12v regulated supplies

[smileypete]

To which saturated V_CE are you referring to that's 0.7v?? Huh?

 

You've got to have about 0.7v across R1 to turn on the pass transistor. Think about it.

 

If you need high current/low dropout it would be better to use the pass transistor with an LDO (low dropout) regulator:

 

http://www.rapidonline.com/Electronic-Comp...-Regulators-Ldo

 

This may be another way to use a pass transistor:

 

 

Just look for a high gain low V_ce(sat) power transistor. (Not a 'Darlington' one though)

 

cheers,

Pete.

Edited April 2, 2008 by smileypete

[chris w]

Yes, but I wasn't envisaging that you use the 78XX for additional current AND run it undervoltage too. Otherwise the output will be no use whatsoever for the proposed voltage limiting application. The 5A version which you found would be the best bet for a higher current in that application.

 

Chris

[Tusses]

hmmm ... perpetual motion .... if only these threads could generate electricity

[Gibbo]

To which saturated V_CE are you referring to that's 0.7v?? Huh?

 

Chris

 

It can't saturate so it isn't a saturated Vce. The negative feedback as a result of the constant current (once the transistor switches on) into the regulator acts a baker clamp and prevents saturation.

 

Look at how the circuit works (if you can even understand it).

 

Initially (at low current, say 0.1 amps) all the output current is through the regulator with it's input through the 3R resistor. The voltage drop across the 3R resistor is 0.1A*3R = 0.3V. The input to the reg is therefore already 0.3 volts low and the transistor is not even switched on yet. This is at just 0.1A output current.

 

Increase the output current to 0.2 Amps. The drop across the resistor increases to 0.6 volts and the transistor starts to switch on. The input to the reg is now 0.6V lower than it would be without the resistor and transistor. As the output current is increased the transistor switches on harder to maintain the voltage drop across the resistor (negative feedback) but it can NEVER saturate and can NEVER increase the output voltage of the regulator above Vin-Rdrop- regulator dropout voltage. Therefore the regulator draws a constant current through the resistor and therefore the resistor's voltage remains constant past about 0.2 amps.

 

Therefore your question "To which saturated V_CE are you referring to that's 0.7v?? Huh?" although said with sarcastic malice actually makes you look rather like you don't understand this very simple circuit.

 

Gibbo

 

You've got to have about 0.7v across R1 to turn on the pass transistor. Think about it.

 

If you need high current/low dropout it would be better to use the pass transistor with an LDO (low dropout) regulator:

 

http://www.rapidonline.com/Electronic-Comp...-Regulators-Ldo

 

This may be another way to use a pass transistor:

 

 

Just look for a high gain low V_ce(sat) power transistor. (Not a 'Darlington' one though)

 

cheers,

Pete.

 

That suffers the same problem. An emitter follower (which is what that is) cannot saturate and still loses about 0.7 volts.

 

Gibbo

[PaulG]

How about this one, folks?

 

http://www.drbob.net/project/powersupply/l...xd/default.html

 

12V regulated supply. Uses a LM336 reference diode and a LM324 op-amp to drive a P-channel mosfet as the pass transistor, so it's very low dropout with low power dissipation.

 

The mosfet is rated at 6.80 A, so it should be fine for running a few LED's.

[Gibbo]

How about this one, folks?

 

http://www.drbob.net/project/powersupply/l...xd/default.html

 

12V regulated supply. Uses a LM336 reference diode and a LM324 op-amp to drive a P-channel mosfet as the pass transistor, so it's very low dropout with low power dissipation.

 

The mosfet is rated at 6.80 A, so it should be fine for running a few LED's.

 

That's better. It's going to be a bit unstable as it is and needs a few caps here and there but it's still better. The major problem is that voltage spikes still get right through it because the negative feedback cannot correct them until *after* they have happened. That's why NPN pass transistors are usually used because they inherently stop this.

 

Gibbo

[chris w]

So you had to double the load to try to prove your point?

 

Just run the sums under a 5A load (which is just as likely a scenario as 2 X 20A) and see what you get. So you just can't say that drops due to internal battery resistance is "more significant". Sometimes it is, sometimes it isn't, so you should take both into account.

