7812 abuse

A 7812 misconnected to a lead acid battery

NT

Reply to
tabbypurr
Loading thread data ...

if the diode conducts from output to input, it never conducts & does nothing. If it points the other way it protects against misconnection but also conducts in normal use, ending any attempt at regulation.

interesting.

NT

Reply to
tabbypurr

y

yes. The 7812 output has been raised to 13.8v.

it's intended for charging to get things up & running rather than as a 24/3

65 power supply. A simpler circuit without regulation would be liable to ca use battery damage if left connected unintentionally - or even intentionall y - I want to avoid that hence the regulator.

yes and no. Yes you can do it, yes we all probably have, but with so little control over end use that sort of set up risks ruining batteries.

that would be ideal if the diode dropped a steady voltage. IRL they don't c ome anywhere near that. It seems to be a common myth among EEs that power d iodes drop 0.65v. They don't, that's the knee voltage only. The 7812 has no way to compensate for that uncontrolled drop.

That's another question that would be good to disuss later.

NT

Reply to
tabbypurr

Why not use a thyristor ?. remember a very simple charger in an old GE thyristor manual from years go. Reverse polarity proof and used just a handful of cheap parts. Cheap, rugged and effective...

Chris

Reply to
Chris

You can apparently extend battery life by designing in a voltage hysteresis, so that the charger is on when the voltage falls below a low set point, then off once it reaches the high set point. Makes sense, as in long term float applications, the damage is done by the continuous float trickle charge current.

There may be patents on this as well, but no details...

Chris

Reply to
Chris

p

lly

ll

m
24/365 power supply. A simpler circuit without regulation would be liable t o cause battery damage if left connected unintentionally - or even intentio nally - I want to avoid that hence the regulator.

ure

r

ttle control over end use that sort of set up risks ruining batteries.

d

ad

rop

e

't come anywhere near that. It seems to be a common myth among EEs that pow er diodes drop 0.65v. They don't, that's the knee voltage only. The 7812 ha s no way to compensate for that uncontrolled drop.

re

A

ere

to

see

or

e

ill

ly

rom

:-)

If there are, they're patenting Kirchhoff's voltage law. If the regulator o pen circuit voltage is 2.3V per cell, 13.8V total, then charging current ta pers to zero when the battery if fully charged. That's called the float cha rge and it can be applied indefinitely.

Reply to
bloggs.fredbloggs.fred

regulator open circuit voltage is 2.3V per cell, 13.8V total,

then charging current tapers to zero when the battery if fully

charged. That's called the float charge and it can be applied indefinitely.

It's an interest area here and at one point had an string of 48 x

6v x 100 a/h gel cell batteries as part of a home energy system. Such a battery will still draw ~30-50mA at float rates, depending on age and load cycle history. To do it right, you also need to compensate for temperature, as the optimum float voltage varies with temperature. Ideally, each battery in the string needs to be independently voltage monitored and corrections applied to maintain balance. With large strings, not an easy task to optimise cell balance and life.

Probably telcos have the most experience. They were all wet cells in the old days, which had a 30+ year design life under optimum conditions...

Chris

Reply to
Chris

y.

That lot applies to batteries always on charge, not so much to a basic char ger just intended to step in on occasion.

BTW I'm looking for a new charger for some SLAs, and not sure what to pounc e on. The chargers I've been offered do 14.4v before throttling back. Some are multistage and try to recover damaged batteries - I don't know how effe ctive that is. If it works it may be worth having.

NT

Reply to
tabbypurr

It's not a good idea to subject a 12v SLA to more than 13.8v since they don't take kindly to the resulting gassing. The problem with SLAs is that the gas bubbles get trapped between the plates and the electrolyte medium (AGM or Gel). I may have this the wrong way round but I think the AGM type tolerates such abuse a little better than the Gel type although such abuse is bad for both. Using a 14.4v charger will increase the internal battery resistance due to gas bubbles becoming trapped within the electrolyte making such chargers totally unsuited for the SLA battery type.

Otoh, a 14.4v charger which throttles back to a lower float charge voltage (13.5v) is better suited to stationary flooded cell LA batteries since the gassing will mitigate the stratification of the electrolyte by the stirring action it creates within the cell.

As I previously mentioned, the use of a gassing charge with flooded cell batteries that can be topped up with distilled water is beneficial when applied in moderation. Not only does it mitigate the stratification effect experienced by stationary batteries, it also balances the charge state of the individual cells within a string, compensating largely for the differing rates of self discharge amongst a string of cells within a battery.

Fortunately, the characteristic of the AGM and Gel SLA electrolytes that prescribes against the use of a gassing charge, also makes the need for stirring up the electrolyte to mitigate the effect of stratification somewhat redundant. However, there still remains the issue of maintaining the individual cells in a charge balanced state.

It's an issue that seems to rely upon the SLA battery manufacturers making sure that all the cells in each battery are extremely well matched in regard of capacity and self discharge characteristics in the hope that each cell ages at the same rate long enough to provide a reasonable service life.

An obvious way to reduce the element of luck in this regard would be to wire a temperature compensated shunt voltage regulator across each cell set precisely to "2.3v" to protect the good cells from the bad cells in the battery string.

That "2.3v" is high enough that an LED indicator could be included to show when the charge balancing/protection is being triggered by the 'bad' or failing cells. The expectation being that one or two indicator LEDs will light up for a few minutes towards the end of charge in a brand new battery before all LEDs extinguish with such balancing/protection events becoming ever more protracted with each charging cycle.

