LG Gram 17 (2021): Some Impressions

[RS4]

21 hours ago, John Ratsey said:

It's now been opened up 3 times. Yesterday I temporarily put back the original SSD as I needed to release the licence on a software package which didn't like the SSD swap (two other packages didn't complain). While I had the base off I also repasted the CPU to see if I could reduce the occurrence of thermal throttling. The original paste was a bit dry. The CPU still hits the thermal limits under sustained load but now it's running a bit cooler under normal conditions. I also put thermal pads on the SSDs and they are keeping cooler.

 

As for the glue, it's still quite sticky. When the feet are pulled off some of the glue comes with them and some stays on the base cover. I try to put back the feet so that the circle of glue is reformed.  However, the computer's feet have an easy time as my Gram 17 is used as a compact desktop and doesn't get put into / taken out of a bag several times a day. Nonetheless, for anyone worrying about the feet getting loose, the fix will be to buy some 15mm circular double-sided tape pads (example here) and use them to replace the original glue when it gets tired.

 

How to make sure that the circle of glue is reformed? Do we have to line up the glue residue on both feet and the base so that they join perfectly?

 

Regarding those 3M tapes, will 10mm size also work? Looking at these two ebay listings:

 

https://www.ebay.com/itm/233290973133?hash=item365139bfcd:g:kqkAAOSwuFxcuBw9&shqty=1

 

https://www.ebay.com/itm/194060824786?hash=item2d2eed30d2:g:0xsAAOSw9idaK2Vd

[John Ratsey]

1 hour ago, RS4 said:

 

How to make sure that the circle of glue is reformed? Do we have to line up the glue residue on both feet and the base so that they join perfectly?

 

Regarding those 3M tapes, will 10mm size also work? Looking at these two ebay listings:

 

https://www.ebay.com/itm/233290973133?hash=item365139bfcd:g:kqkAAOSwuFxcuBw9&shqty=1

 

https://www.ebay.com/itm/194060824786?hash=item2d2eed30d2:g:0xsAAOSw9idaK2Vd

Yes, try to put the feet back so that the glue residue forms a rough circle. I pulled off one of the back feet and took a photo to illustrate how it looks.

 

Your first ebay item will probably be fine. I think LG used something similar but as a ring. The back feet are about 17mm dia and the front feet 15mm but the ring of glue is smaller than the size of of the foot. I can't find the tape circles on ebay UK, only the circular pads (which I'm sure would work fine as they are buried in a hole).

[RS4]

On 7/26/2022 at 1:30 AM, John Ratsey said:

Yes, try to put the feet back so that the glue residue forms a rough circle. I pulled off one of the back feet and took a photo to illustrate how it looks.

 

Your first ebay item will probably be fine. I think LG used something similar but as a ring. The back feet are about 17mm dia and the front feet 15mm but the ring of glue is smaller than the size of of the foot. I can't find the tape circles on ebay UK, only the circular pads (which I'm sure would work fine as they are buried in a hole).

 

Thanks for illustrating with a picture, will be helpful for myself as well as others who need it. I will also place the order from that first ebay link, good price for four sheets of stuff.

[John Ratsey]

One annoyance I've been having with my 2021 Gram 17 is that intermittently it slows to a crawl, becomes unresponsive and takes several minutes before it's willing to start running normally. It's reminiscient of the problem I had around 10 years ago with near-full SSDs which basically told the computer to wait while the SSD shuffled files around to create some usable space from the fragments. However, there's loads of empty SSD space and the disc activity is low. I therefore assume that it's a thermal issue. The i7-1165G7 CPU in my Gram 17 is too willing to get close to 100C (perhaps I need to do another repaste) and even when the CPU hasn't got hot there's other thermal throttling. This is what HWiNFO showed soon after doing a sensor reset. I suspect that if there's been too many thermal alerts then the Gram 17 reckons it's time to walk for a while instead of running.

Setting the Gram 17 to Performance Mode (which needs to be done on each start / restart / resume) does improve the temperatures but is tedious as it's not set-and-forget. Therefore three days ago I got a Cooler Master Notepal U2 plus V2 cooler to help improve the cooling. It's just the right size for the Gram 17 but the default fan speed was too noisy for my liking. That problem has been fixed using one of these to reduce the voltage and hence the fan speed although I not very happy with the raised notebook position (and I've squashed down the back leg of the cooler to reduce the keyboard slope).

