Maker Pro
Maker Pro

Transistor hfe

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Oh ok, some posts back, maybe a page or two, a comment was made, "this is draggng now" Not my comment. Anyway I will keep you updated then.

I do plan to improve layout on a second board, maybe I should have breadboarded this, but it's one item I didn't have yet. Also this simulation software would be useful. I'm not 100% sure, but I think with a base resistor of 100R value, it's not enough to bias the transistor. This is just a thought, I'm no expert just a newbie.

My final goal is a PSU that can comftably run up to 20 Amps, but typically 15 Amps 50% duty cycle. Ideally with about 4 x MJ11015 transistors. But it's a strange thing acquiring an extra 5 Volts with an R7 value of 100R.

I will try 100R for R7 again, a suggestion was made else ware to change the ceramic power resistor to 1R or even 2R value, and this would work with an R7 value of 100R. But I thought then that the power resistor would need i increasing in wattage. But it something to try. thanks for the positive reply.
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
latter today if i can i will swap out the 24R resistor for 100R and put a 1R 10 watt power resistor in there.

see if that works, and measure voltages acorss both resistors, with no load, and loaded to 50 watts.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
I like this sort of thing, I have never seen any thread closed for too many threads, If people are getting bored and they don't want to help you then they can go and ..... well I can't say what I want to say :) This sort of issue is so valuable for people wanting to learn electronics. This is what gains experience.

If you feel under pressure then we can look at this outside of this thread, but I would rather construct an informative thread so everyone that's interested can join in and learn. I think we should start of with some nice close up photos of both side of the PCB. We can then potentially mark up the photo for you and ask you to make some measurements. You ok with that?

Adam
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
I like this sort of thing, I have never seen any thread closed for too many threads, If people are getting bored and they don't want to help you then they can go and ..... well I can't say what I want to say :) This sort of issue is so valuable for people wanting to learn electronics. This is what gains experience.

If you feel under pressure then we can look at this outside of this thread, but I would rather construct an informative thread so everyone that's interested can join in and learn. I think we should start of with some nice close up photos of both side of the PCB. We can then potentially mark up the photo for you and ask you to make some measurements. You ok with that?

Adam
yes Adam sure, that's not a problem. The underside of the veroboard has a few alterations since i started it, veroboard makes a poor breadboard lol.

I can only get pictures to upload on the pc, so will do that later today if i can. Yes i totally agree with your opinion on creating a useful thread, so long as people can see past my not so sharp workings out, where i have got some things wrong. A posative outcome would be ideal.

i will post some pictures later, try and get in as close as possible.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Ok, no rush.... we can then put a plan together to try and understand whats going on.
Cheers
Adam
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Well i have some numbers.
I can't load the output at 17 volts, i don't have enough sacrafical lamps.

R7 100R base, 0.1R 10W power resistor:
Base resistor across it 1.02 Volts.
Power resistor 16.7mV.

R7 100R, base resistor.
Power resistor 1R 10W.

Base voltage across the resistor,
1.21 Volts.
Power resistor volts across it, 165mV.

R7 24R, Base resistor
Power resistor 0.1R 10W,

Base resistor volts across it, 1.18 Volts.
Power resistor volts across it, 6.4mV.

R7 24R Base resistor.
Power resistor 0.1R 10 W.

Loaded to 50 Watts.

Base resistor voltage across it 1.94 Volts.
Power resistor, volts across it 0.457 Volts.

As soon as i can i will upload some pictures of the board, i have carefully checked the board with a powerful magnifier glass.

Hope these numbers throw some ideas up.

Forgot, i did try 2 x 100R in parallel, it dropped the output voltage to 16.5 Volts, but as it seemed hardly any better than 100R did not think it was worth mentioning.
 
Last edited:

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Not the best pictures, but 5MP is the best i have at the moment. I doubt the solder side can be traced out easily, as i have moved some leads on previous resistor tests.
 

Attachments

  • IMG_20160318_220149.jpg
    IMG_20160318_220149.jpg
    269.7 KB · Views: 65
  • IMG_20160318_220119.jpg
    IMG_20160318_220119.jpg
    232.3 KB · Views: 56

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I hate to ask again. But have you measured the quiescent current of the regulator.

