Voltage climb.

[MK14]

As others have said, it could be badly oscillating and you don't have the equipment to develop stuff like that (no scope etc).

As far as I know, there are MANY published (known to be working), PSU designs available, including for very high currents. Why not use or ask for, such a design ?

It sounds to me, like you are just going round and round in circles, and not really getting anywhere.

I think you can buy ready made PSUs (high current), somewhat cheaply. E.g. Around £50, depending on the exact spec.

[MK14]

No load on the output for the following.

100 ohms base Volts across it, 0.55 Volts output will not come down below 17 Volts.

That reading makes little sense to me.

17 Volts means that you will potentially destroy/burn-out/explode/overheat/damage whatever 12V expecting stuff you connect to it. Unless it has been designed to cope with a much higher voltage.

If you disconnect JUST the base of the power transistor, does the voltage change ? (across R7 and/or output)
If you remove the power transistor, does the voltage change ?

But really I think you would be better off, taking one of the steps in my previous post.

[davelectronic]

I like building psu circuits, linear ones for now. I have no intention of connecting anything 12 Volts to a 17 Volt output circuit.
No i don't go round in circles, rather exaust all posabilitys before giving up. I can see this particular transistor is not best suited for this application. But i gave it a try ! Thanks for the help people.

[MK14]

I like building psu circuits, linear ones for now. I have no intention of connecting anything 12 Volts to a 17 Volt output circuit.
No i don't go round in circles, rather exaust all posabilitys before giving up. I can see this particular transistor is not best suited for this application. But i gave it a try ! Thanks for the help people.

Yes, building PSU's can be real fun!

The results you have been getting and the schematic circuit supplied, are NOT really making sense (agreeing).

Hence my suggestion to try partially/fully disconnecting the power transistor, to try and find out why the strange/mysterious 17 Volts is occurring, apparently WITHOUT enough voltage to switch the power transistor on.

tl;dr
One or more things are wired up wrong (different to the schematic) and/or one or more components is bust/faulty. Or some other explanation. Such as it is oscillating e.g. Because 100 Ohms is too high a value for LM78XX series regulators (which is something I suspect may well be the case, BUT I'm NOT 100% sure).

[davelectronic]

Thanks for your help, i've built TIP2955 versions single and double transistors with very good results. I think maybe this transistor the MJ11015 is not best suited to this circuit application. I would like to build a unit with 4 x transistors for HF radio use, wanting to stay with a TO3 package, i'm looking at the PNP MJ2955 its not a darlington, the gain is not huge either, but if its as close to the TIP2955 then i would give the MJ2955 a try.

[MK14]

Thanks for your help, i've built TIP2955 versions single and double transistors with very good results. I think maybe this transistor the MJ11015 is not best suited to this circuit application. I would like to build a unit with 4 x transistors for HF radio use, wanting to stay with a TO3 package, i'm looking at the PNP MJ2955 its not a darlington, the gain is not huge either, but if its as close to the TIP2955 then i would give the MJ2955 a try.

Yes, Darlingtons such as the MJ11015 can make things much more complicated. Their characteristics are harder to deal with (as you probably have just found out), because there can be such a huge variation in how much gain (Hfe), you are going to get, at a particular base voltage and temperature.

I prefer to treat Darlingtons as on/off devices, as their HUGE gain (Hfe), can make them too sensitive/tricky to try and operate in linear mode and/or as a voltage regulator. It can be done though, some people do that. But somewhat often with op-amps, which close the loop so well, that they/it can cope.
That basic circuit (from your schematic, but probably based on the datasheet application circuit, for external pass transistors), was not really designed to cope with darlington transistors. E.g. They need something like DOUBLE the base drive voltage, which probably worsens or even destabilizes, the circuit.

The "standard" datasheet application circuits, all seem to use 3 Ohm resistors. So I would be VERY tempted to stick with 3 Ohms, and "normal" non-darlington transistors. There are probably a number of VERY good engineering reasons, and experimental experience findings, which make the 3 Ohms, a VERY good value to use.

