2N3055 and heat

[spec]

I tried on Google using words like high voltage regulator, etc. The problem is when you don't know what to call the thing you are looking but you know there must be one it ends up as a 'thod it'.
Another one for the goody section of my notes.

Yeah, in general, searches are pretty useless, ebay is one of the worst.

spec

 

[John Potter]

A resistor power rating is a minimum requirement and any resistor that meets the minimum power requirement, or over will be suitable. For example the 500W resistor would be suitable (but it would be inconveniently big and a touch pricey). This is the same with most parameters for most devices. For example, suppose a circuit needs a 30V transistor, any transistor of 36V or over would do, even a 600V transistor.

The absolute value of emitter resistor is not that critical as long as they are all reasonably the same value.

spec

I started out to build a 36V 4.4A power supply. Due to absolute stupidity (1.4141) I ended up way out of my comfort zone. 52V at 3.3A. The Amps don't matter but 52V for a linear supply suddenly became quite a challenge. Maybe that was not a bad thing though. It's more fun doing something that can be put two ways - 'Pushing the envelope' and '****ing Idiot'. I am tending towards the latter. At times there has been some controversy between people who know far more than me. That's not bad for the first thread ever. It does cause the 'if they can't make their minds up I don't stand a chance syndrome though'.

I know this thing is not really practical but in a way that's irrelevant. With help I will finish it and I will use it within it's capabilities.

 

[spec]

I am glad you wrote that. If not an awful lot would have gone right out the window of my tiny brain.

It is painful to see the angst of the inexperienced specifying components and equipment. One OP got into a right tangle about a replacement power supply. He searched all the net to find a 12V power supply that matched the current shown on the label of the equipment. The value was something like 1.255 Amps. Of course, he would never find a power supply with that exact current rating, neither would it be adequate. He had a slightly wider choice when I managed to convince him that any 12V power supply from about 1.5A to infinity amps would be suitable.

spec

 

[spec]

I do appreciate that the data sheet on the 2N3055 is almost a fairy story from any practical point of view.

Data sheets give the parameters of a particular device, but they don't necessarily explain the interaction between the parameters. It is the same with most devices. For example, a car (automobile) may do 40 MPG and have a top speed of 130MPH, but it will not do 40MPG at 130MPH, probably more like 15MPG.

With power transistors it is often thought that if you need a 100W capacity for example, you just select a 100W transistor and that is that. But nothing could be further from the truth, when you take into account maximum junction temperature, SOA, temperature, rate of rise of voltage and current, frequency etc.

The 2N3055 is an old workhorse, and like the 741 opamp amp, has been eclipsed by much better devices, I am afraid to say. But for a less demanding application the 2N3055 would be fine.

spec

 

[John Potter]

Hi John,

Now that i know that you are going for an LM317H as the control element in your PSU, I have knocked out the skeleton schematic below to illustrate the classic current boost circuit for an LM317. Also shown is automatic input voltage switching, which you asked about.

The LM317 has both over-current and over temperature protection, so the overall PSU would have these features.

The bad news is that the circuit needs PNP power transistors rather than NPN which you have.

By the way, if you plan to buy some PNP power transistors please let me know and I will give you the part number for an optimum device.

spec

View attachment 102575​

I initially went down the LM317 route and read the data sheet, then found the LM338 at 5A data sheet and then realised they would both melt. It was then the 2N got added. Then that's not what the data sheet says it is (in the real world) and I got to 4 x 2N. I then needed a 60V regulator as neither the LM317 or LM338 were any good at 60V. Then I found the LM317AHVT.

That's where you came in spec.

The LM317 HVT is 60V and has the same properties as the standard LM317 apart from the higher V, is that relevant ? I take it I can use it in your circuit ? I have a couple of LM317 and 3 LM338's

You do realise you have just made a rod for your own back spec. I really do appreciate it very much but I feel I'm asking an awful lot - even from the ARC Angel.

If you are OK with that I'll go for it. I can use the 2N's for the little 4A 20V (Even that's 80W). The 4A 22V is just a 'something to do'. I understand 2N's can make quite a Joule thief ? I got the first 2N for that.

