Author Topic: Batteroo testing  (Read 169523 times)

0 Members and 1 Guest are viewing this topic.

Offline Ysjoelfir

  • Frequent Contributor
  • **
  • Posts: 279
  • Country: de
  • Collecting vintage stuff and tinkering with it
    • Tinkertubes Lab
Re: Batteroo testing
« Reply #300 on: December 30, 2016, 02:59:57 am »
The actual switching frequency will be high - perhaps 1.5MHz. It would be good if you can see what the frequency actually is. I don't think anyone has actually measured it yet. It is almost certainly between 1MHz and 2MHz and it will probably be most noticeable at a high load.
It is. I attached pictures of the actual switching.

Quote
One of the other things I am interested to see is how quickly the converter wakes. If you switch a high load on suddenly, how low does the voltage drop?

I don't even have a clue how much current the Batteriser can output yet so I don't know if a high current is 100mA or 1A. Due to the very low quiescent current, the wake up circuitry may be very slow, and if the voltage can drop to zero when the load is switched on, that would be a pretty big disaster.
I attached pictures with around 100mA, 500mA and 1A load.

This last waveform shows how the batteriser switches, which is 7 times in this instance during the rising edge. After that de DC-DC converter goes to sleep, because the output voltage got to high.
During the rising edge, between the spikes you see little dents. That is clearly the point where the inductor is charged. The dents are slightly less then 1 us apart, which gives a switching frequency of 1.2Mhz or so.

Try zooming in further on the rising edge, and is should become clearer.
Thanks for clearing that up, my knowledge of switching converters seems to be capable of improvement.

Here I zoomed into the actual switching, which happenes indeed with around 1.2MHz with no load.


another picture, just a bit more zoomed in.


Under load (500mA)


Wake up time with 100mA load ~40µs


500mA, also around 40µs


fully recovered after around 100µs


1A load, 33µs wake up time, dropping down to 750mV


fully recovered after around 60µs but it is struggling to keep up the voltage, idling at around 1.2V (sorry, forgot to move the horizontal cursor..)


judging from those curves, can we assume that 1A load is to much and even 500mA is near the border of "to much"?
Greetings, Kai \ Ysjoelfir
Ysjoelfir @ Tinkertubes Lab
 
The following users thanked this post: sakujo7

Offline FrankBuss

  • Supporter
  • ****
  • Posts: 2031
  • Country: de
    • Frank Buss
Re: Batteroo testing
« Reply #301 on: December 30, 2016, 03:33:39 am »
1A load, 33µs wake up time, dropping down to 750mV


fully recovered after around 60µs but it is struggling to keep up the voltage, idling at around 1.2V (sorry, forgot to move the horizontal cursor..)


judging from those curves, can we assume that 1A load is to much and even 500mA is near the border of "to much"?

Good measurements. This might be another big fail for Batteroo. One of their claims was that it provides smooth output voltage for devices with short peaks of high current, like cameras, so that they run longer. They won't run at all, because the camera shuts down after the first big drop :-DD

My simple light bulb test mentioned here

Thanks to Frank I got a Batterizer in the mail today. Testing with two incandescent light bulbs side by side right now.

showed unexpected problems. The light bulbs powered by the batteriser keep on failing. I thought about a bad batch but until now the batterised bulbs died five times while the non-batterised bulb is not affected at all, keeps glowing as it should. I swapped the bulbs to see if I was just lucky with the still working bulb at the normal battery, but no, the bulb that has never been used on a batteriser blew up after around an hour being used on the batteriser, while the former "batterised" bulb is still glowing. strange. Has already somebody measured the noise coming from those things? are there major switching spikes that could potentially blow up the bulb?

That's strange, light bulbs shouldn't be that picky. A normal battery outputs 1.6 V at the beginning, and 1.5 V rating doesn't mean it fails at slightly higher voltage. This is the article in the Conrad shop. Can you measure the voltage with your scope near the light bulbs?

I wonder what other devices might break because of too high voltage spikes. If there are flashlights with this or similar light bulbs, their owners won't be happy.
So Long, and Thanks for All the Fish
 

Offline drussell

  • Frequent Contributor
  • **
  • Posts: 979
  • Country: ca
  • Hardcore Geek
Re: Batteroo testing
« Reply #302 on: December 30, 2016, 03:34:18 am »
The light bulbs powered by the batteriser keep on failing. I thought about a bad batch but until now the batterised bulbs died five times while the non-batterised bulb is not affected at all, keeps glowing as it should. I swapped the bulbs to see if I was just lucky with the still working bulb at the normal battery, but no, the bulb that has never been used on a batteriser blew up after around an hour being used on the batteriser, while the former "batterised" bulb is still glowing. strange. Has already somebody measured the noise coming from those things? are there major switching spikes that could potentially blow up the bulb?

