Can speed of 2nd harddrive in 6510b DVD-bay be improved

I've put an extra HD in a caddy delivered by newmodeus (converting SATA disk to the "PATA" connector in the DVD-bay in my 6510b - the harddrive is only able to read at 15 mb/s compared to the 100 MB/s when placed in the harddrive-bay (SATA).

Does anyone know if this is normal? is the the HP-connector or the chip in the caddy that is the bottleneck.... Any workarounds to improve the speed?

Ive tested both an intel x25M 80 gb and a Hitachi 7200 rpm 200GB in the caddy with the same slow performance...

 

PATA is slow but it shouldn't be THAT slow...I guess it's a controller limitation of the caddy controller. After all the data goes through the sata controller than as to go through the PATA controller and through the laptop own controller... By the time it get to the rest of the system it would have been slowed down quite a bit.

 

If that is the case a pata-pata controller with af pata hard drive should be faster (newmodeus also sell such a model) - has anyone tested that combination or another brand of hard drive caddy for the DVD-bay with better result or the sata-pata caddy in another laptopmodel with better results?

 

I get the same problem with my 8710w too (exact same 15MB/s transfer rate), so I suspect it's the HP interface limitation because people seem to use this on their Macbook Pro's (first gen PATA optical bay) without such low transfer rates.

 

Is your speed with a newmodeus sata to pata interface or a pata to pata interface?

 

Still interested if anyone have tested a PATA harddrive with a newmodeus-caddy or another brand of caddy in the DVD-bay of a HP notebook - is this connector really limited to 15 MB/s? Newmodeus support hasn´t responded with any useful information...

 

I bought the same PATA-SATA caddy from newmodeus.

 

I tried the link to tweak the registry - however the laptop resets to DMA multiword 2 mode when restarting and the value of masterdevicetimingmode and masteriddatachecksum returns to the values before my tweaking...

I'm running my SSD as primary harddisk and have moved the old harddisk to the DVD-bay, but I still would like more than 15 mb/s of speed for the second HD.

 

I think we are moving in to something that is beyond my computerknowledge

Ive never been using Linux and have no plans of trying it out.

I have no idea of how to deactivate the cardbus bay or cardreader - I can deactivate the optical drive but it didn't seem to change anything - how can I see which devices is sing the ICH8-M primary and secondary channels and if they are acting as slave or master?

 

Nando4 it seems that you know quite a bit about the NewModeUS caddies and the revisions they make - my sata-pata board i 60051 Rev_A and I can see now possibilities for selecting master or slave - do you know if I have an old version (though it´s only 1½ month old....)

 

Ok, looks like there may be a problem with Windows.

Under Linux, hdparm gives:
/dev/sdb:
Timing buffered disk reads: 86 MB in 3.04 seconds = 28.31 MB/sec

Which is at least twice the performance under Vista x64. There is no PIO problem that I can see in Vista x64, neither can I apply any of the PIO/UDMA hard disk fixes mentioned here or on the web because they apply to XP only.

 

hdparm on Windows gives:
/dev/sdb:
Timing buffered disk reads: 40 MB in 3.07 seconds = 13.02 MB/sec

How do I change the drive to slave?

How should I convince newmodeus that their SATA to PATA bridge is broken?

 

Image: master/slave connections to sata-to-pata bridge chip on jumpered Rev C board

Configuring jumperless newmodeus sata-to-pata bridge to work as slave
I can offer the following information to help you along, perhaps while you have discussion with newmodeus on acquiring a master/slave jumpered Rev_C board. Picture above is based on a continuity trace and a visual comparison between the jumpered Rev_C and jumperless Rev_A boards.

• pin 1 jumpered sets slave
• pin 3 jumpered sets master
• If pin 1 and pin 3 are unconnected, it is master (as is on the unjumpered board Rev A board)

What is speculated:
- I can confirm jumpered position is not GND.
- I *believe* then the jumper connects the pin to 3.3V. Speculated because I cannot check a live running board. Usually a digital 0 is GND (or unconnected) and 1 is 3.3V, which is how they set different states of operation.

Warning: if you wish to manually connect pin 1 to a 3.3V source to set it to run as slave, you do so at your own risk!

