How do I make a 17 output relay sequencer with Picaxe

Gazza_AU

Member
PICAXE-28X2 Module and
AXE027 PICAXE USB Download Cable
Cost AU $20.69 each.
Total including shipping was AU $55.46

About half the price I paid for ZX80


Great, will check it out.

Have been looking for solid state alternative for switching 4V DC bidirectional loads of several amps or more.
I believe this might be possible with MOSFET, not something I know about

Found:
Four Channel 4 Route MOSFET Button IRF540 V2.0+ MOSFET Switch Module Arduino D
http://www.ebay.com/itm/3V-5V-Low-Control-High-Voltage-12V-24V-36V-E-switch-Mosfet-Module-For-Arduino-D-/112073228788?hash=item1a1815c1f4:g:zfcAAOSwqfNXmM-M

High-Power MOS FET Trigger Drive Switch Module PWM Adjust Control DC 4V ~ 60V
http://www.ebay.com/itm/High-Power-MOS-FET-Trigger-Drive-Switch-Module-PWM-Adjust-Control-DC-4V-60V-/112229864669?hash=item1a216bd4dd:g:88MAAOSwo4pYSLfc

3V 5V Low Control High Voltage 12V-24V-36V E-switch Mosfet Module For Arduino D
http://www.ebay.com/itm/3V-5V-Low-Control-High-Voltage-12V-24V-36V-E-switch-Mosfet-Module-For-Arduino-D-/112073228788?hash=item1a1815c1f4:g:zfcAAOSwqfNXmM-M

If anybody is up with MOSFET.
I would like to know more about the choices,
for High Current Low Voltage DC Switching
 
Last edited:

Circuit

Senior Member
Have been looking for solid state alternative for switching 4V DC bidirectional loads of several amps or more.
I believe this might be possible with MOSFET, not something I know about


If anybody is up with MOSFET.
I would like to know more about the choices,
for High Current Low Voltage DC Switching
Have you had a look at the PICAXE manual 3; page 8 on MOSFETs? http://www.picaxe.com/docs/picaxe_manual3.pdf
Also http://www.picaxe.com/Circuit-Creator/General-Outputs/FET/

...and there are many many posts on this forum giving MOSFET specifics. Wide range, what you are looking for are the so-called "logic level MOSFETs". These can be switched by the output of a PICAXE, mainly if you are using 5v PICAXE chips. Things get a little more complicated if you are running at 3.3 volts, but still possible. If you need to modulate the switching, fast pulsing as with PWM drives at high rates, then you need to consider MOSFET Drivers. If you just want on-off then this is relatively easy.
An STB80NF03L maxes out at 80 amps (given the correct usage) and can be switched directly by a 5volt PICAXE.
 

Gazza_AU

Member
Have you had a look at the PICAXE manual 3; page 8 on MOSFETs? http://www.picaxe.com/docs/picaxe_manual3.pdf
Also http://www.picaxe.com/Circuit-Creator/General-Outputs/FET/
An STB80NF03L maxes out at 80 amps (given the correct usage) and can be switched directly by a 5volt PICAXE.
Yes, using 5v PicAxe.
I am looking for a MOSFET output that is bidirectional,
A MOSFET arrangement that behaves like a standard relay with contacts.

I will by connecting between Cell Charger, 3.2v 100Ah Lifepo4 Cells and a large 25F Cap. (4v Max)

The Charger, Cell or 25F Super Capacitor will be connected and disconnected from each other at similar voltage.
But in the real world, things might not be so.

Started Vishay Siliconix Account, looking for suitable alternative Bi-directional MOSFET, to request a price
https://www.vishay.com/search?query=Si8900EDB&searchChoice=part

-----------------
Si8902AEDB $2.10 AUD each.
https://www.arrow.com/en/products/si8902edb-t2-e1/vishay?utm_source=octopart&utm_medium=buynow&utm_campaign=octopart
N-Channel 24 V (D-S) MOSFET
3.7 V
5.8A

• Battery protection switch
• Bi-directional switch
-------------------------------------

How would I go connecting a Picaxe output to 2x Si8902AEDB inputs.
http://static6.arrow.com/aropdfconversion/887a16365919e6527779c37f68e162b313730546/554341565919392si8902edb.pdf
Would I need anything between Picaxe Out and Mosfet gates?

Cheers,
Gary
 
Last edited:

premelec

Senior Member
Bi directional MOSFET circuitry isn't that simple - unless you can stand fairly large voltage drop - in such a case you could put the MOSFET as a bridge load and have to deal with the bridge diode drops... otherwise you'd need PMOS and NMOS. A better description of WHAT you are trying to drive would be helpful. Fountains are one thing as rail guns another. Possibly H bridges would work... don't know what you are up to...
 

Gazza_AU

Member
In general, I want to control DC solenoid valves and general logic systems with the Mechanical Relays.

I am purchasing 17x 3.2v Lifepo4 100Ah Cells for DC lighting.
They are to be connected in series and charged to a Max of 60v.

I do not want to go down the rabbit burrow of how cells should or could be used or charged.
Many have argumentative opinion on other forums, please don't bother going there.

Here is the Connection Logic using Mechanical Relays
Solar Charger Relays.jpg
Relays can be connected in a fail safe method, because they are SPDT with no voltage drop or resistance across contacts.

Here is a Connection Logic using MOSFET
Solar Charger2.jpg
I no nothing about MOSFET or their suitability. Best I can do to start.


Part of the experiment: was to put a voltage divider across the Bank Charger outputs whilst in operation.
This creates a reference for "1/17th the Bank Charging Voltage", or the average individual cell voltage in the string of 17.

Bank Voltage Range = 49.3v min to 60v max.
The hypothetical 1/17th bank voltage division would be 2.9v to 3.529v per Cell.

This 1/17th voltage is used to charge a 25F Super Capacitor, which is connected to each cell in sequence.
Note: the original code requirement and the gap between outputs in sequence.


In regards to voltage drop across switch:
If the voltage drop is fixed and known such as diode, it could be compensated for in divider/follower.
An unknown variable voltage drop is unworkable!

With relays, the settled voltage is what it is!
Bi-directional without resistance or voltage drop.

