Already last week we tried to explain the confusion of names surrounding the capabilities of the “Thunderbolt / USB 4” sockets of the new ARM Macs with the information available at the time. In the meantime, the first devices with an M1 processor are in customer hands and some have dismantled their copy immediately. A look inside a Mac Mini, which was posted in the eGPU.io forum, sheds some light on the darkness, but complete clarity looks different.

For example, Apple uses two JHL 8040 R components on the motherboard of the Mac Mini, the Intel specification database ARK as Thunderbolt 4 retimer leads. However, according to ARK, these components have already been available since the third quarter , while Intel itself is available for the first time 2020 talked about Thunderbolt 4 and many details even only in the middle 2020 have been published. In the latter context, Intel has the Thunderbolt 4 chips JHL 8540, JHL 8340 (both controller) and JHL That’s then then one then then or another .

The Mac Mini with M1 SoC contains two Thunderbolt 4 retimer components from Intel.

(Image: @itsage / eGPU.io-Forum)

The solution to the riddle: Retimer modules do not contain any protocol logic, but merely process electrical signals passing through the lines qualitatively. Since the user data rate has not changed between Thunderbolt 3 and Thunderbolt 4, there are no differences for such “stupid” components or an older retimer chip design also works with Thunderbolt 4.

Controller logic This does not apply for the logic of the actual controller. Contrary to what was recently suspected, Apple actually integrated the Thunderbolt controller into its own M1 SoC – chap! The unanswered question, however, is why Apple now only advertises it with Thunderbolt 3 capability.

Possible explanations include that the logic of the slightly modified frequencies in which Thunderbolt 3 and USB 4 / Thunderbolt 3 differ, not supported. A retimer doesn’t care about the difference, for a logic to be certified according to a standard it looks completely different. Topography extensions such as hubs instead of daisy-chaining must also be supported by a Thunderbolt 4 controller.

It would also be conceivable, however, that Intel launched its Thunderbolt 4 program too late for the first M1 products: middle 2020 all the “regulatory” tests for the products that have just been released should have already run. Or it depends on subtleties such as the increased requirements for screen or PCIe bandwidth support compared to Thunderbolt 3, which an M1 SoC may not be able to meet.

Anyway: That the final naming, although crude confused, but completely correct in itself, has already been explained in detail. And also that there are hardly any differences for users because there are hardly any Thunderbolt 4 peripherals and the maximum data rate remains the same.

(mue)