Hangtime

My Feedback: (19)

Hi Bill!!

Assuming the 1000mah single pack impedance is the same or close to the 2000mah single packs impedance, the above statement is accurate, yes!

BillS

OK. Pardon my ignorance. Is it also logical to assume that smaller packs would have smaller impedance?

Bill

JPMacG

My Feedback: (2)

Anyone care to speculate on why the battery manufacturers specify impedance rather than resistance? Seems to me impedance is meaningless information unless you know at what frequency it was measured. Even if you did, what possible use would it be? Do some elaborate calaculation involving the rise time distortion of the servo current waveform?

Hangtime

My Feedback: (19)

Hiya Bill!

Still makin' me work fer my supper, sir? BTW, you have nothin to apologize for.. yah got good guestions, I just hope my answers make sense. If I err here, hopefully Red or one of the electrical engineers here will step in and help fill in the areas where I missed the obvious.

Usually, it works the other way... the smaller the cell dimension; the higher the impedance. In other words 'fat' cells generally (but not always) offer lower impedance than the skinny ones compared in the same technology types. Then generally, a nicad cell of a 'given dimension' will have lower impedance than a nimh cell of the same dimension. There are some exceptions... below is a smattering of cells we commonly use in the hobby and thier impedance specs. Data from Sanyo. (one from GP)

Cell Type/Designation ----Technology -- Dimension ---- Impedance ------------ Notes

Sanyo CP2400SCR --------- Nicad --- Standard Sub-C --- 4.5mOhm-----Big Bird Rx packs, popular in electrics
GP 3300 SCH----------------- NiMH --- Standard Sub-C --- 5.0mOhm----- same as above, ignition
Sanyo CP1700SCR ---------- Nicad --- 4/5 Sub-C --------- 5.5mOhm----- same as above, ignition
Sanyo KR1400AE ------------ Nicad --- Standard 'A' ------- 10.0mOhm---- Common IMAC/Pattern Rx pack
Sanyo HR-AU (2700) ---------NiMH---- Standard 'A' ------- 20.0mOhm---- Common IMAC/Pattern Rx pack
Sanyo KR1500AUL------------ Nicad --- 4/5 Height 'A' ----- 16.0mOhm---- Lighter Pattern Rx, .60 size
Sanyo HR-4/5FAUP (1900) -- NiMH----- 4/5 height 'fat' A - 5.0mOhm------ IMAC & Big Bird Rx, Ignition, electrics*
Sanyo KR-1100AAU ---------- Nicad --- Standard 'AA' ---- 19.0mOhm---- Upgrade Tx, .40/.60 sized Rx
Sanyo KR-800AAE ------------ Nicad --- Standard 'AA'---- 12.0mOhm---- Upgrade Tx, .40/.60 size Rx
Sanyo HR-AAU (1650) -------- NiMH --- Standard 'AA' ---- 25.0mOhm--- Upgrade Tx, smaller pattern/aerobatic
Sanyo HR3U (2100/2300)----- NiMH --- Standard 'AA' ---- 25.0mOhm--- Upgrade Tx, Not reccomended airborne**

* new development cell.. shockingly good performance from a NiMH 'A' cell.
** poor fast charge performance, easy to damage, thin wall case, poor temp resilience.

Ok, as you run down the list, on the left hand column, number's are mah rates. You'll see with only a few exceptions, as the cells get dimensionaly smaller, the impedance trends upward. The 'Notes' are relevant to the way I see them commonly used.

You can also garner (I think) why the guys that run the high capacity NiMH setups and the 'A' sized nicad setups generally run 'em in parallel; cuts that high impedance number down to a lil more comfortable level while bumping up total system capacity & adding a nice redundant saftey factor on switches and connectors. Down tick is it's a bit heavier and field support with out a dual port charger can be problematic.

Bill, Here's a link to a pretty reasonable decription of battery impedance and it's effects from the 'Battery Handbook for Non-Engineers'..

http://www.buchmann.ca/Chap6-page3.asp

Hiya JPMacG!

Sanyo's ratings are done at 1000Hz, but as to why they use an impedance rate instead of a resistance rate.. I dunno. Never thought to ask.. possibly Red can shed some light on it. For our purposes as a 'comparative' number, we can at least evaluate the cells against each other; and on that level the numbers do us some good.