Quote
Sure, lightbulbs suffer greatly from overvoltage, nothing new here, but the ones I used were rated at 1V5 and the batterizer is outputting 1.51V RMS. So that SHOULD be ok, at least if those bulbs are definitely 1V5 rated and not "to use with 1V5 batteries", as I can only read the marking on the bulb itself which states 1,5V.

Even if they were not designed to run at a constant 1V5, they shouldn't just fail after around half an hour, should they? Also, I had two batches, three came with the package from frank, and just for comparison I bought another set of four at the local conrad store.

The actual average running voltage does play a huge role in bulb life, but so can flakey, modulating, dirty power.  If you've ever seen a flakey mains wall switch controlling a standard tungsten bulb lamp, it will burn out bulbs very quickly yet the flickering it causes is usually almost imperceptible even if you know to look for it.

The thermal mass of the filament is actually rather small but sufficient to not be noticed with our eyes' persistence of vision characteristics.  Fully smoothed DC, perhaps with polarity automatically reversed at turn-on to prevent one end of the filament from always breaking first should yield the absolute greatest run time.  Clean AC sine waves should be quite close behind due to their relatively gentle nature.  DIRTY fluctuating, modulating, spikey, dirty crap power will, indeed, put significantly more stress on the filament and can easily cause very short lifespans.

If you compare the batterised bulb to one running off 1.52v from a stable, clean PSU and you'll very likely find that the batteroo sleeve's output is the main culprit, not just simply the elevated voltage.
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1623
  • Country: de
Re: Batteroo testing
« Reply #303 on: December 30, 2016, 03:40:43 am »
Here I zoomed into the actual switching, which happenes indeed with around 1.2MHz with no load.

Based on the voltage ramp during a burst of about 50mV/us and the measured output capacitor of 14uF, the current delivered by the boost converter is about 0.7A.
That matches your measurement.

 

Offline FrankBuss

  • Supporter
  • ****
  • Posts: 2031
  • Country: de
    • Frank Buss
Re: Batteroo testing
« Reply #304 on: December 30, 2016, 03:51:54 am »
The actual average running voltage does play a huge role in bulb life, but so can flakey, modulating, dirty power.  If you've ever seen a flakey mains wall switch controlling a standard tungsten bulb lamp, it will burn out bulbs very quickly yet the flickering it causes is usually almost imperceptible even if you know to look for it.

Interesting, maybe this is another reason they don't recommend using it with passive devices like light bulbs. It not only runs shorter, it actually destroys the devices in some cases.
So Long, and Thanks for All the Fish
 

Offline richnormand

  • Supporter
  • ****
  • Posts: 282
  • Country: ca
Re: Batteroo testing
« Reply #305 on: December 30, 2016, 03:56:05 am »
Would a camera flashgun, a big one like a NIKON SB-700, be a worthwhile test? They are a pretty high drain device but also one that is not a continuous drain so the battery chemistry has some recovery time. It would test a device that photographers would use in a cold climate outdoors so the temperature effect on battery life could be included to see if Batteroo can add some benefit.

Alkaline batteries in flashguns tend to be removed because recycle time gets too long. So the Batteroo would be tested in circumstances where the battery isn't already completely exhausted.

As long as you set the flash to max output every time I would guess. Mine has an automatic adjust feature that quench or limits the light output depending on the camera light and distance reading. In effect each flash energy and recharge is dependent on the scene and position.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 8771
  • Country: 00
Re: Batteroo testing
« Reply #306 on: December 30, 2016, 03:59:53 am »
that SHOULD be ok, at least if those bulbs are definitely 1V5 rated and not "to use with 1V5 batteries"

They're probably "to use with 1V5 batteries". Batteries is the main reason to make "1.5V" light bulbs.

In reality a battery will never run them at 1.5V. The ESR of the battery will probably start them at 1.4V and rapidly drop to 1.2V or less.

(measure it and see...)
 

Offline samgab

  • Frequent Contributor
  • **
  • Posts: 422
  • Country: nz
Re: Batteroo testing
« Reply #307 on: December 30, 2016, 05:37:15 am »
The actual average running voltage does play a huge role in bulb life, but so can flakey, modulating, dirty power.  If you've ever seen a flakey mains wall switch controlling a standard tungsten bulb lamp, it will burn out bulbs very quickly yet the flickering it causes is usually almost imperceptible even if you know to look for it.

