Tag: Der8auer

  • Wire View Pro 2, Der8ench Bench Table, and Thermal Grizzly Gear

    Wire View Pro 2, Der8ench Bench Table, and Thermal Grizzly Gear

    Key Takeaways

    1. Wire View Pro 2 Features: The latest version includes per-pin current sensing, onboard logging, and a compact fan for cooling, helping diagnose and prevent GPU power connector failures.

    2. Diagnostic Capabilities: The device measures current, voltage, and temperature across the 12VHPWR pins, monitoring conditions that could lead to failures, but it cannot guarantee prevention.

    3. Der8ench Test Bench Design: This innovative test bench integrates multiple features like USB hubs, fan controls, and storage mounts, enhancing convenience for overclockers and testers.

    4. Advanced Cooling Solutions: Thermal Grizzly’s new GPU and CPU water blocks feature intricate designs for optimal thermal transfer, aimed at high-performance systems with a focus on durability and usability.

    5. Pricing and Availability: The Wire View Pro 2 is expected to cost around 100 euros, while the Der8ench test bench is projected to be around 200 euros, with production details still being finalized.


    Wire View Pro 2 is the latest version of der8auer’s in-line GPU power monitoring tool. It was first inspired by Elmor’s PMD and designed to look into problems associated with the 12VHPWR connector. Now, this new model includes per-pin current sensing and onboard logging, which helps in diagnosing failures or even preventing them through extended data evaluation.

    Functionality and Purpose

    This device is placed between the GPU and the PSU, measuring current, voltage, and temperature across each of the 12VHPWR pins. It can also monitor sensors on the connector and other external surfaces, such as the backplate or ambient temperatures of the system. The intention is to detect conditions like poor contact resistance or high temperatures before a failure happens. However, der8auer makes it clear that while it can provide diagnostic insights post-failure, it cannot ensure that connector failures will be avoided.

    Enhanced Features

    The improved version boasts onboard storage of up to 32 MB, which can log sensor data at adjustable intervals (for instance, every five seconds). This allows users to track system performance over roughly three months, helping them spot gradual issues like rising power consumption or temperature spikes. A shout-out to Nvidia here—who would’ve guessed that turning power connectors into hotplates would ignite such a wave of innovation?

    A compact 30 mm fan helps keep the connector and PCB cool, managing up to 10 watts of thermal load. This is particularly vital in systems with low airflow, like those that use water cooling. There’s also a fail-safe connector that automatically shuts down the PC when it detects severe problems.

    Release and Pricing Information

    Although the release is set for late 2025, the pricing is still being decided, with estimates around 100 euros. Delays in production are likely due to the need for custom molds for the elongated connectors, which are costly to produce—just the tooling alone can run up to tens of thousands. We thought Nvidia would invest the right amount to get it right the first time, but apparently that wasn’t the case.

    The Wire View Pro 2 is designed for system integrators, diagnostics experts, and high-end users who are worried about the reliability of new high-power connectors. While it is not a cure-all for connector failures, it serves as an essential tool for forensic insight and preventative assessment.

    Innovative Test Bench Design

    The Der8ench table reshapes the expectations of a PC test bench. Initially created for der8auer’s own requirements, this platform merges passive motherboard mounting with active data and power support features. Unlike conventional benches that merely hold hardware, Der8ench provides a semi-integrated USB hub, fan and pump controls, microSD slots, and SATA SSD mounting points—all on a PCB base. It’s like the hardware equivalent of an everything bagel loaded with cheese—full of features, unapologetically jam-packed, yet still practical. Delicious!

    This unit connects directly to a motherboard USB header and offers four USB 2.0 ports (two on each side). It also supports up to four microSD cards, perfect for keeping OS images, drivers, or benchmarking tools. SATA SSDs can be directly mounted onto the bench for affordable, large-capacity local storage.

    Power and Connectivity

    Three fan headers and a pump header allow for controlled power delivery. Each header can be set to 100%, 50%, or respond to external input from the motherboard. Each zone is individually fused, and the table draws power from a 6-pin PCIe connector, making it compatible with standard desktop power supplies.