 

Have a nice day

 

Paul G.

 

I feel you're missing the point Paul and this isn't specifically about your boat.

 

Firstly 20A is really easy to draw. Watch TV (5A), DVD player or satellite box (5A), 8 halogens (galley and saloon at night - not unreasonable) = 8A. There's 18A already. On comes the water pump - another 10A.

 

The point is that, whilst at an individual current level, the loss on the cable may be more or less than the battery internal loss at that instant in time, battery internal loss affects ALL circuits. So for a 5A circuit, the cable loss should be designed to be not more than 0.3v The loss in the battery's internal resistance due to that 5A will be smaller than this at say around 0.1v. But, EVERY, circuit that switches on will affect the original 5A circuit because EVERY circuit that's ON will cause a small voltage drop at the battery which is felt at EVERY appliance. So If I switch on another 5A circuit, the cable loss doesn't change but the battery volts drop is now doubled to 0.2v and so on which is felt at both circuits.

 

So if I switch on 20A and cause a 0.4v battery drop say, that 0.4v is felt even at the original 5A circuit which now has a cable voltage loss of 0.3v as before, plus an additional 0.4v from the 20A load from other devices. So battery voltage drop will ALWAYS be more significant than cable loss. I think you were reading "significant" to mean "greater than". At any instant, the cable drop may be more than the battery drop but battery drop will always have the greater effect and, under normal running (viz: a few circuits ON) the battery voltage drop WILL be the greater of the two.

 

Chris

Edited April 2, 2008 by chris w

[Tusses]

That's better. It's going to be a bit unstable as it is and needs a few caps here and there but it's still better. The major problem is that voltage spikes still get right through it because the negative feedback cannot correct them until *after* they have happened. That's why NPN pass transistors are usually used because they inherently stop this.

 

Gibbo

 

just thinking out loud here - how about a high power zener on the output with a slightly higher V rating to the output. - would this 'soak' up any spikes without being 'in circuit' ?

 

I am sure there is a dedicated component that does this - but I cant remember what it is called.

[chris w]

just thinking out loud here - how about a high power zener on the output with a slightly higher V rating to the output. - would this 'soak' up any spikes without being 'in circuit' ?

 

I am sure there is a dedicated component that does this - but I cant remember what it is called.

 

That wouldn't work because you need a resistor in series with the zener to limit its current. Without the resistor you will destroy the zener. Ergo, the voltage at the circuit output won't be overvoltage limited because, although the zener will kick-in, the rest of the voltage appears across the series resistor so nothing changes at the output.

 

I think you were thinking about VDRs (Voltage Dependent Resistors).

 

Chris

[PaulG]

I feel you're missing the point Paul and this isn't specifically about your boat.

 

Firstly 20A is really easy to draw. Watch TV (5A), DVD player or satellite box (5A), 8 halogens (galley and saloon at night - not unreasonable) = 8A. There's 18A already. On comes the water pump - another 10A.

 

The point is that, whilst at an individual current level, the loss on the cable may be more or less than the battery internal loss at that instant in time, battery internal loss affects ALL circuits. So for a 5A circuit, the cable loss should be designed to be not more than 0.3v The loss in the battery's internal resistance due to that 5A will be smaller than this at say around 0.1v. But, EVERY, circuit that switches on will affect the original 5A circuit because EVERY circuit that's ON will cause a small voltage drop at the battery which is felt at EVERY appliance. So If I switch on another 5A circuit, the cable loss doesn't change but the battery volts drop is now doubled to 0.2v and so on which is felt at both circuits.

 

So if I switch on 20A and cause a 0.4v battery drop say, that 0.4v is felt even at the original 5A circuit which now has a cable voltage loss of 0.3v as before, plus an additional 0.4v from the 20A load from other devices. So battery voltage drop will ALWAYS be more significant than cable loss. I think you were reading "significant" to mean "greater than". At any instant, the cable drop may be more than the battery drop but battery drop will always have the greater effect and, under normal running (viz: a few circuits ON) the battery voltage drop WILL be the greater of the two.

