Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
Part 1 .

Here's the method and bill of materials necessary to nicely upgrade a Mac Pro 5,1 ( 2009-12 ) with a high performance wireless upgrade system . The use of external antennas greatly enhances signal strength and has the advantage of providing semi-directional control for fine tuning a connection . This upgrade provides support for Continuity features .


IMG_0578.jpg


IMG_0606.jpg


Here is the list of parts . Included is the delivered cost in the USA from eBay , amazon , etc .

1 ) Mac Pro . In this case a Mac Pro 4,1 > 5,1 .
2 ) BCM94360CS / BCM94360CSAX model 802.11ac and Bluetooth 4.0 Apple branded Wireless Card . $10 eBay .
3 ) Shaluoman Mini PCI-E WiFi Card Adapter model DZ002 . $13 Amazon .
4 ) Factory Bluetooth 2.1 card's proprietary cable . $00 .
5 ) 3 x Data Alliance brand 20 inch long MHF4 To RP-SMA-female Cable , model MHF4RSf20i . $12 Data Alliance .
6 ) 3 x 5 dBi Hi Gain wireless RP-SMA antennas ( black stubbies ) . $7 eBay .
7 ) A three hole RP-SMA I/O PCIe Slot Bracket . $4 eBay ( shipped from China )
8 ) Some 4 and 10 mm diameter pieces of heat shrink tubing . Nominal cost .
9 ) A length of 1/2 inch diameter PET cable sleeving . Nominal cost .
10 ) Some pieces of 3M brand 467MP adhesive transfer tape . Nominal cost .

Total current cost of upgrade materials for USA users is $46 .

List of tools :

1 ) Screwdrivers , Phillips 0 or Phillips 1 , depending on your choice of wireless card fasteners .
2 ) Micro wire stripper , 22-38 AWG class . I recommend an Engineer brand model PA-14 .
3 ) A deep well 8 mm hex socket T-handle nut driver , I recommend Wiha brand model 336/SW 8.0 x 125 mm ( 33606 ) .
4 ) A 8mm ( or 5/16 inch ) open ended wrench .
5 ) Liquid Electrical Tape with applicator brush , I recommend Star Brite brand .
6 ) Cigarette lighter or hair drier .
7 ) One small awl .

-----------------------------------------------------------------------------------------------------------------------------------------------------------

Our first task to make the custom Bluetooth data cable . This will eventually be attached to the factory's Bluetooth header on the backplane board to the WiFi Card Adapter .

Remove the factory bluetooth 2.1 card from the backplane board . It is located in the bottom right of the backplane board . Retain the bluetooth cable and dispose of the card , as we won't be needing it . The bluetooth cable looks like this , from both sides :

Untitled2.png
Untitled3.png


Taking a wire cutter , cut off the terminal that has gold teeth on both sides until it looks like the image on the right :

Untitled 4.png
IMG_0522 copy.jpg


I also take this opportunity to remove unnecessary wires from the cable . Using an awl , gently push out from the terminal the unnecessary wires . Dispose of the loose unnecessary wires . Now take the bluetooth data cable from the Shaluoman Card Adapter package , it's quite long compared to our modified factory bluetooth cable . If the Shaluoman cable has a large black terminal , cut off that terminal . Retain the small white terminal , as it will eventually attach to the adapter . The cable we want should be in the form on the image in the center . A close up of the Shaluoman cable end is shown in the image on the right .

IMG_0527 copy.jpg
IMG_0531.jpg
IMG_0532.jpg
 

Attachments

  • IMG_0522 copy.jpg
    IMG_0522 copy.jpg
    31.8 KB · Views: 423
Last edited:

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
Part 2 .

Now , we must take our wire stripper and cut away the insulation from the cable ends to prepare for the splicing . All the wires are 30 AWG .