    Two USB Type-C ports can provide power to external devices like thermometers or chargers. Probing points on the PCB offer quick access to PSU lines for external power devices or monitoring.

    For users who frequently swap components for testing, the convenience of integrated I/O and storage is a major advantage over the minimalism found in traditional bench platforms. It’s clearly made for overclockers, system reviewers, and development labs where speed and consistency matter more than looks.

    Pricing and Availability

    The complete package, including aluminum supports and optional accessories, is projected to cost around 200 euros. While specific availability details were not set, initial production will be done in-house before possibly being outsourced.

    Advanced Cooling Solutions

    Thermal Grizzly’s new GPU and CPU water blocks highlight sophisticated machining and fluid design to cater to high-performance systems. The Astral GPU block for Nvidia’s top-tier RTX 5090 and the new CPU block both feature dense microfin structures with 0.2 mm spacing, ensuring a high surface area for optimal thermal transfer.

    Each block goes through a detailed multi-stage CNC process, which includes machining, glass-blasting, and precision finishing. This guarantees a perfect mirror contact surface while maintaining durability and steady thermal conductivity. Joe Robey, Thermal Grizzly’s lead engineer, states that each GPU block takes approximately two hours to produce due to the complex triple-flip machining and tight tolerances.

    Design Priorities

    Key design elements focus on minimizing dead zones to stop air from building up, ensuring steady coolant paths, and enhancing the aesthetic with a clear, symmetrical coolant layout. The CPU block shares a similar look and engineering design, featuring a robust brass body for structural strength and a limited parts count to minimize potential failure points.

    The retail price for the GPU block is set at 500 euros, aimed at high-end users who want both performance and visual appeal in custom loop setups. Details for the CPU block, including its price, weren’t finalized when it was announced.

    Der8auer and Thermal Grizzly keep their focus on addressing practical issues for power users and builders. Instead of chasing trends, they emphasize usability, durability, and thoughtful design based on real-world experience.

    If you’re like me and sometimes find yourself tinkering with your system not because it needs fixing, but because it’s enjoyable—that quiet satisfaction of changing something to see how it responds—then the Der8auer bench is designed with that mindset in mind.

    Truly inspiring work, Nvidia—without your thermal failures, a good portion of this product lineup wouldn’t even be here.

    Source:
    Link


     

  • RTX 4090 Connector Melting Issues: Ongoing Problems Explained

    RTX 4090 Connector Melting Issues: Ongoing Problems Explained

    Key Takeaways

    1. Connector Melting Issue: RTX 4090 users are experiencing melting 12VHPWR connectors, leading to burnt plastic odors and fusing to graphics cards and power supply units (PSUs).

    2. Thermal Imbalance: The lack of current monitoring across the 12V lines results in uneven current distribution, causing thermal hotspots and connector degradation.

    3. Resistance Variability: Significant inconsistencies in resistance values among the power pins can lead to localized heating, generating enough heat to compromise connector integrity.

    4. Vendor Accountability: Inno3D has denied responsibility for connector issues under warranty, while other vendors like ModDIY have shown willingness to assist, highlighting inconsistent vendor responses.

    5. Design Criticism: Der8auer criticizes Nvidia’s power connector design and calls for better protection and support for consumers, emphasizing the need for a redesign to prevent future failures.


    The ongoing issue of melting 12VHPWR connectors has come back into the spotlight, this time impacting a user of the RTX 4090 who reached out to overclocker and YouTuber Der8auer. While enjoying some gaming, the user detected a familiar burnt plastic odor and found that the high-power connector had fused to both the graphics card and PSU ends. The graphics card in question was an Inno3D model that had since been modified with water cooling, operating with an undervolt and connected via a ModDIY cable that used 16 AWG ultra-soft wires. Even with careful assembly and a PSU (SilverStone Extreme 850R) considered suitable for a 4090, the plug became stuck, indicating a thermal issue within the connector. Inno3D’s customer service, however, declined to provide assistance, claiming that the damage was on the cable and PSU sides—an argument that Der8auer finds technically flawed.