When I consider solid-state, the bi-directional effect is far to complex.
The alternative could be the exact same logic where the switching direction is only towards cells!
In this case the capacitor charge voltage would be increased to "compensate for a fixed voltage drop?"
Any cell below the average in series, would incur a small influx from cap.

There is the question of cap charge control, that I will leave for another day.

Regards,
Gary
 
Last edited:

AllyCat

Senior Member
Hi,

It's not clear to me whether you are planning to charge all 17 cells in series and then discharging one at a time, or charging each in turn and then discharging a "60 volt battery". But in either case, my concern is that you are potentially charging/discharging a stack of cells where some have different degrees of charge/discharge to others.

Maybe you have planned a sufficient level of individual cell monitoring, but I know to my own cost that if you accidentally over-discharge a single cell (even the relatively rugged LiFePO4 technology) in a chain of even just a few cells, then you will reverse bias and destroy the cell.

However, returning to your original question: personally I would definitely use relays in preference to that MOSFET arrangement. But I really think that you should be using a basic string of 17 cells, with a Switched-Mode up- or down-converter to handle the low-voltage charge or discharge requirement.

Cheers, Alan.
 

Gazza_AU

Member
Hi,

It's not clear to me whether you are planning to charge all 17 cells in series and then discharging one at a time, or charging each in turn and then discharging a "60 volt battery". But in either case, my concern is that you are potentially charging/discharging a stack of cells where some have different degrees of charge/discharge to others.

Maybe you have planned a sufficient level of individual cell monitoring, but I know to my own cost that if you accidentally over-discharge a single cell (even the relatively rugged LiFePO4 technology) in a chain of even just a few cells, then you will reverse bias and destroy the cell.

However, returning to your original question: personally I would definitely use relays in preference to that MOSFET arrangement. But I really think that you should be using a basic string of 17 cells, with a Switched-Mode up- or down-converter to handle the low-voltage charge or discharge requirement.

Cheers, Alan.
I really did not want to spend time explaining this, which is why I never offered this earlier.
People tend to read what they want or expect into things, you are so far off it is not worth correcting.
The information was provided to the questioner!

The part about using Relay is not overlooked.
Horses for courses

Gary

I have the information I came for, and on to the third code for functions.
 

premelec

Senior Member
In addition to Alan's comments - there are battery management and balancing systems developed for the very large battery packs [up to 70KWHr!] in electric automobiles - A number of integrated circuits devoted to just this job have appeared in the last decade. it's an interesting problem...
 

Gazza_AU

Member
In addition to Alan's comments - there are battery management and balancing systems developed for the very large battery packs [up to 70KWHr!] in electric automobiles - A number of integrated circuits devoted to just this job have appeared in the last decade. it's an interesting problem...
Yes, there are a few dedicated IC TLC3300 and one by Texas instruments that uses a fuel gauge chip.
These circuits get complex for long strings and require additional Monitoring.

If you understand PSOC from application rather then suggestion, it can be a powerful tool in increasing cycle life.
In household situations the load is variable, but in agricultural applications it can be fixed in steps.
Charging Voltage and Low Voltage Disconnect can be narrowed or widened to match the load.
Passive Dissipative balance systems do not work until the cell is mostly charged at about 3.65v.


The idea of Monitoring is not practical in a commercial workplace application.
But the companies are not interested in home builder systems with no monitoring.
A lot of this is driven by Grid Connect Systems and supplier contracts, particularly in Australia, don't know about UK.

The only way something will happen, is if the home builders find a solution.
One issue is voltage measurements need temperature compensation, and don't tell you much about capacity.
The Bank voltage divided by the number of cells, is available to follow any set charge voltage.
In this sense, it has the potential for automation.

The second aspect is the work that a Capacitor could do in cell balance.
The lower the cell below average, the greater the effect.

One way to look at this is connecting cells in parallel, there is no imbalance of cells and cycle life is increased.
You could also charge 17 cells in series with 17 separate fixed voltage chargers, there should be little imbalance.

Now consider a string of cells, the numbers are not important: Bank Voltage Range = 49.3v min to 60v max.
The hypothetical 1/17th bank voltage division would be 2.9v to 3.529v per Cell.

The idea is looking at charging cells in series, and intermittently with a 1/17th divisional per cell voltage.

Cell balance is not correcting or compensation for bad cells.
The cells need to be of good quality and well matched to start with.
The job of balancer with low charge voltages is to nudge cells towards balance.
Anything else, and we are talking another subject.

If you want capacity and fast discharge for 90% of applications, PSOC will make as much sense as Chinese.

Did you have anything to add about the use of Solid State components?

Cheers,
Gary
 
Last edited:

premelec

Senior Member
If i understand what you want to do - sort of flying capacitor - and your original comment on DPST rather than SPST relays I'll note that you can use n SPST relays with a single DPDT polarity reversing relay to capacitor... With careful timing ;-0
 

Gazza_AU

Member
If i understand what you want to do - sort of flying capacitor - and your original comment on DPST rather than SPST relays I'll note that you can use n SPST relays with a single DPDT polarity reversing relay to capacitor... With careful timing ;-0
Tell me more?

The original idea had a flying capacitor whose sole voltage supply came from the cells it was meant to balance.
With a High cell followed by a low cell the drain could continue on the 3rd cell connected.
This idea aims to resolve this issue with floating reference voltage into capacitor, and extended square wave output.
Giving the capacitor time to charge and re-stabilize, between each cell connection.
Once the Capacitor is charged the current flow should not be great, but good tolerances are affordable.

I have been thinking about what Stan said, except that one relay delays its partner in pairs, still work in progress.
Maybe wire the relays so the 1st turns on the next which turns off the previous and it would ripple through the series.
 
Last edited:

premelec

Senior Member
@eclectic - keeper of the vaults - that was only 9 years ago ;-0 and still relevant... reminds me of when I was younger... :-0

@Gazza if you take two points and switch up the line of batteries the polarity of the two point leads reverses each time - so if you put a DPDT reversing relay onto the two points that will reverse the polarity to be correct - and actuate the relay every other time... BEFORE you actuate the 2 spst relays to contact across the nth battery. Timing is essential... no smoke wanted. If need be i can draw a diagram...

In case it's not obvious you can sequence the SPST relays to jump to every other cell so you don't change the polarity relay until every other cell is contacted and then go back and start one cell higher... The big thing is to never connect the wrong polarity! A fuse or breaker in series with the flying capacitor would be appropriate...
 