N.B. I'm getting increasing convinced that making R7 100 Ohms, is a VERY BAD idea. It may well cause terrible instability and NOT give the LM7812C enough current (etc), to operate well/nicely. Is there any evidence that the designer who made the value 100 Ohms, was sensible/experienced at doing circuits like that ?

I would think (as you suggested) the PNP MJ2955, is well worth a try. It is possible that differences between that, and TIP2955's, might be significant, but I would hope NOT!

The circuit should be able to cope with at least a small amount of difference (one would hope).

Anyway, Good Luck!

[Audioguru]

I repeat, the circuit does not use a resistor in series with the base of the transistor or darlington. The circuit with the 100 ohm current sensing resistor has an error.

You want to use a current-sensing resistor of 0.1 ohms so that it does not get as hot as a 1 ohm resistor? Then simply use Ohm's Law to see why it will not work. When the voltage across the resistor is 0.5V to 0.6V then the transistor will begin to turn on. Then the current in the regulator is 0.5V/0.1 ohms= 5A! (impossible with a 1A regulator). If you use a darlington then the regulator current must be 9A!

[MK14]

I repeat, the circuit does not use a resistor in series with the base of the transistor or darlington. The circuit with the 100 ohm current sensing resistor has an error.

You want to use a current-sensing resistor of 0.1 ohms so that it does not get as hot as a 1 ohm resistor? Then simply use Ohm's Law to see why it will not work. When the voltage across the resistor is 0.5V to 0.6V then the transistor will begin to turn on. Then the current in the regulator is 0.5V/0.1 ohms= 5A! (impossible with a 1A regulator). If you use a darlington then the regulator current must be 9A!

What you describe, does not seem to fully match the supplied schematics.

The 0.1 Ohm resistors, seem to be connected between the Emitters, and the raw incoming positive voltage supply, from the main smoothing capacitor.

The 0.1 Ohm resistors, appear to be (load) balancing (current sharing) resistors, to allow slight/partial differences, between the multiple power transistors, base voltages. Without them, multiple power transistors, could be in danger of partial or worse thermal runaway.

tl;dr
Otherwise one of the 4 or 6 power transistors, could have a base which activates some 0.15 volts, BELOW the others. So that transistor, would attempt to take ALL the current (and power dissipation), through itself. Potentially leading to that transistors early failure, due to over-heating.

I.e. They are NOT current sense resistors, in this application.

EDIT: In fairness to yourself, for most of this thread, there was only ONE (single) power transistor (mentioned). So no current/load balancing was needed. Adding to the confusion around 0.1 Ohm resistors. Sorry.

[davelectronic]

This makes absolute sense 100% ! The theory for the 100R base resistor came from a member moderater from another forum ! limited as my knowledge is, i have learned a lot.
Yes i tried no base resistor, all i got was the regulators rated voltage and current, as the transistor did nothing.
Thanks again for all the help !

[MK14]

This makes absolute sense 100% ! The theory for the 100R base resistor came from a member moderater from another forum ! limited as my knowledge is, i have learned a lot.
Yes i tried no base resistor, all i got was the regulators rated voltage and current, as the transistor did nothing.
Thanks again for all the help !

Because you disconnected the base, and left the rest of the transistor connected. We have apparently eliminated it being transistor leakage current.
Also it shows that the 17 Volts IS being triggered by the Base of the transistor. Rather than some other cause.

So either the base is getting a much bigger voltage than you measured, or there are oscillations when the base is connected, and the apparent lowish voltage (across R7), is because the meter is NOT showing the right voltage, because it is fluctuating (oscillating). Tricky to diagnose (Oscillations) WITHOUT access to a scope.

Early indications are that this further shows the 100 Ohms is way too high a value, in this type of circuit. E.g. Terrible oscillations.

