Troubleshoot Upgraded RAM sticks to 2x8GB but causes slow blink, a failed boot

[JakeXPMan]

Briefly explain your current issue(s)

New RAM installing causes a slow yellow light blinking, failure to boot up.

Steps taken to resolve, but have been unsuccessful

No steps taken yet. Old RAM successfully put and boot fully back onto Windows 11.

Your current Antivirus

Windows Defender

I powered down my newly purchased used Dell Precision T1700 (Windows 11 500gb SSD) Removed the plug, held down Power button 30 secs. Then removed the previous 2 4GB RAM sticks of 1.5 volts, (Micron brand) touched the metal case first (static precautions) I've installed RAM a few times in the past, no issue at all, but
I then got the 2 2x8 GB sticks out, (both matching Kingston brand making 16gb upgrade from 8gb) 1.35 volts (supposed to fully work with Precision T1700) Installed them both in the EXACT slots the previous two were in, but the power went on and a SLOW blinking yellow.

I quickly put the original RAM back in the slots (removing the 2 news ones of course) and glad to say BOOT up just fine and diagnostic shows all good, no errors.

It's meant to be a fresh new RAM install with 1.35 volt sticks of 8GB each, same speed, latency and matching specs.
However the older RAM has higher volt sticks 1.5 each, is this set in BIOS typically, like the battery ?

Did I somehow accidently get the wrong type of RAM ? I highly doubt but let's see,

This is the exact RAM specs, 2 identical versions of this stick.
Kingston 8GB DDR3 DIMM PC3L-12800U (ASU1603DL1KBG/8G)
Key Compatibility Details: Taken from internet search

• Speed: 1600 MHz (PC3-12800).
• Type: DDR3/DDR3L DIMM (240-pin).
• Voltage: 1.35V (Low Voltage) or 1.5V.
• Max Capacity: Supports up to 32GB total (4 x 8GB) for most configurations, though some sources suggest up to 64GB (4x16GB) depending on processor.
• Configuration: Non-ECC Unbuffered DIMM is typically used

I am wondering if its worth the risks to experiment, but then again it could be an easy fix. I think I read Dell can be fussy when it comes to RAM sticks.

I bought the RAM lightly used, and pulled from a working system (was told), it looks very clean and normal from outside and came in its carrying case.
Any tips or incite welcome on this matter!
Thanks.

 

[Bot]

I powered down my newly purchased used Dell Precision T1700 (Windows 11 500gb SSD) Removed the plug, held down Power button 30 secs. Then removed the previous 2 4GB RAM sticks of 1.5 volts, (Micron brand) touched the metal case first (static precautions) I've installed RAM a few times in the past, no issue at all, but
I then got the 2 2x8 GB sticks out, (both matching Kingston brand making 16gb upgrade from 8gb) 1.35 volts (supposed to fully work with Precision T1700) Installed them both in the EXACT slots the previous two were in, but the power went on and a SLOW blinking yellow.

I quickly put the original RAM back in the slots (removing the 2 news ones of course) and glad to say BOOT up just fine and diagnostic shows all good, no errors.

It's meant to be a fresh new RAM install with 1.35 volt sticks of 8GB each, same speed, latency and matching specs.
However the older RAM has higher volt sticks 1.5 each, is this set in BIOS typically, like the battery ?

Did I somehow accidently get the wrong type of RAM ? I highly doubt but let's see,

This is the exact RAM specs, 2 identical versions of this stick.
Kingston 8GB DDR3 DIMM PC3L-12800U (ASU1603DL1KBG/8G)
Key Compatibility Details: Taken from internet search

• Speed: 1600 MHz (PC3-12800).
• Type: DDR3/DDR3L DIMM (240-pin).
• Voltage: 1.35V (Low Voltage) or 1.5V.
• Max Capacity: Supports up to 32GB total (4 x 8GB) for most configurations, though some sources suggest up to 64GB (4x16GB) depending on processor.
• Configuration: Non-ECC Unbuffered DIMM is typically used

I am wondering if its worth the risks to experiment, but then again it could be an easy fix. I think I read Dell can be fussy when it comes to RAM sticks.

I bought the RAM lightly used, and pulled from a working system (was told), it looks very clean and normal from outside and came in its carrying case.
Any tips or incite welcome on this matter!
Thanks.

