MacInTouch Amazon link...

USB, FireWire & Thunderbolt

Channels
Troubleshooting, Products
Maybe we should consider the "defect" to be a "feature", as I'd suspect the 100 watts that USB Power Delivery could push could be dangerous?
The official Pi 4 (and recommended) power supply is 15.3W (3A at 5.1V). I assume this was selected in order to satisfy the power requirements of the Pi 4 running at full capacity.

If you were to attach a 100W supply (assuming the board was fixed to properly recognize it), I highly doubt it would try to draw more than 15.3W. I've never heard of a power supply trying to "push" more power than the device is trying to draw. I don't even know if that makes sense, electrically.
 


If you were to attach a 100W supply (assuming the board was fixed to properly recognize it), I highly doubt it would try to draw more than 15.3W
Ever wonder why it is safe to connect an 85-Watt MacBook Pro power brick to an Air that needs only 45? Or how a 45-Watt MacBook Air charger might (eventually) top up a quiescent MacBook Pro?
Ken Sheriff's Blog said:
Macbook charger teardown: The surprising complexity inside Apple's power adapter
There's a lot more circuitry crammed into the compact power adapter than you'd expect, including a microprocessor. This charger teardown looks at the numerous components in the charger and explains how they work together to power your laptop.
Ken Sheriff's Blog said:
Lacking safety features, cheap MacBook chargers create big sparks
Why does a fake charger produce sparks, while a genuine one doesn't? The fake charger constantly outputs 20 volts, so if any metal shorts the connector, it produces a big spark with all its 85 watts of power. On the other hand, the genuine charger doesn't power up until it has been securely connected to the laptop for a full second.
If you were to attach a 100W supply (assuming the board was fixed to properly recognize it),
It's the assumption that the Pi is properly configured that gives pause. That, and the 100-Watt "Power Delivery" USB-C adapter might itself not be configured safely?
 


... I've never heard of a power supply trying to "push" more power than the device is trying to draw. I don't even know if that makes sense, electrically.
No, it does not make sense, electrically.

Technical note: Power supplies we are speaking of are designed as "constant voltage sources" and are not affected by load until the load rises to the capacity of the power supply. Power draw by any device connected to a specification-compliant power supply is entirely under the control of the load device, up to the maximum available power.

Try this analogy: The AC outlet in a typical US home is a nominal 110 Volts AC - a "constant voltage source.”
  • Connect a 100 watt bulb and 100 watts is delivered (~0.91 Amperes).
  • Connect a 10 watt bulb and 10 watts is delivered (~0.091 Amperes).
  • Connect a 1 watt bulb and 1 watts is delivered (~0.0091 Amperes).
  • Connect no bulb and no watts are delivered (can you imagine all those electrons spilling on the floor? ;)
 


I wish I had my notes from November 2018 when I migrated one additional account from an older MacBook Pro with Thunderbolt 2 using the Apple adapter to Thunderbolt 3 for a brand new 2016 MacBook Pro. It was one of the slowest migrations that I have recently done. So the performance numbers you obtained using Thunderbolt between Apple machines are not a surprise.
 


Ric Ford

MacInTouch
I wish I had my notes from November 2018 when I migrated one additional account from an older MacBook Pro with Thunderbolt 2 using the Apple adapter to Thunderbolt 3 for a brand new 2016 MacBook Pro. It was one of the slowest migrations that I have recently done.
It was that exact scenario that first made me aware of the problem years ago. I had two MacBook Pros, each with an internal SSD, and I wanted to clone from one to the other. Naturally, Thunderbolt Target Disk Mode seemed like a great, fast option. But... it wasn't. It was stunningly slow, and that's when I began Thunderbolt Target Disk Mode. It has remained stunningly slow in the years since, through today.
 


