Introduction
At Computex Taipei and other tech conferences earlier this year, new USB4 chipsets from RealTek and Intel were announced. The Sparkle TD-8140 USB4 travel dock is the earliest available Intel JHL8140-based product. JHL8140 (code-named Hoover Ridge) is a cost-reduced version of Intel’s JHL8440 (Goshen Ridge) Thunderbolt controller chip which forgoes downstream 40Gb/s PCIe functions but adds the ability to operate without an external power supply.
Full disclosure: While this is not a sponsored post, the product showcased was provided at no cost for evaluation purposes. Products received in this capacity are destined for teardowns, future device interoperability testing, and/or charitable donations.
Product Specs
- Host Interface: USB4 40Gb/s with 85W PD charging
- USB-C 3.x Gen2 10Gb/s w/DP Alt mode (8K@30Hz)
- DisplayPort 1.4 (8K@30Hz)
- HDMI 2.0 (4K@60Hz)
- 2x USB-A 3.x Gen 2 10Gb/s (5V x 0.9A)
- 3.5mm TRRS combo audio
- Gigabit Ethernet
- USB-C DC power input port (100W PD)
- 80cm detachable USB4 cable
- Dimensions 105x105x33 mm
Unboxing & Physical Characteristics
Unlike the earlier Sparkle products I reviewed (TDX-120GD TB4 dock and TCX-230DA external GPU), the TD-8140 is housed in a simple plastic shell rather than anodized aluminum. While it has a pleasant weight and doesn’t feel cheap, it is not a premium design like the other Sparkle products.
Functional Testing
Monitors
Although there are three receptacles on the dock that can support monitors (DP, HDMI, and USB-C) only two of these can be used at any one time. So if you use the USB-C port for a monitor, either the DP or HDMI port will be disabled. This is a limitation of the Intel JHL series controller chips.
Up to 2x 4K60 or a single 8K30 monitor is supported; I only had 1440p and 1080 monitors available to test and I found no video issues.
Although not specifically disclosed in Sparkle’s materials, the dock is limited to a single monitor output when used with older USB-C laptops – again a limitation of the JHL controller so this was expected.
Power
The hub works with or without an external power supply. Without a power supply, you can exhaust the available 15W from a Thunderbolt laptop port or ~7.5-10W from USB-C laptops quickly. For example, if you attempt to charge a cellphone and run an external hard drive, you may run out of power. In this scenario, it seems that the dock just switches off one of the USB ports. On some budget docks, the whole device can reset instead which can be rather disruptive. Below I’ve gathered some photos I took while testing the most common scenarios. In this testing, I was surprised that my new Plugbale 240W EPR USB-C power meter with OLED screen was able to monitor power while also passing along the 40Gb/s signal. Very impressive!
| Power Scenario | TB4/USB4 mode | USB-C mode |
| bus power + idle + no monitors | ~0.7W | ~1.9W |
| bus power + idle + 2 monitor simulator dongles | ~3.4W | ~2.75W (only 1 dongle active) |
| USB-C downstream charging | ~12W max | ~8W max bus-powered ~12W max AC/Mains powered |
| host charging de-rating at 100W PD input | yes – 85W max | same |
| host charging de-rating at 60W PD input | yes – ~45W max | same |
| USB A ports downstream charging | 2.5W/port max | same |
| Bus power – overload result | single port shutdown, no device reset | single port shutdown, no device reset |
| Bus power – daisy chain after Thunderbolt 3 eGPU | unreliable when plugging in devices | n/a |
I tried connecting the hub to some older laptops that only supported USB-C, not Thunderbolt or USB4. The hub functioned OK, but my laptop (Surface Book 2) only outputs ~10 watts on the USB-C port, so the power available to devices plugged into the hub was even more limited than the Thunderbolt scenario.
With a USB-C power supply plugged in, the dock worked well. The dock accepts 100W PD and did not exceed 85W sent to the laptop. With a 60-watt charger I was able to confirm that laptop power was limited to ~45W so power de-rating is done correctly.
Curiously, I was only able to get ~2.5W from each of the USB A ports rather than the advertised 4.5W. Perhaps the claimed 4.5W is shared between the two ports? I will need to re-test this more carefully to double-check the results.
To-do: deeper comparison of JHL8140 vs VL830 power consumption
USB device performance
I did not test USB device performance, but I had no issues with an external SSD, keyboard, mouse etc. All seemed to work OK.
