This is not just a regular Starlink V5, but a dedicated automotive revision: rev5_pez_auto_proto1. (As usual, I expect production revisions with a _prod* suffix to appear later) The automotive V5 is the same size as the standard V5, but it has 20 fewer antenna elements: 800 instead of 820. The freed-up space is used to accommodate a GNSS receiver. The standard V5 is intended for stationary use and does not include a GNSS module. The automotive version brings back the good old STA8090 GNSS chip. It also tolerates higher operating temperatures before throttling or shutting down (auto vs regular): center_power_cut.t_trip: 107°C vs 99°C center_idle.t_trip: 105°C vs 97°C center_idle.t_clear: 98°C vs 90°C MAC throttling thresholds: approximately 8°C higher The EIRP-versus-temperature derating curve is also different. auto_proto1 loses noticeably less transmit power at high temperatures. For example, at 100°C, the loss is approximately 8.5 dB, compared with 9.97 dB for prod2. One minor difference: the automotive version lacks the RGB LED found on the home version :) #starlink# #tesla#

Starlink Mini 2 is finally entering production, with a built-in battery now confirmed. The current hardware revision is mini2_prod1. The new terminal is slightly smaller than the original Starlink Mini. Its panel PCB measures 225 × 273 mm, compared with 236 × 276 mm for Mini 1. Average EIRP is also slightly lower: 27.4 dBW, compared with 27.57 dBW. The transmit duty cycle has been reduced significantly, from 75% to 11%, which will result in lower uplink speeds. This limitation is most likely driven by the terminal’s power budget. The terminal includes a built-in 4S lithium battery managed by a BQ40Z-series BMS. Battery capacity is currently unknown. Charging is provided via a USB-C Power Delivery port, supporting an input range of 11–24 V. The MAC thermal thresholds have been increased to 95°C and 99°C, compared with 79°C and 83°C on Mini 1. The battery thermal shutdown threshold is set to 80°C. GPS - present. Explicitly configured for in-motion use. The SoC is the same "Catapult". #starlink# #starlinkmini2#

I'm finishing up Stephen Fry’s Odyssey ahead of the movie release. Even though Nolan based his film on an entirely different translation of the epic poem, I wanted to read this anyway. This is my second reading of The Odyssey; the first was nearly 30 years ago! 😀 I've always loved the story and ancient mythology in general, but it can still be a bit of a struggle to keep track of all the minor characters with hard-to-pronounce names.

All night long, a permanent member of the UN Security Council bombed residential buildings in Kyiv. Russian terrorists launched 68 missiles and 351 drones. At least 15 residential buildings are damaged or partially destroyed. 11 confirmed dead.

A very hard night in Kyiv. The city was targeted by a massive barrage of 74 ballistic and cruise missiles and 496 drones. Tragically, 17 deaths have been confirmed, and more than 90 people are injured. Even the Kyiv Zoo was hit, leaving several animals injured. The strikes caused widespread destruction, damaging numerous residential buildings, warehouses, logistics centers, and the infrastructure of some ISPs. I spent the night taking cover in my basement, remaining there until almost 5 a.m. Need to catch some sleep today...

For those who are asking. 1. It's not a new Mini model. Just a regular all-white Mini. Check my post: 2. It's powered by an external power bank. You can see a power cable going to a pocket. 3. It will not work stably when held and moved as in the beginning of the video.

Thanks @bary_t for this find. It's quite unusual to see "_proto" revisions in production these days. I didn't mention "mini_pez" in my post because I thought it was only an internal prototype. Besides the Mini, there are multiple "pez" boards: rev4_pez_proto1, rev5_pez_proto1, rev5_pez_prod1, and rev5_pez_prod2. What is "pez"? It's a codename of the newest Starlink FEM chip, which will replace "panda". The most distinctive feature of the new chip: active snow melt via FEM control bits. The new FEM supports two configuration parameters: snow_cb_idle and snow_cb_rx, which control PA biasing. - In idle mode (snow melting): biases the FEM's PA stage for maximum heat dissipation rather than RF performance, turning the array elements into distributed heaters - In receive mode: normal RX biasing with a residual snow-prevention bias offset According to the board configuration, 120 out of 571 elements are designated as snow-melt-capable FEMs. The snow_melt.trigs.low_snr.enabled = true, thres = 5.0 dB, t_cold = 35°C — snow melt activates when SNR drops below 5 dB, and PCB temperature is below 35°C. It seems that the panel can zone heating to melt snow/evaporate water only where it's really necessary. This makes the system more energy efficient. Previous versions of Starlink terminals for snow-melting were implemented using heater traces integrated into the PCB (rev1, rev2, HP versions) or by regulating overall power delivery to the entire panel (rev3, rev4, non-pez mini). Interestingly, the new chip has the lowest EIRP of all: 29.48 dBm average (panda: 32.38 dBm average). Also, it supports only 4 antenna elements (single port). Older FEMs support two ports with 8 antenna elements in total.