Ric Ford

MacInTouch
I just put a 2017 iMac 5K in Target Disk Mode and connected it over Thunderbolt 3 to a 2018 MacBook Pro. Blackmagic Disk Speed Test says:
Write: 895 MB/s​
Read: 468 MB/s​
That's an awful lot slower than the 40 Gbps (or even 22 Gbps) that Thunderbolt 3 should provide (i.e. ~3000 MB/s).
I’d be interested to see your results if you connected them together by Thunderbolt networking and mount one as a network share (SMB and/or AFP) on the other. I believe you should be able to select a network volume for testing in Blackmagic Disk Speed Test.
I share Todd's curiosity about what numbers one might see from a Thunderbolt 3-connected "point to point" networked drive, served by a machine with a very fast SSD. It would have to be SMB, I think, but if the protocol implementations are highly tuned, it wouldn't completely surprise me if it were faster than target disk mode.
I spent more unsuccessful time yesterday trying to get this to work, and I have no idea what's wrong. Things I tried...
Unable so far to get Thunderbolt 3 file sharing working with the 2018 MacBook Pro (on the latest Mojave), I connected the 2017 iMac 5K (running macOS 10.12) and a 2017 MacBook Air (running macOS 10.13) via Apple's Thunderbolt 3-Thunderbolt 2 adapter, mounted a file-shared volume from the iMac on the MacBook Air, and ran Blackmagic Disk Speed Test. Results were extremely inconsistent (is this due to caching?), but I saw results like this for a file-shared volume over the Thunderbolt 2 network.

AFP:
Write: 150 MB/s​
Read: 621 MB/s​
SMB:
Write: 314 MB/s​
Read: 330 MB/s​
Here's what I get when using the iMac's internal flash drive directly:
Write: 2010 MB/s​
Read: 2286 MB/s​
Here's what I got over Thunderbolt 3 to a Samsung X5 SSD:
Write: 2033 MB/s​
Read: 2627 MB/s​
Here's what I got over Thunderbolt 2 to the Samsung X5 SSD:
Write: 1266 MB/s​
Read: 1293 MB/s​
Of course, Gigabit Ethernet is still worse than Thunderbolt networking:
Write: 75.6 MB/s​
Read: 71.6 MB/s​


#benchmarks
 


Ric Ford

MacInTouch
I just put a 2017 iMac 5K in Target Disk Mode and connected it over Thunderbolt 3 to a 2018 MacBook Pro. Blackmagic Disk Speed Test says:
Write: 895 MB/s​
Read: 468 MB/s​
That's an awful lot slower than the 40 Gbps (or even 22 Gbps) that Thunderbolt 3 should provide (i.e. ~3000 MB/s).
Unable so far to get Thunderbolt 3 file sharing working with the 2018 MacBook Pro (on the latest Mojave) ... I saw results like this for a file-shared volume over the Thunderbolt 2 network.

AFP:
Write: 150 MB/s​
Read: 621 MB/s​

SMB:
Write: 314 MB/s​
Read: 330 MB/s​
#benchmarks
OK, I did a clean install* of macOS Mojave onto an unencrypted volume on the 2018 MacBook Pro and was finally able to get Thunderbolt 3 networking to function, file-sharing from the 2017 iMac 5K.

Blackmagic Disk Speed Test had huge variability, but I got some results like these, varying from run to run:

AFP:
Write: 931 MB/s​
Read: 1086 MB/s​

SMB:
Write: 791 MB/s​
Read: 922 MB/s​



*actually a clone of the virgin Mojave system shipped on the 2017 iMac, ironically...
 


