NetDiag+
NetDiag+ 1.4.11 and Android 1.0.4: steadier Wi-Fi measurements and fewer false alarms
This release adds no new tools. It fixes ways the existing ones could show a problem that was not there: a phone saving power between pings, and servers that never answer ping being counted as loss. It also moves a reminder that was easy to miss. It ships as 1.4.11 on iOS and as 1.0.4 on Android.
Wi-Fi that dozes between pings
To save battery, a phone’s Wi-Fi goes to sleep for short moments whenever no traffic is flowing. A reply that arrives during that sleep waits at your router until the phone wakes up for the router’s next beacon, which most routers send about every 100 ms. The ping then reads longer by a random amount, and no network caused it.
A game or a video call sends packets many times a second, so the radio rarely gets the chance to sleep. A measurement tool that pings less often does give it that chance. In a test on my iPhone on 17 September, holding the radio awake cut the jitter to an AWS region in the UAE from 35 ms to 2 ms, and to Tokyo from 42 ms to 6 ms.
The Latency Dashboard has handled this since September by pinging your router every 50 ms while it measures. Now Traceroute, MTR, the whole Bufferbloat test and the recovery step of Health Score do the same. It only happens on Wi-Fi, only while a measurement runs, and the extra traffic stays inside your home network.
What changes for you:
- Traceroute and MTR: distant hops no longer pick up a random extra delay.
- Bufferbloat: the idle reading no longer includes the delay, so the rise under load may now look larger than before. The new number is the more accurate one. The bufferbloat guide explains how to read the grades.
- Health Score: the check after the load was compared with an idle reading taken while the radio was awake. If the radio dozed during that check, it could report that the line had not recovered when it had.
No packet loss from servers that never answer
Some networks block ping. On them, Health Score could count the reference servers that did not answer as lost packets and flag your provider. Now only servers that actually answered are used for the loss figure, so a blocked ping no longer looks like a problem on the provider’s side.
The reminder moved up
The card that offers a second check at the other end of the day now sits right under the result, instead of at the bottom of the screen, where it was easy to miss.
Latency Dashboard
- Fortnite Middle East is removed. Epic publishes no address to ping, so the row could only show “no reply”.
- AWS Bahrain is now marked as unavailable, which it has been since March 2026.
- The Mumbai row now says what it measures. The game ping guide explains why a cloud region reads differently from the ping a game shows.
Smaller fixes
- On iPhone, Path MTU and MTR could wait for a reply forever in a rare timing case. They now stop on time.
- Traceroute and MTR accept only the error replies that belong to their own probes, so the router’s answers to the keep-awake pings can never be read as a hop.
- Health Score and VPN Check results are no longer sent with usage analytics.
Free · 31 tools · one-time purchase · no subscription