Last edited:

Gazza_AU

Member
@eclectic - keeper of the vaults - that was only 9 years ago ;-0 and still relevant... reminds me of when I was younger... :-0

@Gazza if you take two points and switch up the line of batteries the polarity of the two point leads reverses each time - so if you put a DPDT reversing relay onto the two points that will reverse the polarity to be correct - and actuate the relay every other time... BEFORE you actuate the 2 spst relays to contact across the nth battery. Timing is essential... no smoke wanted. If need be i can draw a diagram...

In case it's not obvious you can sequence the SPST relays to jump to every other cell so you don't change the polarity relay until every other cell is contacted and then go back and start one cell higher... The big thing is to never connect the wrong polarity! A fuse or breaker in series with the flying capacitor would be appropriate...
I am interested in seeing a diagram because I cannot see how an arrangement as described can work to an ends.

I envisage the 1/17th voltage remaining connected to the capacitor.
Intermittent Connection to cells, to prevent any lagging effect on the Capacitor, effecting the next cell.

If the 1/17th reference is set a few Milli-volts higher then a sag adjusted cap, each cell could see higher then 1/17th reference.
In this experiment, the idea was to deliberately increase this reference voltage slightly higher.

Things could be done differently, draw us a diagram.

Cheers
 
Last edited:

premelec

Senior Member
Diagramatic Rlays

My sketch of relays is attached - the capacitor would go on one side of the DPDT polarity reversing relay - I may have a different idea than you as I have worked on this before and am prejudiced... ;-0 In any case this sort of thing is subject to serious possible failures if a relay sticks- welds contacts etc... Note that the cheap relay boards are only rated at 10 amp contacts. However as a trial run you could put a small resistor [piece of wire] in series with each contact...
 

Attachments

Last edited:

AllyCat

Senior Member
Hi,

Ah, so the idea is just for balancing the cells? Sorry, I'd rather lost track of the thread and there appear to have been a few edit/additions.

A great advantage of relays is that their input (coil) and switch (contacts) are completely isolated from each other. So all your coils can be driven in a similar way from the same (low) voltage. But a disadvantage of relays is that they have a finite lifetime and their contacts can be damaged (or even destroyed by welding) if their current or voltage ratings are exceeded. I hope you are including a current-limiting resistor in series with your supercap.

Conversely, a problem with FETs is that the gate (control) voltage must be similar to their Source voltage, so each of your stages would be driven at a different bias voltage. That would need a small (low power) transistor, FET or optocoupler to interface between the PICaxe and each Gate. I can't comment on those particular FETs which appear rather "unusual" (specifically intended for battery protection circuits?). Their schematic symbol shows a diode in series with each Drain, but they can't actually be "built" like that, because the voltage drop would be far too high.

I do agree with the aim to choose your own "ideal" cell voltage operating range. With my own very modest solar/LiFePO4 experiments I soon came to the conclusion that there was little point in charging above 3400 mV because any additional energy stored was quite small, at the expense of potentially reduced lifetime. Similarly I set the "alarm" at 3200 mV in my simple PICaxe 08M2 implementation, which is able to resolve to within a few mV (some details in the code snippetts section).

So the idea is to transfer energy between cells using a capacitor? Have you calculated the amount of energy actually possible? If the voltage difference is only 100 mV or so, then the change in energy, even in a 25 F capacitor isn't very large. Assuming a 0.1 ohm current-limiting resistor, the time constant is 2.5 seconds, so the relays are going to click away quite a lot? Even more so if the voltage and resistance are lower.

Cheers, Alan.
 

Gazza_AU

Member
You are using one relay to switch the capacitor between charge voltage and cells.

The remaining relays used to select the cell number.

Are you trying to sequence the output, there is an easier way if you just want the hopping capacitor with out reference charging?
 

Gazza_AU

Member
Hi,

Ah, so the idea is just for balancing the cells? Sorry, I'd rather lost track of the thread and there appear to have been a few edit/additions.

A great advantage of relays is that their input (coil) and switch (contacts) are completely isolated from each other. So all your coils can be driven in a similar way from the same (low) voltage. But a disadvantage of relays is that they have a finite lifetime and their contacts can be damaged (or even destroyed by welding) if their current or voltage ratings are exceeded. I hope you are including a current-limiting resistor in series with your supercap.

So the idea is to transfer energy between cells using a capacitor? Have you calculated the amount of energy actually possible? If the voltage difference is only 100 mV or so, then the change in energy, even in a 25 F capacitor isn't very large. Assuming a 0.1 ohm current-limiting resistor, the time constant is 2.5 seconds, so the relays are going to click away quite a lot? Even more so if the voltage and resistance are lower.

Cheers, Alan.
There are a lot of welded on parrots when it comes to cells and charging, but you only need the fundamentals without selling pitch to work it out

No problem, there have been a lot of edits, trying to keep it condensed enough to comprehend.
And I am getting pretty grumpy after studying this shit for so long.

My skills are in my hands and logic, your assistance with capacitors is very welcomed.
What do you mean in time constant for 0.1 ohm current-limiting resistor?
And the reference to relays clicking away.
I was thinking the square wave period would be in the order of 5 seconds on and off.

If your calculation is for charging a discharged capacitor to capacity in 2.5 seconds, the capacitor might need to be larger?
And or the charge rate lower.
I was looking at dedicated Capacitor IC's used in Devices with peak current above their cells capability.
This would be bordering on over complex, if resistors could be used.

The question is what capacitor will the switches handle?
It would be difficult to calculate the required charge rate, what is the capacitors state of charge before hand, then what is the time frame we aim for?
What we could predict is the extent of any capacitor discharge to a cell below the reference value.
I would not expect high current loads in this arrangement.

Here is the Bank Divider/Follower Logic to charge capacitor:
BankDividerFollower.jpg
I am no circuit designer, feel free to draw something up with real values or suitable components.

What sort of math could be used for rough approximation?
Operating range of 2.9v to 3.529v per Cell.
Cheers
 
Last edited:

premelec

Senior Member
Alan and I have both considered this sort of thing in detail - but I don't know if your idea is something different - a capacitor can be used as a charge bucket carried around to fill or empty depending on voltage difference - basic equation for capacitors is Q=CV=IT where Q = charge, V = Volts, I = amperes, T = Seconds, C = farads. The current will depend on voltage difference and circuit resistance - ohm's law.

