EEVblog #751 - How To Debunk A Product (The Batteriser)

[Godzil]

800%, even 80%, my arse! 

Are you Probe?

[uwezi]

Wow - 150 pages of discussion, I know that my contribution will drown...

Anyway, I just had the time to read the latest scientific source which the guys at Batterizer cited lately in favor of their view of the world: "Using PSpice to Simulate the Discharge Behavior of Common Batteries" by Steven C. Hageman. It appears that the guys at Batterizer really have a hard time reading, because this manuscript clearly states:

page 261:
- "For pulsed loads with cycle times greater than 10 seconds, the cell gives more total capacity than under a constant load. The rest portion of the pulsed load allows the battery chemistry to recover some of the lost capacity. But as the pulsed load cycle time becomes less than 1 second, the cell does not enough time... In these cases the RMS value of the pulsed discharge current should be used."

-"Cell resistance is a function of the cell's state of charge and ... Alkaline cells show a 2:1 to 4:1 increase in cell resistance from full charge to full discharge."

page 275:
-"When a battery is discharged to a low terminal voltage level, then disconnected from the load, the battery voltage will recover to some higher level in an hour or so. This phenomenon is accounted for in these models when they are discharged at high current levels then left to rest. ... The battery chemistry tries to make a voltage potential difference even if only a few molecules of unused material remain. In this state of discharge the internal resistance of the cell can be an order of magnitude or more than its initial value. If any load is reconnected, the terminal voltage will quickly collapse again to zero."

page 277:
-"Don't simulate pulsed current loads with cycle times less than 5 seconds or so."

[Jay_Diddy_B]

Hi,

@uwezi

I presented an LTspice version of the model described by Steven C. Hageman earlier in this thread. You can find it here:

https://www.eevblog.com/forum/blog/eevblog-751-how-to-debunk-a-product-(the-batteriser)/msg739090/#msg739090

@ drussell

5ky is using two boost converters, one per battery. I supplied to the boost converters to 5ky. They are described here:

https://www.eevblog.com/forum/projects/new-project-a-batterruser/msg739648/#msg739648

and the test results are here:

https://www.eevblog.com/forum/projects/new-project-a-batterruser/msg739968/#msg739968

In addition I measured the efficiency at 100mA, it was 90.5% with Vin=1.25V.

I used the LTC3539 (1 MHz version):

http://www.linear.com/product/LTC3539

Here is a picture of 5ky test:




This is the test data from Batteriser's Video:



The upper trace  (red) shows that as the test continues the battery current rises. This is a characteristic of a switching power supply. The switching power supply is drawing constant power from the batteries. The step is when the Garmin GPS reduces the brightness to prolong battery life.

5ky test results show that the time scale on this graph is wrong. The GPS should run for about 10 hours.



Regards,

Jay_Diddy_B

[drussell]

5ky is using two boost converters, one per battery. I supplied to the boost converters to 5ky.
...
Here is a picture of 5ky test:

I had even watched the video but obviously missed picking up on that.

I stand corrected, thanks!

[f4eru]

Ahh well, you see, that is the magic of this new custom designed IC from the genius Rooparvah brothers...  They've managed to design a custom IC that none of the big IC manufacturers, TI/Maxim/Fairchild/etc were able to achieve. Using Patents, genius, and MAGIC!

No, no nooo, they didn't use Magic, they know better. They use Bullshit.


Hey, why not put the Batterizer guys in touch with the solar roadway guys ? They will design us a nice roadwayzer which will increase the efficiency of solar roadways some 800% ....

[stuarts]

While I think that their claims of up to 800% are a bit difficult to justify, I can see some applications where it might help.
Case in point.

Many years ago I used an original Garmin Etrex whilst gliding.  It would run for about 22 hours on Energizers. If I put in NiMH with a higher nominal capacity than the Energizers, I was lucky to get 2 to 3 hours from them. While I never bothered to check the terminal voltage that the Etrex expired  at, it was obvious that the lower terminal voltage to the rechargables caused the Etrex to die. It would have been interesting to try such a device and see how long it would have run. Even today using Eneloops, 4 hours is about as much as I can get from the Etrex.

[Godzil]

Haaa Bullshit ISO certification 1664 is a really good thing

[drussell]

This is the test data from Batteriser's Video:
...
The upper trace  (red) shows that as the test continues the battery current rises. This is a characteristic of a switching power supply. The switching power supply is drawing constant power from the batteries. The step is when the Garmin GPS reduces the brightness to prolong battery life.