Make certain the factory bluetooth cable terminal data signal type matches the Shaluoman adapter card cable data signal type . For the Shaluoman cable , the USB_D- signal is the black wire and the USB_D+ is the red wire .
IMG_0533 copy.jpg


Once stripped , connect the wires ends by twisting them together .

IMG_0535.jpg


And then apply Liquid Electrical Tape to provide decent insulation . You may have to apply several coats .

IMG_0540.jpg


Once the liquid dries , apply a 4 mm heat shrink tube over the splice and shrink it with the lighter . Once cooled , tug on both ends the tube to verify it was applied properly . Here's what the Shaluoman data cable looks like when finished . Quite a professional looking proprietary cable !

IMG_0541.jpg


Now that our BT data cable has been made , let's attach the Apple wireless card to the Shaluoman adapter . Here's what both sides of the card and adapter look like :

IMG_0552.jpg


IMG_0551.jpg


J0 is a WiFi antenna connector .
J1 is a Bluetooth antenna connector .
J2 is a WiFi antenna connector .

The attachment is self evident , as it can only be installed one way . Here's what the Apple wireless card attached to the adapter and installed in the factory WiFi interface looks like . The WiFi slot is located in the left bottom part of the backplane board . I secured Apple's wireless card to the Shaluoman adapter using 3M adhesive transfer tape . I build my rigs to be shipped and used in mobile environments and that tape is not ever going to come loose - this type of tape is rated to retain skyscraper windows in place , it's that strong . The custom bluetooth data cable is now connected to the adapter . The connection is keyed and can only be inserted one way , but for accuracy the USB_D- is on the left ( black wire ) and the USB-D+ is on the right ( red wire ) as shown below .

IMG_0567 copy.jpg


Now that our wireless upgrade card is installed , let's attach the other end of our custom bluetooth data cable to Apple's factory Bluetooth connector . The connector is keyed and can only be installed one way , but for accuracy the USB_D- wire is on top and the USB_D+ wire is on the bottom .

IMG_0560 copy.jpg
 

Attachments

  • IMG_0551.jpg
    IMG_0551.jpg
    622.8 KB · Views: 304
  • IMG_0567.jpg
    IMG_0567.jpg
    296.9 KB · Views: 302
Last edited:

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
Part 3 .

Next , we attach some additional pieces of 3M transfer tape to secure our custom bluetooth data cable onto the backplane board . This is good cable management and prevents this cable from , say , getting in between the delicate interconnect connections when the CPU Tray is inserted into the Mac . It's a good build practice to avoid placing tape on any exposed surface mounted components ( e.g. resistors , etc . ) , to help avoid any potential electrical shorting . Placing the tape on fiberglass surface is safe .

And we also attach at this time our three MHF4 To RP-SMA-female antenna cables to the wireless card . The terminal ends of these cables are heavy and chunky , so be careful not to let them fall on anything delicate . I chose to install the wireless upgrade project with the backplane removed from the chassis , as it makes things easier to install . The board is placed on a rubber antistatic mat that provides a safe and pretty soft surface to work on . Our progress thus far looks like this :

IMG_0578.jpg


Here's what the terminal looks like on one of the RP-SMA-female cables , in case you are not familiar . It has a washer and an 8mm retaining hex nut .

IMG_0584.jpg


Attach the 1/2 inch cable sleeving around the three long antenna wires and secure with 10 mm heat shrink tubing near both ends . Use your trusty lighter again on the heat shrink tubing and tug on the cooled connections to verify they are solid . Make certain you allow the individual ends of the antennas plenty of freedom , so don't attach the sleeve all the way to the wire ends . A picture of this installed sleeving is seen in the picture of the upgraded wireless board installed in the Mac Pro below . You probably should work with heat shrink tubing away from anything delicate due to the issue of heat being used and possibly some materials dripping away as they melt from using the lighter .

Install the upgraded wireless backplane board back into your Mac Pro , if it is not already installed . It looks like this now . OK , the wireless antennas are already shown installed in the PCIe I/O bracket and we haven't done that yet ...