About Starlink Mini Versions Currently, there is only one Starlink Mini generation: Mini 1. It has eight known production revisions: mini1_prod1 mini1_prod2 mini1_prod3 mini1_panda_prod1 mini1_panda_prod2 mini1_panda_prod3 mini1_panda_prod4 mini1_rugged_prod1 (You can always find a revision of your Starlink in the Star Debug app or the official Starlink app's debug data section.) There are several internal and external differences between these revisions. The mini1_prodX and mini1_panda_prodX revisions use different FEM chips in the antenna array: Bamboo and Panda, respectively. Because of this, these versions are not compatible at the SoC/eMMC replacement level. A SoC/eMMC pair from a Bamboo board can only be installed on another Bamboo board, and the same applies to Panda boards. If they are mixed, the terminal will constantly report low SNR and will never detect a satellite signal. The Panda revisions have higher EIRP on some channels, while the Bamboo power graph is much flatter, with nearly the same output power across all TX channels. All Mini 1 versions use the same Catapult CPU. The built-in router board also has its own compatibility matrix. Router boards from mini1_prod1, mini1_prod2, mini1_prod3, mini1_panda_prod1, and mini1_panda_prod4 are interchangeable, but they are not compatible with mini1_panda_prod2 or mini1_panda_prod3. Likewise, router boards from mini1_panda_prod2 and mini1_panda_prod3 are not compatible with any of the other revisions. The reason is that mini1_panda_prod2 and mini1_panda_prod3 use a direct SGMII connection between the Starlink SoC and the router SoC. All other versions use Ethernet PHYs for the interconnect. Somewhere around the mini1_panda_prod1 revision, SpaceX switched to a single type of plastic for the housing. The more fragile gray plastic on the front panel was replaced with thin ABS plastic, similar to the material used on the rear panel. As a result, the white vinyl sticker is no longer needed either. Interestingly, the mini1_panda_prod3 revision has a slightly higher temperature rating. Its thermal trip point is set to 105 °C, while the other versions are configured for 85 °C. What About “Mini 2”? There are currently no signs of an upcoming Mini 2. This version appears to be stuck at the internal _proto2 revision and does not seem to be progressing. What About the Rumored Battery-Powered Mini Terminal? Starlink gRPC telemetry has included a battery and charger monitoring section for quite a long time, and it does not appear to be tied to any specific hardware revision. Most likely, we will eventually see an additional accessory, similar to the Advanced Power Supply but with battery charging support. Alternatively, this functionality may be added to Mini 2 someday. What is the mini1_rugged_prod1 revision? This is a special revision designed for Starshield. It also uses the Panda FEM and is somewhat close to the early Panda production revisions, but it is still quite different. This version has a newer and better GNSS receiver. (Sorry, I do not want to share the specific part number here) It also includes several hardware improvements for better signal levels and phase-coherent beamforming. Also, no further details here. Packaging and Accessories Later versions of the Mini terminal come in updated packaging with less cardboard and paper. The pipe adapter has also been redesigned. It is now deeper and uses an Allen key, which is included in the box. Power supply and cable are unchanged.

Russian terrorists launched almost 30 missiles on Kyiv. The Dormition Cathedral, a UNESCO World Heritage site, is damaged and burning. There are also damaged residential buildings and infrastructure.

Kyiv is under a massive ballistic missile attack again, with cruise missiles expected in an hour or less. Reports of fire and damaged apartment buildings in the city center. This time, I'm going to a safer place.

SpaceX is preparing the new #Starlink# user terminal - rev5 (Everyone will call it Gen4 anyway). The current revision is rev5_pez_prod2, meaning the terminal should already be in production and likely to hit the market in Q2 2026. The new terminal lies somewhere between the current REV4 and Mini terminal. It is much smaller than the rev4 and slightly larger than the Mini. Panel size is approximately 274 x 350 mm. It uses new FEM chips (codename is "Pez", previous was "Panda"). Only 2 beamformers (similar to Mini) and 820 antenna elements. In comparison, rev4 has 1536 elements, and the latest Mini is 571. The max EIRP is about 24.63 dBm (rev4 is 26.85 dBm), and it seems that the new device has better EIRP stability across a wider temperature range. It's not 100%, but it seems this is the first Starlink terminal without a GPS receiver. The input voltage is 54V with a maximum power of 111W.