There are a few items that I might need help with.
What's :
RL1
Q5 Is the schottky near 39V used as a Volt limiter ?
And as you say Q1 through Q4. (Nasty American things - LOL)
I take it that's a 1.2V to 4a (3 + A) output
Why 2 schottky diodes 'snogging' under R9
I see they still need the emitter resistor, same reason no doubt.

Are you really sure about this ?

One further point. I am VERY patient. If major mods need to be done for any reason - good. I shall be like a pig in..........

If it's any help I do understand how the LM317 type work. Some of it anyway - like the 1.25V between ADJ and OUT, etc. Perhaps the odd protection diode usually referred to as D1 and D2 - IN4007 and the odd little 0.1Uf ceramic cap. I have also absorbed the variable current circuit - bit chunky on the red hot resistors I thought.

I am quite happy with the 'bits' right of the 10,000Uf cap
It was at this point the poo poo hit the fan and I reached for the medication - you.

I have the feeling that I should not ask you this at this point in time. I'm going to anyway.
Suppose I wanted to hang the thing below on the transformer DC 52V.
It's max input is 40V - I think I would be stuffed.
https://www.banggood.com/DC-DC-Step-Down-Adjustable-Power-Supply-Module-p-969200.html?rmmds=search
Just wondered. Don't go into any detail, a simple 'it could be or NO' is fine.
I suppose I could always hang 18 diodes in series - no, I didn't think so.

I have a nice Fluke, old but tested against a new one.

Thanks spec- from your humble minion.

 

[spec]

Hi John- bed time in Somerset- I will post a response in about 9 hours.

spec

 

[John Potter]

It's 02.00 hrs. Hi spec
You have confirmed exactly what I have found out through - I'll use the word 'research' - sounds good. Right at this point in time I think the data on data sheets is about as much use as a fairy tale. Less.
If you are sure you have the time and patience I'll certainly have a go at the circuit of yours.

This is a totally alien subject. If I asked you to design a hydrostatic regenerative steering system for a tank from scratch simply by looking at the Internet - I'm sure you get the point.
Much of what I read was very misleading, all one can do is find other peoples solutions, find the majority and hope it's right. Like this heat problem.
Of course my attempt to make a 160W linear regulator was almost madness - but how was I to know. As one of the other chaps on this thread more or less said - what idiot would choose a linear solution.

The 4 heat sinks have a 0.86 C/W each. They were not overly expensive but it would be nice to use them. They are about the biggest I can fit in a PC power supply case with a 6W external fan. I could improve on that figure by removing most of the bottom of the case and raise it on legs. This would mean a straight pass for the air flow down the fins or even up them. The heat sinks are mounted with their fins vertical. I have almost no rear to the case, it's 90% vent. However after leaving the heat sink fins it only has a 25 mm gap to bend 90 Deg and go out the back. Not exactly a good idea. I have 3 spare cases and the tools to do most things.

At 12V the fan is supposed to produce a flow of 3.2M/s. I could drive it up to 14.4V. It was an automotive fan and 14.4V is considered the maximum a '12V' system runs at. It's not exactly quiet, it looks just like any ordinary PC fan but will helicopter. It would be better at the bottom underneath, less shrapnel if anything gets dropped in it. It's the only one I have. Incidentally if you put 2 fans together the output goes up by about 37%. Most PC fans will run at 20V+ - reduced life of course. I have done a bit on engine cooling systems, it was quite an eye opener. I would love a nice 125mm mixed flow fan.

Time I packed it in as well.
JP

 

[spec]

Hi John,

Your power supply adventures, almost exactly match mine, except there were no chips then.

I came up with this wonderful power supply design with high accuracy, high current and high voltage. It also had precision current control.

One of the experienced engineers at work told me that I would have one hell of a job getting it stable in the frequency domain, and he said that getting rid of the heat would also be a big problem too. Of course, I knew better.

I never did get my perfect power supply working, although I did make a lesser version, which I used for years. But, although the perfect power supply was a failure, it taught me a lot about practical engineering, and that was invaluable.

spec

 

[spec]

If you are sure you have the time and patience I'll certainly have a go at the circuit of yours.

There is no problem with time and patience, unless my missus starts being unreasonable an expects me to do other things: shopping, eating out, talking, fixing the house etc.