Hope all this helps..
<
edit: messed with the columns to get 'em to display better>

JohnMuchow

Internal resistance is often measured with a 1KHz AC discharge current waveform so "impedance" works. There are DC-current internal resistance tests but, I believe, the AC numbers look better as there is less influence from variances in electrode voltage due to the current flowing through it (not just a simple DC voltage drop). There is a lag time before these variances affect the voltage level so testing with AC current lets the total internal resistance/impedance numbers stay lower....and look better for a manufacturer.

DC current tests, depending on the application, may be a better indicator of the cell's true voltage-under-load. This can be source of confusion when one user's experience with the cells doesn't agree with the published internal resistance numbers. It all depends on whether it was measured using AC or DC current and whether the application is pure-DC or AC. Check the data sheets for your cells carefully before buying if the internal resistance numbers are important to you (they should be).

Sanyo's CADNICA Engineering Handbook has a good description of the different cell resistances involved in determining the total internal resistance or impedance and why DC and AC tests are used (Section 3-2, page 11): [link=http://www.sanyo.com/batteries/lit.cfm]http://www.sanyo.com/batteries/lit.cfm[/link]

JPMacG

My Feedback: (2)

Thank you John and Steve.

backfire

I just ran a few quick resistance tests on some common sub-C NiCd cells ranging from 1500 to 2400 mAhr using the direct current method shown on page 12 of the Sanyo engineering manual for NiCads. My results indicate that the DC resistance (most important for our purposes) is at least five to ten times greater than the published impedance specification (that's measured at 1000 Hz) for a given cell. The impedance specification is only useful for calculating the effect of short, low-duty-cycle load pulses. There might be some useful correlation between the published impedance value and the DC resistance but I haven't seen it in the literature so far.

JohnMuchow

This is where I'm still a bit confused as to how to apply the impedance numbers. The typical R/C load is pulsating DC, a varying duty-cycle waveform from 0 to full current draw for the motors (same current as 100% throttle), at up to a several KHz rate depending on the ESC. It's not AC, but it is a series of short pulses and the cell has a bit of time to recover between each pulse (when the current is zero). If you're running full out, then it's pure DC current being drawn from the cells.

IMHO, DC current tests are probably the best way to rate the cells. It stresses them the hardest and is the easiest test to reproduce so just about anyone can do the tests. But, it might represent a worse-than-worst-case scenario and may not accurately represent how a cell will perform when supplying power to an ESC (pulse-width modulated current). IMHO, a lot will depend on how a NiCd/NiMH/LiPo cell responds to, and recovers from, these short pulses. <sigh> Time for more research. :-)

backfire

Here's something else interesting to think about. Although, while digressing from the Rx application a bit, consider an 8-cell pack powering an electric plane at say 25 Amps. Using a typical impedance value of 5 milliohms for each cell, the power dissipation for the pack would be 25 Watts. Given the temperature these packs get to in five minutes, this seems about right. If on the other hand, a value of 50 milliohms is used, the dissipation would be 250 watts. In this case, the pack would probably catch fire in five minutes. It seems apparent that for the purposes of calculating voltage drop for longer duration loads (maybe over 1/4 second), the resistance derived from the DC method should be used. For the purposes of battery power dissipation, use the published impedance value. According to Sanyo, the major 'resistance' component of the DC method is due to cell polarization. It looks like cell polarization is just a change of cell voltage and not resistive power dissipation.

JohnMuchow

Here'a a data point to throw into the mix...

Eveready's method of calculating DC internal resistance of NiCd cells (copied from their on-line guide):
Internal resistance (Re) is calculated using the voltage drop method as described in ANSI C18.2, which states that a fully charged cell rated at less than 5Ah shall be discharged at 10.0C1A(capacity rating at 1 hour rate in terms of amps) for 2 minutes then and switched to 1.0C1A. The voltage shall be recorded just prior to switching and again upon reaching its maximum value after switching. The effective internal resistance, Re shall be calculated as indicated below:

Re = DV/DI where DV = VL - VH and DI = IH - IL

Notations:
Re = Internal Resistance
DV = Delta-V, Voltage Change
DI = Delta-I, Current Change
VL = Voltage recorded after switching
VH = Voltage recorded prior to switching
IL = Current recorded after switching
IH = Current recorded prior to switching

For 50% discharged cells, multiply Re by 1.2 factor.

Their guide to NiMH cells doesn't list a DC resistance measuring method so, IMHO, the method can be used for both (perhaps all) cell chemistries.