 

I've therefore revisited the power options. A while back I used a utility (name now forgotten) to add some of the missing items in the advanced power options. Putting the Maximum Processor State down to 99% turns off turbo mode which is more drastic than I want (but sensible when running on battery) but I've found that setting the Maximum Processor Frequency to 4500 MHz does cap the CPU speed to 4500 MHz. This will bring down the maximum CPU core temperature but does not stop the other thermal throttling alerts shown above. I need to do some testing to find out the sensible maximum CPU frequency.

 

If only I could buy the longer heatsink and dual fan unit that LG do make then this would be a much better solution to the thermal problem. The i7 CPU is more than the standard LG thermal solution can handle for more than a few seconds.

[Steerpike]

10 hours ago, John Ratsey said:

One annoyance I've been having with my 2021 Gram 17 is that intermittently it slows to a crawl, becomes unresponsive and takes several minutes before it's willing to start running normally. It's reminiscient of the problem I had around 10 years ago with near-full SSDs which basically told the computer to wait while the SSD shuffled files around to create some usable space from the fragments. However, there's loads of empty SSD space and the disc activity is low. I therefore assume that it's a thermal issue. The i7-1165G7 CPU in my Gram 17 is too willing to get close to 100C (perhaps I need to do another repaste) and even when the CPU hasn't got hot there's other thermal throttling. This is what HWiNFO showed soon after doing a sensor reset. I suspect that if there's been too many thermal alerts then the Gram 17 reckons it's time to walk for a while instead of running.

Setting the Gram 17 to Performance Mode (which needs to be done on each start / restart / resume) does improve the temperatures but is tedious as it's not set-and-forget. Therefore three days ago I got a Cooler Master Notepal U2 plus V2 cooler to help improve the cooling. It's just the right size for the Gram 17 but the default fan speed was too noisy for my liking. That problem has been fixed using one of these to reduce the voltage and hence the fan speed although I not very happy with the raised notebook position (and I've squashed down the back leg of the cooler to reduce the keyboard slope).

 

I've therefore revisited the power options. A while back I used a utility (name now forgotten) to add some of the missing items in the advanced power options. Putting the Maximum Processor State down to 99% turns off turbo mode which is more drastic than I want (but sensible when running on battery) but I've found that setting the Maximum Processor Frequency to 4500 MHz does cap the CPU speed to 4500 MHz. This will bring down the maximum CPU core temperature but does not stop the other thermal throttling alerts shown above. I need to do some testing to find out the sensible maximum CPU frequency.

 

If only I could buy the longer heatsink and dual fan unit that LG do make then this would be a much better solution to the thermal problem. The i7 CPU is more than the standard LG thermal solution can handle for more than a few seconds.

May I ask, what are you doing with your laptop when these peaks are occurring?  Hopefully you are doing something that is stressing it somehow! 

 

I'm having to run all day, every day on battery at the moment so I've been really focusing on reducing power consumption. I don't want to sacrifice screen brightness, so I've been playing around with Processor Power Management / Max processor state, and setting it to 40% !!!!  Surprisingly, I'm not really noticing any difference for casual web page browsing, email handling, casual spreadsheet use. Not sure if it's really helping much with actual battery, but it's the lack of impact on real-world 'use' that is surprising me! 

 

Also, where are you getting this setting from: "setting the Maximum Processor Frequency to 4500 MHz". I did whatever I needed to do to get 'max processor state' back, but I wasn't aware of a frequency setting.  I'm aware of the setting  "Processor performance boost mode" (to enable this setting, go to regedit and navigate to:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Power\PowerSettings\54533251-82be-4824-96c1-47b60b740d00\be337238-0d82-4146-a960-4f3749d470c7 )

[John Ratsey]

11 hours ago, Steerpike said:

May I ask, what are you doing with your laptop when these peaks are occurring?  Hopefully you are doing something that is stressing it somehow! 

 

Also, where are you getting this setting from: "setting the Maximum Processor Frequency to 4500 MHz". I did whatever I needed to do to get 'max processor state' back, but I wasn't aware of a frequency setting.

I'm not intentionally putting the computer under load. Perhaps it's the innumerable tabs in Firefox but the two thermal throttling alerts always re-appear within a few seconds of resetting the HWiNFO sensors. At the moment it's showing core thermal throttling when none of the cores have exceeded 75C. Perhaps there's another sensor which HWiNFO isn't directly reading.