Ric Ford

MacInTouch
I just put a 2017 iMac 5K in Target Disk Mode and connected it over Thunderbolt 3 to a 2018 MacBook Pro. Blackmagic Disk Speed Test says:
Write: 895 MB/s​
Read: 468 MB/s​
That's an awful lot slower than the 40 Gbps (or even 22 Gbps) that Thunderbolt 3 should provide (i.e. ~3000 MB/s).
Time for Carbon Copy Cloner to clone a vanilla macOS Mojave partition from a 2018 MacBook Pro to an empty iMac 5K volume in Thunderbolt 3 Target Disk Mode:

29.2 GB in 5 min. 40 sec. (omitting subsequent startup cache update time at end) is about 86 MB/s if my math’s right.
 


Time for Carbon Copy Cloner to clone a vanilla macOS Mojave partition from 2018 MacBook Pro to iMac 5K in Thunderbolt 3 Target Disk Mode:
29.2 GB in 5 min. 40 sec. (omitting subsequent startup cache update time at end)
If the target drive was clean, the CCC (file copy) operations had essentially no seek delays and very little conversation between systems - [CCC] just keeps a stream going and going, one file after another. Doing a restore or migration most likely involves much additional overhead chatter. So, again, your results are not surprising.

Thanks for documenting all these test cases. I've seen many, but the desired end result was always good configurations, not shortest time, so I never made much note of the times involved, especially since, for the most part, they are better times than those achieved a decade or two ago.
 


Ever wonder why it is safe to connect an 85-Watt MacBook Pro power brick to an Air that needs only 45? Or how a 45-Watt MacBook Air charger might (eventually) top up a quiescent MacBook Pro?
Every power supply from every manufacturer works this way.

The spark from a short circuit is because a short circuit is, effectively, an infinite load. So it will drain everything the supply can produce (burning it out, if there is no current-limiting circuitry in place).

An 85W power supply may (and hopefully does) have a current-limiting circuit to prevent it from trying to deliver more than 85W - that will prevent it from overheating and burning out if it is overloaded (e.g. attaching it to a computer that's trying to draw 100W), but that would be completely unnecessary when attaching it to a lower-power device.

The power delivery protocols on USB are not designed to keep a power supply from forcing more current into a device than it can handle, they are designed to prevent the device from drawing more current than the cable can deliver without overheating.
 


... Connect no bulb, and no watts are delivered (can you imagine all those electrons spilling on the floor? ;)
James Thurber told this story:
My grandmother lived the latter years of her life in the horrible suspicion that electricity was dripping invisibly all over the house. It leaked, she contended, out of empty sockets if the wall switch had been left on. She would go around screwing in bulbs, and if they lighted up, she would fearfully turn off the wall switch and go back to her Pearson's or Everybody's, happy in the satisfaction that she had stopped not only a costly but dangerous leakage. Nothing could ever clear this up for her.
 


I have a 2017 MacBook Pro 15" that had the keyboard replaced when that recall went out. Because the top of the computer is one unit, it also includes the battery, so that was also replaced, but the USB-C/Thunderbolt assemblies, two on each side, were reattached to the motherboard and to the side of the computer with glue, I believe. They are not as solid as they once were, and 3 out of the 4 ports would fail regularly/intermittently.

I finally convinced Apple to take a look at it, even though we are out of warranty as I missed the 30-day cutoff to purchase Apple Care. A Genius person ran diagnostics and told me that the hardware was fine, the issue was in the software, and that the best solution would be to wipe the computer and replace everything manually - not to use a backup for migration - and then add each item in Documents back, and then reinstall all the software, one by one.

Frankly, I don't have the weeks that effort would take, and asked about modifying the preferences and extensions. He had little idea about that process but, with caution, thought it might work.

I went into the Libraries (all three) and removed items that had been around for more than a few years and were specific to peripherals and software that I no longer used. I was quite brutal about the culling, and even though I had to password-accept each deletion, I did remove several hundred older items.

Well, what do you know, the four ports are now working like a charm. I am also getting the speed back that I was missing in the file transfers, and things have been that much faster.

My advice is to take a look of the older preferences and delete them — that may be the cause of your slowdown. Worked for me.
 


Ric Ford

MacInTouch
... My advice is to take a look of the older preferences and delete them — that may be the cause of your slowdown. Worked for me.
That's a great story, but as far as my own test results are concerned, it's clearly irrelevant to the problem, since both the internal drive and all other external drives are working at top speed - the only thing that's slow is Target Disk Mode, and that's been true for many years and many tests on many different Macs and Mac operating systems.
 