 

Chris

 

I was just using my boat as an example. I didn't say that it was typical. I don't have a TV or DVD player. All my lights are fluorescent. So, in your example, that just leaves the water pump!

 

Anyway, I think I need to "agree to disagree" with you at this point as I am sure everyone else is getting bored by this exchange.

 

Feel free to have that last word if you feel the need.

 

Cheers

 

Paul G

[chris w]

I accept your graceful retirement from the fray

 

Chris

[PaulG]

An interesting surge supression IC.

 

http://www.automotivedesignline.com/howto/205101011

 

As they say in the article, something similar could be built in discretes as a "front end" to the regulator design I posted earlier. Over to you, Gibbo - are we getting anywhere near?

[Gibbo]

An interesting surge supression IC.

 

http://www.automotivedesignline.com/howto/205101011

 

As they say in the article, something similar could be built in discretes as a "front end" to the regulator design I posted earlier. Over to you, Gibbo - are we getting anywhere near?

 

Indeed. I looked into this chip for a current design project. Built a few test setups with it. It does indeed do what it says on the tin. We eventually decided against it because it's a bit expensive for mass production, it isn't second sourced and if LT decide to discontinue it (something they seem to specialise in) we get left up a creek. We've built a discreet version of it instead. Of course these issues aren't of any relevance to someone who wants to build one of something for himself.

 

It's nice to see that some people realise what a major problem it is to crack and that clearly a zener or 7812 isn't the answer. If it was, why would a company of LT's standing have invested major time and resources into developing a solution?

 

A quick google on "load dump protection" will reveal what a problem it is that the uninitiated don't realise.

 

A good spike protection system involves a p channel depletion mode mosfet and an ldo. That works great. But there's still no regulation. Just spike prevention.

 

Trying to clamp or swamp the spikes is NOT the answer. That leads to tears. One has to stop them getting through in the first place.

 

Gibbo

Edited April 2, 2008 by Gibbo

[PaulG]

Indeed. I looked into this chip for a current design project. Built a few test setups with it. It does indeed do what it says on the tin. We eventually decided against it because it's a bit expensive for mass production, it isn't second sourced and if LT decide to discontinue it (something they seem to specialise in) we get left up a creek. We've built a discreet version of it instead. Of course these issues aren't of any relevance to someone who wants to build one of something for himself.

 

It's nice to see that some people realise what a major problem it is to crack and that clearly a zener or 7812 isn't the answer. If it was, why would a company of LT's standing have invested major time and resources into developing a solution?

 

A quick google on "load dump protection" will reveal what a problem it is that the uninitiated don't realise.

 

A good spike protection system involves a p channel depletion mode mosfet and an ldo. That works great. But there's still no regulation. Just spike prevention.

 

Trying to clamp or swamp the spikes is NOT the answer. That leads to tears. One has to stop them getting through in the first place.

 

Gibbo

 

So I guess the answer to Tusses' original question (remember that?) is "No", then.

 

Pity. I was rather hoping that it would be "42".

 

Paul

[Tusses]

Pity. I was rather hoping that it would be "42".

 

Paul

 

 

[benjamino]

So is there a solution to this then? I have one of these and it has a 12 volt 4amp ac-dc power brick with it - I would love to be able to run it straight off my 12 volt supply on the boat instead of via the inverter for obvious reasons but I'm not sure on it's tolerances to the voltage so think it should be a regulated supply. Because it uses 4amps that rules outs those regulators from maplin which are 1 amp right? Would the amperor stabliser work?

Edited August 26, 2008 by benjamino

[MikeV]

So is there a solution to this then? I have one of these and it has a 12 volt 4amp ac-dc power brick with it - I would love to be able to run it straight off my 12 volt supply on the boat instead of via the inverter for obvious reasons but I'm not sure on it's tolerances to the voltage so think it should be a regulated supply. Because it uses 4amps that rules outs those regulators from maplin which are 1 amp right? Would the amperor stabliser work?