 

I found a useful bit of software called Power Plan Settings Explorer which lists all the options and allows the user to unhide the hidden settings (the list of ticks shows what is hidden).

 

Last evening I was doing some more research by running wPrime and looking at HWiNFO's CPU sensor data. If Performance mode is enabled then PL1 stays at 28W and the fan works as needed. In Optimal mode PL1 quickly dropped below 28W and, at one stage, went as low as 11.5W. That's for the whole CPU package and around 6W is used by parts other than the CPU which dropped to around 800MHz (like winding the CPU performance back by 10 years or more). I also explored the advanced BIOS settings and read, and need to re-read, the archived discussion about advanced tweaking.

 

The thermal limitations mean that turbo is only any use if there's load on only a single core. I need to try wPrime 32M running 1 thread and seeing the power behaviour. Turbo is easily disabled by setting the maximum processor state to 99% which is my default for battery operation.

 

If only someone sold the dual fan and heatsink unit but I've failed to find it on the www. I wouldn't mind the extra weight and slightly increased power consumption.

[Steerpike]

On 8/20/2022 at 6:26 AM, John Ratsey said:

...

This is what HWiNFO showed soon after doing a sensor reset. I suspect that if there's been too many thermal alerts then the Gram 17 reckons it's time to walk for a while instead of running.

 

 

I haven't used HWInfo for years, but I just installed it now for grins.  As I mentioned in my other response, I'm currently experimenting with aggressive CPU throttling to see what makes a difference, and what affects battery usage.  I'm currently running with 'max processor state' at 30% (down from 40% earlier!) and my general usage seems unaffected (~40 tabs in Edge, ~40 tabs in Chrome, 5 excel spreadsheets, 1 word doc, Outlook with ~10 emails being read/written, Quicken, Foxit Reader, ~5 notepad sessions ...) (the only thing that was noticeably slower was a complex PDF of a map took a while to scale/repaint). I 'reset' the triggers, as you did, and then walked away from the laptop. I just returned and did a few minor things, and ... those two items you have triggers on are 'Yes' for me too - Core Thermal Throttling and Package / Ring Thermal Throttling.  So it would appear those two items may just get tripped on the slightest provocation. 

 

Interesting that my Max values are way higher than yours!  FYI, my cpu is i7-1195G7, while yours is i7-1165G7. I finally figured out how to resize the columns to show that, after getting the screenshot. 

 

EDIT TO ADD - I just sat here looking at my HWInfo figures, after resetting the triggers, and noticed that the two triggered items were 'no'.  I then 'locked' the screen (windows key + L) and unlocked, and ... the two values went to 'yes'. Why, I can't explain but you may try that on your side ... If you see that they are both 'no' for a while, then walk away and let the timer lock your screen (or sleep, or whatever) then unlock, it may be the cause for you too. 

Just did it again - reset triggers; observe 'no'; lock; unlock; values are 'yes'. 

 

EDIT TO ADD 8/22 - I just did a few more 'reset triggers', and then lock/unlocks. I was able to do 2 quick lock/unlocks without seeing the triggers, but on the third, I saw the two triggers. Note how low all my temperatures are: 

 

 

Nothing higher than 63C. So whatever those triggers are, they are certainly sensitive! 

 

[John Ratsey]

I turned off Turbo mode by setting maximum processor state to 99% and reset the HWiNFO sensors. A couple of hours of typical usage hours and the maximum processor temperature is 66C and there are no throttling alerts (this is with the computer on the Cooler Master cooler). I'm wondering if those those throttling alerts are linked to the fan trip point temperatures.

If you can keep everything below 70C then perhaps everything is happy. What would happen if the Active Trip Point 0 is set to, say, 60C.

 

Performance and esponsiveness seems fine with Turbo disabled. Those bursts of higher speed don't make much difference in overall speed in normal usage (might be noticeable if there's a single-threaded process running) but create the power and heat and spikes which have repercussions.

[Steerpike]

3 hours ago, John Ratsey said:

I turned off Turbo mode by setting maximum processor state to 99% and reset the HWiNFO sensors. A couple of hours of typical usage hours and the maximum processor temperature is 66C and there are no throttling alerts (this is with the computer on the Cooler Master cooler). I'm wondering if those those throttling alerts are linked to the fan trip point temperatures.