These shunt regulators only need handle a low bypass current load to save their cells from excess gassing voltage (mere milliamps being enough to make a 'good cell' go high resistance from trapped bubbles or electrolyte loss) and can include a bi-coloured LED to show when such protection has reached its limit and the battery condition has finally deteriorated to the point of impending failure.

--
Johnny B Good
Reply to
Johnny B Good

you must be thinking of some other circuit.

I mean like this:

in ---[7805]-------+-- | | +--[200R]-+ | | +-->|-----+ | [340R] | --+--

so long as in is less than 20V the 7805 should see acceptable voltages on all pins in the fault condition.

--
  When I tried casting out nines I made a hash of it.
Reply to
Jasen Betts

Ah. 2 problems though. This reg needs to run off a variety of PSUs, and thyristor regulators don't. Also lack of a reliable supply of thyristors.

NT

Reply to
tabbypurr

Something like this?

formatting link

Rather than protecting the 7812 from reverse output, it just passively pulls the 7812 ground below system minus in case of polarity reversal.

The common is still ca. 2 V above the "+" output in this case, but it's weakly driven by a 10 K or higher resistor, to limit unwanted currents.

The 1.2 V diode is an IR LED of arbitrary cheapness as commonly found in TV remotes. Your supply channels should have those. The others are plain small signal silicon diodes of 0.7 Vf nominal. Output nominally 13.9 V.

Note that if you weakly pull up the ground slightly above the output of a 7812, the 7812 (at least the ones I've tested) typically turns off. You may need to experiment with the voltage levels and possibly add a diode directly between 7812-GND and 7812-OUT if the -2 V offset from the above circuit happens to be too much and wreaks havoc somewhere.

Reply to
Dimitrij Klingbeil

wered 7812 see -12v on its output pin. The -12v is from a lead acid, and ye s I know they don't charge at 12v etc etc. I could put up with an occasiona l failure, but not a lot obviously, it's a low cost consumer app. Has anyon e tried this, or have insight into the expectable results? Moving to a much less common regulator isn't really workable on this one.

Don't overlook the possibility of using a cheap mechanical relay. A couple diodes and a relay may do the trick for you (but it would require the (batt ery?) voltage to be above the pickup voltage of the relay before it would e nergize the output). Modern car battery chargers sometimes have this 'featu re', they will not attempt to charge a battery that is completely flat.

--sp

Reply to
speff

I like that one. Ta.

NT

Reply to
tabbypurr

powered 7812 see -12v on its output pin. The -12v is from a lead acid, and yes I know they don't charge at 12v etc etc. I could put up with an occasio nal failure, but not a lot obviously, it's a low cost consumer app. Has any one tried this, or have insight into the expectable results? Moving to a mu ch less common regulator isn't really workable on this one.

e diodes and a relay may do the trick for you (but it would require the (ba ttery?) voltage to be above the pickup voltage of the relay before it would energize the output). Modern car battery chargers sometimes have this 'fea ture', they will not attempt to charge a battery that is completely flat.

Yes. I considered that but the inability to charge a totally flat battery i s a real problem. It doesn't matter if the product is a gimmick for rich fo lk who can call a man in if something doesn't work, it really does matter i f the end user is relying on it.

A relay is cheap as chips when you're building a >$1000 piece of testgear, but would double the components cost on something as simple as this. I do m icrocar-like design now, how can we hack off every penny and still have it work and be fairly reliable?

NT

Reply to
tabbypurr

I ran a few tests here, to confirm my assertions, but eventually succeeded in killing some 7812s. The earlier confidence was mainly based on accidents occuring with other negative sources that had their own current limit. A battery doesn't, of course, and such a fault does kill the 7812. So you'll have to be a bit more component-intensive, with at least one series limiting element, if you want it proof from that fault.

Even a crowbar circuit will only protect if it's blowing a series limiting fuse.

RL

Reply to
legg

unpowered 7812 see -12v on its output pin. The -12v is from a lead acid, an d yes I know they don't charge at 12v etc etc. I could put up with an occas ional failure, but not a lot obviously, it's a low cost consumer app. Has a nyone tried this, or have insight into the expectable results? Moving to a much less common regulator isn't really workable on this one.

ttery is made by people outside of our control, so connector hardware won't solve it. I could fit a fuse & crowbar diode, but that takes time to blow & the 7812 still sees the -12v, just not for long. And a charger that fails to work because a fuse needs replacement is not a good feature here.

risk.

hat pin is already included, it's easy protection. But a diode on the outpu t pin would work at 1A, thus drop much more than 0.65v & have an uncontroll ed & uncompensated voltage drop. There lies the problem with a series diode .

rete regulator is not impossible in principle, but a simple 7812 would be m uch preferable.

ed to going in the right direction with a reversed battery, though the term inal voltages would be way out. The voltage drop is affordable. I guess tha t's the way forward.

Thanks.

On the subject of crowbars... the current output of a lead acid can of cour se be huge. Fuses take finite time to blow, even at 100A or more. And crowb ars don't have zero resistance. During that time I'm far from confident the y would provide the 78 enough protection. Plus a fuse is something end user s will curse and abuse, neither of which is good.

I ended up with a diode on the input and another on the ground pin. The wor st a misconnection could do is feed current through the preceding psu circu itry and provide the 78 with +12 on it's input pin. The input side diode v drop affects headroom only, not V_out.

Appreciate the ideas everyone.

NT

Reply to
tabbypurr

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.