This is perhaps the single most important measurement you can make right now.

connect the common leas to ground, and use your multimeter in a low A range between the input and your filtered DC.

On another subject, "leakage" in your transistors will reduce (not increase) the voltage across and the current through R7.

A diving analogy is a leak in an air line prior to the regulator. is this likely to increase or decrease the pressure? (it's not a perfect analogy but I hope it helps)

There is also a small chance that the 2A version of the 7812 will operate slightly differently in this circuit than the normal 7812 but I can't see why.
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
So open the circuit on the input side of the regulator and measure in series idle no load ? Does that sound right. Not sure how it will help in getting that 17 Volts output down. But i suppose it will show if the regulator is contributing to the not so regulated output. I doubt it is contributing, as the voltage would be as per the regulator if it was.

It will be another day now ( battle with long term illness) But yes i will put 100R back in circuit and open the circuit on the regulator input and measure the current. Must remember to dissconect the cooling fan lol.

Just a theory, but i think 100R is to high to bias the transistor well enough to kick it into action. Just the lower voltage across the base resistor, compared to the voltage across a resister 1/4 of that value.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Well i have some numbers.
I can't load the output at 17 volts, i don't have enough sacrafical lamps.

R7 100R base, 0.1R 10W power resistor:
Base resistor across it 1.02 Volts.
Power resistor 16.7mV.

R7 100R, base resistor.
Power resistor 1R 10W.

Base voltage across the resistor,
1.21 Volts.
Power resistor volts across it, 165mV.

R7 24R, Base resistor
Power resistor 0.1R 10W,

Base resistor volts across it, 1.18 Volts.
Power resistor volts across it, 6.4mV.

R7 24R Base resistor.
Power resistor 0.1R 10 W.

Loaded to 50 Watts.

Base resistor voltage across it 1.94 Volts.
Power resistor, volts across it 0.457 Volts.

As soon as i can i will upload some pictures of the board, i have carefully checked the board with a powerful magnifier glass.

Hope these numbers throw some ideas up.

Forgot, i did try 2 x 100R in parallel, it dropped the output voltage to 16.5 Volts, but as it seemed hardly any better than 100R did not think it was worth mentioning.
Thats good work, I will have a look at this later today.
Adam
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Thank you Adam. Theory only nothing solid, but i think 100R is to high to get the transistor conducting forward properly. With 100R i think its bypassing the regulator leaving it vertually redundant, otherwise if it was functioning it would regulate that 12 Volt output. With a 24R value for R7 i get a normal idle voltage. Yet when i connect the 50 watt load the voltage it increases by 0.7 volts for about 1 minute before settling back to normal 12.20 Volts. Anyway i'm prepared to follow the R7 100R trail to its conclusion, what ever that may be.
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
I have over looked some thing... Whilst, and all the readings i have taken, where with the cooling fan running with its 47R series dropper resistor.. doh, looks like i've got to remeasure with it disconnected. A compleate oversight on my part.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Ok I have managed to simulate your findings. But I dont know why its doing it at the moment. If I connect a 6K8 resistor on the input to the reg, I get 17 Volts on the output. If I then change the 100R to 47R it works. It maybe a start up thing but not sure. The 100R should be able to supply the 8 mA needed for the reg.... need to look a bit deeper.
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Well after some more testing i'm even more puzzled.
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.

24 ohms base Volts aross it, 144.2 mV . Its unstable at light loads, gaining at least 2 Volts with a 20 watt load, 0.7 Volt increase with a 50 watt load. But will drive the cooling fan and its series resistor with stable voltage.

10 ohms base Volts across it 60.2 mV . When loaded to 50 watts its stable with 0.3 voltage drop. 20 watt load sends the output between 11.50 and 11.98 Volts seems unstable, drives the cooling fan and its series resistor stable output.