Interesting, maybe this is another reason they don't recommend using it with passive devices like light bulbs. It not only runs shorter, it actually destroys the devices in some cases.

I wonder if the switching just happens to be at a certain harmonic frequency which makes the filament vibrate, which causes it to break quickly?
 

Offline JonF

  • Contributor
  • Posts: 5
  • Country: gb
Re: Batteroo testing
« Reply #308 on: December 30, 2016, 05:59:28 am »
On flashguns:

If they would kill a batteriser in fewer than 10 firings then I think that'd be a good test.
I think most flashguns have manual mode so scene dependant output power shouldn't be an issue (all my flashguns can do this).
There are cheap manual guns from the likes of Yongnuo that could be used to avoid damaging anything good with the spikes etc.
 

Offline photon

  • Regular Contributor
  • *
  • Posts: 231
  • Country: us
Re: Batteroo testing
« Reply #309 on: December 30, 2016, 07:44:37 am »
Kudos to those replying to this topic.
 

Offline samgab

  • Frequent Contributor
  • **
  • Posts: 422
  • Country: nz
Re: Batteroo testing
« Reply #310 on: December 30, 2016, 07:50:33 am »
I just watched Daves latest live stream of the Bateroo testing in the trains.  It wasn't until Dave was wrapping up and he mentioned testing the Batteroo in an active device, which I take it was the MP3 test, and a passive device which I presume is the train that it dawned on me that in order to light the LED on the front of the train there must have been some sort of active circuitry in the - passive - train to allow a AAA battery to light an LED.

Likely it's just an LED without any active circuitry. Single cell, 1.6-0.9V, no active circuitry needed to run a basic LED for the purposes of this toy.

Edit: I suppose it depends on what you mean by "active" circuitry... Do a torroid, a transistor, and a resistor count as active? I suppose the transistor does make it active, technically?

Update, as shown in Dave's photo of the train's innards, it's a bulb, not an LED (makes a lot of sense with it being simple and running on 1xAAA), so totally passive, no active circuitry at all.
 

Offline samgab

  • Frequent Contributor
  • **
  • Posts: 422
  • Country: nz
Re: Batteroo testing
« Reply #311 on: December 30, 2016, 08:27:55 am »


From a side-by-side testing of the toy train, the velocity was pretty constant due to the Batteriser.
But surprisingly the train without the Batteriser was constant too.



Here's the same test, but charting the full runtime off Batteroo'd versus NON-Batteroo'd. I think it tells the full story more clearly:
 
The following users thanked this post: cowana, Kean, sakujo7, Marcel_X, dr_frost_dk, newbrain, JonF

Offline samgab

  • Frequent Contributor
  • **
  • Posts: 422
  • Country: nz
Re: Batteroo testing
« Reply #312 on: December 30, 2016, 08:44:41 am »
Here's another way to look at the data: Laps per minute over time:

Edit: Slight fix to zero lap with no batteroo.
« Last Edit: December 30, 2016, 08:56:19 am by samgab »
 
The following users thanked this post: Marcel_X, ve2mrx, Blocco

Offline cowana

  • Frequent Contributor
  • **
  • Posts: 305
  • Country: gb
Re: Batteroo testing
« Reply #313 on: December 30, 2016, 08:47:28 am »
Very clear representation of the data there, good job :)
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 8771
  • Country: 00
Re: Batteroo testing
« Reply #314 on: December 30, 2016, 09:03:53 am »
Very clear representation of the data there, good job :)

Wonder which one of the set Batteroo will publish?

 

Offline djos

  • Supporter
  • ****
  • Posts: 847
  • Country: au
Re: Batteroo testing
« Reply #315 on: December 30, 2016, 09:04:56 am »
Here's another way to look at the data: Laps per minute over time:

Edit: Slight fix to zero lap with no batteroo.

Nice graphing, shows their product for the snake oil that it is!
The impossible often has a kind of integrity which the merely improbable lacks.

Visit my Tindie store for Tandy 1000 Adapters for EX, HX, SX, SL, TX & TL etc
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 27198
  • Country: au
    • EEVblog
Re: Batteroo testing
« Reply #316 on: December 30, 2016, 09:12:29 am »
From a side-by-side testing of the toy train, the velocity was pretty constant due to the Batteriser.
But surprisingly the train without the Batteriser was constant too.
That's something Batteroo could have posted, leaving out the flat line after the battery with the sleeve died, I'm sure they will quote you. It shows clearly that it is much better with the sleeves. Breaking news: "famous blogger David Jones admits that the Batteroo sleeve increases the performance, see the video and the the graph he created". How much did they pay you for this? :)

Yep, agreed. I'll repost it, so have removed links. The x-axis was wrong anyway.
 

