This will be a long post. Followed by another long post. Sorry.


If anyone reading this (I dont expect its many people) invests in stocks, then you may have been on a message board or two. Yahoo! has them, Google has them, and there is Raging Bull. These generally have two parties, Pumpers and Bashers, otherwise known as longs and shorts. The long beleive in the company, be it the technology, the management, the market, or whatever, and the shorts simply dont for whatever reason.

I recently got involved on the Yahoo! board for MDTL, a company called Medis. Of all the message boards I have been on, this is one of the most entertaining with quite a cast of characters. I reccommend going there and taking a look. I am impressed with the logic presented by some, the faith presented by others, and the total lack of racism, or any sort of predjudice that is to common on other boards.

Some people do get frustrated and start calling people names and mock their intelligence (mostly people who are low in that resepct). But it is still rather fun.

Anyway, Medis has spent many tens of million of dollars developing a power pack. They claim this power pack will be better than anything out there. Some on the board claimed it would be "dozens" of times better. That would be pretty amazing.

Here is a graph of how much energy they were able to suck out of one. It shows that when they discharged it at a constant rate of .65 watts, they were able to pull out 15 Watt-hours. First, some notes about this graph.

I suspect that they picked 0.65 W as a baseline discharge rate because there are many articles such as this one, that say cell phone draws 180 mA on average and a capacitor is used to smooth out the spikes in current. A Lithium Ion battery runs at around 3.6V for most of its life (It starts at 4.2 and can go down as far as 3V, not in a linear fashion). So this means that a cell phone uses .65W. So they chose a good number for power draw.

However, then they choose to use 5.6V as the voltage to run at. No cell phone will tolerate this (The Li+ battery should not be exposed to voltages over 4.2V), which means a DC converter would have to be used which will reduce the total energy out since it will provide ineffeciency in the conversion. Lets say that this is 10% inefficient, so the real energy output is only 13.5 Wh.

Lets also look at this graph. They took the entire energy output and counted it towards the total. An underpowered cell phone will drop a call. It wont work below 0.5W average. Looking at their graph, this takes you to only 13.5Wh, then take of the 10% again and you are down to 12.1 Wh out of their device. BTW this Medis Power Pack (MPP) device is slated to be 12 to 20 dollars (this number changes with time).

One thing to note, by their data, it looks as if you could make a 17 hour phone call. This is cetainly better than the battery in the phone!

But lets compare to the incumbent technology, alkaline batteries. Lets get rid of all the problems with their data and just assume the 15Wh is correct. I set up an experiment:

From an idea someone on the Yahoo message board had, I took 6 AA batteries, hooked them in series and discharged them first at constant current, then at constant power (or best I could)

Cost to me for this power pack was
6 AA batteries 60 cents x 6 = $3.60 (test 1 and 2 used 4 year old batteries, test 3 use new)
2 battery holders 1.50 x2 = $3

If it were a real power pack there would a a DC convertor circuit that probably would cost about 2 dollars in quantity. I think it is safe to say that if a 6xAA device were made in quantity it would cost way less than 10 dollars to make.

Setup
Here are some pictures of the setup

The power pack
(used Energizer MAX batteries)



















The circuit





















I used a Blue Earth BE485 microcontroller to gather data. These are quite versatile and feature packed. Check out Blue Earth. The BE485 can only measure voltage between 0 and 5 volts. So I used a potentiometer to reduce the battery voltage down to that range and scaled the readings in the program.The scaling was 9:5 (9 volts on the 6AAPP was scaled to 5 volts)



















Here is a picture of it in its interface board and the computer setup taken some time into the first run:



















In order to compare apples to apples as much as possible, I am going to assume that this device (the 6xAAPP) should run at the same power as the MPP. In order to do that, it would require a step down DC-DC convertor. Here is a datasheet for one that shows you can get up to 90% efficiency doing that. Lets assume 85%. So, if my load is at 3.6V and 180 mA= 0.65W, then the 6xAAPP would have to run at 0.8W to accomodate the expected 15% loss. The average voltage as a AA battery discharges is 1.3V. So the correct current through a stack of 6 AA batteries is 0.8W/(1.3V*6)= 120 mA.

RESULTS

Here is the result of the first run (click to expand):












I was able to extract 16Wh out of the batteries. This phone would run for 18 hours. This result includes presumed DC-DC convertor losses. It also ran a bit high in power, the average power here is 0.85W. It is clear I had some noise on my lines I was sampling. The microcontroller imparted most of it as they are prone to do. Since the current is shown on the right axis, it looks magnified. I also had a bit of trouble in the beginning with data collection (lost connection between BE485 and computer) and there are some gaps, but battery discharge never stopped.

I ran the same test at 180mA. This resulted in less energy taken out of the 6xAAPP. Only 13 Wh was extracted but it was run at 1.3W! This means this device could work on even highly featured devices like a Treo. It ran for 10 hours at this power.

I then ran the same test again (brand new batteries), but this time I tried to do it at a constant power draw as the DC convertor would do. I didnt have any DC converter chips around so i did it by hand (poorly) by adjusting the current through the 6xAAPP to a level based on its voltage.













It is pretty obvious where the places are where I made adjustments. The average power here is 0.77W. Assuming the same 85% efficiency, that would provide 0.65W of power to a device. Also I lost some energy in the calculations becuase for a new battery the voltage converter was no longer set to show the true voltage on the battery. so while it shows 9V its was actually as high as 9.3V!

Here I got 16.4 Wh. The extra bit could have been because of the different dicharge profile, or it could have been becuase these were new batteries. It doesnt really matter. The key here is that at 0.77W, this device ran for 21 hours!

One more little thing to note is about the size. Here are the two sizes compared to each other

AA power pack Medis Power Pack
1.9"x2.5"x1.4" 3.7" x 2.6" x 1.4"

While there is no DC converter in the 6xAAPP, there is space between the two battery packs for a surface mount DC-DC convertor, so I think it is fair to use the measured envelope size.

So it seems that the 6xAAPP is cheaper (expecially when you go to buy the second one!), smaller, and lasts just as long or longer.

In the next post i'll go into why these are not the reasons I think that the company will not work out. Its a market matter, not just the fact that this device isnt technically superior to the incumbent technology. Yes, you can consider me a 'basher' for this stock. When the time is right I do plan to short it.

I will be happy to answer any questions or comments.

File Under:
Post a Comment