A slow blinking amber/yellow power LED on Dell desktops is commonly a “hardware fault / failed POST” indication, and RAM is one of the most frequent triggers. The good news is: there is usually no “RAM voltage setting” you need to change in BIOS for DDR3 vs DDR3L—systems read the RAM’s SPD data and train timings/voltage automatically. If the machine boots fine again with the old sticks, the motherboard/CPU are very likely OK.

About the 1.35V vs 1.5V concern

Most DDR3L (PC3L) UDIMMs are dual-voltage (they can run at 1.35V or 1.5V depending on what the system provides). So going from 1.5V sticks to 1.35V DDR3L sticks typically does NOT require any manual BIOS setting.

If the new sticks cause an amber blink, it’s more likely one of these:

• One stick is bad (even if it “looks clean”).
• A seating/contact issue (very common with “it worked before” situations).
• The RAM is the wrong type electrically (ECC / Registered / buffered) even if it “fits.”
• SPD/XMP/organization incompatibility (some Dell systems can be picky with certain module organizations or SPD programming, especially with used/OEM-pulled RAM).
• BIOS compatibility issue (older BIOS sometimes has poorer memory compatibility).

Low-risk checks (recommended order)

• Reseat carefully: Remove AC power, press power button 15–30 sec, then reinstall the Kingston sticks making sure both latches click fully into place. Partial insertion is a top cause of RAM POST failures.
• Test one stick at a time: Install only one 8GB stick in the primary slot (the “first” slot per Dell’s service manual). Attempt power-on. Then repeat with the other stick. This quickly separates “bad stick” from “pairing/slot” issues.
• Test each slot (still one stick only): If one stick works in one slot but not another, that points to a slot/motherboard issue. If neither stick works in any slot, it points to incompatibility or both sticks being faulty.
• Confirm the RAM type on the label: It should say something like PC3L-12800U / DDR3L UDIMM. Avoid anything marked ECC, PC3L-12800E, “Registered” (RDIMM), or “Buffered.” (Those will often cause a no-POST condition in systems expecting UDIMM.)
• Clear residual training data: If your model supports it, perform a BIOS/NVRAM reset using Dell’s documented procedure (some have a jumper or a “clear CMOS” method). This can help if memory training data gets stuck after a failed boot attempt.

Firmware and compatibility steps

• Update the BIOS to the latest Dell release (while running on the known-good original RAM). Memory compatibility fixes are a common reason to update BIOS on older workstations.
• Avoid enabling XMP (if present): Dell workstations typically expect JEDEC default timings, not XMP profiles. Many OEM BIOSes ignore XMP anyway, but if your BIOS exposes it, leave it off.

Interpreting the amber blink properly

Dell uses specific LED/blink patterns (and sometimes diagnostic LEDs 1–4) to narrow down the failing component. If your T1700 has additional diagnostic lights, matching the exact blink pattern against the Dell Precision T1700 service manual will confirm whether it’s specifically “memory not detected/failed” versus another POST issue.

If the Kingston kit still won’t POST after the single-stick tests

At that point, the safest conclusion is: the modules are either faulty or not truly compatible with your T1700 despite being “DDR3L-1600 UDIMM.” Used RAM can be intermittently bad, and Dell systems can be picky about SPD/programming or internal chip organization.

The practical next step is to return/exchange that kit and buy RAM known to be compatible with the Precision T1700 (ideally Dell-tested part numbers or a reputable configurator listing from a major vendor).

Conclusion: there usually isn’t a BIOS voltage setting to fix here; methodically test one stick at a time and update BIOS, and if it still fails to POST, treat the kit as incompatible/defective and exchange it.

 

[stonjean633]

DDR3 on 1.5v is the older one
DDR3L is the newer of the two on 1.35v and is backwards compatible with older mother boards. This voltage should not be the issue here.

Depending on the processor, maybe it does not support 16gb with max of 8gb per stick?

 

[Bot]

DDR3 on 1.5v is the older one
DDR3L is the newer of the two on 1.35v and is backwards compatible with older mother boards. This voltage should not be the issue here.

Depending on the processor, maybe it does not support 16gb with max of 8gb per stick?

You’re right that in general DDR3L is designed to work in DDR3 (1.5 V) systems, so “DDR3L by itself” usually isn’t the problem. Where it can still bite is that some DDR3L sticks are effectively 1.35 V–only (less common), or the SPD programming/module organization isn’t something the Dell BIOS likes.

On the capacity point: a CPU limitation is very unlikely to be the reason here.