The circuit is not mine; it is straight out of one of the National Semiconductor (now Texas Instruments) three terminal regulator (TTR) application notes, and is pretty standard anyway. The only bit that is my design is the automatic switching of the transformer tap, but that is not exactly anything new. By the way, the sketch of post #40 is only an outline for discussion. It has not been analyzed in detail, so it is not necessarily a circuit you could build, at the moment.

spec

 

[spec]

Much of what I read was very misleading, all one can do is find other peoples solutions, find the majority and hope it's right.

This is a major problem, especially with the internet. A lot of the information is either confusing or simply wrong.

One of the best sources of information is the maintenance manuals for off-the-shelf equipment, but even then there are typos, and even gross errors.

The number one problem with any feedback system, especially stabilized power supplies, is frequency stability. That is always the killer. It is especially troublesome with stabilized power supplies because you never know what the load is going to be- it could be anything.

spec

 

[spec]

I have 3 spare cases and the tools to do most things.

One further point. I am VERY patient. If major mods need to be done for any reason - good. I shall be like a pig in..........

This is very good news. The #1 reason for failure to complete a home-built electronics project is the mechanics, so if you have mechanical skills and tools, that is a major bonus, persistence too.

The 4 heat sinks have a 0.86 C/W each. They were not overly expensive but it would be nice to use them. They are about the biggest I can fit in a PC power supply case with a 6W external fan. I could improve on that figure by removing most of the bottom of the case and raise it on legs. This would mean a straight pass for the air flow down the fins or even up them. The heat sinks are mounted with their fins vertical. I have almost no rear to the case, it's 90% vent. However after leaving the heat sink fins it only has a 25 mm gap to bend 90 Deg and go out the back. Not exactly a good idea.

At 12V the fan is supposed to produce a flow of 3.2M/s. I could drive it up to 14.4V. It was an automotive fan and 14.4V is considered the maximum a '12V' system runs at. It's not exactly quiet, it looks just like any ordinary PC fan but will helicopter. It would be better at the bottom underneath, less shrapnel if anything gets dropped in it. It's the only one I have. Incidentally if you put 2 fans together the output goes up by about 37%. Most PC fans will run at 20V+ - reduced life of course. I have done a bit on engine cooling systems, it was quite an eye opener. I would love a nice 125mm mixed flow fan.

This is also very good news. You obviously know about heatsinks and fans - the bigger the better.

No doubt you know that the optimum orientation for a heatsink is for the fins to run vertically, so that convection air currents can flow freely from the bottom of the heatsink and through the fins.

From what you say, there is little doubt that you will succeed with your monster PSU, in spite of your modesty.

spec

 

[spec]

What's :
[1] RL1
[2] Q5 Is the schottky near 39V used as a Volt limiter ?
[3] And as you say Q1 through Q4. (Nasty American things - LOL)
[4] I take it that's a 1.2V to 4a (3 + A) output
[5] Why 2 schottky diodes 'snogging' under R9
[6] I see they still need the emitter resistor, same reason no doubt.

[1] RL1 is a single-pole, two-way relay: a 12V automotive type would suit- they are cheap and freely available.
[2] No, it is a 39V Zener diode, and is the voltage reference that tells Q5 when to energize the relay to reduce the voltage drop across the power transistors and TTR.
[3] The US nomenclature is much better because it is shorter (VT= Q, IC= N etc). You are talking to an Americanophile here- if it wasn't for the Yanks, the UK would be in a sorry state, in my opinion that is.
[4] I was a bit sloppy and just wrote 4A as opposed to 3A+. The output voltage will be 1.25V to whatever maximum voltage the circuit will do, probably about 42V. I have not worked it out. Without a bit of extra circuitry the LM317 family of TTRs cannot go any lower than their internal voltage reference, which is nominally 1.25V.
[5] They are Zener diodes (just one bent end on the symbol). Along with the resistor, they form a snubber to absorb the back EMF (voltage) from the relay coil (inductor) when the relay turns off. This not only protects the relay coil but also the transistor controlling the relay coil.
[6] Yes, the emitter resistors are for current sharing between the power transistors, as previously discussed. But the emitter resistors also have another function in the circuit of post #40. Consider the four parallel power transistors as just one big power transistor (which functionally is what they are). It is important to get the current ratio between the TTR and the power transitors right, so that the TTR, current limiting and over-temperature protection works for the power supply supply as a whole. That is why the emitter resistors are a higher resistance than would be needed just for current sharing alone. I expect the emitter resistor value to increase even more when the detailed analysis is done.

spec

 

[spec]

Yet another post. This time about the overall project and way forward.