IMG_0593.jpg


Here is a close up of the cable management of the upgrade once the backplane board is reinstalled into the Mac Pro . When you reinstall the Processor Cage back into the Mac Pro , remember to keep all cables from the fastener points ( screw holes with a white square printed on the PCB ) or something could get pinched . As I believe the center antenna connector ( J1 ) of the Apple wireless card is for bluetooth , I colored both ends of this MHF4 To RP-SMA-female antenna cable with blue ink . This is so the center external Black Stubby antenna can be designated the bluetooth antenna in case the user needs to know what to adjust for improved reception . The other two being used for WiFi .

IMG_0594.jpg


Install the three hole RP-SMA I/O PCIe Slot Bracket in slot 4 of the Mac Pro .

Attach the three antenna terminals onto this bracket the way shown in the pictures below . The nut goes on the outside ( outside of Mac ) of the bracket while the washer goes on the inside ( inside of the Mac ) . Remember to install the bluetooth antenna in the center hole ( you did remember to color both ends of the BT antenna cable terminals with ink , right , to differentiate from the other two WiFi antenna cables ) ? Use the 8mm ( 5/16th inch ) open ended wrench to hold the inside nut while using the deep well 8 mm hex socket T-handle nut driver to tighten the outside nut . You want to do a really tight job as users often adjust the black stubby antennas and you don't want anything coming loose .

IMG_0608.jpg


IMG_0607.jpg


Here's what it should look like now from the exterior . The factory that made these brackets didn't align the holes as well as they could ...

IMG_0596.jpg


Now attach your black stubby RP-SMA wireless antennas and you are good to go !
 
Last edited:

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
Here are some screen grabs of the wireless portion of the System Report . We now have a WiFi signal to noise ratio ( SNR ) of 34 with the external antennas . The SNR alone doesn't do justice to the upgrade as the external antennas can punch through obstructions better than the stock internal antennas , through obstructions like the copper pipes and hardwood floors I have in my shop ) . Using the stock factory WiFi card and the internal wireless antennas we would obtain a SNR of around 19 to 20 . Higher is better and 20 is considered the bare minimum for an usable signal . I have no means of testing Bluetooth strength , except subjectively . All the Continuity features are activated .

Screen Shot 2020-04-26 at 11.27.20 AM.png

Screen Shot 2020-04-26 at 11.35.52 AM.png


Comments are now welcome and thank you for your patience .
 
Last edited:

hwojtek

macrumors 68020
Jan 26, 2008
2,274
1,276
Poznan, Poland
I have it made exactly the same way, great writeup.
A couple of remarks:
- the card you've used is the "lesser" of Apple's AC cards (tops out at 867Mbit IIRC) with three antenna terminals in MHF4 connector standard. The "holy Grail" is the card with four antenna terminals (3x wifi, 1x BT) in UFL standard.
- the MHF4 cable connectors are extremely hard to fix onto the card terminals, make sure you are someone with 20/20 vision or a huge magnifying glass and a very steady hand ;)
- if using the 4-port versions, the existing UFL antenna connectors may be reused, with an exception of BT - an extension cable is needed, as the original BT cable doesn't reach the new card.
- external antennas are a much more smart idea though, genuine Mac Pro antennas are not exceptional.
- if you want to procure your own PCI bracket for antennas, you need to bore/stamp three holes 6.75 mm (0.265 inch) in diameter for the big RP-SMA connectors.
 

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
"- the card you've used is the "lesser" of Apple's AC cards (tops out at 867Mbit IIRC) with three antenna terminals in MHF4 connector standard. The "holy Grail" is the card with four antenna terminals (3x wifi, 1x BT) in UFL standard."

I know . I was fortunate to obtain an 2017 iMac wireless card for just $30 last year and installed it in a Mac Pro 5,1 . The card is model BCM943602CDPAX_2 . It was a dream to install using the cMP's factory internal u.FL antenna connectors instead of messing around with MHF4 connectors so small they'd break if you looked at them the wrong way . The card was incredibly "snappy" in a cMP , too . It's subjective , but it really surprised me how instantaneous everything was . I also obtained a SNR of around 35 using the factory internal antenna system with this card . I can only imagine how much nicer it would have been with external antennas .

Here is a picture of the card I used . Evidently , Apple glued it in rather firmly and the previous owner used his teeth for removal or something ... I'm shocked it actually worked as it looked like it survive a major combat action .

1.jpg

2.jpg




"- the MHF4 cable connectors are extremely hard to fix onto the card terminals, make sure you are someone with 20/20 vision or a huge magnifying glass and a very steady hand ;)"

My eyesight is not exactly 20/20 , but I do have access to some nice jeweler's loops and if those can't do the job , an USB microscope . Truth be told , it was not that hard to install MHF4 antennas to the card in this instructional because the backplane board was not installed in the chassis . Once that card is inside the chassis , it is very hard to attach these type of antennas . I suspect the engineers that developed the MHF4 standard played around with just the card and antennas on a bench all day , attaching and disconnecting cables all day and thinking , "weee , it's easy !" Of course , they didn't actually try connecting anything in a real world environment ( like something buried in a chassis ) , or reality would have hit them in the head like a bag of bricks .

"- if you want to procure your own PCI bracket for antennas, you need to bore/stamp three holes 6.75 mm (0.265 inch) in diameter for the big RP-SMA connectors."

Thanks ... that's good to know .
[automerge]1587925366[/automerge]
I have it made exactly the same way, great writeup.
A couple of remarks:
- the card you've used is the "lesser" of Apple's AC cards (tops out at 867Mbit IIRC) with three antenna terminals in MHF4 connector standard. The "holy Grail" is the card with four antenna terminals (3x wifi, 1x BT) in UFL standard.
- the MHF4 cable connectors are extremely hard to fix onto the card terminals, make sure you are someone with 20/20 vision or a huge magnifying glass and a very steady hand ;)
- if using the 4-port versions, the existing UFL antenna connectors may be reused, with an exception of BT - an extension cable is needed, as the original BT cable doesn't reach the new card.
- external antennas are a much more smart idea though, genuine Mac Pro antennas are not exceptional.
- if you want to procure your own PCI bracket for antennas, you need to bore/stamp three holes 6.75 mm (0.265 inch) in diameter for the big RP-SMA connectors.


By the way , can you verify which of the connectors is the bluetooth one , for the BCM94360CSAX card I used in this instructional ? I really should use a 2 dBi antenna with that connection .
 
  • Like
Reactions: Meester X

hwojtek

macrumors 68020
Jan 26, 2008
2,274
1,276
Poznan, Poland
I believe the bottom one (when looking on an opened Mac Pro standing vertically) is the BT. I will double check by the end of the week, because I need to adjust a couple of things in my setup so it's gonna be dismantling time (again).
 

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
I believe the bottom one (when looking on an opened Mac Pro standing vertically) is the BT. I will double check by the end of the week, because I need to adjust a couple of things in my setup so it's gonna be dismantling time (again).

I'd appreciate that . And if you could , please tell me which connector by name ( J0 , J1 , J2 ) is the BT .
 

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
Try J1. Easy to test anyway. Unscrew the external J1 antenna, if BT stop working (signal strength can be monitored by BT Explorer), then it's the BT antenna.

I just installed a copy of the Apple written Bluetooth Explorer found in the additional tools for Xcode folder for the particular version of Xcode I am using onto this System .

To everyone reading this , there is also a third party paid written program called Bluetooth Explorer available at Apple's App Store . This should not be used for Bluetooth diagnostics . Why Apple is tolerating this is unbelievable .

Inspired by Apple's non enforcement of their own IP ( let that sink in ) , I am considering developing an alternative macOS called Snow Catalina . I don't know how to code , but then I guess that other guy can't either .

Anyways , I learned more about and performed some more tests with this wireless card upgrade .

The BCM94360CSAX model wireless card utilizes a model BCM20702 Bluetooth chip from Broadcom , which uses Class 1 Part A technology . This transmits only half ( 10 dBm ) the power of a true Class 1 ( max 20 dBm ) device .

This makes an interesting situation for upgrading the wireless in our cMPs . Should we use true Class 1 Bluetooth USB dongles , instead ? No one makes a 20 dBm class USB dongle for macOS Systems to my knowledge .

The highest performing USB BT dongle I can find for Macs is the Bluetooth 4.0 Class 1 Startech USB device , which uses a Qualcomm CSR8510A10 chip that is rated for just under 10 dBm . So it seems as if the Startech device also acts like a Class 1 Part A device . It's not a bad dongle , BTW and it supports Continuity features .

With Apple's Bluetooth Explorer using a negative scale , I discovered that this Mac transmitting to my stereo's Elsinkin BT audio receiver that the :

BT RSSI is 0 line of sight at a distance of 20 feet using a 5 dBi antenna , which is great .

BT RSSI is -5 through a wall of eight inches of hardened plaster and steel wire mesh using a 5 dBi antenna , which is also great .

The Raw RSSI of the BT Esinkin audio receiver device used for these tests is -66 .

I'm not certain I know the difference between Raw RSSI and straight RSSI .

Screen Shot 2020-04-30 at 1.51.31 PM.png


On the WiFi side , things improved when I reattached the antennas on both ends of the cable . We are up to a SNR of around 40 and a transmission rate of 1300 Mbps . That's good enough for professional video conferencing . Unfortunately , the cable service I have with my local ISP maxes out at around 250 Mbps . The test was at a distance of only ten feet from the Wireless Router , but the signal had to go through eight inches of hardened plaster and four layers of steel wire mesh wall . This was deliberate in order to provide a serious obstruction to test signal strength .

Screen Shot 2020-04-30 at 9.59.47 AM.png


I also ran some WiFi bandwidth tests with the latest browser versions for the latest version of Mojave in this cMP and discovered that Apple is up to its old optimization tricks again . By optimization , I mean they want to encourage you to buy a new Apple computer because it seems there is insufficient performance running Safari with our cMPs .

Screen Shot 2020-04-30 at 10.19.05 AM.png


Screen Shot 2020-04-30 at 10.23.06 AM.png


Safari v 13.1 had a downstream of 123 Mbps and an upstream of 13 . Latency was 17 ms .
Firefox v 75 had a downstream of 243 Mbps and an upstream of 12 . Latency was 21 ms .

Firefox more closely matched the maximum bandwidth I am provided by my local ISP ( 250 Mbps downstream ) and should be used as the browser of choice with this System configuration .

Safari was only half as fast .

I like Apple workstations and macOS . I am not really a big fan of Apple apps .
 

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
I am not a Safari fan. My primary browser is Brave, but Safari's speed is fine.
Screenshot 2020-05-01 at 3.44.22 AM.png


I think this result should be good enough to show "Safari (same version as yours) can fully utilise a 1000Mbps network".
 

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
What wireless card are you using ?
You can click the link in my signature.

Anyway, that speed isn't from Wifi, but LAN.

For browser speed testing, it's better to use a network have stable performance. Wifi's performance can fluctuate a lot, not a good tool to test browser speed.
 

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
You can click the link in my signature.

Anyway, that speed isn't from Wifi, but LAN.

For browser speed testing, it's better to use a network have stable performance. Wifi's performance can fluctuate a lot, not a good tool to test browser speed.

This entire thread is about wireless networking , so your results are invalid .

I know what the factory 1 GbE performance would be with this machine .

The cMP being discussed has the BCM94360CSAX wireless card installed in the factory WiFi interface transmitting to an Apple 802.11ac Base Station with macOS Mojave build 18G4032 . I tested the WiFi with the very latest versions of Safari and Firefox . Safari is obviously not optimized for this .

Since Apple is the only company that writes 802.11ac drivers , driver support is crucial for these older machines .

Sometimes I'll try out a third party USB 802.11ac WiFi dongle with drivers provided by the manufacturer . Even installing what they call "ac" drivers only provides an "n" connection . And even if we were to obtain "ac" performance , Continuity feature support requires an Apple branded card in the appropriate interface .
 
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
This entire thread is about wireless networking , so your results are invalid .

I know what the factory 1 GbE performance would be with this machine .

The cMP being discussed has the BCM94360CSAX wireless card installed in the factory WiFi interface transmitting to an Apple 802.11ac Base Station with macOS Mojave build 18G4032 . I tested the WiFi with the very latest versions of Safari and Firefox . Safari is obviously not optimized for this .

Since Apple is the only company that writes 802.11ac drivers , driver support is crucial for these older machines .

Sometimes I'll try out a third party USB 802.11ac WiFi dongle with drivers provided by the manufacturer . Even installing what they call "ac" drivers only provides an "n" connection . And even if we were to obtain "ac" performance , Continuity feature support requires an Apple branded card in the appropriate interface .
My result about "Safari is fine" of course is valid. That's about "browser speed", has nothing about wireless or not.

If you want to compare browser speed, you should use a network that can produce stable result. Otherwise, you have no idea that you are measuring the browser speed, or the wifi speed at that moment.

Or if you want to know "Safari can only perform well with cable network, but not up to full speed with wifi", then it's another area to explorer. And this will require to use Wifi for testing.
[automerge]1588282215[/automerge]
Anyway, Safari is fine for Wifi, same network speed as other browser.
Screenshot 2020-05-01 at 5.27.10 AM.png


I use Brave as control
Screenshot 2020-05-01 at 5.29.02 AM.png
 
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
As I said, Wifi performance is too unstable if you want to measure the browser's speed. It's very pointless to compare Wifi network speed between different browsers. Too many variables at the same time.

Same computer, same network, and the wifi speed suddenly dropped during the upload test. Should I blame the browser?
Screenshot 2020-05-01 at 5.25.19 AM.png


Of course not, because clearly it's not the browser causing this speed drop.
[automerge]1588283289[/automerge]
I also ran some WiFi bandwidth tests with the latest browser versions for the latest version of Mojave in this cMP and discovered that Apple is up to its old optimization tricks again . By optimization , I mean they want to encourage you to buy a new Apple computer because it seems there is insufficient performance running Safari with our cMPs .
To make it clear, my reply was mainly for this paragraph.

Obviously that doesn't fit my results.
 
Last edited:

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
OK , I ran some wired and wireless tests with both Safari and Firefox with this upgraded cMP .

I ran them a total of three times each and got basically the same results .

I also alternated between the browser test runs just to be conservative ( ran safari , then firefox ,then safari , etc . I also cleared the caches each test ) .

Here are the results :

Safari ( "ac" wireless )

safari wireless.png


Safari ( factory Gigabit Ethernet )

safari wired.png


Firefox ( "ac" wireless )

firefox wireless.png


Firefox ( factory Gigabit Ethernet )

firefox wired.png


Safari is clearly crippled with both wired and wireless .

Firefox is behaving as expected .

This is a clean install of Mojave - installed within the last week so it's still high performance . Safari of course , as included in the OS , is a clean install as well .
[automerge]1588287036[/automerge]
As I said, Wifi performance is too unstable if you want to measure the browser's speed. It's very pointless to compare Wifi network speed between different browsers. Too many variables at the same time.

Same computer, same network, and the wifi speed suddenly dropped during the upload test. Should I blame the browser?
View attachment 910981

Of course not, because clearly it's not the browser causing this speed drop.
[automerge]1588283289[/automerge]

To make it clear, my reply was mainly for this paragraph.

Obviously that doesn't fit my results.

I've installed a number of these "ac" wireless cards in the cMP and when I saw the poor results in Safari , I went into a cold sweat . I thought I screwed up the upgraded wireless System install somehow ( Maybe there was some EMI involved from another component , like an USB Card , which has been known to happen . It's why you see copper insulation tape installed in some Apple products , to reduce EMI ) . Nope . The wireless System upgrade is fine . Safari is at fault and Apple is responsible for writing all the code .
 
Last edited:
  • Like
Reactions: Nguyen Duc Hieu

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
OK , I ran some wired and wireless tests with both Safari and Firefox with this upgraded cMP .

I ran them a total of three times each and got basically the same results .

I also alternated between the browser test runs just to be conservative ( ran safari , then firefox ,then safari , etc . I also cleared the caches each test ) .

Here are the results :

Safari ( "ac" wireless )

View attachment 910996

Safari ( factory Gigabit Ethernet )

View attachment 910997

Firefox ( "ac" wireless )

View attachment 910998

Firefox ( factory Gigabit Ethernet )

View attachment 910999

Safari is clearly crippled with both wired and wireless .

Firefox is behaving as expected .

This is a clean install of Mojave - installed within the last week so it's still high performance . Safari of course , as included in the OS , is a clean install as well .
[automerge]1588287036[/automerge]


I've installed a number of these "ac" wireless cards in the cMP and when I saw the poor results in Safari , I went into a cold sweat . I thought I screwed up the upgraded wireless System install somehow ( Maybe there was some EMI involved from another component , like an USB Card , which has been known to happen . It's why you see copper insulation tape installed in some Apple products , to reduce EMI ) . Nope . The wireless System upgrade is fine . Safari is at fault and Apple is responsible for writing all the code .
So you 100% ignore my result and conclude that Safari crippled the network performance?

How about your test site's coding doesn't work well with Safari?

My test results obviously busted your theory already.
[automerge]1588313143[/automerge]
Just did some tests with speedof.me. I really suggest you go for a better speed test site.
Screenshot 2020-05-01 at 2.00.40 PM.png


Speedtest.net already provided that both Safari and Brave are fine for LAN / Wifi. But speedof.me show very inconsistent result. Therefore, I can only assume it's speedof.me unable to check network speed properly.

Anyway, Safari can reach 943.68Mbps, obviously not crippled.
 
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,628
8,557
Hong Kong
Just in case you still argue about Safari crippled only wifi. Here is the test result
Screenshot 2020-05-01 at 2.08.47 PM.png


Safari is fine. It's speedof.me produce erroneous result on different browsers. (N.B. Brave + wifi works perfectly in sppedtest.net. So both Brave and wifi is proved fine.)
[automerge]1588313796[/automerge]
Just in case you wart to know. There is no magic in Firefox. This is the LAN result, pretty much same as Brave. It's speedof.me can't measure the network properly.
Screenshot 2020-05-01 at 2.14.09 PM.png

[automerge]1588314066[/automerge]
And this is Firefox with wifi
Screenshot 2020-05-01 at 2.18.24 PM.png


Should I conclude that Apple only optimised the network (wifi-ac) performance for Safari? Obviously not.
 
Last edited:

hwojtek

macrumors 68020
Jan 26, 2008
2,274
1,276
Poznan, Poland
I believe the only way to accurately measure the impact of various browsers on the transmission speed would be to test on a connection to/from a local server located on the same segment of the network, connected with a wired 1Gbit (preferably 10G) Ethernet. This way you measure wifi, and not the ISP performance.
If I wanted to test the wifi throughput, I'd use iPerf anyway.
 

Snow Tiger

macrumors 6502a
Original poster
Dec 18, 2019
854
634
I believe the only way to accurately measure the impact of various browsers on the transmission speed would be to test on a connection to/from a local server located on the same segment of the network, connected with a wired 1Gbit (preferably 10G) Ethernet. This way you measure wifi, and not the ISP performance.
If I wanted to test the wifi throughput, I'd use iPerf anyway.