The power resistor i used was 0.1 ohms 10 Watts. For 1 ohm resistor i would need it to be at least 20 Watts. I don't have one to hand at the moment. I tried 10 ohms base resister and a 1 ohm 10 watt power resistor. Volts across the 10 ohm base resistor was 5 Volts, i know way to much for half watt resistor. The voltage across the 1 ohm resistor was 4.1 Volts, again to much for that resistor. It was stable with all loads until the 10 ohm base resistor failed.

I think from this a base resistor has to be higher than 10 ohms, maybe 15 ohms 3 watts, and the power resistor of 1 ohm 20
watts. A 10 ohm resistor and 1 ohm power resistor does not limit the voltage below 5 Volts with a 50 Watt load. Hope some of this makes sense. Thanks Adam !
modify_inline.gif
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
I think maybe there is oscillating going on maybe, but wih no scope i can not confirm that. I've a strong feeling this MJ11015 transistor is not best suited to this application.

Its internal structure i think means it will behave badly in this emitter follower type circuit. I can't see any resistor value combination working properly.

So whilst i've not thrown the towel in compleatly on this transistor, maybe i should look for another PNP variant, non darlington maybe. I would really like the TO3 style package.

The MJ2955 transistor comes to mind, but not sure if its gain is good enough for use in this type of emitter follower regulator circuit. As i say i have not given up on the MJ11015 transistor yet, but will look at alternatives.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
So open the circuit on the input side of the regulator and measure in series idle no load ? Does that sound right.

Do it like this

regqui.png

We need to know what this current is -- and have nothing else connected!!

And if you're having problems with oscillation (or suspect you are), try placing 0.1uF capacitors between the output and ground, and the input and ground -- although in this case, between the left side of R7 and ground may be better. (see below)

Ensure the distance between the capacitors, R7, and the 7812 (in fact, everything inside the grey box below) is kept as short as possible. Use caps rated at about 50V (that is likely to be the lowest you'll find -- higher is OK too) and that are "ceramic". R7 should be 1/4W or higher (if 22R or higher).

regvolt.png

If you suspect oscillation, use your multimeter on an AC range and read the output voltage. Compare this with the measurement of the DC voltage prior to the regulator. On most good meters, the AC reading will b the approximate RMS equivalent of and AC component. Ideally this should be near zero on the output (even under load), and near zero on the filter capacitors (at zero load) rising as load is added.

If your meter reads something similar to the DC voltage in position A above, then place a capacitor (1uF would be fine) in series with your meter lead.

If you can give us the DC and AC voltages at point A and B under no-load and under higher load that would be great. The same measurements with the lower R7 would also be interesting. Oh and throw in the voltage across R7 (both AC and DC).

If the caps (the ones labelled 0.1uF) seem to improve things, you should also see the AC voltage at point B fall (assuming the problem is oscillation)
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
Thanks Steve. I will have to put another test circuit together. There is only so far you can breadboard on veroboard lol, its not constructed for that.

I've cast a net far and wide, and the general view is these high power darlingtons are best versed in a closed loop circuit with some feed back probably.

I could put another together, i had a feeling that was how you measured the idle current of the regulator. Thanks for the tip on multimeter use to try and spot osallating, that's one i will remember.

I don't think i can tame these beast transistors, so for now will use an alternative. Its many opinions over a few forums that think unless i compleatly rethink this circuit (closed loop) I will continue to chase my tail.

I don't have infinate time on my side (long story) But have wanted to construct a high current linear power supply (On my bucket list) Thanks for all the help here, and to Adam for trying this in simulation. For now its time for me to move on forward.

Thanks again for everyones help !
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I've cast a net far and wide, and the general view is these high power darlingtons are best versed in a closed loop circuit with some feed back probably.

That's exactly what this circuit is doing. :) However darlingtons have both higher gain and greater phase shift than a single transistor, so it's quite reasonable that these could lead you to instability. Low dropout regulators use a PNP pass element (like you're doing here) and they are notorious for being unstable.

If I get time, I'll create one of these myself using high power darlingtons and see what happens.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Oh, and I just noticed that I swapped emitter and collector in the image in post #176. Oops.
 

sureshot

Jul 7, 2012
234
Joined
Jul 7, 2012
Messages
234
yes that would be very interesting if you have time to put a working model together.
Thanks again.
 
Top