JohnMuchow

Another bit of info from Eveready's NiMH guide (with an interesting difference between NiCd and NiMH cells pointed out):

For purposes of electrical analysis of the battery cell, the Thevenin equivalent discharge circuit shown in Figure 7 is often used. This models the circuit as a series combination of a voltage source (Eo), a series resistance (Rh = the effective instantaneous resistance), and the parallel combination of a capacitor (Cp = the effective parallel capacitance) and a resistor (Rd = the effective delayed resistance).

See Figure 7 below: Equivalent Discharge Circuit for a Nickel-Metal Hydride Cell

Eo = effective cell no-load voltage
Re = (Rh + Rd) = total effective internal resistance
Rh = effective instantaneous resistance
Rd = effective delayed resistance
Cp = effective parallel capacitance
E = cell termination voltage

For steady state purposes, the cell voltage at a given current is Eo - iRe, where Re, the effective internal resistance, is the sum of Rh and Rd. The transient response is shown in Figure 8 where the initial voltage drops immediately to Eo - iReh and then transfers exponentially (with a time constant = Cp *Rd) to the steady-state voltage. Obviously the process reverses when the load is reduced or removed. For many application the steady-state voltage is adequate for describing cell performance since the time constant for most cells is small: usually less than 3 percent of the discharge time. When compared to a nickel-cadmium cell, the steady-state voltage for the nickel-metal hydride cell will be reduced since, although the instantaneous resistance is comparable, the delayed resistance will be on the order of 10 percent higher.

See Figure 8 below: Example of Transient Voltage Profile for a Nickel-Metal Hydride Cell

[Edit] The left-hand graphic is the equivalent circuit (Figure 7) and the right-hand graphic is the discharge curve (Figure 8).

BillS

Gosh from an engineering perspective the information is great.

The original post was in search of greater field safety. How can the recent discussions be translated into greater field safety for everyone?

Bill

Hangtime

My Feedback: (19)

Bill, yah got me... I'm kinda slack-jawed myself. Like I think I mentioned earlier, the wealth of engineering expertiese available here is truly astounding. The good news is I'm learning something, bad news is, yer thread's been utterly hijacked.

Hang on tho.. something good may come from all this; possibly you can glean some of the up-thread discussion on checking a pack at the field into a short article for you clubs newsletter? Lotta good info from a bunch of posters.. could save some planes among the 'battery clueless' group.

For the 'battery reckless' croud, might I suggest yer club purchase a shot gun and a skeet launcher for the field saftey officer?

JohnMuchow

Bill, I think there are some great ways to distill some of this information down into practical recommendations good for everyone. We have already indentified two new sections to add to the Tech Tips section of our web site.

But, yea, we got pretty technical.

There's always stuff going on behind the scenes too. <hint hint>

BillS

Hangtime,

After much consideration the �two is better than one concept’ has a lot of merit especially as it relates to batteries.

Extrapolating the concept might lead one to postulate that: Two 125lb females are better that one 250lb female. It would appear that applying the concept to females is going to be an even more difficult and complex sell.

Someday Granddaddy will just have to get over the fact that he doesn’t like the fuel gage. Ironically and unfortunately those who most need voltage information least understand why. A battery less airplane coming through the pits is still dangerous.

After watching many great flyers yesterday at a warbird fly in the �voltage check drill’ seems even more ridiculous.
1. Stop the car.
2. Get out of the car.
3. Plug the gas gage in.
4. Rev the engine.
5. Check the gas gage.
6. Make decision.

On the drive to the event the automobile gas gage was probably checked several hundred times. Discussion later with a passenger revealed that he also checked my gas gage.

While waiting for manufacturers and AMA for a SAFETY wake up maybe someone will … Shot gun suits me but consensus among club members is difficult. Talkiewalkie’s maybe. Maybe flying a spare voltmeter.

JohnMuchow’s web site is a good read.
http://www.camlight.com
Thanks for your comments John.

Also unfortunately 'battery clueless' will get glossy eyed before completing the great read presented here. It is still a people issue.

Jyrki

I have been happy about battery capasity in my Sig Something Extra with standard servos. With a 1400mAh Nimh I can fly as much as I want in a day without going near empty. The first pack was 600mAh Nicd. That battery I flew until 4.2V!! or so without problems! The onboard led column corresponded quite well the battery condition. I think the discharge curve of NiCd is more linear than Nimh curve (?)