If you can keep everything below 70C then perhaps everything is happy. What would happen if the Active Trip Point 0 is set to, say, 60C.

 

Performance and esponsiveness seems fine with Turbo disabled. Those bursts of higher speed don't make much difference in overall speed in normal usage (might be noticeable if there's a single-threaded process running) but create the power and heat and spikes which have repercussions.

Did you see my 'edits' to my original post immediately before yours? I am able to trip those two triggers 'on demand', just by locking / unlocking the screen - and I'm currently running with max processer state at 35% as an experiment! 

[John Ratsey]

10 hours ago, Steerpike said:

Did you see my 'edits' to my original post immediately before yours? I am able to trip those two triggers 'on demand', just by locking / unlocking the screen - and I'm currently running with max processer state at 35% as an experiment! 

Yes, I saw that and can replicate the behaviour. Perhaps it's an HWiNFO bug.

 

In previous times I did a lot of experimenting with the max processor state. This graph is one I prepared 6 years ago:

The percentage labels are the maximum processor state and anything below 50% had no effect. The corresponding speed and power consumption have to be deterermined by looking at the HWiNFO results after a standard process (eg wPrime). It would be interesting to see the corresponding graph for recent CPUs. If trying to minimise overall power consumption one has to balance the reduction in processor power against the the longer time to run the process before the CPU and get back to sleep.

[Steerpike]

5 hours ago, John Ratsey said:

Yes, I saw that and can replicate the behaviour. Perhaps it's an HWiNFO bug.

 

In previous times I did a lot of experimenting with the max processor state. This graph is one I prepared 6 years ago:

The percentage labels are the maximum processor state and anything below 50% had no effect. The corresponding speed and power consumption have to be deterermined by looking at the HWiNFO results after a standard process (eg wPrime). It would be interesting to see the corresponding graph for recent CPUs. If trying to minimise overall power consumption one has to balance the reduction in processor power against the the longer time to run the process before the CPU and get back to sleep.

Interesting! The huge drop from 100% down to 99% is, presumably, due to the fact that going from 100->99 ALSO disables turbo mode? 

 

[John Ratsey]

4 hours ago, Steerpike said:

Interesting! The huge drop from 100% down to 99% is, presumably, due to the fact that going from 100->99 ALSO disables turbo mode? 

 

Yes, 100% processor state enables Turbo mode, 99% processor state disables Turbo mode. That other power option I found  recently which can set the maximum processor speed would allow filling that gap in the curve. However, it's not worthwhile doing that for the Gram 17 where the fully turbo speed seems to last no more than a second.

[Steerpike]

@JohnRatsey: You've had your LG Gram 17 (2021) for some time now, I believe. How is your keyboard holding up?  I've had my model since April 2022, and I've used it almost every day. The spacebar in particular, and the 'alt' key to a lesser degree, is showing significant signs of wear (I use 'alt-tab' quite a bit). I'm not really bothered by it, but when I compare it to my last Samsung laptop (NP930X5J) which I used almost every day from 2014 to 2022, this seems unusual. 

 

The keys seem to have some sort of texture added to the surface, and that texture has already worn off in the area where my thumb hits the spacebar. I had to angle the camera carefully to capture this effect; it's not obvious from all angles, but - at the correct angle, it's quite noticeable.  The 'R' and 'T' keys are also quite noticeably worn when viewed at different angles. 

[John Ratsey]

@Steerpike my 2021 Gram 17 is currently resting pending rehoming but I've got it out and photographed the keyboard. There's a shiny area on the spacebar, the touch pad looks used and you can see where my palms rest on the palmrest. The other keys look new.

After taking the photo I've tried to clean the marks using some isopropyl alcohol on a cloth. The touchpad now looks much better but the other wear marks persist.

 

I've also looked at the keyboard of my Galaxy Book Pro which I bought in January 2022. It hasn't had as much use as the LG and there's a shiny area on the space bar but no other obvious wear.

[Steerpike]

6 hours ago, John Ratsey said:

@Steerpike my 2021 Gram 17 is currently resting pending rehoming but I've got it out and photographed the keyboard. There's a shiny area on the spacebar, the touch pad looks used and you can see where my palms rest on the palmrest. The other keys look new.