    Current Monitoring Issues

    Der8auer elaborates that the main problem focuses on the absence of current monitoring across the individual 12V lines in the 12VHPWR interface. It seems that RTX 40 and 50 series graphics cards are unable to identify uneven current distribution among their six power pins. This imbalance, often caused by varying contact resistances in the wires or pins, can create thermal hotspots. In this particular instance, Der8auer managed to take apart and replace the damaged connector on the GPU side himself, using an H++ version of the 12VHPWR plug. His measurements showed significant inconsistencies in resistance values—one pin was as low as 12 milliohms, while another reached 43 milliohms, and one had no connection whatsoever. These differences directly lead to uneven current flow, which results in localized heating and ultimately causes connector melting and degradation. Even with a modest power draw, this can generate up to 4.6 watts of heat concentrated in a small area—sufficient to compromise the connector’s integrity over time.

    Vendor Responses

    Adding to the frustration is the inconsistent response from vendors. ModDIY displayed some readiness to assist, including an offer to pay for repair costs, but Inno3D consistently denied any responsibility, even though the card was still under warranty. Der8auer emphasizes that only the GPU-side design can effectively reduce these risks, as neither the cable nor the PSU can monitor load distribution. He notes that while high resistance can effectively “cut off” a wire, it causes more current to flow through the remaining wires, which leads to a cycle of overheating. Ultimately, Der8auer criticizes Nvidia’s poor power connector design and the lack of accountability from its partners. With high-end graphics cards priced over 2,000 euros, consumers deserve better protection—and quicker support—when failures happen. Until Nvidia revisits this design, users face ongoing risks, regardless of how meticulously they assemble or undervolt their systems.

    Source:
    Link

  • Melted RTX 5090 Power Connector Due to Circuit Design Flaw

    Melted RTX 5090 Power Connector Due to Circuit Design Flaw

    Key Takeaways

    1. A user experienced a dead RTX 5090, initially thought to be user error due to third-party cables.
    2. Der8auer investigated and found severe burns on both connectors and cable shielding, indicating a serious issue.
    3. One cable was pushing nearly 260 Watts (22 Amperes) through a single wire, far exceeding the recommended 5-6 Amperes.
    4. Thermal images revealed a hot spot over 100 degrees near the PSU, suggesting potential for cable failure under extreme conditions.
    5. Caution is advised with power setups for the RTX 5090, and it’s too early to blame the hardware without further investigation from Nvidia.


    Getting an RTX 5090 turned out to be quite a challenge, and for the few lucky ones who did get their hands on one, there was an unfortunate incident where a user ended up with an extremely costly piece of junk. Recently, on Reddit, one user shared their troubles with a dead RTX 5090, and many thought it was just a case of user fault since they were utilizing a third-party cable.

    A Deeper Investigation

    However, Der8auer connected with the user to check out the damaged GPU and its cables. His findings tell a completely different story. In a video he shared, you can see that both sides of the connectors were entirely burnt. The cable’s shielding also suffered burns from the intense current flowing through it. Notably, this issue occurred with just one wire on the 12V-2×6 cable, suggesting that it carried more current than the others.

    Shocking Power Draw

    Using a cable provided by Nvidia, Der8auer discovered that one cable was pushing nearly 260 Watts (22 Amperes) through a single wire, which is drastically higher than the advised range of 5-6 Amperes. Thermal images taken during the test confirmed the presence of a hot spot exceeding 100 degrees near the power supply unit (PSU). This indicates that any cable, even those supplied by the original equipment manufacturer (OEM), can fail under such conditions.

    Caution Advised

    While there may be real concerns regarding how the RTX 5090 draws power from the PSU, this is just one documented case, and it’s too soon to start blaming the hardware. Nvidia has not yet provided an official statement on the matter. It might take some time for them to respond, as they are likely preoccupied with addressing black screen problems that have arisen from a recent driver update.