Amperor should work. I have one the delivers 5amps to my LCD TV.

 

The Rolls Royce solution is one of these things:

Orion DC/DC Converter

They do a model with 9-18 volts input and 12.5V output up to 30 amps. This could then be used to supply multiple 12V devices with a regulated supply. Pricey though, but they do occasionally come up on eBay. I picked one up for £10 although I've yet to fit it! //Mike

[benjamino]

Amperor should work. I have one the delivers 5amps to my LCD TV.

 

The Rolls Royce solution is one of these things:

Orion DC/DC Converter

They do a model with 9-18 volts input and 12.5V output up to 30 amps. This could then be used to supply multiple 12V devices with a regulated supply. Pricey though, but they do occasionally come up on eBay. I picked one up for £10 although I've yet to fit it! //Mike

30 amps? Looks like it says 3 amp limit and thats if its getting 18 volts? http://www.marinemegastore.com/product-pro...R1351203000.htm

[MikeV]

30 amps? Looks like it says 3 amp limit and thats if its getting 18 volts? http://www.marinemegastore.com/product-pro...R1351203000.htm

That's a different model. They do a whole range suited to different aplications. The 30 amp model is the Orion 12/12-360W. If you look at the link to the PDF I posted it includes the specs of all the models.

[benjamino]

That's a different model. They do a whole range suited to different aplications. The 30 amp model is the Orion 12/12-360W. If you look at the link to the PDF I posted it includes the specs of all the models.

Ah ok so it is! Ideal - thanks for the pointer

[Clivo]

Is there a range of 12v regulated supplies ?

 

it seems to come up on here quite often - 12v equipment and LED lighting etc.

 

is there a gap in the market ? (kerching:) )

or are the suitable solutions on the market already.

 

I am thinking along the lines of , for instance, a fuse board type box that takes the raw 12v (upto 15v ish (17v? on desulphation ?) ) and limits any over voltage situation on the output (fused) side.

 

Is this any good:-

 

DC to DC Convertor

 

Not my idea I hasten to add, suggested by an electrical boffin friend of mine when I posed the question. Not exactly cheap but with LED bulbs at around a fiver a pop, possibly worth it before they do pop.

 

Clivo

[Sheepshank]

I've been attempting to follow this thread but have got completely bamboozled with mosfets and such like. Am I correct that the basic problem is the variable voltage produced whilst charging by running the engine, with the over-voltage affecting LEDs and other 12V equipment such as TV's?

 

I'm considering an alternative solution, whereby I divide the domestic batteries into two groups and use the engine to charge one group while using the second group for the equipment. When the voltage of the second group falls below a set level, there's a warning given and with the flick of a double pole switch, I'd be able to set group 2 charging and use group 1 for equipment. Hence the voltage should never exceed the full battery level.

 

Do you think this would work?

 

Adrian

[Chris Pink]

I've been attempting to follow this thread but have got completely bamboozled with mosfets and such like. Am I correct that the basic problem is the variable voltage produced whilst charging by running the engine, with the over-voltage affecting LEDs and other 12V equipment such as TV's?

 

I'm considering an alternative solution, whereby I divide the domestic batteries into two groups and use the engine to charge one group while using the second group for the equipment. When the voltage of the second group falls below a set level, there's a warning given and with the flick of a double pole switch, I'd be able to set group 2 charging and use group 1 for equipment. Hence the voltage should never exceed the full battery level.

 

Do you think this would work?

 

Adrian

 

Smiley Pete proposed such a 2 bank electric supply system for reasons of efficient charging, have a search but be warned you could spend a long time wading through interminable electrical arguments.

 

But the principal is sound, I don't run LEDs whilst charging and haven't destroyed any yet.

[BlueStringPudding]

But the principal is sound, I don't run LEDs whilst charging and haven't destroyed any yet.

 

Neither did we, but ours still died!

[Chris Pink]

Neither did we, but ours still died!

 

That's rather odd as one of the selling points about LEDs is their longevity, what kind were they?

 

Any other possible causes of voltage surges?

did you fit DC-DC converter?

did that solve the problem?