 

I don't have the appropriate tools to do it. Might just shoot newmodeus an e-mail instead. This is such a tedious expense. Should have just waited out for the next EliteBook refresh and upgrade my aging (and well-used) 8710w.

 

No, I'm not upgrading to the current 8730w. My sources have indicated that there will be an EliteBook refresh corresponding to the new Intel next gen platform.

There's other reasons for wanting the upgrade, namely Quad Core (for development) as well as RAID capabilities, which the current setup won't support.

 

A rev C board should be on its way in the mail to Denmark - I´ll let you know the result when I receive the board...

 

After installing the new board with the disk as master I get a speed of 13mb/s using "HDPARM - t --direct hda" and multiword dma 2 mode - Using the new board with the jumper as slave I´m only able to get a speed of 3.3 MB/s (Pio-mode I expect)... so absolutely no success so far - any good ideas? I´ve tried to reset the mode to UDMA6 using the editing of the registrationdatabase but after reboot I´m back to the described modes....

 

I've now tried manually setting the mode to UDMA5, UDMA4 and UDMA3 manually with the -X-parameter and confirmed it with the -i parameter both with the board as slave and master.
In both cases the first read with the -t parameter takes a long time and results in a very slow read of 5-10 kB/s.
After the test the disk is reset to multiword dma2 by the computer (confirmed by the -i parameter) and the next test with the -t parameter gives a read of 10-15 MB/s with the disk as master or 3 MB/s with the disk as slave....

so it seems that the computer not just reverts the mode to multiword dma2 at booting but at every read atempt.....


Is it of any interest that I have to call the drive for hda and not sdb as you suggest?

 

Troubleshooting transfer mode issue with newmodeus sata-to-pata optical bay caddy

I hope you are Chrixx are being accurate with your description of what you are seeing on your end, otherwise it's just an exercise in chasing one's own tail, or in this case, adding noise to this thread.

I wonder if your bios hard sets the interface speed to mdma2 mode for optical drive speeds, rather than the faster udma5? Running in pio or mdma2 mode is likely reason you are seeing poor performance. Want to run hdd in optical bay caddy in UDMA5 100MB/s interface speed mode, the highest speed possible on a ICH8M PATA interface. So suggest proceed to:

Change drive interface mode to run UDMA5 mode
1/ Alternate the caddy between Master and Slave configurations to see if there is a difference in performance. Easy to do with the Rev_C jumpered board. If have the jumperless Rev_A board, consider this workaround.

2/ check the UDMA mode you are running, perhaps applying this fix. Sounds like Windows may have defaulted you to a lower speed, as it does when encountering an error.

3/ confirm running UDMA/PIO mode as shown with astericks below using hdparm-for-Windows. Note: your drive may be /dev/sda or /dev/sdb. Running 'hdparm -i /dev/sda' and 'hdparm -i /dev/sdb' may help you determine the drive of interest by the brand identifier returned, eg: WD/Seagate/Hitachi..

hdparm -i /dev/sdb
/dev/sdb:

PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5​

4/ Perform read timing test:

hdparm -t --direct /dev/sdb​

5/ manually set the HDD in the optical drive bay to run UDMA5 mode:

hdparm -X udma5 /dev/sdb

/dev/sdb:
setting xfermode to 69 (UltraDMA mode5)
​

6/ Perform read tests again to see if there is an increase in transfer rate performance

hdparm -t --direct /dev/sda
/dev/sdb:
Timing O_DIRECT disk reads: 104 MB in 3.00 seconds = 34.67 MB/sec​

7/ If successful, add hdparm command to a batch file to run in startup folder. Conveniently can set standby idle timeout and immediately spindown the HDD to conserve battery life when running a off a primary SSD or HDD.

hdparm -S 12 /dev/sdb
hdparm -y /dev/sdb
​

If no success changing mode
As it appears to be, then something is switching the drive to mdma2 mode. Need to isolate if this is an os specific or some hardware issue. I recall reading somewhere that IDE would switch to lower modes if the cabling was of poor quality, but clearly that doesn't apply here. A couple of suggestions from which you could pick and choose what to do:

1/ tried to bootup in Safe Mode, delete ALL the drives (HDD/optical drives) and IDE controllers, restart, and let it pnp the devices, including the 2.5" SATA HDD in optical drive bay?[/I]

2/ try a different brand of SATA HDD to do the test. I can confirm a 80GB SATA Fujitsu works fine with the newmodeus sata-to-pata bridge as master OR slave as shown in my review of the newmodeus 9.5mm JAE50 sata-to-pata optical bay product.