 

After taking the photo I've tried to clean the marks using some isopropyl alcohol on a cloth. The touchpad now looks much better but the other wear marks persist.

 

I've also looked at the keyboard of my Galaxy Book Pro which I bought in January 2022. It hasn't had as much use as the LG and there's a shiny area on the space bar but no other obvious wear.

Thanks for taking out the laptop and taking the photo!  Isopropyl alcohol has been my 'go to' cleaner for a long time, but given the 'fragility' of this keyboard I've been restrained in using it in case it acts as some sort of solvent. I used a very small amount of diluted liquid hand soap to clean mine, and it did a great job on the trackpad, like yours, but no change to the space bar/Alt key.  

 

I paid more attention to how I place my hands, and I noticed that I will often rest my left thumb on the alt key, so in addition to actually using it a lot, it gets even more wear from that. My right thumb sits permanently on the space bar right where the wear is seen.  Since the space bar has no lettering on it, it's not an issue but I wonder whether the white lettering is vulnerable. I guess not, since the white lettering is actually semi-transparent to facilitate the back-lighting, so I guess that is 'solid' rather than a surface treatment. 

 

Speaking of keyboard backlight - I noticed that on my Samsung Galaxy Pro, the backlight is 'ambient light sensitive'; it only kicks in when it detects a dark environment. But on the LG Gram, it simply comes on when the keyboard is 'in use', regardless of ambient light. This means, the backlight is on a lot!  So I've taken to turning it off unless I really need it (Fn + F8) to further save battery. Not sure yet if this is making any difference! 

 

Edit to add - I love your neat labeling above the side sockets!  What are you using for that?  I'm big on labeling and have a dedicated cable-labeler and a more general purpose 'Brother' labeler, but I don't see any reflections around your labels so I'm intrigued!  

 

Also, I should note that I bought the 'quick disconnect' USB-C power adapters based on your recommendations way back. I now have 4 of them and may buy more! 

[John Ratsey]

4 hours ago, Steerpike said:

Edit to add - I love your neat labeling above the side sockets!  What are you using for that?  I'm big on labeling and have a dedicated cable-labeler and a more general purpose 'Brother' labeler, but I don't see any reflections around your labels so I'm intrigued!  

 

Also, I should note that I bought the 'quick disconnect' USB-C power adapters based on your recommendations way back. I now have 4 of them and may buy more! 

The lettering is one of these (I found it on UK ebay) but I perpetually wonder why notebook manufacturers don't print some symbols on the keyboard surround during production. The challenge was finding white lettering. I first tried using lettering created by my Dymo LetraTag label maker but black lettering on a white background didn't look very elegant (although that's what is on my little Samsung). I haven't bothered with putting the lettering on my new Gram 17 as I'm much more familiar with the port locations.

 

The lettering on the keys is moulded translucent plastic and won't wear out.

 

One advantage (in addition to convenience and reduced port wear) of the QR USB-C connectors is that pulling a cable won't casue the computer to take a trip to the floor. I'm using the right-angle connectors to route the power cables to the back of the computer although this configuration can't coexist with using the HDMI port.

[Steerpike]

@John Ratsey (and others) - I searched this thread for "NVMe" and didn't find the answer, but I think you posted in this thread somewhere the exact model number of SSD that you installed in your LG Gram. I've been putting this off for almost a year now, and the time has come to buy a 2TB SSD.  So what brand did you buy, and what brand would you buy today (if different)? 

 

I'm not interested in a 'high performance' SSD - any NVMe is going to be way faster than the SATA 3 SSD I have in my old laptop, and that's fast enough for my purposes (photo storage).  But I do want to minimize extra power consumption, as I have to live on battery every day.  The SSD will see very little read/write activity once populated so 'idle power' is what matters. 

 

I can get a Crucial P3 Plus (Gen 4) (CT2000P3PSSD8) for $112, a Kingston NV2 (Gen 4) (SNV2S/2000G) for $105, a Western Digital WD_BLACK SN770 (Gen 4) (WDS200T3X0E) for $119, Western Digital Blue SN570 (Gen 3) (WDS200T3B0C) for $109, or a Samsung 970 EVO Plus (Gen 3) (MZ-V7S2T0B/AM) for $139. I can even get a PNY CS1030 (Gen 3) (M280CS1030-2TB-RB) for $99 (that's just the brands I'm familiar with). 