IC Voltage regulators are somewhat notorious for oscillating at the best of times, anyway. Hence the (often) need for significant decoupling capacitors across their inputs and outputs.

[Audioguru]

PLEASE post the schematic you are using with its parts numbered.
If the base of the PNP transistor is not connected to the input of the voltage regulator then the transistor does nothing.
If you are using a darlington then please SHOW a darlington and the increased value of the current sensing resistor for it.

[davelectronic]

You sure know what your talking about, and i'm no expert. For now i'm not going to continue to try and tame these monster transistors. It makes more sense to use a single transistor alternative PNP TO3 package.
I did have bypass caps on the regulator itself, mounted with the regulator on a small three hole piece of veroboard.
A tip was give  to me to use a meter on an AC range to try and spot osallation, its one i will remember, but its time for me to move on, for now at least.

Thanks again for all the help !

[MK14]

Update:

I seem to have found an older/same? version of your circuit, but complete with technical explanations on how it is suppose to work.

http://www.zen22142.zen.co.uk/Circuits/Power/1230psu.htm

http://www.eeweb.com/blog/circuit_projects/lm7812-high-current-power-supply-by-tip2955-pass-transistors




I can believe the original circuit can work out ok, and give 12 volts.

But I think where things have gone wrong, is by your substitution of the standard power transistors, with the darlingtons.

The original circuit (it confirms this in the explanation), was relying on the significant base currents, in order to provide the regulator with a decent load.

But when you put in the darlington(s), they have a considerably higher gain (Hfe), of about 100 times higher. Maybe even more. Therefore that will potentially pass considerably LESS current, back into the regulator. Which could be worsening its tendency to misbehave/oscillate etc. I.e. Darlington(s) may make the R7/100 Ohms unsuitable/incorrect, in that situation.

Without much more extensive analysis etc, I CAN'T be 100% sure, exactly how it will behave in practice. I've played about with some simulations, but that has only partially improved my understanding.

tl;dr
Probably best to use the specified (or similar) power transistor type. The TO3 version, should still be ok as well, as it should give broadly similar base currents. But even that change is a small risk.

Reportedly (as detailed on the web page(s), linked to above), it DOES WORK, when built EXACTLY as specified.

Changing the transistors to darlingtons, is apparently a step too far.

[Delta]

Mate, you have seriously fucked something up if your base resistOr is overheating.  There should be next to no current flowing into the base of a transistor at DC, especially a Darlington pair.

Stop what you are doing, and start again.  An outboard boost transistor across a regulator should not have a base resistor anyway!

[davelectronic]

Your spot on MK14 ! Yes that is a circuit i refferenced to build this circuit, and can confirm with the transistors in the schematic it works well. It was my idea to try this high powered darlington. Another friend has simulated it, and added an RC network gives a stable output with 100R base resistor.

I did explain some wild voltage fluctuation to him on inital loading of the circuit with these darlington transistors. But for now with no scope i'm going to try the MJ2955 transistor instead, as with no scope i can not trouble shoot with the darlingtons in circuit. I might go back to this circuit with darlingtons once i have a scope.

Along with the wild voltage fluctuations, i was also getting some odd buzzing from the cooling fan, but it was not constant, rather intermitent on connecting loads. Lighter loads seemed worse. Thank you for taking time to look into this, at least i have a better understanding of what was going on in the circuit with darlington transistors. Thanks again.

[davelectronic]

Mate, you have seriously fucked something up if your base resistOr is overheating.  There should be next to no current flowing into the base of a transistor at DC, especially a Darlington pair.

Stop what you are doing, and start again.  An outboard boost transistor across a regulator should not have a base resistor anyway!

I am not continuing with the darlingtons for now, i'm aware i overloaded a resistor be it for a very short time, i replaced it once i relised what i had done. For now i'm continuing with more modest non darlington power PNP transistors. Thanks for the input.