Yeah, I'm still a bit perplexed as to exactly what is happening there on their graph, that result has never made any sense at all to me.  Beyond completely fudging it on their part with extra resistance or starting with partially discharged batteries (can't tell since we see only current, not the starting cell voltage) or somesuch, my best guess is that because their screen tapper wasn't over the OK button on the alkaline battery feature notification like it was during 5ky's tests, the GPS runs until the message is displayed, then because the OK is never pressed it eventually times out and powers down or something.

Perhaps 5ky can shed some light on that for us as to what might have been happening there. 

5ky, Does the unit eventually shut itself off or go into some sort of standby mode if you don't press the OK button when the backlight reduction notice appears?  According to their graph it looks like it took about 25 minutes from the reduction in current point.  (If anything on that graph can be believed, that is...)

5ky test results show that the time scale on this graph is wrong. The GPS should run for about 10 hours.

Yeah, something certainly makes no sense with that upper, red graph.  Either they completely fudged something to do with the test itself or the timescale on the results or something like my hunch above must be occuring.  If they really measured that first test with fresh cells and didn't fudge the timescale, they must have some very significant extra resistance in there for the GPS to have seen such low voltages at its input so quickly.

As for total runtime, since they were using the Alkaline setting they should have got more like 17 hours total with the reduced backlight, not 10 hours, shouldn't they?  The time to backlight reduction should be more like 6 hours according to 5ky's data rather than the hour and a half or so that they seem to have observed.

[drussell]

While I think that their claims of up to 800% are a bit difficult to justify, I can see some applications where it might help.
Case in point.

Many years ago I used an original Garmin Etrex whilst gliding.  It would run for about 22 hours on Energizers. If I put in NiMH with a higher nominal capacity than the Energizers, I was lucky to get 2 to 3 hours from them. While I never bothered to check the terminal voltage that the Etrex expired  at, it was obvious that the lower terminal voltage to the rechargables caused the Etrex to die. It would have been interesting to try such a device and see how long it would have run. Even today using Eneloops, 4 hours is about as much as I can get from the Etrex.

That is a good example of a product that is designed for alkaline batteries and doesn't play well with rechargeables due to the chemistry's lower output voltage where a boost converter (either in the product itself or when necessary, a Batterizer) would be beneficial.  There used to be a lot more products that were like this than the typical consumer would be likely to be using today, but they certainly still exist.

The problem is that if you try to use a NiCd / NiMH in the current model Batterizer designed for alkaline cells you will be repeatedly over-discharging your rechargeable batteries and likely significantly shorten their cycle life because the Batterizer is designed to really, really heavily discharge an alkaline cell.  This is why they state that they're planning a separate model Batterizer for rechargeables in the future but that you are not to use the current model with rechargables.

... which is so silly because one for rechargeables would have been a far better product to begin with (but much more difficult to try to fool the public into buying with grandiose claims of always increasing battery life many times over) for those cases where you want 1.5 volts instead of 1.2 volts nominal from your cells... 

Once again, we're doing Batteroo's engineering for them, but what else can 'ya do? 

This has been discussed previously by others and I mentioned it in post #2075, for example:

...
The most useful kind of "Batterizer" would be one which changes the discharge curve of a NiCd or NiMH to mimic an alkaline battery to make them a drop-in replacement for an alkaline.  The rechargeable battery companies could even build that INTO these special rechargeable cells without any problem (they could shrink the actual cell size to make room for the electronics) kind of like a typical rechargable Lithium chemistry battery pack will have control electronics in it.  It can include short and deep-discharge protection, adjust the voltage output to vary like an alkaline so battery meters would still function, etc, etc.
...

[tree]

I found another product that I have at home that doesn't play well with rechargeables: my hand soap dispenser. It will actually stop functioning within a few weeks of inserting freshly charged eneloop batteries. After the soap dispenser stopped functioning, I charged my batteries with one of those smart chargers and I only managed to store ~300mAh of charge when they normally can do >2000mAh. Had batteriser come out with a completely different claim, such as providing a guaranteed 1.5V output voltage they would have been successful, but they did not. They're sticking to their ridiculous claims and trying to justify them when science is NOT on their side.

[Probes]

800%, even 80%, my arse! 

Are you Probe?

My name is Probes T. Monkey.

Pleased to meet you!

[Fungus]

I found another product that I have at home that doesn't play well with rechargeables: my hand soap dispenser.

You have a battery powered soap dispenser??   

[tree]

I found another product that I have at home that doesn't play well with rechargeables: my hand soap dispenser.