The mechanical construction and layout of your power supply is essentially independent of the power supply architecture and detailed electronic design.

The overriding mechanical factor is the cooling of the pass elements (power transistors and TTR). So what I suggest is for you to design and make the mechanical side which will form a good foundation for the development of the power supply circuit. It will also be a good test-bed for any future experiments that you may undertake.

Just one piece of advise: make sure that you minimize wire lengths as much as possible and allow for massive conductors for the high current paths, especially 0V.

Also design for full accessibility, so that you can get at all the power supply components easily. Lack of accessibility is the second most common cause for project failure. For example, if you were housing some electronics in a box, always mount the electronics on the lid and not buried in the box. And always have the connections to any printed circuit board along one edge only, so that the PCB can hinge up for access to both sides of the PCB.

I expect you will be using panel-mount terminals (plus, minus, and earth) on the front panel for connection to the power supply. Panel-mount terminals are shown @ https://www.electro-tech-online.com/articles/core-components.782/

In parallel with the mechanical design and construction, we can thrash out a suitable circuit to do the job.

spec

 

[John Potter]

Hi John,

Your power supply adventures, almost exactly match mine, except there were no chips then.

I came up with this wonderful power supply design with high accuracy, high current and high voltage. It also had precision current control.

One of the experienced engineers at work told me that I would have one hell of a job getting it stable in the frequency domain, and he said that getting rid of the heat would also be a big problem too. Of course, I knew better.

I never did get my perfect power supply working, although I did make a lesser version, which I used for years. But, although the perfect power supply was a failure, it taught me a lot about practical engineering, and that was invaluable.

spec

I still miss the glow of an El34. Would I build another one - of course I would.

 

[John Potter]

There is no problem with time and patience, unless my missus starts being unreasonable an expects me to do other things: shopping, eating out, talking, fixing the house etc.

The circuit is not mine; it is straight out of one of the National Semiconductor (now Texas Instruments) three terminal regulator (TTR) application notes, and is pretty standard anyway. The only bit that is my design is the automatic switching of the transformer tap, but that is not exactly anything new. By the way, the sketch of post #40 is only an outline for discussion. It has not been analyzed in detail, so it is not necessarily a circuit you could build, at the moment.

spec

I'm glad you have the same problems I have. I have one more. Please God make me the person my dog thinks I am. That was very kind of you to use the word 'discussion', especially as I know nuffinc.

 

[John Potter]

This is very good news. The #1 reason for failure to complete a home-built electronics project is the mechanics, so if you have mechanical skills and tools, that is a major bonus, persistence too.


This is also very good news. You obviously know about heatsinks and fans - the bigger the better.

No doubt you know that the optimum orientation for a heatsink is for the fins to run vertically, so that convection air currents can flow freely from the bottom of the heatsink and through the fins.

From what you say, there is little doubt that you will succeed with your monster PSU, in spite of your modesty.

spec

Once you start to use fans the orientation of the thing you are trying to cool isn't anywhere near as important as it is with just convection - until you switch off. I expect the same applies to transistors to some extent. I didn't know it was a monster PSU. Now, if we were to use this 40V - 80A old welder I have or the 15v 60A battery charger / starter .............. I jest.

 

[John Potter]