3/ try different Intel ATA drivers to see if this switching to mdma2 mode might have been corrected in newer version. Now I have seen on XP systems the possibility of using either an Intel or a Microsoft ATA/IDE driver. Could use drivermax to let it find the drivers for you. There is also the alternative Universal ATA XP driver as a last resort?

4/ bypass your existing OS install altogether. Download and install microXP (100mb) to see if it's something to do with your OS install and work from there.

5/ Ask Chrixx to chime in here, who has Linux, to use 'hdarpm -X udma5 /dev/sdb' then 'hdparm -t --direct /dev/sdb' to isolate if it's os or hardware.

My machine runs in UDMA5 mode with the optical drive caddy, but using 'hdparm -X' I can switch to lower UDMA4/UDMA3/pio modes and 'hdparm -t --direct' does then show the slower read timings.

Conclusion
The sum total steps above combined with those already performed meaning every reasonable attempt has been made to have the optical bay caddy work with a SATA HDD. I hope you and Chrixx have followed instructions carefully above as I otherwise see no reason for it to not work.

Shortest path to a working solution
If this is a genuine issue, dealing with vendors in getting such a solution like this to work might be more frustration than it's worth. Perhaps by design. Easier then to just buy the alternative caddy:

• 9.5mm PATA and SATA-to-PATA caddies are listed here.
• $19US delivered 12.7mm PATA caddy
• $19US delivered 12.7mm SATA-to-PATA caddy

The rear multibay connector unscrews on the NC6000-version caddy to reveal the JAE50 connector required as shown in this post. At $19US it is worth the gamble to see if can get it working. If it does work then can add 6510B/8710W to the list at DIY: Adding SSD and/or HDD storage using an optical bay caddy

 

Hi Nando4
Thank you for your new suggestions - and yes, I have been very carefull about checking the information I wrote in my last post using hdparm after every step.

I have now performed the same procedures using af fujitsu 160 GB 5400 rpm harddrive - the results were identical. Low speed and the computer resetting the mode to mdma2 after using the -t parameter. When the second harddrive is not installed the ssd in the primary (SATA) harddisk bay is callede hda by hdparm, but when the second drive is installed this drive is called hda and not sdb as you write in your suggestions for the hdparm parameterization - does that indicate any problems?

The only alternative driver I have testest so far is the newest from intels homepage (5.1.2600.5512 instead of 5.1.2600.2180) but it made no difference. I´ll let you know if I succeed in installing other drivers or OS.

After using -X paramater to set mode to udma5:


Using -i parameter after runing the -t parameter:


with dvddrive in the dvdbay:

 

No, only to delete the ide channel (not in safe mode) and then reboot with the sata drive installed in the dvd bay - on first reboot it didnt find the second hd, but on second boot it found it and connected it in mdma2 ... I will try that...

 

Under Linux, it refuses to switch to any other mode and only stays in UDMA2.

 

UDMA2 not multiword DMA 2 ?

 

Yes UDMA2, but MDMA2 under Windows. In Windows, I can set -X to UDMA without errors, but it yields terrible performance exactly like you described.

 

[.deleted.]

 

Ok, calling up HP with a truly technical question like that is out. HP's support (even enterprise level) is unable to handle questions like this and neither will they have the answer for it from my experience. You will need to speak to an internal engineer working with the PSG notebook design team, but you can never get to them through regular support channels because the engineers do not deal with customers.

 

Easiest way to save oneself a stack is frustration is to seek Shortest path to working solution. $17US-delivered alternative optical bay caddy worldwide. If it works, please post working solution here and perhaps also at
DIY: Adding SSD and/or HDD storage using an optical bay caddy thread.

HP could help you here, by selling you a 6530P or 8730W. Though they'd be helping themselves first.

 

Ok I have done some more testing on my 8710w now...

To cut a long story short, it still doesn't work at anything above MW-DMA2 (15MB/s). This is after many hours of tweaking, trying all the tricks posted in various threads, master/slave hardwiring, registry tweaks, hdparm, even under knoppix, it is always stuck at a max of MW-DMA2. I can set it lower using hdparm, but when set higher it always baulks at the first test, then resets to MDMA2.

The 2 caddies I currently have look very similar to the ones on eBay, the PATA caddy has no bridge chip, and the SATA caddy has a Marvell 88SA8040.

My initial posts on this subject were from my memory of trying to get this to work early last year, so I have some corrections/updates to make (like how I thought the PATA caddy had a bridge chip).

So I have tried the 2 caddies again in an old nw8000 in the proper multibay slot (using the adapter plate off a CD drive), and initially they were also at MDMA2, which is what I previously reported. But this time I also found that an actual genuine HP multibay HDD adaptor was stuck at MDMA2, so I knew something else was at play.

So I used the registry tweaks detailed elsewhere, and I can now confirm that both caddies are capable of UDMA5 (SATA) and UDMA3 (PATA) on the nw8000 multibay - probably limited by the HDDs themselves. The speeds were confirmed with hdparm.

So back to the 8710w, I also tried a bit of hacking with BAR-Edit, but it didn't work - I don't have a lot of spare time, so I could easily have missed something. I also set the main HDD to non SATA native mode in the bios, and that seemed to change the order of the drives and devices under XP, so the main HDD is the Primary IDE channel, and the caddy is Secondary IDE channel. Still no change to the speed issue.

I think I'm ready to admit defeat at this point.

nando4 I think you should change your other summary thread where it says it works on the 8710w. You could add a note that it only works at 15MB/s. I hope you realise this is pretty much a useless speed these days. It means real-world transfers of 8MB/s if you are lucky. So while it is technically working, it definately isn't a viable solution.

I'm currently thinking about using the spare mini-PCIe slot pads (no connector soldered on it) to rig something up. Maybe a small SSD connected via the USB2 pads in there. Or maybe a proper SATA raid card via a mini-PCI to PCIe slot adapter. It could be wired to go into the optical bay (but you would loose the optical bay permanently). Also there wouldn't be much room left for SSDs. So maybe an expresscard SATA wired to the mini-PCIe would be better space/power wise, but I don't know if they are bootable. Anyone else have any thoughts?

 

I would take the description given by Chrixx, stumo and diskfreak with a grain of salt as believe they are inaccurate.

Technically, the PATA controller will not switch to UDMA2/MDMA2 mode unless instructed to do so, which may come from the driver if in encountered bad signalling between itself and the drive, if it had say a bad cable or connection. Something that HP would need to rectify if it was the case and something you would already have noticed with sporadic optical drive response.

In the interim, others who may wish to try the $17US caddy are welcome to do so. Please refer to DIY: Adding SSD and/or HDD storage using an optical bay caddy. Worthwhile noting the point in the linked thread:

A SSD+HDD setup can work out, on balance, to be a better-bang-per-buck performance upgrade than a system upgrade. Storage usually being slowest component of a system.

 

Hi Nando4 - I´m very thankful for your interest in the "project" of getting som speed on the hp DVD-bay. But I don´t really understand what you think is "inaccurate" about the descriptions of the problems.

I browsed the web yesterday and found a lot of discussions on different Linux-forums with people trying to get DVDs run at more than UDMA2/MDMA2 under Linux on HP laptops from this time period. With absolutely no success despite the DVD running at higher speeds on other non-HP notebooks. Several of this discussions ended up concluding that HP seemed to have restricted the possibilities via the BIOS which give close to zero possibilities to tweeking any thing.

The Hitachi 7k200 disc i first tried to run in the DVD-bay i now in a external enclosure with USB and Esata and i´ve bought a AKE BC188 Esata/SATA PCMCIA card which doesn´t protrude from the PCMCIA slot for the 6510b. I get max transfer rate of 70 MB/s and 52 MB/s on average using HDtach - exactly the same average rate as when testing the drive internally in my desktop (though max transfer then passes 100 MB/s....). The external 2.5" casing even has a dock, which I have mounted in a 3.5" diskslot in my desktop so it is very ease to share data between my notebook and my number-crunching desktop. I think I will stick with this solution until I get another notebook with more speed on the DVD-bay....

 

My 2510P's UJ-852S DVD runs at a max of MDMA2 for a reason. It's a low speed device. It cannot be set any faster, nor does it need to be. If there are issues of the DVD running at 15MB/s, that would show up with sporadic optical drive performance. The Windows driver down steps speed if it encounters errors.. I have gone through all of this, yet it seems yourself, stumo and chrixx would like to continue with innaccuracies. I cannot understand what one hopes to gain by doing that.

For the other readership, a G.Skill Falcon primary SSD giving night-and-day differences in os loading, app loading and system response, supplied HDD goes into optical bay caddy and can be hotswapped with the optical drive on demand. The advantages are almost like a too-good-to-be-true list, but it is true.

If you don't want nonsensical inaccuracies, but do want bang-per-buck ($17US delivered caddy) performance, refer to DIY: Adding SSD and/or HDD storage using an optical bay caddy

 

nando4, I believe you are being ridiculous here.

While I appreciate your very detailed posts on adding a hard disk to the optical bay drive, you must remember that it only worked in your 2510p (if it worked on 8000/6000 series laptop, then please disclose your methods). I have tried both the caddies on eBay and they result in the same performance limitation that I am seeing. I've tried tweaking with the ata and ide modules on Linux and I'm still seeing the same problem.

Now, the not so good news is that on Linux, the hard disk in the optical bay can simply disconnect itself randomly, thus freezing the system (if you installed your system files or home directory on that drive). This indicates a problem with the interface itself (could be BIOS or a hardware problem, I don't know). I cannot apply the fixes that you mentioned in Windows because the registry values do not exist in my registry and on top of that, I'm not running XP.

Now, there is nothing "inaccurate" about any of diskfreak, stumo or my comments (do you even know what that word implies?). I am truly insulted. We have been trying to get this working with the time we have (I don't sit around all day trying to get this to work because I actually have to work, as do others). We're not trying to discredit any of your posts, so why slander us? I've only been stating the different things I've tried and the results I've gained so far. If you're actually in Sydney (assuming you're the same guy on WP) and would like to take a look at my laptop and caddy, then you're invited to by all means. I'd just be as happy to get this working because this is definitely a cost-effective solution, as you have said it yourself.

My further tests on a Toshiba Satellite A105 yields the following results under Windows XP:

hdparm -tT /dev/sdb
/dev/sdb:
Timing cached reads: 60 MB in 2.00 seconds = 30.00 MB/sec
Timing buffered disk reads: 92 MB in 3.06 seconds = 30.04 MB/sec

hdparm -i /dev/sdb
/dev/sdb:
Model=Hitachi HTS722020K9A300, FwRev=DC4OC76A, SerialNo=080416DP0420DTGEZSXP
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
BuffType=DualPortCache, BuffSize=15203kB, MaxMultSect=16, MultSect=off
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5 udma6
AdvancedPM=yes: mode=0x80 (128) WriteCache=enabled
Drive conforms to: unknown: ATA/ATAPI-2 ATA/ATAPI-3 ATA/ATAPI-4 ATA/ATAPI-5 ATA/ATAPI-6 ATA/ATAPI-7

We can clearly see that it is performing much better on a non-HP device. Next up, I'll search for a dv6000 to test this.

 

I too am offended by these accusations. Just because it is plug 'n play UDMA on the 8510p doesn't mean it is universal. For a start, doesn't that laptop have both devices on a single IDE channel? That is a significant difference right there. The 8710w has the optical drive on its own IDE channel.

When I was having a quick go with the BAR-Edit software and the ICH8 datasheet, I could see quite clearly that the chipset registers for the IDE channel in question were not set up for UDMA. These registers get set up by the BIOS. So I'm comfortable saying it has been configured to be locked to MW-DMA by the BIOS in our case. I'm not saying it can't be unlocked, just that I can't do it with the resources I have available.