I don't know if you looked at the LTC6802 data sheet I ref'ed 9 years ago - it's interesting... and expensive...
 
Last edited:

AllyCat

Senior Member
Hi,

The current which flows in a capacitor is almost totally dependent on the value of the Resistance connected in series with it and the Voltage across that resistor (Ohms Law, I = V/R). With no resistance, there's a fair chance that any relay contacts connecting the capacitor (to another low-impedance component like a battery) will get welded together by excessive current.

The "Time Constant" is calculated simply by the product of the Capacitance and the Resistance; we don't often have the luxury of working directly in the base units of Farads and Ohms, but here we do, so the TC might be 25 x 0.1 = 2.5 seconds. It gives us a "feel" for how long things might take to happen (or to stop happening). It's the time that it would take the capacitor to charge or discharge to its "final" voltage (in the case of a resistor directly across a capacitor the final voltage would be zero) if the current continued at its initial value. But as soon as the voltage changes, the current falls, so the voltage doesn't reach its final value in one time constant (or theoretically ever). It's basically the same concept as the "half life" of a radioactive element.

In the case of a R-C network the voltage actually gets to 63% of the "final" value in one TC and about 99% in five TCs (at which point we can usually assume that nothing much more useful is going to happen). But my main "concern" here (if I understand the idea correctly) is that if the capacitor starts with almost the same voltage as any particular cell, then very little current will flow and not much "balancing" is going to happen.

Of course the capacitor might be charged to a higher voltage from "somewhere else", but then I don't see that it's very different from any normal (resistive) balancing arrangement.

Cheers, Alan.
 

premelec

Senior Member
I was aiming to track cell voltages both charging and discharging with no concern to voltage ref which would be a separate issue...

The current max will depend on voltage and resistance - the wiring resistance and supercap internal resistance [.03 ohm?] in series.
If you have .1 ohm and 10 amp that implies maximum voltage of R x I = 1 volt. if only the cap internal resistance then .03 x 10 or 300mv.

A bigger issue is what the timing of the switching is - in short how big a voltage drop do you get in any cell in X seconds at what current - then this gives you an estimate of how much charge the capacitor must store [Q=IT]. No simple answer - empiricism required... The faster you switch the cap around the less voltage difference and hence current - and relay noise / wear... Another approach is to 'watch' the cell voltages and only start cap flying routine when some particular difference in cell voltages is measured. This could be done with separate circuitry or slow cycling and measuring the cap voltage and speed up when it becomes larger [storing sampled lowest and highest values being required to make the decision].
 

Gazza_AU

Member
#Capacitor calculations at 3.529v
Note: I am calculating based on an unlimited current supply to Capacitor before the resistor.
Because my Bank Charger is 50 Amps max, this approach should be conservative!
-

To choose Capacitor size based on relay capability, this data sheet might help:
https://www.sparkfun.com/datasheets/Components/TS12S-R.pdf

Single 30F Capacitor Specs
Rated voltage 2.5VDC
Surge Voltage 3.0VDC
Rated Capacitance 30F
Rated Current 5300 (3.5A continuous)
Max. Current 16800 (16.8A Peak pulse at 3.0VDC)
Max. Internal Resistance 31 (3A)

All ratings below have been calculated for 2 Series Caps
15F - 2x 30F Capacitors in Series?:
Rated voltage 5VDC
Surge Voltage 6VDC
Rated Capacitance 15F
Rated Current 2650 (2.65A continuous)
Max. Current 8400 (8.4A Peak pulse at 6.0VDC)
Max. Internal Resistance 62
-

45F - 2x 90F Capacitors in Series?:
Rated Current 7600 (7.6A continuous)
Max. Current 23800 (23.8A Peak pulse at 6.0VDC)
Max. Internal Resistance 42
-

60F - 2x 120F Capacitors in Series:
Rated Current 10500 (10.5A continuous)
Max. Current 33500 (33.5A Peak pulse at 6.0VDC)
Max. Internal Resistance 42
-

50F - 2x 100F Maxwell "BCAP0100 T01" Capacitors in Series:
http://www.mouser.com/ds/2/257/Maxwell_HCSeries_DS_1013793-9-341195.pdf
Rated Current 5A
Max. Current 18A
Max. ESR 30 mΩ
Note: ESR acts like a resistor in series with a capacitor (thus the name Equivalent Series Resistance).



Working off the 50F - 2x 100F Maxwell Series calculations
What size Resistor would limit inrush into the fully discharged Capacitor,
to < then peak pulsed current, for the approximated relay rating at 3.529v?

Relay Board Module Optocoupler LED for Arduino PiC ARM AVR
Relay Data Sheet: https://www.ghielectronics.com/downloads/man/20084141716341001RelayX1.pdf
Note: SRD-05VDC-SL-C relay ends with "C" for Form C.
Making these "Ebay Board Module" relays Contact Ratings actually: 7 Amp Resistive or 3 Amps Inductive at 28V DC.

Volts x Amps = Watts
28v x 7 Amp = 196W DC
3.529v x 55.5A = 196W DC, (is this correct?)

3.529v / 50Amp = 0.07058 ohm resistor.

#Charging Time Constant at 50F 3.529v
50F x 0.07058 Ohms = 3.529 Sec



There are many things that can be done to improve your passive capacitor version.
Lets help each other to test all these ideas!
 
Last edited:

Gazza_AU

Member
Looking at the diagram, I don't think a simple shift register is the answer.
A shift register will turn the 'next' bit on at the same time as turning the 'current' bit off, with no gap between.
The diagram shows that between each pulse there is a 'dead time' with no pulse active.

A counter going from 0 to 33, with output pins controlled by counts at 0,2,4,6 .... would give the required pattern, with a couple of pauses to determine pulse width and 'dead time'.

There a plenty other ways, but this is probably easiest to code.