Try this analogy: The AC outlet in a typical US home is a nominal 110 Volts AC - a "constant voltage source.”
  • Connect a 100 watt bulb and 100 watts is delivered (~0.91 Amperes).
  • Connect a 10 watt bulb and 10 watts is delivered (~0.091 Amperes).
No argument with that statement. But I have an inexpensive bedside swingarm lamp which is clearly labeled, "Caution: To Reduce the Risk of Fire Use MAX 40 Watt Type A." The lamp is dumb, and depends on humans using it to read and follow the instructions. Screw a 100-watt bulb into that lamp, and it will work, until it dangerously overheats.

On the electronics front, and by the kind of equipment, you'll understand this was some years ago, we were moving offices and our own maintenance crew plugged the power supply from a DSL modem into the input port for a then very expensive LinkSys WiFi router. Pffffft. The barrel plugs were identical, the power supplies looked the same....

The USB-C connector may just be a replacement plug for old-fashioned 500mA 5V 4.5W USB 2, or it can go all the way to USB 3.0 Power Delivery capability of 100 watts / 20 volts. "Power Delivery" is supposed to be "intelligent" with chips on each end that communicate to each other what level of power is safe and necessary. Bad programming, bad chips, bad cables, and it can go wrong.
Wikipedia said:
USB-C

... Many cables claiming to support USB-C are actually not compliant to the standard. Using these cables would have a potential consequence of damaging devices that they are connected to. There are reported cases of laptops being destroyed due to the use of non-compliant cables.

Some non-compliant cables ... (cause) a device connected to the cable to incorrectly determine the amount of power it is permitted to draw from the cable. ... Cables with this issue may not work properly with certain products, including Apple and Google products, and may even damage power sources such as chargers, hubs, or PC USB ports,
 


No argument with that statement. But I have an inexpensive bedside swingarm lamp which is clearly labeled, "Caution: To Reduce the Risk of Fire Use MAX 40 Watt Type A." The lamp is dumb, and depends on humans using it to read and follow the instructions. Screw a 100-watt bulb into that lamp, and it will work, until it dangerously overheats.

On the electronics front, and by the kind of equipment, you'll understand this was some years ago, we were moving offices and our own maintenance crew plugged the power supply from a DSL modem into the input port for a then very expensive LinkSys WiFi router. Pffffft. The barrel plugs were identical, the power supplies looked the same....

The USB-C connector may just be a replacement plug for old-fashioned 500mA 5V 4.5W USB 2, or it can go all the way to USB 3.0 Power Delivery capability of 100 watts / 20 volts. "Power Delivery" is supposed to be "intelligent" with chips on each end that communicate to each other what level of power is safe and necessary. Bad programming, bad chips, bad cables, and it can go wrong.
Or, to paraphrase all that ... using inappropriate or outside-of-specification equipment can be hazardous.

Since humans live by "If it fits, it is OK", hazards abound.
 


Just ran into the Mac Mini 2018 (6-core i7) problem where an unpowered device or hub (AmazonBasics USB 3.1 Type-C to 4-Port Aluminum Hub) randomly disconnects.

I was attempting to transfer a large number of multi-gigabyte files from one USB 3 hard drive to another. I got disconnected part way through 3 times... I then connected the drives directly to the Mini and the copy process had no issues.

Wasn't this a known issue that was to be fixed by Apple with an update ? (I'm running Mojave 10.14.5.)
 