I tried my best (given the limited time I had) to identify and reconfigure the registers in question. There are a lot of registers to go through and you really need far more experience and knowledge of configuring the ICH8 than I to have a chance of getting this working.

So I can honestly say I will give USD$200 to the first person who provides the method I can use to get an optical bay HDD caddy working at UDMA3 or greater speeds on my 8710w. This is using the kind of caddies available in nando4's numerous links and excludes physical hardware mods to the laptop itself.

This is no joke, I will gladly pay that money and be forever greatful to that person, even if it shows that I'm some sort of idiot by not being able to get this simple thing woking myself (which has been eluded to by a certain member), it is worth far far more to me to have this working.

If enough of us 8710w (and other) users get together there could be a serious wad of cash waiting for that person - and they would deserve every penny.

 

This is the most strange thread I have ever participated in.....

Dear kind Nando - I can´t see that I have done anything but thanking you for sharing your knowledge and asked you kindly to tell me which information you felt was inaccurate.

Personally I think the difference you describe between the 6510b and the 2510 - that the 2510 also use the IDE channel for the primary HDD - could be the whole explanation because it has forced HP to allow the IDE to run in UDMA when they wrote the BIOS - is their any method to "read" a bios and compare the BIOS between the 2510 and the 6510b or other of the problematic HP machines?

As a professional scientist peer reviewing lots of publications in natural science I know one very important thing when interpreting ones results - don´t ever generalize your experience outside the actual experimental set up.... I feel you are doing this when generalizing from the 2510 to the 6510, 8710 etc - let´s go back to working scientifically and finding the reason why three people with "comparable" machines (HPs with ICH8M with MDMA2 DVD bay and primary HDD on SATA-bridge) can´t get more speed while the ICH8M with primary HDD on IDE (2510) can.....


Even if you don´t write in this thread again, then please remember the above scientific principle - live well, and once again thank you for your interest

 

I too am greatful for nando4s input and very useful information and well documented recipe for a working solution.

Unfortunately his current recipe has been confirmed by 100% of all contributers so far that it doesn't work properly on the 8710w and 6510b. So hopefully someone can come along and help us further, as we are now well out of our depth - which in my case is a reasonable amount, having been into computers for over 15 years.

 

I can also confirm that the HP nx6320 is locked to mdma2.
Exact same symptoms as 8710w...
hdparm allows it to be set to udma, confirmed by -i, but on first speed test run -t it takes a long time, and either spits out a very slow rate eg. 100kB/s, or it gives 15MB/s. In both cases it has reverted back to mdma2 with the next -i.

This is also the case when adjusting the XP registry as detailed elsewhere.

This is also the case when hardwiring the caddy to master as detailed elsewhere.

This laptop has the optical bay on its own IDE channel just like the 8710w, but in this case it has a 945 chipset with ICH7, not 965/ICH8.

I only wish it was as easy to get this working as has been suggested, and so there could be an easy USD$200 up for grabs.

 

Please see Slow optical bay caddy likely due to HP setting PATA optical interface to 33Mhz, rather than 66Mhz mode. Included are two software workaround for that problem as well as the interface being set to UDMA2 (33MB/s) mode. If found to be the solution, is the USD$200 reward offer still valid?

 

I have just sold my last 8710w, so no, there will be no $200 reward. I already tried using baredit to setup the ICH registers ages ago. It didn't fix the issue then, so I doub't it will fix it now.

Even then, you are talking about changing UDMA2 to UDMA5. We can't even get UDMA2!!! UDMA2 would at least be usable in real life. We are stuck with MW-DMA2, which has a theoretical max of 15Mb/s. And more like 5MB/s real world.

 

33MB/s is still extremely poor, regardless. I'd rather cut my losses and get a decent notebook with real RAID support. I couldn't be happier with my current setup having sold off the 8710w. Fast, reliable writes are crucial for my DB dev work.

 

Sure.. I can certainly see the improvements in the newer tech, particularly the use of sata optical drive interface.