[MK14]

Your spot on MK14 ! Yes that is a circuit i refferenced to build this circuit, and can confirm with the transistors in the schematic it works well. It was my idea to try this high powered darlington. Another friend has simulated it, and added an RC network gives a stable output with 100R base resistor.

I did explain some wild voltage fluctuation to him on inital loading of the circuit with these darlington transistors. But for now with no scope i'm going to try the MJ2955 transistor instead, as with no scope i can not trouble shoot with the darlingtons in circuit. I might go back to this circuit with darlingtons once i have a scope.

Along with the wild voltage fluctuations, i was also getting some odd buzzing from the cooling fan, but it was not constant, rather intermitent on connecting loads. Lighter loads seemed worse. Thank you for taking time to look into this, at least i have a better understanding of what was going on in the circuit with darlington transistors. Thanks again.

Assuming we are both right, that the problem is caused by the changeover to using darlingtons. You and/or your friend and/or someone else, can fix the new circuit, so that it DOES work with them, eventually, some time in the future. (As you said, a scope would help see what is going wrong, in better detail).
Darlington transistors, can be (in theory) slower than standard transistors, which probably doesn't matter in a (somewhat slow) DC power supply. But it might (e.g. transients, stability etc).

I guess after all this difficulty with them. It makes it an interesting challenge, to solve any issues with them, and make a darlington power transistor based PSU, work well.

On the other hand, I DON'T think darlington transistors, offer any real advantages (in this application), to using standard power transistors. Their big advantage (for some applications), is their huge gain (Hfe), of e.g. 10,000 (typically). Hence needing minimal driving requirements. BUT that does NOT help with the circuit  you are using. Because it already can drive, normal power transistors.

But as you have just said, for the time being, going back to the MJ2955 (hopefully similar enough to the TIP2955's), is the way to go for now.

I see getting a scope as a sort of investment (in Electronics), for the next 5 to 10+ years. You learn a lot more with a scope, and it does things that other test equipment can't really help you much with. By seeing EXACTLY how it is misbehaving, you can get a much better idea as to what needs to be done, to fix it.

As you said previously, you can use AC mode on multimeters (to some extent), and other tricks. But it is a lot of messing about and hassle, just to get a tiny fraction of the information, that a decent scope would give you, in seconds.

[davelectronic]

I chose these MJ11015 darlingtons as i wanted a more higher maximum power devices, and less voltage sag or drop on the output, my oversight was not looking into darlington behavior in a psu type circuit first. I have looked up the MJ2955 and TIP2955, and compared them, there is little to no difference at all, other than higher power dissipation 90 watts for the TIP2955, and 115 watts for the MJ2955's All other specifications look the same, including hfe. So hopefully should be ok.

The aim is to build a final 20 Amp linear psu for cb radio and linear amplifier use. I will need 15 Amps 12 Volts on about a 50% duty cycle, so figure 20 Amps for a final build should be plenty.  When i did have the darlingtons running, for the same load on the output things run much cooler. But for now the instability is a problem with out that scope. But i should be able to utalise the MJ2955's hopefully.

[MK14]

How about THIS ?



STMICROELECTRONICS    2ST2121  Bipolar (BJT) Single Transistor, PNP, -250 V, 25 MHz, 250 W, -17 A, 80

http://uk.farnell.com/stmicroelectronics/2st2121/trans-pnp-250v-17a-bipo-to-3/dp/1751964

£2.36 + VAT each, from Farnell.

I'm NOT 100% sure about compatibility, but a quick glance looks good.

250 Watts!

[Audioguru]

Yes, you can use the obsolete transistor. If the input is 17V, the output is 12V and you limit the output current to 17A then it will heat with (17V - 12V) x 17A= 85W and it will need a huge heatsink with a fan.
You can use a 3 ohms current sensing resistor and NO series base resistor. Use the capacitors shown on the circuit I posted.   

[davelectronic]

That looks like an interesting transistor. My neivity... Why does it have - before all the listed parameters, other PNP data sheets i've seen don't, or is that normal.