You have a battery powered soap dispenser??   

Yes. Stick your hand to break the IR beam and you get soap. Black magic!

[Fungus]

I found another product that I have at home that doesn't play well with rechargeables: my hand soap dispenser.

You have a battery powered soap dispenser??   

Yes. Stick your hand to break the IR beam and you get soap. Black magic!

Until the sensor gets gummed up and it empties all the soap over the floor.

[SeanB]

Not for long, those things break down due to the unprotected electronics getting wetted by spilt soap. the sensor is smart, it only dispenses once with a hand approach, it does not carry on till empty.

[tree]

I found another product that I have at home that doesn't play well with rechargeables: my hand soap dispenser.

You have a battery powered soap dispenser??   

Yes. Stick your hand to break the IR beam and you get soap. Black magic!

Until the sensor gets gummed up and it empties all the soap over the floor.

Actually it's quite smart. Sensor always gets gummed up and never leaks on the floor. It triggers on edges I think rather than hi/lo values.

[drussell]

You have a battery powered soap dispenser??   

Yeah, a no-touch soap dispenser. they're actually a pretty nifty.    A great idea.

Until the sensor gets gummed up and it empties all the soap over the floor.

...
the sensor is smart, it only dispenses once with a hand approach, it does not carry on till empty.

Like SeanB says, they won't just continuously squirt soap.  Those dispenser doodads (or at least any reasonably well desgned one ) does a single-shot burst of soap when the beam is broken but then does hysteresis so it won't activate again until the IR sensor beam is re-established for some period of time and then broken again for another squirt.

I don't own one but the ones I have used work great.

[drussell]

My name is Probes T. Monkey.

Pleased to meet you!

Huh...
He came to life...
Good for him!

(Just like Lard Lad) 

[Fungus]

Until the sensor gets gummed up and it empties all the soap over the floor.

the sensor is smart, it only dispenses once with a hand approach, it does not carry on till empty.

Like SeanB says, they won't just continuously squirt soap.

Maybe you meant you've never seen one continuously squirt soap. Not yet.

[tree]

Until the sensor gets gummed up and it empties all the soap over the floor.

the sensor is smart, it only dispenses once with a hand approach, it does not carry on till empty.

Like SeanB says, they won't just continuously squirt soap.

Maybe you meant you've never seen one continuously squirt soap. Not yet.

Slip on a soaperiser and get 800% more soap!


Shut up and take my money!

[5ky]

Minor disclaimer, in that picture that Jay posted of the setup on the last page, I had the voltage measurement hooked up wrong.  (I was measuring battery voltage, before the boost converters, which is what I should be doing, only I had the DMM hooked up across BOTH batteries as if they were in series, but they weren't--their outputs are what was in series)

I was on very little sleep, and had a brain fart and was measuring the voltage of the batteries as if they're in series, which didn't immediately throw red flags since that meant it was actually the voltage of one boost converter output in series with the other battery, so the voltage looked correct on the DMM.  I unfortunately didn't notice for an hour that the batteries weren't in series, but instead the outputs were, so I stopped and had to restart the test with fresh batteries, measuring only one battery.  I took measurements of both batteries during the test and they were VERY parallel--I got very lucky that the batteries didn't stray at all during the test.  I just doubled the voltage in excel to compensate for only measuring one.  The difference between the sum of both battery voltages versus my doubling method would be negligible.  (good news is--I found an excuse to go pick up a second bench DMM to avoid problems like this in the future)

Perhaps 5ky can shed some light on that for us as to what might have been happening there. 

5ky, Does the unit eventually shut itself off or go into some sort of standby mode if you don't press the OK button when the backlight reduction notice appears?  According to their graph it looks like it took about 25 minutes from the reduction in current point.  (If anything on that graph can be believed, that is...)

From what I can tell (it's sitting on that warning message as we speak and it's been 20+ minutes), it's just a warning or reminder that the screen is being dimmed and that you should consider lithium or nickel batteries.  It has not shut off (yet), and I don't imagine that it will.  (it would be annoying for it to shut off in the middle of your golfing round just because you didn't click "ok" to confirm that you read the warning message)

Oh and for those discussing those battery soap dispensers--they're awesome!  I find them to be handy when I'm putting rub on some pork ribs.  I can just use my wrist to turn the faucet on and break the IR beam to get my soap.  No need to get anything messy in the process.  (well, I still manage to make a huge mess when using the smoker)

EDIT: words and stuff

[drussell]

The difference between the sum of both battery voltages versus my doubling method would be negligible.  (good news is--I found an excuse to go pick up a second bench DMM to avoid problems like this in the future)

A few op-amps rigged up to be two (or more) differential amplifiers that buffer the two (or more) cell voltages at unity gain across the + and - terminals and then added together (which could easily be extended to n number of cells) would easily give a reasonably accurate sum of the voltages of the individual cells.