[Steerpike]

On 7/25/2022 at 10:51 AM, RS4 said:

... Regarding those 3M tapes, will 10mm size also work? Looking at these two ebay listings:

 

https://www.ebay.com/itm/233290973133?hash=item365139bfcd:g:kqkAAOSwuFxcuBw9&shqty=1

 

https://www.ebay.com/itm/194060824786?hash=item2d2eed30d2:g:0xsAAOSw9idaK2Vd

 

On 7/26/2022 at 2:08 PM, RS4 said:

 

Thanks for illustrating with a picture, will be helpful for myself as well as others who need it. I will also place the order from that first ebay link, good price for four sheets of stuff.

What did you end up buying and ... did it work? Was 10mm a good fit?  I notice that @John Ratsey's link was for 'pads', 1mm thick, while your second ebay link above was for 'foam', at 1.5mm thick.  Also, the first link no longer finds anything. 

 

I can't find any such product selling in the US, so my options are to order from China or UK.  This looks to be the same product that John Ratsey recommended, but from Ebay US (still shipping from UK). 

 

3M DOUBLE SIDED STICKY PADS Strong Heavy VHB Adhesive Mounting Tape Square Round | eBay

 

Did you cut out a hole with a knife ahead of installation (before removing the covers)?  I'll probably order the 'pads' above, but I'm curious to know if you think 10mm is the best size. 

[RS4]

9 hours ago, Steerpike said:

 

What did you end up buying and ... did it work? Was 10mm a good fit?  I notice that @John Ratsey's link was for 'pads', 1mm thick, while your second ebay link above was for 'foam', at 1.5mm thick.  Also, the first link no longer finds anything. 

 

I can't find any such product selling in the US, so my options are to order from China or UK.  This looks to be the same product that John Ratsey recommended, but from Ebay US (still shipping from UK). 

 

3M DOUBLE SIDED STICKY PADS Strong Heavy VHB Adhesive Mounting Tape Square Round | eBay

 

Did you cut out a hole with a knife ahead of installation (before removing the covers)?  I'll probably order the 'pads' above, but I'm curious to know if you think 10mm is the best size. 

 

I bought the one sold by seller Tamanala but never got the product, forgot about it. I have only opened the laptop once for adding 2TB SK Hynix P31 SSD, so the rubber feet glue are still in good condition. John would be able to guide better on this matter as he is the only one with experience on this.

[John Ratsey]

On 3/6/2023 at 12:02 AM, Steerpike said:

@John Ratsey (and others) - I searched this thread for "NVMe" and didn't find the answer, but I think you posted in this thread somewhere the exact model number of SSD that you installed in your LG Gram. I've been putting this off for almost a year now, and the time has come to buy a 2TB SSD.  So what brand did you buy, and what brand would you buy today (if different)? 

I've got a WD SN850 in slot 1 and WD SN770 in slot 2. They both seemed to offer best value for performance / money at the time of purchase and I have no regrets. Before each purchase (the SN770 was bought first) I searched for the quality reviews (eg Tom's Hardware and Anandtech) and checked for any adverse comments for the SSDs I was considering. Prices seem to be steadily dropping and 4TB will soon be affordable.

 

I think idle power consumption is no longer an issue but the system power settings may be a factor. Go to the power options via Control Panel and check the Link State Power Management under PCI Express settings is set to Maximum Power Savings.

 

As for the question of new sticky pads, I've not yet felt the need to buy any. If you prise the rubber feet and screw covers off carefully then the adhesive remains and is still slightly sticky and will hold things in place when you put them back. Id that's not enough, you could use a very small amount of a glue (eg UHU) which remains flexible when it sets. The pre-made sticky pads speeds up manufacturing by avoiding the problem of the glue getting in the wrong places and needing to be removed.

[extremecarver]

If 1TB is enough - the Samsung 980 has very low power usage in both idle and use. Sadly no 2TB model. The Samsung Pro series NVME always have higher power usage.

[Steerpike]

On 3/9/2023 at 3:01 PM, John Ratsey said:

I've got a WD SN850 in slot 1 and WD SN770 in slot 2. They both seemed to offer best value for performance / money at the time of purchase and I have no regrets. Before each purchase (the SN770 was bought first) I searched for the quality reviews (eg Tom's Hardware and Anandtech) and checked for any adverse comments for the SSDs I was considering. Prices seem to be steadily dropping and 4TB will soon be affordable.