[1] RL1 is a single-pole, two-way relay: a 12V automotive type would suit- they are cheap and freely available.
[2] No, it is a 39V Zener diode, and is the voltage reference that tells Q5 when to energize the relay to reduce the voltage drop across the power transistors and TTR.
[3] The US nomenclature is much better because it is shorter (VT= Q, IC= N etc). You are talking to an Americanophile here- if it wasn't for the Yanks, the UK would be in a sorry state, in my opinion that is.
[4] I was a bit sloppy and just wrote 4A as opposed to 3A+. The output voltage will be 1.25V to whatever maximum voltage the circuit will do, probably about 42V. I have not worked it out. Without a bit of extra circuitry the LM317 family of TTRs cannot go any lower than their internal voltage reference, which is nominally 1.25V.
[5] They are Zener diodes (just one bent end on the symbol). Along with the resistor, they form a snubber to absorb the back EMF (voltage) from the relay coil (inductor) when the relay turns off. This not only protects the relay coil but also the transistor controlling the relay coil.
[6] Yes, the emitter resistors are for current sharing between the power transistors, as previously discussed. But the emitter resistors also have another function in the circuit of post #40. Consider the four parallel power transistors as just one big power transistor (which functionally is what they are). It is important to get the current ratio between the TTR and the power transitors right, so that the TTR, current limiting and over-temperature protection works for the power supply supply as a whole. That is why the emitter resistors are a higher resistance than would be needed just for current sharing alone. I expect the emitter resistor value to increase even more when the detailed analysis is done.

spec

I just saw a tiny light at the end of the tunnel, only very tiny mind you. May I use the word 'trying' here - good. Basically you are just trying to make one BIG LM317.
I can guess what a TTR is (I think it's the LM317) but if you could possibly use the full title just once it might save me thinking it's short for - Through The Rhubarb.
I have read that if you want 0V just add a couple of diodes on the end. Probably wrong of course. The minimum I will need will probably be the 2V for a red LED. That sounds silly I know but it's a fact.
Sorry but I think I had better find out what these circuit symbols mean. I was OK for resistors, plain diodes and simple stuff like that but once you start adding 'wiggly bits' I'm lost.
I do apologise for using technical terms like 'wiggly bits'.
Please don't worry about rounding numbers. One chap wanted everything to 3 decimal places and then argued about the third place.

 

[spec]

POST Issue 8 of 2016_11_27

Hi again, again John,

Below is an outline schematic, for your consideration, of a classic bench power supply with the addition of a negative voltage rail, which acts as a voltage reference and also allows the PSU output voltage to be adjusted down to a true 0V. This approach has other advantages too.

The symbol with two interlinking circles is a constant current generator, which comprises either a single constant current diode, or some other circuitry.

The PSU also has output current control which can be adjusted.

I suggest that this approach may be a good choice for your bench PSU.

spec

COMPONENT LIST
R1: 100R
R2: 0R47, 1W, min
R3: 0R47, 1W, min
R4: 0R47, 1W, min
R5: 0R47, 1W, min
R6:
R7
R8
R9: 470R
R10: 470R
R11: 470R
R12: 470R
RV1: 10K, Multiturn
RV2: 100R multiturn
C1: 10,000uF, 62vV min
C2: 1,000uF, 62V min
C3: 1,000uF, 62V
C4: 1,000uF, 62V
D1: MBR10100
D2: MBR10100
D3: MBR10100
D4: MBR10100
D5: MBR10100
D6: 1N4007
D7: 1N4007
Q1 2N3055
Q2 2N3055
Q3 2N3055
Q4 2N3055
Q5 2N3055
Q6 2N3055
N1: LM337-5
N2; LM358



DATASHEETS
(1) LM337 Negative three terminal regulator: https://www.ti.com/lit/ds/symlink/lm137.pdf
(2) MBR10100, diode, Schottky: https://www.onsemi.com/pub_link/Collateral/MBR1080-D.PDF
(3) IN4007, diode rectifier: https://www.vishay.com/docs/88503/1n4001.pdf
(4) 62V or greater, 10,000uF res capacitor:
(5) LM358 dual operational amplifier (only one opamp used): https://www.ti.com/lit/ds/symlink/lm158-n.pdf
(6) xxx, 10 turn potentiometer:

 

[spec]

I just saw a tiny light at the end of the tunnel, only very tiny mind you. May I use the word 'trying' here - good. Basically you are just trying to make one BIG LM317.

Correct.

I can guess what a TTR is (I think it's the LM317) but if you could possibly use the full title just once it might save me thinking it's short for - Through The Rhubarb.

I did! (see post #49)

I have read that if you want 0V just add a couple of diodes on the end. Probably wrong of course. The minimum I will need will probably be the 2V for a red LED. That sounds silly I know but it's a fact.

Not correct. A true low impedance 0V output for a bench power supply is absolutely essential.

Sorry but I think I had better find out what these circuit symbols mean. I was OK for resistors, plain diodes and simple stuff like that but once you start adding 'wiggly bits' I'm lost.

