub]

There are different algorithms used for caching.


Sometimes temporal based (ie data most recently used is likely to be used

again), sometimes locality based (ie data close together is likely to be

used at the same time), sometimes strictly read-ahead (ie your disk is

heavily reading sequentially)


Sometimes the algorithms will adjust based on read patterns. Every

company will try to tweak their algorithms to look best in the benchmarks

and also to try and improve real world performance.


w/o knowing how these algorithms were tweaked I wouldn't feel

comfortable making that statement. For example it wouldn't be hard

to imagine they split the 16MB cache into 2 parts, 8MB for temporal

recently used data, and 8MB for predictive read ahead, thereby, for

single thread video playback, resulting in roughly 8MB of data on

either side of the current video position sitting in cache.

 

[adone36]

One of the main features of the Quickview series is that the algorithm used is specifically designed for streaming. The reason it is only 2 meg is that any more is simply useless. A PC on the otherhand is called upon to seek files all over the drive. A larger cache here and an algorithm which closely predicts future seeks and maintains previous seeks increases performance.

 

[icecow]

Quote:

Originally posted by sfhub There are different algorithms used for caching. Sometimes temporal based (ie data most recently used is likely to be used again), sometimes locality based (ie data close together is likely to be used at the same time), sometimes strictly read-ahead (ie your disk is heavily reading sequentially)

Those are old textbook os design terms I'm familiar with (temporal/local).


I'm just losing my memory I'm smarter then ...i can remember.

 

[rm -rf *.*]

Quote:

Originally posted by pu82 Sorry, I don't have an SPL meter to do your sound measurements in my nonexistent anaechoic chamber. I can give you a qualitative evaluation. These 2 Hitachi drives are quieter than the Maxtor Maxline II 300GB Quickview drives. You can't hear either model unless you are really close to the drive anyway.

No need for you to apologize - I'm the twit that posted the question before the cranial reflexes reminded me of the fact that unlike me, not everyone has easy access to an anechoic test chamber... *oops* Sorry.


Your qualitative assessment is probably more helpful/useful anyway since most everyone reading this can relate to it.


In an attempt to bring your assessment of the noise level a little bit closer to home, since I don't know "how loud" the Maxtor Maxline II 300Gb drives are when installed in an RTV, could you give a comparison to something nearly all of us know/have experienced, say the stock 40Gb drives? (or whatever ReplayTV-Stock drive your units might have come with).


TIA.

 

[Matiero]

Did you get a chance to try this test?:

Quote:

Originally posted by pu82 After I complete the 4000 and 4500 testing, I will reformat both drives and swap machines to see if the failure is to the drive itself or to the drive model.

I'm interested in trying out a single Hitachi 400GB in a 55XX ReplayTV unit myself, and am interested to know if you had a failed drive, or if there's truly an incompatibility issue.

 

[Matiero]

Oops. Forget my post. I just did a re-read and got my answer. Sorry.