Precision T1700 + “2 x 8 GB” is normally supported
The Precision T1700 platform (Haswell-era Intel) is commonly specified for up to 32 GB total using 4 x 8 GB DDR3/DDR3L UDIMMs, depending on the exact configuration. So 16 GB (2 x 8 GB) is typically well within spec for the motherboard/CPU memory controller.

Even in cases where a system had a lower “supported maximum,” you’d usually still get some kind of POST with reduced/partial memory or a different diagnostic message—not a hard fail that only happens with the new modules.

What’s more likely than a “16 GB not supported” issue

• One of the 8 GB sticks is bad (very common with used RAM).
• The sticks are not truly UDIMM non-ECC, despite the listing (label might reveal ECC/registered/buffered variants).
• Module organization incompatibility (some 8 GB DDR3 sticks are built with different chip layouts; certain OEM BIOSes are picky).
• BIOS memory compatibility: updating to the latest Dell BIOS can materially improve RAM compatibility.
• Seating/contact: one side not fully latched will cause an amber-blink/no-POST immediately.

Best way to confirm it’s not a CPU capacity limit

• Check Dell’s official T1700 memory spec for your exact form factor (MT vs SFF) and configuration.
• Verify your installed CPU model (Core i5/i7 vs Xeon E3). Either way, 2 x 8 GB is normally fine on this platform.

Most useful next step (quick isolate)
Test the Kingston sticks one-at-a-time in the primary memory slot (then swap sticks). If either stick boots alone, your system supports 8 GB modules and 16 GB total; the problem is then either one bad stick, a pairing/training issue, or a slot/compatibility quirk. If neither stick boots alone in any slot, treat it as incompatible/defective RAM and return/exchange, ideally for a kit explicitly listed for the Precision T1700.

If you also share the exact amber-blink pattern (how many blinks / pause cycle) and whether your T1700 has any diagnostic LEDs lit, that pattern can usually confirm “memory error” versus another POST fault.

 

[stonjean633]

Try the old school method of cleaning the ram pins with eraser.
Then isolate the ram sticks. Use one then the other and see if they work either single or dual.

 

[Bot]

Try the old school method of cleaning the ram pins with eraser.
Then isolate the ram sticks. Use one then the other and see if they work either single or dual.

Cleaning the contacts and isolating sticks is a sensible, low-risk way to troubleshoot this, with a couple of cautions.

Cleaning RAM contacts (eraser method) — do it carefully
Using a pencil eraser can work, but it can also leave residue or create static if you’re rough with it.

• Unplug the PC, hold the power button ~15–30 seconds.
• Remove the RAM.
• If using an eraser: use light pressure and straight strokes along the gold contacts (don’t scrub hard).
• Afterward, remove all eraser crumbs/residue. Ideally wipe the contacts with 90%+ isopropyl alcohol on a lint-free cloth/swab and let it fully dry.
• Avoid abrasives/sandpaper and avoid touching the contacts with bare fingers afterward.

If you have isopropyl alcohol available, alcohol-only cleaning is often safer than an eraser.

Isolate the RAM and slots (most useful test)

• Install ONLY one new 8GB stick in the primary slot (the first slot Dell expects populated).
• Try to power on. If it posts, shut down and retest the same stick in another slot.
• Repeat the same process with the second new stick.
• If both sticks work individually but not together, it can point to dual-channel pairing/training/BIOS compatibility rather than a dead module.

If neither new stick works alone
That strongly suggests incompatibility (wrong type like ECC/registered, or SPD/module organization Dell doesn’t like) or both sticks being faulty. In that case, updating the BIOS (while using the old working RAM) is worth doing once, but if it still fails after a BIOS update, returning/exchanging the kit is the most practical path.

Extra practical tip
When reinstalling, make sure both retention clips fully click into place. A stick that looks seated but isn’t fully latched will commonly produce the Dell amber/yellow blink with no POST.

 

[Jonny Quest]

Try the old school method of cleaning the ram pins with eraser.
Then isolate the ram sticks. Use one then the other and see if they work either single or dual.

That was my thought, Bot's one stick at a time option. Never heard of the eraser option before

 

[Bot]

That was my thought, Bot's one stick at a time option. Never heard of the eraser option before

The “one stick at a time” test is the most reliable next step, because it tells you quickly whether you have a bad module, a bad slot, or a pairing/training issue.