For others interested, the fix to improve the performance on the 2510P slave interface involves changing the IDE timing from a 33Mhz to a 66Mhz base (I believe it doubles it internally to 66/133Mhz). In any case, my 30MB/s writes on the slave interface went to 50-60MB/s. Linked are the very easy instructions to do this.

I've found the problem is introduced by the Microsoft IDE driver as the 2510P bios does set the slave channel correctly to the 66/133Mhz operation (0/31/1 54h=0x3033), which the driver then changes to 0x1011. I'm trying to find an older intelide.sys driver as was delivered with ICH7M systems to see if it corrects the problem. Surely there must be a registry setting somewhere to undo what the Microsoft driver is doing to slow down the interface.

 

6910p and 6710b/6510b schematics show optical drive pin 45 is used to set a MULTIBAY device. Try isolating that pin on a optical bay caddy and see if that improves HDD performance.

 

8710w owner here, and i can confirm everything chrixx and stumo is experiencing...

please share if you have gotten a caddy to work on 8710 or 8510 at any UDMA5 settings...

 

I have a 2510P and was able to flash a 6710p bios on it as described here to enable the SATA port. Looks like you guys want to do the opposite, use the 2510P bios to enable booting off the HDD in the DVD bay. The 2510P bios does allow that AND it's at full UDMA5 speed. Just note to NOT allow a 2nd pass to update the bios if testing the 2510P bios.

Problem then is the 2510P bios has disabled the sata port, so no booting off your sata HDD. You'd need to manually enable it as described in the same link above.

We can see that the best solution lies in disassembling the 2510P and 6510b/6710b/6910b bios and seeing the differences in the data they write to the PATA port/device. Still facing an uphill to merge the changes as a mod since these bios are further compressed making opcode changes not so easy.

Has any owner engaged HP in delivering a BIOS to fix this problem? They really should be involved.

 

Anybody tried flashing 2510p BIOS to 8710 machine as nando said?
Speed problems were fixed?

 

I have a HP Compaq 8710p. This solution might also work for 8510p, 8510w, 8710w or 6510b or any laptop with SATA-to-PATA caddy (not SATA-SATA caddy).

I've ordered this caddy - 2nd HDD / SSD caddy (SATA-to-PATA) for HP (replace optical) [Model: OBHD-SATA12-BU] - $42.00


(the image is clickable just to show interface of the caddy).

The problem is - bios thinks that caddy is a CD-ROM and resets HDD mode to MDMA2 (so HDD speed is around ~17mbytes/sec)

I managed to get UDMA6 on the drive in the caddy with this method:

Guide is HERE

 

so can you do a benchmark on your disk speed at the optical bay nbruser?
For example ATTO?

Thanks fo the info

 

my HDD is slow (it's a Fujitsu HDD supplied with Notebook).

I've tested with UDMA-5 speeds are better around 36-46mb/sec.

Without it only 14-15mb/sec.

Need to upgrade it to 1TB WD drive and test it properly out. But I'm sure this hack gives performance.

 

Thank you nbruser

Only thing suspicious is the write speed at atto > 2048
So after some days of using this workaround you consider
it stable with noo problems?
I consider buying a hard disk optical bay to try it if you give
me a green light...

 

Yes, it's quite stable, bios no longer resets to MDMA2. Also changed the value written into Bar-Edit from 1011 to 1033, works OK (speed almost the same maybe a little bit more). To do that you need to go to UDMA-Fix\Bar-Save, open up Bar_Edit.ini with notepad and change:

Code:

[PCI]
Nb=1
Addr1=8000F954
Val1=00001011

to

Code:

[PCI]
Nb=1
Addr1=8000F954
Val1=00001033

The crappy speed on benchmark is due slow HDD, HDD works on UDMA5. But 45mb/sec reads much better than 14-15 mb/sec at UDMA2 anyway.

 

Installed at 8710w newmodeus adapter
Installed UDMAfix from nbruser with the latest fixes:
[PCI]
Nb=1
Addr1=8000F954
Val1=00001033

Also changed the /dev/sdb --> /dev/sda

The speed was fixed at UDMA5

The speed of my hard disk went from 15MB/sec --> 75MB/sec

THANK YOU nbruser

Any idea how to put the speed to UDMA6?