Something like this, it should run on any PICAXE that has enough pins ....
Code:
[color=Navy]#picaxe [/color][color=Black]28X2[/color]


[color=Blue]symbol pulsetime  [/color][color=DarkCyan]= [/color][color=Navy]1500[/color]
[color=Blue]symbol gaptime    [/color][color=DarkCyan]= [/color][color=Navy]1300[/color]

[color=Blue]do
      if [/color][color=Purple]b0 [/color][color=DarkCyan]=  [/color][color=Navy]0 [/color][color=Blue]then pulsout a.0[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif    
      if [/color][color=Purple]b0 [/color][color=DarkCyan]=  [/color][color=Navy]2 [/color][color=Blue]then pulsout b.7[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]=  [/color][color=Navy]4 [/color][color=Blue]then pulsout b.6[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]=  [/color][color=Navy]6 [/color][color=Blue]then pulsout b.5[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]=  [/color][color=Navy]8 [/color][color=Blue]then pulsout b.4[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]10 [/color][color=Blue]then pulsout b.3[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]12 [/color][color=Blue]then pulsout b.2[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif    
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]14 [/color][color=Blue]then pulsout b.1[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif    
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]16 [/color][color=Blue]then pulsout b.0[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]18 [/color][color=Blue]then pulsout c.7[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]20 [/color][color=Blue]then pulsout c.6[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]22 [/color][color=Blue]then pulsout c.5[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]24 [/color][color=Blue]then pulsout c.4[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]26 [/color][color=Blue]then pulsout c.3[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif    
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]28 [/color][color=Blue]then pulsout c.2[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]30 [/color][color=Blue]then pulsout c.1[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif
      if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]32 [/color][color=Blue]then pulsout c.0[/color][color=Black], [/color][color=Blue]pulsetime [/color][color=Black]: [/color][color=Blue]pause gaptime [/color][color=Black]: [/color][color=Blue]endif    

      inc [/color][color=Purple]b0
      [/color][color=Blue]if [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]34 [/color][color=Blue]then
            [/color][color=Purple]b0 [/color][color=DarkCyan]= [/color][color=Navy]0
      [/color][color=Blue]endif
loop[/color]

Cheers,

Buzby


Unique Combination Sequencer - Gazza_AU
Grid contains every combination of 17 outputs meeting with, or from another output.
17 rows of sequences read from left to right;
0,1,0,2,0,3,0,4,0,5,0,6,0,7,0,8,0,9 ETC.
_ 1,_ 2,1,3,1,4,1,5,1,6,1,7,1,8,1,9 ETC.
_ _ _ 2,_ 3,2,4,2,5,2,6,2,7,2,8,2,9 ETC.
There are &#8220;if&#8221; and &#8220;skip&#8221; rules to prevent the same output combination happening more than once within the 16 rows of sequential numbers.

Unique Combination Sequencer - Gazza_AU.jpg

Code:
#picaxe 28X2
symbol pulsetime = 1000
symbol gaptime = 500
main:

if b0 = 0 then pause gaptime : pulsout a.0, pulsetime : pause gaptime : endif
if b0 = 1 then pause gaptime : pulsout a.1, pulsetime : pause gaptime : endif
if b0 = 2 then pause gaptime : pulsout a.2, pulsetime : pause gaptime : endif
if b0 = 3 then pause gaptime : pulsout a.3, pulsetime : pause gaptime : endif
if b0 = 4 then pause gaptime : pulsout a.4, pulsetime : pause gaptime : endif
if b0 = 5 then pause gaptime : pulsout b.0, pulsetime : pause gaptime : endif
if b0 = 6 then pause gaptime : pulsout b.1, pulsetime : pause gaptime : endif
if b0 = 7 then pause gaptime : pulsout b.2, pulsetime : pause gaptime : endif
if b0 = 8 then pause gaptime : pulsout b.3, pulsetime : pause gaptime : endif
if b0 = 9 then pause gaptime : pulsout b.4, pulsetime : pause gaptime : endif
if b0 = 10 then pause gaptime : pulsout c.0, pulsetime : pause gaptime : endif
if b0 = 11 then pause gaptime : pulsout c.1, pulsetime : pause gaptime : endif
if b0 = 12 then pause gaptime : pulsout c.2, pulsetime : pause gaptime : endif
if b0 = 13 then pause gaptime : pulsout c.3, pulsetime : pause gaptime : endif
if b0 = 14 then pause gaptime : pulsout c.4, pulsetime : pause gaptime : endif
if b0 = 15 then pause gaptime : pulsout c.5, pulsetime : pause gaptime : endif
if b0 = 16 then pause gaptime : pulsout c.6, pulsetime : pause gaptime : endif


if b0 = 16 then skip

if b0 = 0 then skip0
if b1 = 0 then pause gaptime : pulsout a.0, pulsetime : pause gaptime : endif
skip0:

if b0 = 1 then skip1
if b1 = 1 then pause gaptime : pulsout a.1, pulsetime : pause gaptime : endif
skip1:

if b0 = 2 then skip2
if b1 = 2 then pause gaptime : pulsout a.2, pulsetime : pause gaptime : endif
skip2:

if b0 = 3 then skip3
if b1 = 3 then pause gaptime : pulsout a.3, pulsetime : pause gaptime : endif
skip3:

if b0 = 4 then skip4
if b1 = 4 then pause gaptime : pulsout a.4, pulsetime : pause gaptime : endif
skip4:

if b0 = 5 then skip5
if b1 = 5 then pause gaptime : pulsout b.0, pulsetime : pause gaptime : endif
skip5:

if b0 = 6 then skip6
if b1 = 6 then pause gaptime : pulsout b.1, pulsetime : pause gaptime : endif
skip6:

if b0 = 7 then skip7
if b1 = 7 then pause gaptime : pulsout b.2, pulsetime : pause gaptime : endif
skip7:

if b0 = 8 then skip8
if b1 = 8 then pause gaptime : pulsout b.3, pulsetime : pause gaptime : endif
skip8:

if b0 = 9 then skip9
if b1 = 9 then pause gaptime : pulsout b.4, pulsetime : pause gaptime : endif
skip9:

if b0 = 10 then skip10
if b1 = 10 then pause gaptime : pulsout c.0, pulsetime : pause gaptime : endif
skip10:

if b0 = 11 then skip11
if b1 = 11 then pause gaptime : pulsout c.1, pulsetime : pause gaptime : endif
skip11:

if b0 = 12 then skip12
if b1 = 12 then pause gaptime : pulsout c.2, pulsetime : pause gaptime : endif
skip12:

if b0 = 13 then skip13
if b1 = 13 then pause gaptime : pulsout c.3, pulsetime : pause gaptime : endif
skip13:

if b0 = 14 then skip14
if b1 = 14 then pause gaptime : pulsout c.4, pulsetime : pause gaptime : endif
skip14:

if b0 = 15 then skip15
if b1 = 15 then pause gaptime : pulsout c.5, pulsetime : pause gaptime : endif
skip15:

skip:


inc b0
if b0 = 17 then seq
goto main
seq:

inc b1
if b1 < 17 then seq2
Let b1 = 0
Seq2:

Let b0 = b1
goto main
 

premelec

Senior Member
As a caution I would always issue ALLoFF to relays connected to batteries before ANY relay ON command and use small gauge wire that will fuse rather than cause battery failure if inadvertent short occurs. Technically PULSOUT works fine... however i'd make sure it starts with the right initial condition as well turning off all relays... hope it works...

"The pulsout command generates a pulse of length time. If the output is initially low, the pulse will be high, and vice versa. This command automatically configures the pin as an output, but for reliable operation you should always ensure this pin is an output before using the command (using high, low or output)."
 
Last edited:

Gazza_AU

Member
As a caution I would always issue ALLoFF to relays connected to batteries before ANY relay ON command and use small gauge wire that will fuse rather than cause battery failure if inadvertent short occurs. Technically PULSOUT works fine... however i'd make sure it starts with the right initial condition as well turning off all relays... hope it works...

"The pulsout command generates a pulse of length time. If the output is initially low, the pulse will be high, and vice versa. This command automatically configures the pin as an output, but for reliable operation you should always ensure this pin is an output before using the command (using high, low or output)."

The code is offered as a going concern for anybody wanting to experiment with Passive Balance Capacitor systems.
The logic graph is offered in case someone can reduce the algorithm and improve code efficiency/timing

I cannot see any benefit to changing the code in the aspect you suggest, the default start condition is a logical reset and always starts with pause.

Standard chaser sequence on passive Capacitor:

Low cell can only be topped through multiple Cap hops from a high cell that could be many cells away in sequence.
This Algorithm solves that problem by sharing every cell, with every other cell in sequence from both directions.

I could increase the efficiency a little with some modification.
If you have ideas or improvements, modify the code and re-post.
-


Working off the 50F - 2x 100F Maxwell Series calculations
What size Resistor would limit inrush into the fully discharged Capacitor,
to < then peak pulsed current, for the approximated relay rating at 3.529v

Relay Board Module Optocoupler LED for Arduino PiC ARM AVR
Relay Data Sheet: https://www.ghielectronics.com/downloads/man/20084141716341001RelayX1.pdf
Note: SRD-05VDC-SL-C relay ends with "C" for Form C.
Making these "Ebay Board Module" relays Contact Ratings actually: 7 Amp Resistive or 3 Amps Inductive at 28V DC.

Volts x Amps = Watts
28v x 7 Amp = 196W DC
3.529v x 55.5A = 196W DC.

3.529v / 50Amp = 0.07058 ohm resistor.

#Charging Time Constant at 50F 3.529v
50F x 0.07058 Ohms = 3.529 Sec
-

Hi,
my main "concern" here (if I understand the idea correctly) is that if the capacitor starts with almost the same voltage as any particular cell, then very little current will flow and not much "balancing" is going to happen. Of course the capacitor might be charged to a higher voltage from "somewhere else", but then I don't see that it's very different from any normal (resistive) balancing arrangement.
Cheers, Alan.
What you say is true, the capacitor must be large enough to extend discharge.
The voltage has to be high enough above a low cell to extend the time it remains so.
I have further logic to add, once the divider/capacitor are sorted.

As a piece of research equipment, my interest is to use the biggest Capacitor that is realistically possible.
I apologize to yourself and anybody else I may have been a bit short with.
I started this project from the ground up based on cell characteristics and my application.
This application is unique, so much so that countless people across forums have replied with preconceived ideas relevant to their design need.
This is great, but the energy and time spent explaining that my system is unique and why, takes up 10 pages.
Often I leave with no relevant answers after spending hundreds of hours studying and sharing the details.
You have also helped, thank you.
-

If this is correct and the relay can handle 50Amps at 3.529v?
This condition can only occur once each cycle in normal use!

What is the largest capacitor to meet this condition with maximum effect?
I am thinking 100 or even 200F might be better?

What size capacitor bank in Farad would you choose, and why?
 
Last edited:

Gazza_AU

Member
Relay Board Module Optocoupler LED for Arduino PiC ARM AVR
Relay Data Sheet: https://www.ghielectronics.com/downloads/man/20084141716341001RelayX1.pdf
Note: SRD-05VDC-SL-C relay ends with "C" for Form C.
Making these "Ebay Board Module" relays Contact Ratings actually: 7 Amp Resistive or 3 Amps Inductive at 28V DC.

Volts x Amps = Watts
28v x 7 Amp = 196W DC
3.529v x 55.5A = 196W DC.


Which means I am looking for Capacitor bank rated at 50Amps.
50 Amps is also the charge rating for 100Ah cells!
That sounds like a big capacitor, where do we find 2 capacitors rated at 100Amps?

600 Farad Capacitor XV3585-2R7607-R
2.6 ESR m&#937;
33 Amp current rating
320Amp peak

AUD $75 each, makes the decision easier!
So it looks like a mater of shopping around for the best deal
 
Last edited:

AllyCat

Senior Member
Hi,

I'm in no way an expert on Lithium cell technology (but the forum has at least one member who is) so I won't say any more on that subject, particularly as I still don't know/understand what you're trying to achieve. However, at the currents that you're now discussing, I believe you should be considering the use of FETs not relays.

But I do have some experience in interpreting data sheets, which rarely tell you what you really want to know (and are sometimes written more as an "advertisement" than to impart serious technical information). So, for the relay data sheet that you've linked:

The maximum (initial) contact resistance is quoted as 100 milliohms. I've no idea how realistic that is (or how it may change over life), but it represents 1 volt drop at 10 Amps. That's also 10 watts, so I hope it doesn't last too long.