Ric Ford

MacInTouch
Just ran into the Mac Mini 2018 (6-core i7) problem where an unpowered device or hub
(AmazonBasics USB 3.1 Type-C to 4-Port Aluminum Hub) randomly disconnects.
...
Wasn't this a known issue that was to be fixed by Apple with an update ? (I'm running Mojave 10.14.5.)
There were some USB bug fixes, but they may be different problems:
Apple said:
What's new in the updates for macOS Mojave

macOS 10.14.4
  • Improves the reliability of USB audio devices when used with MacBook Air, MacBook Pro, and Mac mini models introduced in 2018.
macOS Mojave 10.14.2 Update
  • Resolves an issue that prevents displays from working when connected to MacBook Pro models introduced in 2018, if certain third-party USB graphics devices are also connected.
More FYI about hubs
Apple said:
Using USB devices with your Mac
Hubs that conform to what is commonly referred to as the "2008 version 1 USB 3.0 specification" are supported.
Another potential problem is electrical interference, e.g. between cables or cables/hub or cables/drive/hub, etc. Changing cables and spacing and orientation may help. The high clock speed of USB 3 seems to create far more of these problems vs. slow USB 2.

Also... I'd be very reluctant to use an unpowered USB hub to host storage devices—that seems like it might be asking for trouble, especially with multiple devices attached.
 


Just ran into the Mac Mini 2018 (6-core i7) problem where an unpowered device or hub
(AmazonBasics USB 3.1 Type-C to 4-Port Aluminum Hub) randomly disconnects....
It's a lack of sufficient power. I don't recommend connecting bus-powered hard drives to a hub, even if it's a powered hub. I always connect them directly to the Mac.
 


It's a lack of sufficient power. I don't recommend connecting bus-powered hard drives to a hub, even if it's a powered hub. I always connect them directly to the Mac.
I, too, have had problems with unpowered or even powered USB hubs in the past.

Since I started using this Anker powered hub this year I’ve had no disconnects or any problems.

I attach up to 5 bus-powered USB 3 drives without any problems.

Bought on Amazon through Ric’s MacInTouch Amazon link, of course.
 


FWIW, I used to have disconnects of backup USB drives with a powered USB hub when I was using it with my 2010 Mac Mini running on an external (FireWire) drive but have not had them since upgrading to a 2018 Mac Mini last month and moving the USB drive to it (with the 2018 Mini running from its internal solid-state drive). So insufficient power sounds plausible to me.
 



It's a lack of sufficient power. I don't recommend connecting bus-powered hard drives to a hub, even if it's a powered hub. I always connect them directly to the Mac.
Actually, both drives were powered, and I even tried it with only one drive hub-attached. My conclusion is that unpowered hubs are only useful for keyboards, mice, printers, etc. I didn't think that the power requirements for an unpowered hub to do its hub-stuff would be overrun when using powered devices. Lesson learned. (I have a powered hub on order, report to follow.) By the way, the hub was attached via USB C thru a Thunderbolt 3 port.
 


I wish I had my notes from November 2018 when I migrated one additional account from an older MacBook Pro with Thunderbolt 2 using the Apple adapter to Thunderbolt 3 for a brand new 2016 MacBook Pro. It was one of the slowest migrations that I have recently done. So the performance numbers you obtained using Thunderbolt between Apple machines are not a surprise.
This began as an issue [for us] as soon as the T2 chip laptops came out. We had been migrating users to new machines via Thunderbolt Target Disk Mode for ages. It went at reasonable speed, and we generally had a migration done in an hour (with SSD drives on both sides).

Once the new T2 machines came out, that Target Disk Mode speed dropped [drastically]... coming from a Thunderbolt 2 to Thunderbolt 3 machine. Migrations that used to take an hour went to 12 hours (!).

We finally abandoned that workflow and now simply clone the old machine to an external SSD, and then migrate to the new machine from that. Back to 1 hour.... but with the added initial clone time.

If anyone has any insight into what happened or otherwise how to resolve without that work-around, I'd love to hear.
 


