Discussion in 'Power Electronics' started by stube40, Mar 10, 2010.

1. ### stube40

85
0
Feb 9, 2010
I've recently been doing some work with lead-acid batteries. As most of you know, these things are nuts in terms of the amount of power they can deliver - effectively they are a power supply that can deliver hundreds of amps in a small space of time. The upper limit is only defined by the limits of the chemical equation going on inside, coupled with the melting point of the matterials used to make the battery.

I'm wondering if there's any existing or new battery technologies that have an inherently more stable chemical equation that results in some sort of integral virtual current limiting or current control?

It's been a while since my high-school chemistry lessons, but maybe there are some chemical equations that cannot be speeded up? The assumption here is that the speed at which the chemical equation takes place is proportional to the power delivery.

Can anyone can shed any light on this?

2. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
Generally the last thing you want is any form of internal current limiting because that will result in power loss inside the battery leading to heating of the battery and less power able to be delivered to the load.

Which battery is best depends on the application. There is not a perfect battery for all applications (or indeed perhaps for any application).

3. ### stube40

85
0
Feb 9, 2010
Thanks. I'm currently using a bank of LA batteries with some series high-power resistors in order to create a really crude PSU for a test load. The resistors are there purely for current limiting purposes, but they also have an overall effect on a transient step change (which I also create crudely simly using a battery isolation switch to complete the circuit) due to a virtual RL circuit that is created based on the internal inductance of the load and the resistor bank - this is also combined with the internal resistance of the battery and the load (both are negligible in this context).

So, I guess I was looking for the silver bullet solution in which the LA batteries and resistor bank could be replaced by a type of battery technology that is inherently current-limited to some arbitrary value by the very nature of the chemical reaction that is inside the battery. I imagine I am dreaming though that such a technology actually exists or that it could be this simple.

4. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
What sort of current limit, and at what voltage?

If the power requirements are moderate then you could use an active circuit with a MOSFET (or several MOSFETs) to control the maximum current. There would still be some series resistance but ir would probably be far smaller.

What is the load and what are you trying to achieve?

5. ### 55pilot

434
3
Feb 23, 2010
When you do resistive current limiting, whether by adding a fixed resistor or by adding a MOSFET (the controller makes the MOSFET act as a variable resistor to achieve the desired current), you have to worry about heating in the resistive element. As soon as you get past a few ten's of milliamps, heat sinking the resistive element is going to become you biggest problem.

---55p

6. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
I envisaged a mosfet and appropriate other circuitry to limit the current by sensing the current through a small value resistor and then preventing the current from exceeding some set point.

Whilst I agree that heatsinking would be the major issue, at least the supply voltage would not droop before the current limit was reached.

If I understand the problem correctly, the change in voltage with current (caused by the simple current limiting resistor) is an issue. An active current limiting circuit would help -- maybe. I suggested a mosfet rather than a bipolar transistor as it would have a lower voltage drop at moderate currents.

7. ### stube40

85
0
Feb 9, 2010
The power level of the circuit makes everything tougher - try 150V into 3.75 Ohm (40 Amps) to give a whopping 6kW overall.

I have designed an IGBT-based H-bridge circuit to control this, but at the simplest level could be a current control device if the IGBTs could be switched fast enough.

Why is current sensing via an in-circuit series resistor the best way? What about a hall-effect sensor like the Allegro ACS758? Maybe hall effects have an unacceptable lag associated with them that make them unsuitable for such a control loop?

8. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
Yeah, they're better, but considering your initial method of current limiting was a series resistor...
Eeek! At a power level like that, I'll pass on this

I would be thinking about switch-mode regulation or something current limited like a saturating core transformer.

9. ### stube40

85
0
Feb 9, 2010
I only mentioned it because I monitor the actual current with the ACS758

10. ### 55pilot

434
3
Feb 23, 2010
Hall effect current sensors are not as accurate or fast as a good precision resistor. They have HORRIBLE performance at low currents with absolute no redeeming values.

At high currents they become a realistic option as long as you are willing to live with their low accuracy and other poor characteristics. You also have to worry either about noise or about delay and low bandwidth, if you filter out the noise. Most importantly, you REALLY have to know what you are doing to get a good reliable implementation.

As bad as hall effect current sensors are, when you get to really high currents, resistive shunts are even worse. So hall effect becomes the preferred choice because everything else is either worse or far worse.

I have been looking at the multiple threads you have posted about various aspects of this project. What exactly are you doing? The big picture please. What is your knowledge and experience level and what kind of resources do you have access to accomplish your goals?

---55p

11. ### 55pilot

434
3
Feb 23, 2010
Active current limiting would improve the problem greatly by adding only a small current sense resistor and a small Rds_on for the MOSFET during non current limited operation.
Power bipolar transistor is not really a good choice for this. Many people still use it because it is much easier to understand than a MOSFET which makes it easier to design with and control. For low voltages, the power MOSFET is the best choice. For higher voltages, the IGBT becomes the device of choice.

---55p

12. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
Seriously, if this was true you would have more resources available to you than a forum on the Internet.

OK, so you have a bank of 12 x 12V Lead acid batteries providing a nominal 150VDC rail to some unspecified load.

Presumably the series resistor is not required for the load's normal operation.

The rest of your magic is applying transient voltages to the load, and for this you have a current limit on your power supply to partially isolate the transient source from the battery.

Am I right so far?

What you want to be able to do is superimpose these transients (being spikes in either direction) onto the power supply without them being absorbed by the battery (you're testing the load not the battery).

I think you have suggested that the load requires 40A at 150V, although I am not certain if you were referring to the short circuit current limit or the static load there.