I don't think I've ever thought about that one before but that actually sounds like a handy apparatus to have around.  I think I just might solder one up with, say 8 differential inputs, on some proto-board with some junkbox op-amps just to have one around for adding (or subtracting, if you reverse an input +/-) things on the bench.  As long as the op-amps have a reasonably small offset voltage with the inputs shorted (any unused inputs can be disabled by shorting them, preferrably to ground, I suppose) the resultant output should even be fairly accurate.

...
5ky, Does the unit eventually shut itself off or go into some sort of standby mode if you don't press the OK button when the backlight reduction notice appears?  According to their graph it looks like it took about 25 minutes from the reduction in current point.  (If anything on that graph can be believed, that is...)

From what I can tell (it's sitting on that warning message as we speak and it's been 20+ minutes), it's just a warning or reminder that the screen is being dimmed and that you should consider lithium or nickel batteries.  It has not shut off (yet), and I don't imagine that it will.  (it would be annoying for it to shut off in the middle of your golfing round just because you didn't click "ok" to confirm that you read the warning message)

I thought it might possibly conclude that the user wasn't paying attention so an inactivity timer would activate but hadn't thought about the fact that this is intended for a round of golf so some people might well just have it in a pocket or whatever and not even look at the GPS at all until the end of the round to get some statistics rather than peeking at it during the round.  That would be quite an anti-feature, I suppose... 

Hmmm...  That still leaves me wondering just what the heck they were doing to appear to get such horrible battery life on the alkalines directly. 

I wonder how much resistance would have to be in series with that thing to make it shut off at a battery terminal voltage of 2.65-2.7 volts (approximately where your alkaline tests were after 2h, not that your set-up was exactly a 0 ohm set-up either, but still...)  That could be tested with a variable power supply and a small-value rheostat or some discrete small-value resistors, I suppose, just to try to approximate how bad their test setup might have had to have been...  Perhaps?  Maybe?  I'm at a loss on that one.

I just can't seem to fathom how on earth they could have observed such terrible battery life if the thing really did shut off.

[drussell]

Maybe you meant you've never seen one continuously squirt soap. Not yet.

Your quote conveniently left out my next sentence where I stated:

(or at least any reasonably well desgned one )

There may well be poorly designed ones out there but I haven't seen one personally. 
Have you actually seen one that misbehaves like that or are you just theorizing?

[5ky]

A few op-amps rigged up to be two (or more) differential amplifiers that buffer the two (or more) cell voltages at unity gain across the + and - terminals and then added together (which could easily be extended to n number of cells) would easily give a reasonably accurate sum of the voltages of the individual cells.

I don't think I've ever thought about that one before but that actually sounds like a handy apparatus to have around.  I think I just might solder one up with, say 8 differential inputs, on some proto-board with some junkbox op-amps just to have one around for adding (or subtracting, if you reverse an input +/-) things on the bench.  As long as the op-amps have a reasonably small offset voltage with the inputs shorted (any unused inputs can be disabled by shorting them, preferrably to ground, I suppose) the resultant output should even be fairly accurate.

A similar solution was my next plan of attack had the battery voltages strayed at all from each other, but the night-of, I was far too tired (and lazy) to come up with a better solution 

For anyone wanting to look at the raw data, here's a link to the spreadsheet.  It's 16 MB so I wasn't able to attach it directly to this post: http://s000.tinyupload.com/?file_id=35560814306500935701

For those who do look at the spreadsheet, the graphs are to the right of the data at the very top.  Top one is just the boost converter versus regular alkaline.  Graph below it includes the lithium and NiMH batteries.

To make the lines appear thinner for visibility, I added a weighted rolling average, dynamically, so you can just set both weights to 0.5 if you want to see the raw data in the graphs.  (it will look the same, but the noise will be greater, but it also makes it easier to see the voltage dips too since the filtering lessens those by quite a bit)

My wife is making me go to the dog park with her since it's nice out today, but once I'm home, I'm going to make another quick video with these results.  The world needs to know that Batteriser will probably give less battery life in "most" cases than not using them at all. 

[AmmoJammo]

My wife is making me go to the dog park with her since it's nice out today...

Do you have a dog...?