Hidden "feature" in 1.66.645.3...

Chat about Hammerhead's Karoo in general.
Post Reply
dthempel
Site Admin
Posts: 320
Joined: Tue Mar 06, 2018 3:35 pm
Location: Cape Cod, MA
Contact:

Hidden "feature" in 1.66.645.3...

Post by dthempel » Thu Jul 18, 2019 1:47 pm

Folks should probably notice battery improvement in this version. For some reason, they didn't announce it to the ATG group and we "found" it, and they then admitted that they had been tweaking it and hoped we noticed. Seems a weird way to "test", but I guess it's one way to do it. Reports are that hourly usage has dropped to around 7 or 8% depending on number of sensors, navigation, etc., but it does seem improved for most/all of the ATG users.

I'm not sure why they didn't add that as a bullet item in the public release at it would seem to be something they'd want to toot their horns about...

JohnJ80
Posts: 261
Joined: Tue Apr 03, 2018 4:55 pm

Re: Hidden "feature" in 1.66.645.3...

Post by JohnJ80 » Thu Jul 18, 2019 3:03 pm

Battery consumption in electronics is highly dependent on the application, the application software, and even small changes in the application usage. In the semiconductor industry, it’s the hardest spec to generate and it seems like it’s always wrong. Often when you look at the power specs of a part that is under software control, the specs are ridiculously wide/generous because it’s too hard to predict.

So that’s probably why they didn’t mention it. They thought they had it but the only way to know is when it’s out there in volume.

Based on my testing, looks really solid. I’m at 7.1% per hour with a full suite of sensors. Prior to this I was at 9.5% to 10.3%. That’s a 33% improvement and crushes their 10 hour spec presuming it’s linear over the whole battery capacity.

dthempel
Site Admin
Posts: 320
Joined: Tue Mar 06, 2018 3:35 pm
Location: Cape Cod, MA
Contact:

Re: Hidden "feature" in 1.66.645.3...

Post by dthempel » Thu Jul 18, 2019 4:30 pm

And yet they just sent a "glossy" email touting the improvements...

JohnJ80
Posts: 261
Joined: Tue Apr 03, 2018 4:55 pm

Re: Hidden "feature" in 1.66.645.3...

Post by JohnJ80 » Thu Jul 18, 2019 10:06 pm

Why shouldn’t they tout it?

User avatar
Steve
Site Admin
Posts: 582
Joined: Tue Mar 06, 2018 2:56 pm
Location: Surrey, UK

Re: Hidden "feature" in 1.66.645.3...

Post by Steve » Fri Jul 19, 2019 7:07 am

JohnJ80 wrote:
Thu Jul 18, 2019 10:06 pm
Why shouldn’t they tout it?
I think that's Doug's point. Why didn't they put it in the changelog?

JohnJ80
Posts: 261
Joined: Tue Apr 03, 2018 4:55 pm

Re: Hidden "feature" in 1.66.645.3...

Post by JohnJ80 » Fri Jul 19, 2019 12:39 pm

Because they weren’t sure until they had field reports confirming it in quantity. That’s the problem with electronics and power measurements on chips.

They’ve been bitten by over promising and under delivering. I’d give them credit for being cautious for a change.

dthempel
Site Admin
Posts: 320
Joined: Tue Mar 06, 2018 3:35 pm
Location: Cape Cod, MA
Contact:

Re: Hidden "feature" in 1.66.645.3...

Post by dthempel » Fri Jul 19, 2019 1:28 pm

Sorry, but they should have told the ATG group. We could have specifically done some testing instead of spending half the week speculating on "is it better?"

I'm not unhappy with the improvement -- far from it -- I'm just baffled by their continued communication stumbles. The "glossy" email came out just a few hours after the update notice. They *knew* they were going to send it and had details from the ATG group confirming their suspicions. They didn't wait for general users' data to confirm.

Bottom line for me: It should have been in the patch notes. That's what patch notes are for. They're perfectly welcome to toot their horns with the extra communication detailing the improvement, but it should have been noted. If I "snuck" something into my code and then announced it a few hours later after going into my production environment, I'd get a tongue lashing... or worse.

User avatar
Steve
Site Admin
Posts: 582
Joined: Tue Mar 06, 2018 2:56 pm
Location: Surrey, UK

Re: Hidden "feature" in 1.66.645.3...

Post by Steve » Fri Jul 19, 2019 4:14 pm

JohnJ80 wrote:
Fri Jul 19, 2019 12:39 pm
Because they weren’t sure until they had field reports confirming it in quantity. That’s the problem with electronics and power measurements on chips.

They’ve been bitten by over promising and under delivering. I’d give them credit for being cautious for a change.
Doesn't make sense to announce it in the marketing emails if they don't feel confident announcing it in the changelog.

JohnJ80
Posts: 261
Joined: Tue Apr 03, 2018 4:55 pm

Re: Hidden "feature" in 1.66.645.3...

Post by JohnJ80 » Fri Jul 19, 2019 4:24 pm

Rule #1 of marketing - if you've got it, talk about it.

I don't see how this issue is worth beating them up for it. However they did it, it's a good thing. They look like they figured out the battery consumption issue absent evidence to the contrary now. That's a great thing because at 7.1% per hour that translates into around 14 hours of battery life presuming that's consistent across capacity. If so, that's downright awesome and for that, they deserve kudos. Changelog vs market email issue = inconsequential.

Meanwhile, Apple is doing silent updates to fix major hacks. That's not showing up in a changelog either.

User avatar
Psyclist
Posts: 43
Joined: Fri Jul 06, 2018 1:26 am
Location: Vientiane, Laos

Re: Hidden "feature" in 1.66.645.3...

Post by Psyclist » Tue Jul 23, 2019 4:38 am

Strangely enough, last week's update was the first one ever that gave me any battery weirdness. None of the bugs people reported ever affected me, and now I had a ride where I started at 80%, it was down to 50% relatively quickly and then turned off. After I switched back on, it came up with 25% and took about 5 minutes to "drain" that to zero while the device didn't even get warm, so it was definitely some battery gauge bug, not a stuck process or so that would have caused extreme CPU load (not that it would have been able to drain the battery *that* fast anyway)

Post Reply