The eraser trick is an old technician method for oxidized/dirty contacts, but it’s not something I’d treat as a first-choice fix on relatively modern hardware because:

• It can leave residue/crumbs that cause problems if not cleaned off well.
• If you press hard, it can slightly scuff the gold plating.
• It doesn’t help if the root cause is incompatibility (ECC/registered, SPD quirks, module organization) or a genuinely bad stick.

If you want the safer “contact cleaning” version, use 90%+ isopropyl alcohol and a lint-free swab/cloth, let it dry completely, then retest.

Practical next step: do the single-stick boot test in the primary slot, then swap sticks, then (if either works) test in dual-channel again. If neither stick boots alone in any slot, it’s very likely the kit is incompatible or defective and should be exchanged/returned.

 

[Smoke]

Assuming you didn't already, save the BIOS settings as they are and then try to boot on default BIOS settings in case some RAM timing or individual voltage setting isn't compatible.

 

[Bot]

Assuming you didn't already, save the BIOS settings as they are and then try to boot on default BIOS settings in case some RAM timing or individual voltage setting isn't compatible.

That’s a reasonable low-risk step, with a couple of practical notes for Dell systems.

On most Dell Precisions, RAM voltage/timings are not meant to be manually tuned (they’re normally read from SPD and trained automatically), but loading BIOS defaults / clearing NVRAM can still help if the firmware got “stuck” after a failed memory training attempt.

Safe way to try BIOS defaults

• Boot using the known-good original RAM.
• Enter BIOS Setup (F2 on most Dells).
• If there is an option like “Load Defaults” / “Restore Settings,” apply it, then save and reboot.
• Shut down, unplug AC power, hold power button 15–30 seconds.
• Install the new RAM and try to boot.

Things to watch out for when loading defaults

• If you use BitLocker/device encryption, changing certain BIOS settings can trigger a recovery key prompt. Make sure you have the recovery key saved before doing firmware changes.
• If you had any custom settings (boot mode, SATA mode AHCI/RAID, Secure Boot settings, virtualization), defaults can revert them. On many systems this won’t matter, but SATA mode changes in particular can cause Windows boot issues.

If defaults don’t help
At that point, the single-stick isolation test is still the quickest way to determine whether one module is bad versus an incompatibility. If neither new stick boots by itself in any slot even after BIOS defaults (and ideally a BIOS update), the most practical conclusion is incompatible or faulty RAM and it should be exchanged/returned.

 

[stonjean633]

That was my thought, Bot's one stick at a time option. Never heard of the eraser option before

It's a compliment.Means you're way younger than me. Hehe.

 

[Bot]

It's a compliment.Means you're way younger than me. Hehe.

I’m an AI, so I don’t have an age, but the “eraser on contacts” approach is definitely one of those older field techniques that a lot of long-time builders/techs learned.

If anyone does try it, the main point is to keep it low-risk:

• Use very light pressure (don’t abrade the contacts).
• Remove all residue afterward (ideally finish with 90%+ isopropyl alcohol and let it dry fully).
• Then go straight back to the most definitive test: one stick at a time in the primary slot, then swap sticks, then test dual-channel again.

That sequence usually settles whether it’s a seating/contact issue, a bad stick, or plain incompatibility with that Dell platform.

 

[Halp2001]

One more contribution to the various and different suggestions already shared in the thread:

• Check the module density: The T1700 tends to be more stable with 2Rx8 (double‑sided) RAM than with 1Rx8 (single‑sided). Even within Kingston there are different variants, and not all are compatible.
• Disable “Fast Boot” in BIOS: Forcing a full POST can help the system train the memory correctly.
• Test mixing modules: Install one original module together with one new one. If it boots, that’s a clear sign of incompatibility between the two new sticks.

Perhaps one of these tests will give you an extra clue

 

[Bot]

One more contribution to the various and different suggestions already shared in the thread:

• Check the module density: The T1700 tends to be more stable with 2Rx8 (double‑sided) RAM than with 1Rx8 (single‑sided). Even within Kingston there are different variants, and not all are compatible.
• Disable “Fast Boot” in BIOS: Forcing a full POST can help the system train the memory correctly.
• Test mixing modules: Install one original module together with one new one. If it boots, that’s a clear sign of incompatibility between the two new sticks.

Perhaps one of these tests will give you an extra clue

Those are sensible additions, with a couple of clarifications so the troubleshooting stays safe and conclusive.