The problem you have (I think) is that the power supply has a very low impedance and your attempts at introducing transients in parallel with the batteries results in the transient being mostly absorbed by the batteries rather than being delivered to the load.

If this is the case, then you need to look at some method of adding the transient voltage in series with the battery rather than in parallel with it, or by isolating the battery from the transient using a diode.

The attached schematic shows what I mean. In all cases I show a capacitor charged by a battery and then switched so as to cause a transient. I show a capacitor because in cases I've seen a transient is measured by both the peak voltage and the energy -- a capacitor easily qualifies for this, but you may choose an alternate arrangement. Also no resistors to limit the charge (or discharge) current are shown.

The capacitor across the load delivers positive transients -- they are isolated from the power supply by the diode, so they are delivered to the load. The battery charging the capacitor must have a voltage equal to the power supply voltage PLUS the transient voltage (e.g. for a 150V PSU and a required 50V transient, you would require a 200V source here)

The capacitor in parallel with the diode creates negative transients. Being in series with the main power supply means that the transient voltage is applied to the load. Any voltage that is applied to the capacitor will create a transient. If the voltage is less than the PSU, the transient will not go to ground. If the voltage is > PSU you will get negative going transients.

It is unlikely that you would use a switch, or batteries, or possibly even capacitors, but this method of applying transients will not cause the current issues *I think* you're having with the power supply.

I could, however, have it all wrong

File size:
13.9 KB
Views:
119
13. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
Wow. That was weird. My reply ended up before the post I replied to!?!?

14. ### stube40

85
0
Feb 9, 2010
The bigger picture? I could tell you but I'd have to track you down and kill you!

The purpose of the setup I've described over many posts is to be able to fire high-power transients at a test load and monitor the response. Whilst this thread is specific to the power source, some of the other threads relate to the control of this source via an IGBT-based H-bridge that can pulse the power, short it and reverse it. The onboard micro-controller allows us to do this is various sequenced fashions. By doing this we can understand more about our test loads.

15. ### 55pilot

434
3
Feb 23, 2010
You still have not answered the question about what is your knowledge and experience level and what resources to you have to accomplish what you are doing.

Second point: If it is so confidential that you have hunt people down and kill them if you tell them, you should not be posting about it in any form. That is why I call BS on that one.

---55p

16. ### stube40

85
0
Feb 9, 2010
I was only having a laugh - sorry, but I didn't mean to antogonise you. I'll tread more carefully from now on.......

Knowledge and experience? That's not a simple answer. Here's the brief answer - I completed a 4-year honors degree in digital electronics at 21, then finished a PhD at 25. I've spent the next 12 years in industry.

The PhD was related to the design and control of an 8-legged walking robot for the nuclear industry. This focussed mostly on software and not electronics (writing firmware for the embedded CPUs, software for the supervisor PC and proprietory serial communications protocols).

Since then I've been employed by 4 companies as both an electronics designer and software programmer. Recent years have focussed more on electronics design in small signal areas such as high-precision sensor data acquisition. The most complex PCBs I've designed have used FPGAs and high spec 12-bit serial ADC chips running at 720mbps.

I've recently changed companies and been thrown in at the deep end with my first power-electronics project, hence the reason I've been spending so much time on this forum.

Regarding resources, we are a very small team of 5 with limited apparatus but a huge (almost limitless) budget.

17. ### stube40

85
0
Feb 9, 2010
Man, that was weird - I totally missed your reply as a result and have only just noticed it now.

Thanks for the useful information - what you've said puts a very different slant on things for sure.

The series resistor is there only for one purpose - to define the current that goes through the load, with the load itself having zero resistance.

At present, the power supply IS the transient source. A step change is created in the most crude and fundamental way - an isolation relay. When the H-bridge PCB is completed, this will do the switch on/off, whilst allowing pulsing and also polarity reversal - this is where things will get very interesting. For now we put up with doing simple step changes using the relay, resistor array and the LA battery bank.

18. ### 55pilot

434
3
Feb 23, 2010
Thanks for sharing that. In summary you are someone who either knows what he is doing or can pick up the knowledge with some guidance. Also, you are someone who has the right tools or can get them if you realize that they are needed. Most importantly, you are with it enough to realize that you are dealing with some dangerous power levels that can seriously hurt or kill you and know what you need to do to stay alive and in one piece.

What I was trying to avoid was once again getting involved with a yahoo that wants to charge 1.6F worth of caps to 330V from 220V mains through a resistor divider and discharge them into a coil to make a coil gun. But he does deserve credit for working really hard to get a Darwin aware. Sorry for directing some of that frustration in your direction.

Quite a bit has been posted in this thread in the last few hours and I need to read it all before commenting on the issues.

I look forward to working more you on this and other things.

---55p

P.S. The website certainly seems to be having some issues. The post order is really messed up.

19. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,412
2,780
Jan 21, 2010
If your load has zero resistance then there is no need to have a 150V power supply. All of that will be dropped across your current limiting resistor.

You are better off using a single 12V lead acid battery (it will be more than capable of supplying 40A).

If your load really has a zero resistance then all transient conditions can do is vary the magnetic field surrounding the conductor as far as I can tell. That may be an issue if your device is superconducting or if it is constructed in such a way that excess magnetic fields may affect it in some mechanical way.

<shrug>

Still, you could use a similar arrangement to introduce transients if you replace the battery with your battery and current limiting resistor combination.

20. ### stube40

85
0
Feb 9, 2010
Cool, no worries. I'm certainly safety conscious, although that didnt stop me arc-welding on two occasions whilst wiring up the battery array. I'm now using thick rubber gloves that are suitable for checking an elephants prostrate.