Other notes
When I daisy-chained the Sparkle hub after my external Thunderbolt 3 GPU, I noticed inconsistent results with the hub resetting whenever I plugged or unplugged a device. I suspect this was a power issue with my eGPU but I’ll have to investigate further.
The Dock is shown in the Thunderbolt control center, but is missing from the built-in Windows 11 USB4 domains page [Settings – Bluetooth & device -> USB -> USB4 Hubs and Devices].
Teardown
PCB Details
[click for hi-res photos]
Component Topology
Functional chips:
- TB4/USB4 upstream port
Supporting chips:
- Cypress CY7C65219 Dock management controller
- Cypress CYPD5225 USB-C PD controller
- Cypress CYPD3177 DC barrel jack replacement controller
- 6 flash ram chips by Winbond and GigaDevice
Analysis
Sparkle makes good use of PCB real-estate with the top side fully packed with chips and other components. The rear is sparsely populated with just the Realtek Ethernet chip, 5 flash chips for firmware, and some mosfets – components that don’t get very hot and don’t need a heatsink.
There are 3x 10-pin headers with some of the pins routed to the FlashROM chips which are probably used for firmware programming and testing. In addition, there are two un-populated headers (a 4-pin and a 5-pin) next to the C-Media audio chip. These are probably used to attach a speaker, microphone, and volume/mute buttons for the Conference Dock model showcased at Computex:
Similar to the other Sparkle devices, each ports gets its own filtering and protection diodes. But the overall design lacks robust EMF shielding seen in some of my other device teardowns. In the case of the Sparkle desktop TB4 dock and eGPU, the chassis was metal, but the TD-8140 is just plastic. I don’t want to mislead – with good PCB design, you don’t necessarily need shielding to meet minimum requirements – but adding shielding can’t hurt and makes the device more premium. Even just adding sticky-metal foil to the inside of the plastic case and attaching a conductive foam to make contact with the PCB ground could help. But maybe Sparkle’s PCB is just excellent and wouldn’t improve with any of that – I have no way to test it with my equipment.
After examining the signal traces coming from the JHL8140 chip, I thought it looked familiar, so I compared it to the Sparkle TDX-120GD which has the JHL8440 chip. The chip size is identical as is the signal routing. Are these actually the same chip? Its possible, but Intel makes it clear in the MDDS documentation that the JHL8140 is 0.226 grams and the JHL8440 is 0.45 grams so something is different internally. I don’t think Intel is simply fusing off the PCIe logic on these – its probably a different design but re-using the functional blocks from the JHL8440.
Conclusion
Overall, I like the Sparkle Dock. There is only one product I can compare it to directly: the Hyper HD583 USB4 Mobile Dock.
Sparkle TD-8140 vs Hyper HD583
| Feature | Sparkle USB4 Travel Dock TD-8140 | Hyper USB4 Mobile Dock HD583 |
| Price ($USD) | $179.99 ($138 on sale) | $139.99 ($77-120 on sale) |
| USB4 chipset solution | Intel JHL8140 | Via VL830 |
| Cable | 80cm detachable ✅ | 30cm attached |
| Legacy USB-C system monitor support | 1x 4K60 | 2x 4K60 (MST)✅ |
| Mac monitor support | 2x 4K60 or 1x 8K30 ✅ | 1x 8K30 (MST) |
| TB3/TB4 monitor support | 2x 4K60 or 1x 8K30 | same |
| USB-C video out | 1x 8K30 or 4K60 ✅ | n/a |
| HDMI | 2.0b (4K60+HDR) | 2.1 (8K30) ✅ |
| DisplayPort | 1.4a (8K30) | same |
| USB-C downstream data ports | 1 (combined with video) | 2 (when no PSU connected) ✅ |
| Size | 105 x 105 x 33mm | 85 x 85 x 19.5mm ✅ |
| Weight | 243 grams | 131 grams ✅ |
| EMF shielding | n/a | sheet metal ✅ |
| Internal heatsink | aluminum ✅ | n/a |
| Carrying case | included ✅ | n/a |
| Other ports | 1GbE, 3.5mm audio, 2x USB A 10Gb/s | same |
| Power | 100w PD (85W to laptop) | same |
Which is better?