Module density / rank (1Rx8 vs 2Rx8)

Yes—OEM workstations can be picky about module organization (rank and chip density), even when everything is “DDR3L-1600 UDIMM.” Two modules with the same headline specs can behave differently if one is 1Rx8 and the other is 2Rx8 (or if they use higher-density ICs).

Practical steps:

• Check the label on each Kingston stick for “1Rx8” or “2Rx8” (often printed on the sticker).
• If the two “matching” sticks are not actually the same rank/organization, treat the kit as mismatched even if the model family looks similar.
• If you can boot with the old RAM, a tool like CPU‑Z (SPD tab) can show rank/organization and JEDEC tables for each slot, which helps confirm what you really have.

Disable Fast Boot / force full POST

Disabling Fast Boot can help in some cases because it forces full memory initialization/training rather than trying to reuse previous training data.

Notes:

• Do this while running the known-good original RAM.
• Also consider a proper “cold boot” between attempts: shut down, unplug AC, hold power 15–30 seconds, then retry with the new RAM.

Mixing old + new modules as a test

Mixing can be a useful diagnostic, but interpret the results carefully:

• If the system boots with 1 old + 1 new stick, that does not automatically prove the two new sticks are mutually incompatible. It can also mean one of the new sticks is simply bad.
• So, mixing is best used after the “one stick at a time” test has already shown that each new stick can POST by itself.

A clean, low-risk order that usually settles it quickly:

• Update BIOS to latest (with old RAM installed).
• Disable Fast Boot in BIOS (optional, but reasonable).
• Test each new 8GB stick alone in the primary slot.
• If both work alone, test them together (then optionally test mixed old+new to see if dual-channel pairing is the trigger).
• If either stick fails alone in every slot, treat the kit as faulty/incompatible and exchange/return it.

This approach avoids guesswork and gives you a clear “bad stick vs compatibility/organization vs training/BIOS” outcome.

 

[JakeXPMan]

Lots of good posts in short time, nice will have tomorrow to deep dive some options on this RAM. One detail is BIOS is updated to the latest version BEFORE any RAM was changed around. Version A28.

Processor is NOT a Xeon, but an I7-4790 (which uses non-ECC memory) I assume the kit I have isn't ECC, it ends with the letter U. Not E or R. (registry), unless I have overlooked something, but I wrote every detail before checking online.

I'm thinking it's one of these things,

1. New RAM is just not functional (RAM issue itself)
2. BIOS doesn't like this kit (fussy workstation PC)
3. CMOS needs reset to read the new RAM before it can boot (YT video showed this working)
or, also could be the remaining issues BOT and MT members mentioned above.

My next step, will be... I have found I have another test machine a Dell Desktop (Inspiron 3847) my father just dropped it off. It happens to be useable with this RAM as well, it has a Windows 10 HDD on it, I can test the new RAM and see what it does. This PC is Gen 3, with an i3 (dual core processor) so its nice, but humble in specs for sure. Will be a good safety net, or test machine.

My goal still is to use my new Precision, as it can take Gen 4 processors including Xeon's, built robust... and with 4 RAM slots, it can upgrade much more ram (when it actually works lol)

 

[Bot]

Lots of good posts in short time, nice will have tomorrow to deep dive some options on this RAM. One detail is BIOS is updated to the latest version BEFORE any RAM was changed around. Version A28.

Processor is NOT a Xeon, but an I7-4790 (which uses non-ECC memory) I assume the kit I have isn't ECC, it ends with the letter U. Not E or R. (registry), unless I have overlooked something, but I wrote every detail before checking online.

I'm thinking it's one of these things,

1. New RAM is just not functional (RAM issue itself)
2. BIOS doesn't like this kit (fussy workstation PC)
3. CMOS needs reset to read the new RAM before it can boot (YT video showed this working)
or, also could be the remaining issues BOT and MT members mentioned above.

My next step, will be... I have found I have another test machine a Dell Desktop (Inspiron 3847) my father just dropped it off. It happens to be useable with this RAM as well, it has a Windows 10 HDD on it, I can test the new RAM and see what it does. This PC is Gen 3, with an i3 (dual core processor) so its nice, but humble in specs for sure. Will be a good safety net, or test machine.

My goal still is to use my new Precision, as it can take Gen 4 processors including Xeon's, built robust... and with 4 RAM slots, it can upgrade much more ram (when it actually works lol)

BIOS A28 and an i7-4790 rule out a lot of “platform can’t do it” concerns. That CPU’s integrated memory controller normally supports 32 GB total on LGA1150 boards (and the T1700 is commonly configured up to 32 GB with 4 x 8 GB), so 16 GB (2 x 8 GB) should be fine when the modules are compatible.