 

I think idle power consumption is no longer an issue but the system power settings may be a factor. Go to the power options via Control Panel and check the Link State Power Management under PCI Express settings is set to Maximum Power Savings.

 

As for the question of new sticky pads, I've not yet felt the need to buy any. If you prise the rubber feet and screw covers off carefully then the adhesive remains and is still slightly sticky and will hold things in place when you put them back. Id that's not enough, you could use a very small amount of a glue (eg UHU) which remains flexible when it sets. The pre-made sticky pads speeds up manufacturing by avoiding the problem of the glue getting in the wrong places and needing to be removed.

I finally installed the new WD SN770 2 TB NVMe SSD in slot 2 last night!  It wasn't that hard to open the case, but it wasn't a pleasant experience either!  I found that pushing inwards on the big feet, causing the circle to deform, was the best way to detach the glued-on feet. Once it starts to deform slightly, you can hear the glue detaching, then it's easy to lift off. The little screw cover caps were weird; I attacked those with a very narrow flat-head screwdriver, and they just 'popped' right off.  Removing the screws was effortless. Removing the base was the usual nerve-wracking experience, wondering if you are going to break something!  Following your notes above, I started at the front but there was no easy way to get that first clip to 'pop', but I somehow managed it using the very narrow flat-head screwdriver (should have used a 'spudger' but couldn't find it).  There are a total of 18 miniature 'prongs' on the panel that clip to recesses in the frame (6 front; 4 left, 4 right, 4 rear), and you just never know if you are going to break one or not.

 

I took a lot of pictures, and may create a dedicated thread just about the case stuff.  Good news is, there's still 9 screws holding the panel to the frame, so even if several clips failed, it wouldn't really matter. Note - saw a great online video (after the fact) where they used a suction cup on the base to remove it (this video, at about 6:10 mark. https://www.youtube.com/watch?v=9fTL4lzH4Bs - not our exact model but close enough! ). 

 

 

As for the SSD itself - I downloaded and installed "Western Digital Dashboard"; seemed pretty basic, but did confirm I have the latest firmware. Just about the only setting I saw was for 'gaming mode', which I left off.  I also confirmed that I had "Link State Power Management under PCI Express settings is set to Maximum Power Savings."

 

Thanks for the guidance! 

 

[Steerpike]

@John Ratsey I seem to recall, you are hyper-sensitive to fan noise, as am I. Or more accurately, you are hyper-sensitive to slight changes in noise, such that a steady faint 'hum' of a fan is much better than a stop/start fan? 

 

OK, So - after I installed the second SSD in my LG Gram, I noticed a strange, high-pitched, random sound that I find hard to describe. Basically, it's high-pitched 'electrical' noise, the type of noise a circuit gives off when there's some high-frequency processing going on. Also similar to what is heard from battery chargers, where they are doing lots of voltage switching (though this tends to be more steady/predictable). Actually, best description may be - the sound you USED to hear from a hard drive, when the head was 'seeking' - very random, high pitched. 

 

I thought at first it was the fan - perhaps I introduced some debris into the case, but that wasn't the case.  Upon closer inspection, I can hear it most clearly if I put my ear right above the ASDF keys (which sits right above CPU / SSD area) OR - if I turn the unit upside down and place my ear over the 'vent' area (note that only a small portion of the long row of vent-holes is actually vented - most of it is closed off). The 'open' (unobstructed) vent holes sit directly above the CPU. So - my current theory is, this is new noise from the newly installed SSD; or, could it be simply noise from the CPU or RAM (which also sit in that vicinity) that I didn't notice before?  I'm loath to removing the SSD just to test ... but that will be my next step if I can't figure out where it's coming from!   

 

Anyway - just wondered if you have noticed this noise on your LG Gram. Note that this is VERY faint ... but I happen to sit in a very quiet room, and don't have any TV/music playing during the day, and can literally hear a pin drop! Plus, I have the most sensitive hearing ... more of a burden these days 🙂 . 

 

PS I also found a nice way to remove the base of the laptop - you can insert your thumbnail between the removable lid and the surrounding frame at the very front/center, and push inwards towards the center of the laptop, while ever-so-slightly lifting with your nail.  The inward force slightly buckles the lid such that it causes the latches to disengage, thus allowing you to lift up the lid. Once you have the first latch detached, the rest is relatively easy. Hopefully this shows the idea - just forcing the thumbnail in there is enough to cause the latches to disengage and let the lid rise. 