The minimum "Electrical Life Expectancy" is specified at 10^5 operations at "nominal coil voltage" (sic). I presume that's an error for nominal contact current, but perhaps they're implying that the drive (and thus the switching speed) must be as specified. However, much the same lifetime seems to be indicated in the adjacent graphs. If the relay operates once per minute, then 10^5 corresponds to just under 3 months of operation.

"Large" capacitors generally have a wide manufacturing tolerance and one of your linked data sheets specifies -20% + 80%, which is a ratio of over 1:2 (and there's also a reference to 30% change over life). So if you are "unlucky", two capacitors in series won't "share" the applied voltage, but one could see around twice that on the other (e.g. 2.4 volts : 1.2 volts). That looks to be only just within tolerance for a purely passive LiFePO4 balancing circuit (but perhaps the higher voltage capacitor will just become "leaky" and help balance their voltages?). Also, one of the data sheets appeared to indicate a "lifetime" also of 10^5 (full) cycles, which no-one would expect from a "normal" capacitor. With my only foray into using (much smaller) "supercaps", I was really disappointed with their series resistance, so high that they were completely useless for my applicaton.

Finally, the basic calculations are really rather simple, just Ohms Law (to indicate the current and/or the acceptable resistance, depending on which is "known") and the Time Constant (to indicate appropriate cycle times). Both of these need to "know" the total Effective Series Resistance: That will be the sum of the resistances of the Cell, the "Switch" (contacts/FET), the Capacitor and all the wiring (assuming that there are no nasty plugs and sockets as well). As I've already said above, I would expect the cells "normally" (or ideally) to be operating always in a range of 3.2 to 3.4 volts (for good lifetime).

Cheers, Alan.
 

Gazza_AU

Member
Hi,
So, for the relay data sheet that you've linked:

The maximum (initial) contact resistance is quoted as 100 milliohms. I've no idea how realistic that is (or how it may change over life),
but it represents 1 volt drop at 10 Amps. That's also 10 watts, so I hope it doesn't last too long.

The minimum "Electrical Life Expectancy" is specified at 10^5 operations at "nominal coil voltage" (sic). I presume that's an error for nominal contact current, but perhaps they're implying that the drive (and thus the switching speed) must be as specified. However, much the same lifetime seems to be indicated in the adjacent graphs. If the relay operates once per minute, then 10^5 corresponds to just under 3 months of operation.

Cheers, Alan.
In regards to Lifepo4 cells, and only if you want to understand beyond plugin and use:=
I seriously suggest you ignore everything you have been told or gleamed for universal use.

Download all the Scientific PDF papers you can find and read.
Or just do and believe what benefits you.
Much to do, and I don't intend to offer my knowledge on cells here.

The other points you make are valid, and a lot of work has already been done.

The relay Specifications are a real concern!
Starting with the voltage drop across contacts.

Pretty sure you would understand that the current will be a lot less then 10Amps after initial capacitor charge, once per start up?

Missed this: PERFORMANCE (at initial value) when ever something is not easy to understand, something is hidden, because someone smart enough to design a good relay knows good English. Thats been my approach with Sellers, and working out quite well.

Going by the data, you are saying that for each set of contacts in series "a voltage of 10 amps passes" it would lose 1v?
So 10v 10Amps would be zero after passing 10 sets of relay contact?
That is how I would read this!

3 months life span switching once per minute, sounds like the product should not be on the market.
Not much to say, unless someone else whats to chip in?

I am interested in what FET devices you would use for source and sink.
This test aims to top cells as the primary function.
So could work in only one direction.
The bidirectional approach may be better, but simplicity for more reliability rules!

Share a link to specs so we can discuss
 
Last edited:

Gazza_AU

Member
@AllyCat

Allan, think I might know why you are harping on the cell thing.
It doesn't always pay to take things to literal, or read extra into things, I make the same mistakes.

If I explain every last detail, my fingers fall off, people go to sleep, and 19 years later the posters question is unanswered.

2.9v to 3.529v is not the voltage I charge my cells.
It is the range that I want the balance designed for.
Add: As a piece of test equipment, it is what it turns out to do.

All good,
gary
 

AllyCat

Senior Member
Hi,

you are saying that for each set of contacts in series "a voltage of 10 amps passes" it would lose 1v?
So 10v 10Amps would be zero after passing 10 sets of relay contact?
No I am most definitley NOT saying that. And certainly not that the unit of Voltage is Amps. :eek: :

The manufacturer states in his data sheet that the Maximum contact resistance (of a new device) is 100 mohms. That means that if 10 Amps flows through the contacts then the Voltage Drop might be up to 1 volt. That's what the data sheet tells us; any further inference (except that 10 Amps through 0.1 ohm represents 10 watts of instantaneous power) is "guesswork", which as an engineer I have always preferred to avoid.

10^5 is a perfectly acceptable number of cycles for many applications (particularly mechanical); it happens to be also the minimum specification for writing to the EEPROM memory cells in a PICaxe. It's up to the system designer to work with (or around) the value, or choose an alternative component/technology.

It was you who mentioned currents aroiund and above 10 amps. My "concern" was more that if you are balancing cells to within a few mV (which IMHO is quite practical with a simple "active" circuit if you wish to do so) then the currents which flow though a resistance of a few hundreds of mohms can be only a few tens of mA. Time-shared over 17 cycles (or worse 17^2 cycles), the net (average) current flow isn't going to be very much at all.

Cheers, Alan.
 

Gazza_AU

Member
Hi,


No I am most definitley NOT saying that. And certainly not that the unit of Voltage is Amps. :eek: :

The manufacturer states in his data sheet that the Maximum contact resistance (of a new device) is 100 mohms. That means that if 10 Amps flows through the contacts then the Voltage Drop might be up to 1 volt. That's what the data sheet tells us; any further inference (except that 10 Amps through 0.1 ohm represents 10 watts of instantaneous power) is "guesswork", which as an engineer I have always preferred to avoid.

10^5 is a perfectly acceptable number of cycles for many applications (particularly mechanical); it happens to be also the minimum specification for writing to the EEPROM memory cells in a PICaxe. It's up to the system designer to work with (or around) the value, or choose an alternative component/technology.