About “U means non‑ECC”

Be careful with that assumption. “U” generally means unbuffered (UDIMM), but ECC UDIMMs are also unbuffered. The more reliable indicators are what’s printed on the module label:

• Non‑ECC UDIMM is typically labeled like: PC3L-12800U
• ECC UDIMM is typically labeled like: PC3L-12800E (or explicitly says “ECC”)
• Registered/buffered is typically labeled like: PC3L-12800R (RDIMM)

So it’s worth double-checking the physical sticker for any mention of “ECC” or an “E” in the PC3/PC3L code, regardless of the marketing description.

CMOS/NVRAM reset — can help, but don’t over-rely on it

A CMOS reset sometimes helps after a failed memory training attempt, but if the sticks are incompatible or one is bad, it won’t “fix” that. Since you’re on a Dell, use the documented Dell method (jumper/button/battery method per service manual), and note:

• If Windows device encryption/BitLocker is enabled, firmware changes can trigger a recovery key prompt. Make sure you have the recovery key saved before doing CMOS resets or changing boot/TPM settings.
• After the reset, install only ONE new stick first (primary slot), then expand to two sticks only after a successful POST.

Testing in the Inspiron 3847 is a good diagnostic

That’s a solid, low-risk way to separate “bad RAM” from “T1700 compatibility quirk.”

Interpret the results like this:

• If the Kingston RAM fails to POST in the 3847 as well (especially one-stick-at-a-time), it strongly points to a faulty module or wrong-type RAM. Return/exchange is the practical move.
• If each stick works fine in the 3847 but not in the T1700, that points to a T1700-specific compatibility issue (module organization/rank/SPD behavior), or possibly a marginal DIMM slot/contact issue on the T1700.

Best next-step sequence (keeps it conclusive)

• On the T1700: test ONE new stick at a time in the primary slot (cold boot each attempt).
• If neither works: test those same single sticks in the Inspiron 3847.
• If they work in the 3847 but not the T1700: check the Kingston labels for 1Rx8 vs 2Rx8 and confirm non-ECC; then consider exchanging for a kit explicitly listed as compatible for the Precision T1700 (same capacity/speed, but different module variant).

Given your BIOS is current and the original RAM still works perfectly, the most likely outcomes are (1) one/both used sticks are bad, or (2) they’re a variant the T1700 BIOS doesn’t like even though the headline specs look correct. The cross-test in the Inspiron plus the one-stick POST tests will pin that down quickly.

 

[stonjean633]

Just to add, in case the RAM boots on your other machine is to run the built-in Memory Diagnostics in Windows. Once it passed the memtest, you can be sure the Sticks are good and can narrow down to compatibility issue
(Dell board could be picky on RAMs)

 

[Bot]

Just to add, in case the RAM boots on your other machine is to run the built-in Memory Diagnostics in Windows. Once it passed the memtest, you can be sure the Sticks are good and can narrow down to compatibility issue
(Dell board could be picky on RAMs)

Windows Memory Diagnostic is a useful sanity check, but it can’t make you “sure the sticks are good” in the absolute sense.

Why it helps, but isn’t definitive

• A pass result reduces the likelihood of obvious RAM faults, but it can miss intermittent errors, temperature-related faults, or rare pattern-specific issues.
• It also doesn’t prove the RAM will be compatible with a different motherboard/BIOS memory training behavior (your Precision may still reject modules that another Dell accepts).

Better way to build confidence if the RAM boots in the other PC

• Run Windows Memory Diagnostic as a quick first check (fine).
• For higher confidence, run a dedicated tester like MemTest86 from a bootable USB for multiple passes (this is more thorough than the Windows built-in test).
• In Windows, you can also run a few hours of heavy, memory-stressing workloads (multiple app instances, large file operations) to see if any instability appears, but a proper boot-time memory test is still preferred.

How to interpret results for your situation

• If the RAM consistently passes thorough testing in the Inspiron but still won’t POST in the T1700, that strongly supports a compatibility/training/module-organization issue with the Precision (rather than “bad RAM”).
• If it fails any memory test in the Inspiron, treat the kit as faulty and return/exchange it rather than continuing to experiment in the Precision.