The Sparkle feels like a higher quality product but is also twice the size and weight. Inside, ~40% of the volume is empty space. At first, I thought this was because they would use the same chassis parts for the TD-8140S Conference dock model and just cut a hole in the top for the speaker. But looking carefully at the photos from Computex, the Conference dock is even thicker. So I’m not sure why Sparkle left so much empty space. I think it would be better to slim-down the dock and add air vents for the heatsink. Afterall, most business-class laptops are less than 20mm thick but this dock is 33mm so it would put a bulge in anyone’s laptop case.
As for features, the docks are similar. On both products, I think including 1Gb/s Ethernet instead of 2.5Gb/s is a big miss. The BOM cost increase of the better Realtek RTL8125 chip is small and the rest of the PCB remains the same so I don’t know why these products must go with the dated RTL8153.
Because of the MST chip in the Hyper, it is the better choice for legacy USB-C laptops while the Sparkle is better for Mac and Thunderbolt-equipped PCs. The USB-C port with DP alt mode on the Sparkle makes it a more versatile and future-proof product as you can daisy-chain another dock that can consume USB+DP, connect to a USB-C monitor directly, or even a VR headset. No such luck with the Hyper.
So can I recommend the Sparkle Dock? At $179.99 list price it is rather expensive compared to the $139.99 Hyper dock. There are a handful of other vendors planning to launch JHL8140-based products in October/November so hopefully pricing will drop below $150. In any case I’ll be keeping watch and will update the master Thunderbolt 4 / USB4 hub/dock list accordingly.
Update: Both the Sparkle and Hyper have been heavily discounted – below $138 for the Sparkle and down to $77 for the Hyper.
Appendix
Some excerpts from the JHL8140 spec that leaked last year – I’m not sure what I wanted to say about these, but here they are in case of interest:
Dan S. Charlton 
Great to see another innovation and iteration of hardware. The added ability to power the dock from the host makes it that much more functional for sure. One thing which I think would be interesting to look at is the USB C charge input port, where you connected the PD supply. If a USB C monitor were connected, which can provide host power, would it also function in DP-alt mode on that port? I believe that has some similarity to your test with the eGPU, albeit that is not utilizing DP-alt mode.
LikeLike
Good thinking – but sadly that connector is power-only.
I suppose its possible to have role-switch PD and data+video on two connectors 1 for monitor, and one for laptop, but I suspect the standard Cypress PD solution does not allow for that. ALl the chips on the board align to Intel’s simplest reference designs for JHL8440.
LikeLike
BTW The top down picture confused me initially as appears to be a barrel connector power supply instead of a 3.5mm audio jack. The OE site (
https://www.sparkle.com.tw/en/Product/Detail/79) has two versions, one of which shows as 3.5mm audio cable (https://www.sparkle.com.tw/Uploads/ProductBanner/3794ec2c-3570-4e5f-8e96-4d309400d92b.jpg) the other has the barrel connector (think that is the incorrect version).
LikeLike
Oh wow I totally missed that. Someone in photoshop land goofed. I’ve swapped the image on my blog. I will let the Sparkle PM know about the goof and see if I can get you a t-shirt or something 🙂
LikeLike
That is also one of the reasons I thought about that USB PD input port for use as a data port. And then my brain did a double-take when I read about the audio jack. Black medium please! 😀
LikeLike
Any idea if this dock supports DSC?
LikeLike
That’s a yes by virtue of it not being dependent on the dock itself. Laptop iGPU and monitor itself need support for DP1.4. Keep in mind that not all monitors implement DSC when they are DP1.4. I’ve experienced this on an LG 4K with USB C and DP 1.4 support.
LikeLiked by 1 person
Is it actually true that the hub doesn’t need to support DSC? I have an Anker 556 USB4 hub, and when I hook it up to my 144Hz 4K monitor over HDMI, DSC does not work and I get YCbCr422. Whereas if I connect the monitor directly to my Mac over USB-C, I get DSC and full RGB. (Haven’t tried using DisplayPort to the hub yet, though.)
LikeLike
You’re not wrong. The challenge is that DP1.4 includes DSC, which both docks state is what they are running. But the practical application can be problematic with DSC not being used. So it’s certainly relevant to ask the specific question if DSC is possible on the hardware. Making matters worse, there is no such thing as a configurable param to force DSC state. The thing that potentially gives it away is the the max supported video output is listed as 8K@30Hz. This implies it isn’t utilizing DSC as would otherwise be 8K@60Hz. I’ve only been able to force DSC on hardware from Dock Case, but that’s USB 3.2 Gen 2 and not USB 4 / TB4.