I have heard of 'wiggly amps' but never 'wiggly bits'. You are not ex Royal Navy are you?

spec

 

[John Potter]

Yet another post. This time about the overall project and way forward.

The mechanical construction and layout of your power supply is essentially independent of the power supply architecture and detailed electronic design.

The overriding mechanical factor is the cooling of the pass elements (power transistors and TTR). So what I suggest is for you to design and make the mechanical side which will form a good foundation for the development of the power supply circuit. It will also be a good test-bed for any future experiments that you may undertake.

Just one piece of advise: make sure that you minimize wire lengths as much as possible and allow for massive conductors for the high current paths, especially 0V.

Also design for full accessibility, so that you can get at all the power supply components easily. Lack of accessibility is the second most common cause for project failure. For example, if you were housing some electronics in a box, always mount the electronics on the lid and not buried in the box. And always have the connections to any printed circuit board along one edge only, so that the PCB can hinge up for access to both sides of the PCB.

I expect you will be using panel-mount terminals (plus, minus, and earth) on the front panel for connection to the power supply. Panel-mount terminals are shown @ https://www.electro-tech-online.com/articles/core-components.782/

In parallel with the mechanical design and construction, we can thrash out a suitable circuit to do the job.

spec

I have been collecting various bits. I have about 2000 0.25W resistors but only a few caps mainly ceramics. I have just about enough to make a couple of little LM317 regulators with variable pots.
I sent away for and have :- heat shrink (a lot different sizes) , 5 prototyping board (just circular copper pads), a 4 digit V & A meter, an assortment of LMXXX and not much more. I saved a lot of the coloured wire from the PC power supplies.
I am going to send away - my second order to Banggood - for 10 pairs of leads with crocodile clips (I only have 2), 10 pairs of banana plugs and sockets, a selection of fuses, fuse holders. 2 cheap V & A meters for the 4A 20V thing, 10 knobs with 1 - 10 on them. Banggood also have a rotary switch that looks OK.

I was going to make this thing roughly modular. The transformer, rect. for the 36V and 12V (fans and meter supply) are in one case. This is up and running. Attached to this is the second case for the regulator heat sinks. I think I will have to take it apart and rebuild with a third one on the end. No 'walls' between them. It's held together on 15mm aluminium channel underneath. It looks better than it sounds. Should I add some photo's and maybe we should start a new thread?
I have just seen your latest circuit. I wanted variable Amps but didn't like to ask for it. Everything I have seen before has been either very good at starting fires or 'core look at all them bits'.

I think I can get my head around your circuit. I know I can.

The short wires and accessibility can conflict a bit. I was thinking of putting the heat sinks in one case and the control in a third case. I was going to run the power transistors metal to metal with paste and insulate the heat sinks form everything. Heat sinks in a star formation as a single unit. I take it an 8way connection, to this 4 slice toaster, for easy removal is not on ? It would be nice to unplug and swap.

I am going to have 'big thinks' about this spec.
The annoying this is I have some lovely light gauge anodised sheet steel and no way of bending it or cutting it properly. Enough to cover a garage floor. A lathe, stick welder, Dremel type thing, little pillar bench drill, snips, etc., etc. and a lovely expensive soldering iron I can't get tips for. I'm getting a reasonable one from Banggood. The big buck thing has still not arrived.

I have just had another look at your circuit. I was fine until I saw the words 'negative regulator'. I'll get there don't worry. I'll concentrate my thoughts on the hardware and get them back to you.

When you say 'massive conductors' do you mean the sort of wire in a PC PSU - 20A stuff or are we talking copper busbars. I know you are not.

Well spec, thanks again. I have a feeling this winter is not going to seem as long as it once did.
The only negative thing is the awful wait for bits from my girlfriend Ruth at Banggood. I don't mind using CPC for some things but the China stuff is pretty good (so far) and a quarter the price. Perhaps we should discuss getting some 'stock' in. Strangely enough the difference between CPC and China for the 2N's was not that bad. Other things however were. BG do assortment bags of bits. I don't know what I need though. I have a feeling I am about to find out.

If you could possibly add a few 'for example' names to the components it might help me look that type up and understand a bit more. I understand everything is up in the air - just a rough idea might help.