Forgive my newbie questions, where in the circuit would the sence resistor go ? My seem obvious to some, but i've not built many transistor circuits, mostly followed known designs in previous builds i have done. In fact i've mostly used the lm338T and the lm338K and the 2 Amp 78S12cv. Also converted a few ATX psu's and server power supplys, although some are noisey electrically and don't lend themself well to powering HF radio gear.

So i've done very little with transistos really, would have a link to an example schematic diagram ? Thank you.
That transistor is about what i paid for the MJ11015 transistors, x 6 lol redundant for the moment.
So this first two versions of the  TIP2955 transistor circuits are the only ones i have built to date.
First a single transistor model, second with two transistors, that one is powering my cb and linear amp, and has been for about a month now. Max current drawn from the two transistor version is about 7.2 Amps full load.
But i would be interested in this transistor and circuit for it, but for where does the current sense resistor go in the circuit, also the transistors base emitter and collector, are they as per the original schematic that used the 6 x TIP2955's thanks again,

[MK14]

The PNP transistors, are sort of complete opposites of NPN ones. So since NPN has got positive values on the datasheet. You can (if you want to be really strict), make them negative values (as they are opposites). But leaving them positive, makes them easier to understand.
E.g. A cars top speed might be 125 MPH, so we could say the top speed (while in REVERSE), is -25 MPH. But most people would say the top speed in reverse is +25 MPH.

Current sense resistor (On Diagram, Marked Rsc=Resistor Short Circuit) example follows, see page 20 on the data sheet, or image below:

https://www.fairchildsemi.com/datasheets/LM/LM7812.pdf

The new (250W) power transistors, pin configuration, is shown (towards the top) on the datasheet.

[davelectronic]

Ok, thank you for posting that, i understand the reason you explained about the data sheet now, thanks.
Are both these transistors the 2ST2121 or is one a driver for the 2ST2121 ? And the RSC is that like 0R or a jumper link etc.
Its certainly a circuit i could build. I will check out page 20 thanks.
Ok after looking at page 20 there are sone interesting circuits there, i'm guessing as they both say TIP 42 that both transistors are the same. Does that configuration represent a darlington pair, i'm not sure. But it must be good as its in the data sheet.

[MK14]

The schematic was mainly meant to be an EXAMPLE to give you an idea, as to how to do it. Rather than an EXACT suggestion.

Q1 represents the 2ST2121.

Q2 (TIP42 may be ok) detects the overcurrent and limits the output current, accordingly. The resistor value depends on what output current you need, and maybe other factors.
At VERY high currents, it would be something like 0.05 Ohms (suitably high wattage as necessary, SHUNT resistor).

There is a formula on that schematic, but I appreciate that the formula is NOT easy to quickly understand.

If you are NOT 100% sure how to design that part of the circuit, then I guess you can ask here, or somewear, or your friend.

The maximum output current (limit) you want will need to be known. The design of the rest of the PSU, probably needs to be fixed as well, as that may change the exact design of the overcurrent circuitry. Then it probably needs to be tested.
At REALLY high currents, it can get tricky. As a problematic current limit, could overheat the output transistors.

[MK14]

Designing the current limit circuitry, can be deceptively complicated.

For such HUGE currents (which I think you intend), it makes foldback current limiting, or maybe better still, something which switches the unit off (disables the output), somewhat important.

https://en.wikipedia.org/wiki/Foldback_(power_supply_design)

The "foldback" concept is especially important. Otherwise the power transistors, could be badly overloaded (overheated), in certain current overload situations.

A quick glance at the circuit/schematic I supplied above, seems to show that it is NOT a foldback type design.

But I'm NOT 100% sure, because the regulator HAS got its own limiting capabilities built in, and it would take me longer, to determine the full circuit operation.

tl;dr
Probably NOT a foldback current limit type, unless there is a complicated interaction with the regulator, which I have NOT spotted yet!