But now my heli with S9252 digital power eaters Is somehow troubled. Battery is 2500 Nimh and I have ran them empty TWICE!! Both time I was very lucky to have GV-1 shutting down throttle at 3.8V and I have had just enough power to make an autorotation landing. After a short rest, the voltage climbs up as ready to fly!! (but its not ready ofcourse). Because of the nonlinear discharge curve and cold weather here in Finland I plan to turn back to NiCd even if it has a shorter capasity per weight.

I got my lesson and now Im quite paranoid about battery maintenance. I very much appreciate idea of the original beginner of this thread. Battery maintenance should be developed. We should advance information about battery behavious (as here I have seen), help people select and use them correct way.

But also we should take seriously all efforts of developement and ideas around. There are practical battery maintenance products suitable to rc use, components like http://www.maxim-ic.com/quick_view2.cfm/qv_pk/4232 which I actually have here on my desk as samples from maxim All the cell phones and laptops have something like that. Rc batteries are really getting old-fashion. Theory of calculating the remaining charge is actually developed in a such a low price chip. I have just been too busy or lazy to wire it up. I know if I wait a couple of years, those will be integrated to best rx boxes. I would like very much if my throttle gets gradually slower when the battery drains ;o .. or any other indicator? ..plane waving ailerons "please land me soon, I have trouble"

For example the history of Rc heli gyro was began from a prototype of an enthusiast and now they are such high level products!

Jyrki

JohnMuchow

Thanks for you nice words Bill!
I think you nailed it with the above statement. One of the hardest things for us to decide is how much information to present when talking about cell care and discharging. We have a wild mix of customers, from first-timers to R/C and robotics to 20 year veterans and just one document/web page won't serve either audience well. Looks like, as we expand our Tech Tips section, that we'll be leaving an introduction/overview of each topic on the web page and get alot more detailed and technical in the PDF version.

This might be an option for local clubs...create a basic handout that describes stuff a newcomer on a very limited budget can do but also have a URL to a much more detailed web site or PDF/Word document. This will keep the newcomers from getting glossy-eyed but hopefully allow those with more interest to really dig into the more technical stuff.

I'm wondering if maybe we'd like to all get together (in another thread?) to create that basic handout to make available to anyone (or any club) who wants to use it? We'd be happy to put it up on our site (no ads or mention of CamLight in the document, just good info) if that would help get the word out. Worth doing as a first step?

BillS

I have a friend who is a good flyer and a very good builder and has very nice airplanes (much nicer than mine). Every time he is approached with any information about batteries he says: “I am not an electrician. I don’t want to be an electrician.” He will check the voltage fairly regularly.

It is difficult to convey even the smallest amount of understanding to my friend. He probably does have an intuitive sense that the battery charge is about used up at 5v.

backfire

Ref. Post #109

I'm surprised that nobody commented on the fact that, another way to know that the resistances I determined using the DC method wouldn't be valid for high currents is the fact that; to get 250 Amps from the battery I described, the battery would have to be shorted (no output). We all know that a fresh 2400 mAhr, 8-cell battery can supply 25 Amps with only 1-2 volts drop (5-10 milliohms per cell). I ran my DC resistance tests between 0.5 and 1.0 Amps which is too low (the Sanyo manual is somewhat vague in this respect, I think something is lost in the translation). I ran another test on one of the cells using higher currents as described in the Eveready method (JohnMuchow Post # 110). Indeed, the measured resistance was closer to the published AC method impedance value. I think, however, that the tests I ran at the lower currents reveal a characteristic of NiCds that are important for our Rx application. Exercising servos and watching the magnitude of battery voltage fluctuations never agrees with fluctuations calculated using the published impedances. It's clear now that for low currents (lightly loaded servos) that the apparent 'resistance' is much greater and causes this. The low current measurements include, among others, the effect of cell polarization as depicted in the cell equivalent circuit. As the servo currents rise, however, cell polarization effects diminish and the effective resistance becomes much lower. At this point, further voltage drop is reduced.

JNorton

My Feedback: (2)

This has been a very interesting and informative thread.

I for one am going to be a lot more careful using digital high (spike) draw servos. I will follow Hangtimes advise and use parallel NiCad packs to lower the impedance. I'll wire them following Red's advise for twin switches and plug both batteries into the receiver. If I ever get to 1/3 scale I'll use a seperate buss and spend the money on an in-flight recorder.

Isn't RCU great.

John