That's an excellent suggestion . I've used iPerf in the past to verify bandwidth between 10 GbE PCIe adapters .

But for years , I've used Speed Of Me as a quick means to verify bandwidth of wireless and factory ethernet interfaces in these cMPs . When the hardware is working properly , the results are consistent . Now , suddenly , this method is failing with Safari .

Speed Of Me is considered reliable by ISP field techs in my area , even over in house solutions .
 
  • Like
Reactions: Meester X

hwojtek

macrumors 68020
Jan 26, 2008
2,274
1,276
Poznan, Poland
Still, using an external service, while shows the "real-world" performance does not really help in tuning the wifi, because it introduces a lot of unnecessary variables. For example:
- my Ubiquiti USG router does not route >300 Mbit/s in software mode. I've spent a couple of hours investigating why even a wired client does not iperf to my ISP @1 Gbit/s as it should. Switch to hardware routing and it's 1 Gbps.
- my access points (again, Ubiquiti's AC Lite and LR) top out @867 Mbit/s by design, so I wouldn't even bother with checking iperf to my ISP.
- Speedof.me is wildly inconsistent for me and I would not consider it reliable enough (then again, I'm used to speedtest.net). The nearest speedof.me server is located about 180 miles from me, usually I get real transfer speeds from other services located on this server farm closing to 650 Mbit/s. Two tests on speedof.me and one shows 200 Mbit/s, the other 400. Now to make things even funnier, a speedtest.net server in Sweden, twice as far, shows close to 800 Mbit/s on Ethernet. And yet it's only about my satisfaction, since it says exactly nothing about my wifi.
This is why I test whatever I may control within my environment, and tune the ISP performance afterwards (not that there is a lot of things to tune there, but sometimes dropped packets may happen due to cabling, misconfigurations etc, so it's worth checking and reporting to the ISP).
 
  • Like
Reactions: Meester X

Meester X

macrumors newbie
Jul 18, 2020
2
0
Hi All,

I just came across this incredibly helpful thread. I wish I'd found it earlier as I soldered my bluetooth USB header wires to the Mini PCI Adapter and every write-up I'd seen used red for negative and black for positive and it really bugged me but I did it that way anyway to avoid future confusion.


I had a question (and please forgive me) I reused a photo from hwojtek - please forgive me. I just wanted to double check - I have pretty much the same card and wanted to confirm this is the correct way of wiring it. Wifi Performance is OK. Bluetooth performance is awful but I'll try an external antenna tomorrow.

Many thanks.

2.jpg
 
Last edited:

quattro4ever

macrumors member
Nov 25, 2019
38
2
Poland
hello gentlemen, my question concerns the card whose pictures I posted,
is there any way and a chance to make the card work in mac pro 5.1?
Mainly I mean if it can work connected to PCIe,
I know that I would have to connect the BT power from the motherboard, but is it enough?
 

Attachments

  • Bez-nazwy-2.jpg
    Bez-nazwy-2.jpg
    553.1 KB · Views: 330
  • Bez-nazwy-3.jpg
    Bez-nazwy-3.jpg
    873.7 KB · Views: 369
  • Bez-nazwy-1.jpg
    Bez-nazwy-1.jpg
    586 KB · Views: 362

cdiepenheim

macrumors newbie
Apr 7, 2020
19
3
Hi all,

We can help me out. I installed like you all ( I suppose) a BM 94360Cd in my 5,1 with an adapter kit and connected USB. I removed the CPU case and fans for better sight and working.
System info shows I have BT 4 and Wifi AC, Wifi I can use.
But the problem i have that I cant use BT, cant find any device in the area .
I connected a external SMA Antenna with extension cables.
What is maybe wrong, who can help me debugging this?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.