LikeLiked by 1 person
Thank you for the context! This field is so incredibly confusing right now, and your website is an absolute treasure trove. I feel bad about it, but I feel like the only way to figure these things out for sure is to buy the hardware, check to see if your expected resolution/refresh rate works at full RGB, and return it if it does not. There’s just not enough information provided on the marketing pages, and even devices with the same underlying chips seem to support different subsets of features.
By the way, do you know if USB4/TB4 is actually required for (full RGB + non-DSC) 4k@120Hz? I see that some docks from Dock Case are using DP 1.4, which suggests 4k@120Hz support, but I don’t know if a USB 3.2 Gen 2 connection provides enough bandwidth on the DP side for that. (They’re listed as only supporting 4k@60Hz FWIW, but I don’t know how definitive that is.)
LikeLike
Oops, I guess you’re not the author! Well, feel free to ignore that part. 😅
LikeLike
How does the math work out for “USB-C 3.x Gen2 10Gb/s w/DP Alt mode (8K@30Hz)”? My understanding was that USB4 hubs using tunneling for DP could get 4-lane performance (meaning approx. 25.92Gb/s, or barely enough for 8K/30Hz without DSC) concurrently with 10Gb/s USB speeds, while alt mode necessarily dropped you down to 2-lane performance (12.96Gb/s, or only enough for 4K/30Hz without DSC). So how could an alt mode port possibly be doing 8K@30Hz? Moreover, if that port is streaming video, wouldn’t that limit the remaining ports to 20Gb/s total, meaning (probably?) the equivalent of 2-lane tunneled DP performance? Or is there some tunneling magic to the USB-C port here that I don’t understand?
LikeLike
In 4 lane mode USB is basically useless (USB 2 over signaling). In 2 lane mode USB is solid and limited to 5Gbps USB 3.2 Gen 1. This is not 10Gbps as is more about the protocol with the bandwidth calculation being derived from that.
Running multiple protocols on a single wire is an interesting process to figure out scenarios of with a given list of things connected, how will it work.
For the newer Dock Case hubs (I have one) you can force 2 or 4 lane mode which is how DSC is triggered for 2 lane mode. The Dock Case menus call this Extreme (4 lane) vs Balanced (2 lane) if I remember. I don’t have a display that is higher than 4k@60Hz to test higher resolultions or refresh rates.
TB4 has the capability to get more stuff to the hub for further breakout. And technically can support three displays via alt-mode via the downstream TB4 ports. I believe it includes support for one physical DP port and one PCIex1 device natively. This is why you’ll often see a 2.5G port. Those with a 1G port utilize USB protocol. TB4 doesn’t boost the DP 1.4 bandwidth per port though and abides by the same 2 lane / 4 lane principal for that physical port. I haven’t recently researched the subtraction formula for bandwidth remaining from TB4 itself.
Bottom line, if your goal is 4 lane mode and to have USB devices, then a TB4 (USB 4) dock is needed. If 2 lane mode is needed (with or without DSC) and to have USB devices then a USB 3.2 Gen 2 dock is sufficient.
LikeLiked by 1 person
Hello, and thank you for this detailed and helpful teardown.
I am a bit of a newbie at this so I may be mixing up specs. Could you tell me how much power this dock can provide via its USB-C 3.2 Alt DP port to a portable monitor? Going by the table in this teardown I understand up to 12W (with mains power to the dock). Your page on USB-C Splitters lists this dock as well, but there with 15W for that particular spec. The manufacturer spec sheet on the other hand seems to indicate just 5W (“1x USB-C 3.2, 5V/1A Device Charging”).
Basically I want to run a single cable to the monitor, data and power in the same cable. I can do this when I connect the monitor directly to my laptop or my phone, but I am worried the dock will form a bottleneck. 5W would be very little, 15W sounds plenty, 12W I am not sure.
So I am a bit confused how much I can expect coming from this dock – as I said, I may be mixing up things.
LikeLike
Hi I was able to get 12W from the USB-C downstream dp-alt-mode port.
I will double-check this. 5W is definitely not correct but I don’t think 15W is either. I reckon 10W is what the manufacturer intended and 12W is just a bonus.
LikeLike