 

 

UPDATE: Oh boy, I'm not imagining things: Check this out - Do SSD Make Noise? (Explained and Solved!) – UltimatelyTech.com "

 

Given the presence of this component (i.e. an inductor), a phenomenon called ‘electromagnetically induced acoustic noise’ occurs. This sound is what is getting termed as high pitch noise. This noise is also known as coil whining.

On its own, an SSD can produce this type of sound which is so quiet, it’s only just audible when it occurs."

 

And a video, with sound ... SSDs are NOT SILENT! SSD M.2 making high pitch noise or crackling/whirring - YouTube

 

 

Many, many more links online to this phenomenon! 

 

So - the next question is, was this introduced by the WD SSD, or was it there before with the primary SSD, and/or, the RAM / CPU circuitry? Or maybe the new SSD is ADDING to the existing noise, making it more noticeable to me? Maybe I need to wear earplugs 🙂

[John Ratsey]

11 hours ago, Steerpike said:

@John Ratsey I seem to recall, you are hyper-sensitive to fan noise, as am I. Or more accurately, you are hyper-sensitive to slight changes in noise, such that a steady faint 'hum' of a fan is much better than a stop/start fan? 

 

OK, So - after I installed the second SSD in my LG Gram, I noticed a strange, high-pitched, random sound that I find hard to describe. Basically, it's high-pitched 'electrical' noise, the type of noise a circuit gives off when there's some high-frequency processing going on. Also similar to what is heard from battery chargers, where they are doing lots of voltage switching (though this tends to be more steady/predictable). Actually, best description may be - the sound you USED to hear from a hard drive, when the head was 'seeking' - very random, high pitched. 

 

Anyway - just wondered if you have noticed this noise on your LG Gram. Note that this is VERY faint ... but I happen to sit in a very quiet room, and don't have any TV/music playing during the day, and can literally hear a pin drop! Plus, I have the most sensitive hearing ... more of a burden these days 🙂 . 

 

PS I also found a nice way to remove the base of the laptop - you can insert your thumbnail between the removable lid and the surrounding frame at the very front/center, and push inwards towards the center of the laptop, while ever-so-slightly lifting with your nail.  The inward force slightly buckles the lid such that it causes the latches to disengage, thus allowing you to lift up the lid. Once you have the first latch detached, the rest is relatively easy. Hopefully this shows the idea - just forcing the thumbnail in there is enough to cause the latches to disengage and let the lid rise. 

 

So - the next question is, was this introduced by the WD SSD, or was it there before with the primary SSD, and/or, the RAM / CPU circuitry? Or maybe the new SSD is ADDING to the existing noise, making it more noticeable to me? Maybe I need to wear earplugs 🙂

I haven't noticed my Gram 17 making annoying high-pitched whines but that may be because any sound is drowned by the continouus purr of the cooling fan or maybe because my ears are old enough that they struggle to pick up the highest frequencies. Furthermore, I put thermal pads (2mm, I think) on the SSDs because I had some in the drawer. The pads should benefit the cooling but could also dampen any vibration of the electronic components.

 

Thanks for the tip about removing the base. I'll have to remember that for when I next need to get inside.

[Steerpike]

9 hours ago, John Ratsey said:

I haven't noticed my Gram 17 making annoying high-pitched whines but that may be because any sound is drowned by the continouus purr of the cooling fan or maybe because my ears are old enough that they struggle to pick up the highest frequencies. Furthermore, I put thermal pads (2mm, I think) on the SSDs because I had some in the drawer. The pads should benefit the cooling but could also dampen any vibration of the electronic components.

 

Thanks for the tip about removing the base. I'll have to remember that for when I next need to get inside.

SSD Thermal pads - something like this? 

https://www.amazon.com/Thermal-pad，-70x20x1-5mm-Pcie-Heatsinks/dp/B07QMWB4W6/ref=sr_1_11

 

Intuitively you'd think a 'foam'-like product would insulate rather than radiate, but - obviously not!  a pad like that would appear to be just the right thing!  

 

Did you apply to both sides? The SSD I bought was single sided, so I would assume there's space on both sides for padding. Unsure whether the pre-installed SSD is two sided or not.