This began as an issue [for us] as soon as the T2 chip laptops came out. We had been migrating users to new machines via Thunderbolt Target Disk Mode for ages. It went at reasonable speed, and we generally had a migration done in an hour (with SSD drives on both sides). Once the new T2 machines came out, that Target Disk Mode speed dropped [drastically]... coming from a Thunderbolt 2 to Thunderbolt 3 machine. Migrations that used to take an hour went to 12 hours (!). We finally abandoned that workflow and now simply clone the old machine to an external SSD, and then migrate to the new machine from that. Back to 1 hour.... but with the added initial clone time. If anyone has any insight into what happened or otherwise how to resolve without that work-around, I'd love to hear.
I don't have any technical explanation for the T2 Target Disk Mode slowdown. (I can speculate that performance was not considered an essential goal.)

Since I insist that everyone I assist must implement Time Machine backups and perform regular Carbon Copy Cloner full bootable backups, it has since been easy to forgo Target Disk mode and just use Migration Assistant from either type of backup. Since Carbon Copy Cloner only copies changed files, the pre-migration cloning is not particularly costly in time. Users capable of following a checklist procedure can initiate the required backup.

For a complex upgrade — for example, replacing an old Mac Mini with a Mojave-capable version and upgrading multiple applications in the same process — it is well worthwhile to verify that the Time Machine backup is current (using TheTimeMachineMechanic) and to repeat the Carbon Copy Cloner clone backup.

The waiting period is useful to review the list of planned changes with the user. Some changes are not part of macOS or application upgrades. Where fixed IPv4 addresses are used, the DHCP reservations have to be upgraded in the 'router'. This review and external changes can easily cover the backup interval.
 


Ric Ford

MacInTouch
In case anyone wondered about this, Thunderbolt 3 cables apparently work for USB, too. I just plugged in a Samsung T5 to the USB-C/Thunderbolt 3 port of a 2017 iMac, using a Thunderbolt 3 cable, and it all worked, even though the Samsung T5 is a USB device, not a Thunderbolt device.
 


In case anyone wondered about this, Thunderbolt 3 cables apparently work for USB, too.
Sometimes the obvious hides in plain sight?

For a lengthy exegesis:
The Broadcast Bridge said:
Confusion Still Reigns Over USB-C/Thunderbolt Cables
It is hard to believe the vast amount of confusion that continues over cables with USB-C/Thunderbolt connectors. These cables may be the future of computing, but right now they are a big, sloppy confusing mess. The problem is cables that handle USB-C, Thunderbolt 3 and USB 3.1 all have the same connector and look alike but do different things.
Less is sometimes more, and this does seem in agreement.
Paul Grimshaw on Quora said:
What is the difference between a Thunderbolt 3 vs USB-C cable?
USB-C is a connector, not a protocol in its own right ... Passive Thunderbolt 3 cables support USB 3.1 Gen 2 over USB-C, but a USB-C cable does not necessarily support Thunderbolt 3, even if it supports USB 3.1 Gen 2. To get the full 40Gb/sec you either need a copper cable 0.5m long or less, or an active or optical longer cable....
 


I have a question involving FireWire. A few years back (~2011?) there was a spirited discussion about transferring LP's to digital format here on MacInTouch. At that time, FireWire devices were the dominant hardware for prosumer audio: Echo Audofire 2 or 4, Focusrite Saffire (various models), Behringer FCA202, etc.

These devices are still available used on eBay for prices ranging from $30 to $90, and my question is, are they still supported in Mojave? I know that Echo is out of the audio production business, so their last update was for Yosemite. On the other hand, I've always heard that FireWire devices from the past somehow magically still work without custom drivers. Still true?

I contacted Sweetwater, and their recommendation was to go with a Focusrite Scarlett 2i2 3rd-gen USB unit; has anyone used one of these under recent macOS releases? My main requirement is to get 24-bit input, so that I can scale my waveforms properly before converting to 16-bit CD audio. In the past I've had a lot of fiddling on each and every LP to adjust the gain so that signal is maximized yet not clipping.
 



Amazon disclaimer:
As an Amazon Associate I earn from qualifying purchases.

Latest posts