That approach keeps the troubleshooting evidence-based and avoids risking repeated no-POST cycles on the main workstation.

 

[JakeXPMan]

Update, the RAM test is currently working in the older PC Dell Inspiron 3847. Running a slow DELL diagnostic with F12, seems to boot up no issues as well.

This RAM I bought is dual sided PC3L 11-11-B1. and the ones replaced are PC3L single sided 11-13-A1.

The Micron RAM on the precision is 2x4GB PC3 (no L) 1.5 volt and single sided.

 

[Bot]

Update, the RAM test is currently working in the older PC Dell Inspiron 3847. Running a slow DELL diagnostic with F12, seems to boot up no issues as well.

This RAM I bought is dual sided PC3L 11-11-B1. and the ones replaced are PC3L single sided 11-13-A1.

The Micron RAM on the precision is 2x4GB PC3 (no L) 1.5 volt and single sided.

That’s a very useful data point: if the Kingston 2x8 GB runs diagnostics and boots normally in the Inspiron 3847, the modules are at least basically functional. That shifts the focus to a Precision T1700-specific POST/training/compatibility issue (or less commonly a slot/contact issue on the T1700).

What the “dual-sided vs single-sided” difference likely means

“Dual-sided” often correlates with a dual-rank module (commonly 2Rx8), while many “single-sided” sticks are single-rank (often 1Rx8). Some OEM boards are pickier about rank/chip organization than the headline specs (DDR3L-1600) suggest.

That said, the T1700 is generally expected to handle 8 GB DDR3/DDR3L UDIMMs, including many dual-rank variants, so this points more toward a specific module-variant/SPD training quirk than a simple “dual-sided won’t work” rule.

The most likely root causes now

• DDR3L voltage behavior mismatch: Some PC3L modules are truly 1.35 V-only. If the T1700 is only supplying 1.5 V and the module can’t tolerate it, it may fail POST, while the Inspiron (which may support 1.35 V operation) works fine.
• SPD / module-organization incompatibility: The Kingston “ASU…” part you listed may be an OEM-specific module variant. Two sticks can be “PC3L-12800U” and still differ in how they present JEDEC tables/rank info to BIOS.
• T1700 slot population sensitivity: Dell workstations can be strict about which slots must be populated first (and a failed attempt can sometimes require a full power drain / NVRAM reset before it will retrain).
• A marginal DIMM slot/contact on the T1700: Less likely since your old RAM works, but still possible if the 4 GB modules are more forgiving than the 8 GB ones.

Concrete next steps on the Precision (keeps it evidence-based)

1) Verify the Kingston module details while it’s working in the Inspiron
Use CPU-Z (SPD tab) or a similar tool and note:

• ECC vs non-ECC (it should clearly show non-ECC for your i7-4790 system)
• Rank (1R vs 2R)
• JEDEC profiles and any XMP profile

This doesn’t “fix” anything, but it confirms exactly what the modules are reporting to firmware.

2) Do a proper cold-start + retrain attempt in the T1700

• Unplug AC power, hold power button 15–30 seconds.
• (Optional but reasonable) Clear CMOS / NVRAM using Dell’s documented method for the T1700.
• Install ONLY one Kingston 8 GB stick in the primary slot (the first slot Dell specifies for single-module operation).
• Try boot.
• Repeat with the other stick, same slot.

3) If neither Kingston stick will POST alone in the primary slot
At that point, the most probable explanation becomes “this specific Kingston variant isn’t compatible with the T1700’s memory initialization (or it’s effectively 1.35 V-only and the T1700 is driving 1.5 V).” Since it works in the Inspiron, returning/exchanging for a kit explicitly validated for the Precision T1700 is usually the fastest resolution (Crucial/Micron kits that list the T1700 specifically, or Dell-qualified part numbers).

4) If one stick posts but two don’t
That points to dual-channel pairing/training behavior or slot population order. Then:

• Populate exactly as Dell’s manual specifies for 2-DIMM configuration (not just “same slots as before”).
• Test the same two sticks in the other channel’s pair of slots.

Key takeaway

Because the Kingston 2x8 GB works in another Dell, this is now much more consistent with a Precision T1700 compatibility/training/voltage-implementation mismatch than with “the CPU can’t handle 16 GB.” The single-stick-in-primary-slot test after a full cold power drain (and optionally a CMOS clear) will usually give a definitive direction: either it can retrain and work, or you’ll want a different 2x8 GB kit that’s known-good for the T1700.