It was you who mentioned currents aroiund and above 10 amps. My "concern" was more that if you are balancing cells to within a few mV (which IMHO is quite practical with a simple "active" circuit if you wish to do so) then the currents which flow though a resistance of a few hundreds of mohms can be only a few tens of mA. Time-shared over 17 cycles (or worse 17^2 cycles), the net (average) current flow isn't going to be very much at all.

Cheers, Alan.
Could we talk about my question relating to FET and what you would use?
I can answer your questions if you can acknowledge mine!

There is no guarantee that this experiment will do anything, perhaps create some smoke :)

There are two common types of balance method, dissipation of high cells, or active non-dissipative.
Dissipative through resistors only works when the variation in cell voltages reaches a certain level.
Even if Dissipation could be done continuously, it makes no sense to do so, for obvious reasons.

By the time you can start balancing, the cells are close to full capacity.
This factor also reduces the time left to balance.
This is why Dissipative balance is not used in true PSOC.

I have spoken to manufactures of banks that limit cell voltage to 3.6v for increased cycle life, and the balance system supplier.
No surprise that this 3.6v setting is a limitation of the Dissapative balance system.
It is no problem to discharge to 2.9 or even higher if you don't need the capacity.
Likewise you can also charge to even lower cell voltages for what ever envelope is required.

Regardless of the outcome of this experiment, I will install active balance and it may need to be Taylor made.
There is a business behind this experiment that continues regardless of the balance investigation.

So active balance has the potential to suit true PSOC application.
The Linear/Analogue product with fly-back transformer could be used if I can modify the 12 cell plans, this would cost a lot of money.
Alternatively I would have to design the PCB from scratch, the parts are already listed, but building such a large PCB system from novice skill set is a concern!

Their system is difficult to understand, the Engineer was not helpful in saying the controller needs to tell the Monitor system what to do.
The company wants to promote its mates in pcb design rather then supply product support and promote sales.

In short, you cannot compare dissipative with active.
Even Active systems might have limitations on minimum voltage at which it begins.
Or there control system might not run independent, so all settings can be made from efficient solar controller.

So now we come to your question, or concern about how much such a capacitor could do?
As it stands, you would have to say very little.
The plan is to maintain a 1/17 baseline reference through the design process, and include gain adjustment for testing.

Some claim Active balance has to work both ways, which is arguably true based on logic and efficiency.
But this is not the problem I am trying to solve.

The question is, what balance system would do any balancing through the middle of a charging cycle?
Where can I buy it, and what does it cost?
I should add that this lighting system is otherwise extremely efficient from electrical stand point.
Also, reducing the charge and discharge levels increases the life cycle in oversized systems dramatically.

I think the proposed modification of passive capacitor is worth trying.
That is not to say a passive or charged capacitor system will work for my application.
Clipping high cells with resistive dissipation is sort of the opposite of topping cells with a higher voltage from Capacitor.

I would like to discuss what FET or alternative component you would use to replace relays.
Many people suggest using FET based on the fact it is done, rather then understanding how!

Only have one way flow from cap to cell.

Cheers
 
Last edited:

AllyCat

Senior Member
Hi,

It's not MY project, I'm just offering a little (free) advice (which others seem to have stopped doing). ;)

I've hardly ever used FETs myself and am certainly no expert. The type/rating will depend whether you want to use them "locally" for a single cell where the ~3 volts will require a very low Threshold voltage (perhaps like you linked earlier) or from the PICaxe supply (where a "Logic Level" FET will do). But if you want it to replace the isolation provided by a relay, then you will need a 60+ volt FET and suitable drive circuitry.

There was a recent thread where it was said that it's not possible to determine the state of charge (a "fuel gauge") from a Lithium cell's terminal voltage. That would imply that you could only balance when the cell is approaching an "end stop".

Actually, the sequential balancing concept seems an interesting idea, it's the passive/supercap aspect that I "dislike". An isolated, regulated, current-limited dc-dc converter (perhaps input from the cell stack, output of ~3.4 volts) could be switched to each cell in turn. Turning it off (i.e. to zero current) whilst the relay(s) switch could extend the relay rating to 10^7 cycles (i.e. 25 years at a 1 minute cycle). Also, by monitoring the charge/discharge current, the relays need only switch when "balancing" each cell in turn has been successful or is unnecessary. Logging the charge/discharge data for each cell might also provide useful information.

Cheers, Alan.
 

Gazza_AU

Member
Hi,

It's not MY project, I'm just offering a little (free) advice (which others seem to have stopped doing). ;)

I've hardly ever used FETs myself and am certainly no expert. The type/rating will depend whether you want to use them "locally" for a single cell where the ~3 volts will require a very low Threshold voltage (perhaps like you linked earlier) or from the PICaxe supply (where a "Logic Level" FET will do). But if you want it to replace the isolation provided by a relay, then you will need a 60+ volt FET and suitable drive circuitry.

There was a recent thread where it was said that it's not possible to determine the state of charge (a "fuel gauge") from a Lithium cell's terminal voltage. That would imply that you could only balance when the cell is approaching an "end stop".

Actually, the sequential balancing concept seems an interesting idea, it's the passive/supercap aspect that I "dislike". An isolated, regulated, current-limited dc-dc converter (perhaps input from the cell stack, output of ~3.4 volts) could be switched to each cell in turn. Turning it off (i.e. to zero current) whilst the relay(s) switch could extend the relay rating to 10^7 cycles (i.e. 25 years at a 1 minute cycle). Also, by monitoring the charge/discharge current, the relays need only switch when "balancing" each cell in turn has been successful or is unnecessary. Logging the charge/discharge data for each cell might also provide useful information.

Cheers, Alan.
What you have outlined in your concept is already being done.

An isolated, regulated, current-limited dc-dc converter (perhaps input from the cell stack, output of ~3.4 volts) could be switched to each cell in turn.
This has been done and could work for a standard charge cycle, if you ever needed such application, it is worth trying!
But you would just buy a balance product of the shelf, why the hell would you bother when there is no problem to solve.

The Monitoring you mention, Quite a few people are saying the same thing.
As far as I know it has not actually been done, because it still requires you to design or purchase a monitoring system.
It Just offers you another solution for something you can already buy off the shelf.

It doesn't address the problem I am working on.

I was right about you, your just an ............
 
Last edited by a moderator:
Top