Offline samgab

  • Frequent Contributor
  • **
  • Posts: 422
  • Country: nz
Re: Batteroo testing
« Reply #317 on: December 30, 2016, 09:13:50 am »
Cheers. I suppose the question is, would little Johnny notice the difference in speed for that first hour and a half or so (with vs without Batteroo), and would Johnny want to keep playing with the toy in the second hour when it's slowed to 2-3 laps per minute, vs 3-4 laps per minute? Or would he ask mum for new batteries at that point anyway?
Also, I wonder what this would look like if the testing was done in bursts, 5 mins on, 60 mins off, repeat. To simulate real world usage more closely? This type of testing would obviously have to be automated somehow, as no-one would have the time or patience to do that sort of testing manually. This would give the chemistry in the Alkaline cells time to recover between bursts. I think in that scenario, the one without a sleeve would outperform the sleeved one by an even greater margin.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 8771
  • Country: 00
Re: Batteroo testing
« Reply #318 on: December 30, 2016, 09:25:25 am »
Cheers. I suppose the question is, would little Johnny notice the difference in speed for that first hour and a half or so (with vs without Batteroo), and would Johnny want to keep playing with the toy in the second hour when it's slowed to 2-3 laps per minute, vs 3-4 laps per minute? Or would he ask mum for new batteries at that point anyway?

The kid won't be paying attention after half an hour. He may leave it going but he's not counting the laps.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 27198
  • Country: au
    • EEVblog
Re: Batteroo testing
« Reply #319 on: December 30, 2016, 09:29:41 am »
Here you go, much more informative.
I'll release this in a short unnumbered video on my main channel with a small bit of commentary rather than bury away on my 2nd channel, I think it's a very important passive device test.

 
The following users thanked this post: samgab, Fungus, Watth

Offline amspire

  • Super Contributor
  • ***
  • Posts: 3633
  • Country: au
Re: Batteroo testing
« Reply #320 on: December 30, 2016, 09:44:09 am »
another picture, just a bit more zoomed in.

The period of the major spikes seems to be about 400nS so that would mean a 2.5MHz switching frequency.

The ringing at about 80MHz is concerning, although it could be in the probe circuit rather then the Batteriser. Not sure how another UL emissions test would go with a full load. I don't think Batteroo will be rushing to do the test, especially as that sleeve could make quite a good antenna. Looks like about 300mV p-p of switching noise. When you add it to the 200mV p-p burst mode saw tooth at low loads and the 10-20uS wake-up delay out of the sleep mode, the output of the Batteriser is starting to look pretty ugly, especially if you put it next to a plain battery.
« Last Edit: December 30, 2016, 09:45:48 am by amspire »
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 27198
  • Country: au
    • EEVblog
Re: Batteroo testing
« Reply #321 on: December 30, 2016, 09:44:36 am »
And even more informative again, including the difference in velocity in percentage.
What kid is going to care about a 10% velocity difference vs almost double the run time and 150 laps extra?

 
The following users thanked this post: Kean

Offline EEVblog

  • Administrator
  • *****
  • Posts: 27198
  • Country: au
    • EEVblog
Re: Batteroo testing
« Reply #322 on: December 30, 2016, 10:43:28 am »
It turned into a pretty decent video with the commentary, so I'll now upload this a numbered video on the main channel.
Running out of time to do a full edited summary video on the testing before I head off on holiday, but I'll try.
 
The following users thanked this post: SeanB

Offline Muttley Snickers

  • Supporter
  • ****
  • Posts: 1828
  • Country: au
Re: Batteroo testing
« Reply #323 on: December 30, 2016, 10:58:22 am »
Now that you have a number of datum reference points and a handy little setup could you please run an Energizer Ultimate Lithium and perhaps an equivalent alkaline to the copper tops in another brand, this information might come in handy sooner or later, anyway have an enjoyable break.

The entire concept was loco from the get go and ended in a derailed train wreck.   :)  :-+
One smart cookie, better make that two for good measure.
 
The following users thanked this post: SeanB

Offline rollatorwieltje

  • Supporter
  • ****
  • Posts: 567
  • Country: nl
  • I brick your boards.
Re: Batteroo testing
« Reply #324 on: December 30, 2016, 11:28:49 am »
It would be nice to do the test with a NiMH rechargeable as well, with and without Batteriser. I think the train is an interesting test, it's right on the edge where NiMH and Alkaline should perform similarly (poor internal resistance for alkaline versus overall lower capacity for NiMH).
 
The following users thanked this post: samgab, sakujo7


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf