2N3055 and heat

[John Potter]

Hello spec. If you were a bit nearer I would buy you a pint - of whisky. Perhaps cider?
Please don't take this to heart (everyone on here) , I do have a rather strange sense of humour.

Why didn't some bu**er tell me before. (because I never asked)

I have other Forums I use.
I seemed to be boring people but it was that the threads were getting too long and deviating from the original. I've never done anything like this before.
One night I thought 'thod it' and joined 2 more Forums. This is one of them.
One problem is that most replies are in some alien language way above my head. I'm just a thick mechanical engineer.
It must be very frustrating and quite difficult to come down to my level. It's like writing the instructions for something you understand very well. They are usually crap to those who don't.

Is 'spec' short for 'speculum'? (Sorry but I used to test drive them)
Your explanation of ripple was good.

It looks like I shall have to have 'Big Thinks' again. The switching will be no problem but the number of 2N3055's to use and cooling rears it's ugly head again.
I'm inclined to leave things as they are in case I use the wrong setting and go into 'frying tonight' mode. I might even want to pull 100W.
Perhaps I could use just 2 with a thermal cut out?

I see what you mean about capacitors - sorry about the language.

I needed a power supply. This simple linear regulator type offered a way of making one and getting back into electronics and modern components.
However there are not many 160W ones around. Now I know why, all of a sudden it's not so simple.

I was making a second one to 'practice' on It's a transformer from an old car battery charger. It had the old plate rectifier and an output of 17V at 4A. I guess this was because those old rectifiers dropped volts.
I added a few more turns on the output and got it up to 22V.
I thought I could use it for the lower end of my requirements but this new 2 stage idea rather makes it redundant. I'll finish it though.
When I opened this old charger I found most of the connections were just twisted together or wrapped around. The thermal cut out which should have been on the transformer was lying at the bottom of the case. The plate rect. had been replaced by a modern 35A bridge rectifier. The internal fuse had been wired across. At 17V DC the Ammeter usually went to the case. The transformer looks and works well.

I'll have a go at 'quoting' after I have had a good look at your reply spec.

Thanks a lot spec.

Incidentally - I have a 60Amp continuously rated 100+Amp boost car starter/charger. It's never used now.
Are you thinking what I'm thinking ? I jest.

 

[spec]

Hi John,

I needed a power supply. This simple linear regulator type offered a way of making one and getting back into electronics and modern components.
However there are not many 160W ones around. Now I know why, all of a sudden it's not so simple.

While building a power supply is good fun and experience, if you need a power supply for your electronic work just buy one, or three. They are cheap and could be in your hand in a few days. I have three of these, 0V to 30V adjustable at 0A to 10A adjustable, which are recommended. Remember that, like bank accounts, cars, and houses, power supplies can never be too big (in capacity that is, not physically). **broken link removed**

spec

 

[spec]

Is 'spec' short for 'speculum'? (Sorry but I used to test drive them)

It is just my initials re ordered.

Your explanation of ripple was good.

Thanks: most subjects are simple to handle if you boil them down to the basics and approximate. Once you get the basic design right you can then move on to all the fancy calculations to optimize/tolerance a design- if necessary that is.

spec

PS: you live in a nice part of the country.

 

[John Potter]

Hi John,



While building a power supply is good fun and experience, if you need a power supply for your electronic work just buy one, or three. They are cheap and could be in your hand in a few days. I have three of these, 0V to 30V adjustable at 0A to 10A adjustable, which are recommended. Remember that, like bank accounts, cars, and houses, power supplies can never be too big (in capacity that is, not physically). **broken link removed**

spec

Your right, they do look good value. I think Father Christmas might get a letter up my chimney.

 

[John Potter]

Hello spec. It worked.
I have had everything delivered from Banggood except the 60V&A regulator, trust that to be last.
Everything was fine but the packaging is thin on the ground. Well wrapped up in bubble wrap for delicate items.
I still don't know how they can do it free delivery.

I am sure they will come a time when I need something a bit more professional than I can build for the same price or even know how to.
I have spent more than they cost already - but that's not the point.

These power supplies, etc. I am doing are mainly to get me back into electronics. Quite where I go from there is open.
I would like to build an Audio Amp, no chips though. I have a good one but that's not the point.

My wife and I had a car each so I went out and got kit cars to build.
The first was a Dutton Sierra and the second was an EagleSS. I then rebuilt another Dutton Sierra from the chassis up.
Most people never even finish 1 kit car. The Dutton lasted 16 years and the Eagle 10. I sold them. The last Dutton I sold soon after I finished it.
As one 'matures' one finds new hobbies - or goes back a long way to old ones.

I do have an Ebay account somewhere, I really should pull my finger out and use it.
It's a good job your not Bob Jones.

What is your opinion on the number of 2N3055's - should I stick to 4?

Thanks spec.

 

[spec]

Hello spec. It worked.
I have had everything delivered from Banggood except the 60V&A regulator, trust that to be last.
Everything was fine but the packaging is thin on the ground. Well wrapped up in bubble wrap for delicate items.
I still don't know how they can do it free delivery.

I am sure they will come a time when I need something a bit more professional than I can build for the same price or even know how to.
I have spent more than they cost already - but that's not the point.

These power supplies, etc. I am doing are mainly to get me back into electronics. Quite where I go from there is open.
I would like to build an Audio Amp, no chips though. I have a good one but that's not the point.

My wife and I had a car each so I went out and got kit cars to build.
The first was a Dutton Sierra and the second was an EagleSS. I then rebuilt another Dutton Sierra from the chassis up.
Most people never even finish 1 kit car. The Dutton lasted 16 years and the Eagle 10. I sold them. The last Dutton I sold soon after I finished it.
As one 'matures' one finds new hobbies - or goes back a long way to old ones.

I do have an Ebay account somewhere, I really should pull my finger out and use it.
It's a good job your not Bob Jones.

What is your opinion on the number of 2N3055's - should I stick to 4?

Thanks spec.

Building a kit car is something I have always wanted to do, but I did completely rebuild a crashed mini in the 1960s.

Yes you must get ebay set up- you need to set up a Paypal account too.

As I have said before, the prices on ebay are amazing. I did not believe the 99p plus free delivery adverts were real, but you just hit 'buy' and sure enough 5 weeks later the item pops through the letter box all nicely packed too. Or you can often get the same item for about £2 with three days delivery from a UK supplier.

About the power supply. If you view it as an experiment it is an excellent circuit to start with. To advise about how many 2N3055s to use we would need a performance specification first then thrash out a definitive circuit- neither would be a big deal.

spec

 

[John Potter]

Hi spec. Thanks for that it'll help a lot.

Well done on the mini, it's easier to build from new. The second Dutton that had been built badly and I rebuilt was twice the task.

I have a Paypal account somewhere, I've even found the details. I bought some batteries for a Bosch drill that were crap , English company. Never bothered after that. I'll find something for 99P and see what happens. It's not been used for years.

Everything from Banggood has turned up in 3 weeks apart from the 60 V regulator (the bit I want to see most). The circuit boards, heat shrink, meter and 10,000uF capacitor came yesterday. The capacitor does not appear to have a blow out bottom, no rude comments please. That goes with the 1500 resistors, soldering iron tips and 100 LED's.

I have started on my next order. 10 Pcs 50cm croc leads £2.69, a couple of cheap meters.
I'm rather fond of a 600W boost converter. 10V - 60V to 12V - 80V with V and A adjust, at £11, not much more than a packet of fags. I found my old XP welder. I bought a proper welder and it got hidden away. It's 40V at 80A, more like 40V at 20A as a transformer? The 600W boost convertor would tack on the end of that nicely. And before you ask I have absolutely no idea what I would use it for. I think I would change the trimmer style resistors for off board standard ones though.

One of the problems with Banggood is finding kit. The 60V 400W regulator I found is a good example. I was looking under 'power supply' and 'buck converter', etc. and although very many came up it's a 'digital control adjustable dc step down regulator'. Not many 60V ones about

In regard to the spec for the power supply.
I have the 52V at 3.33A power supply done. 10A bridge with 10,000uF (maybe) cap and fused. Output to be switched to 0V - 26V at 3.33A and 0V - 52V at 3.33A. It has a separate 14V x 1.5A DC supply for fans and meter supply.
Note. I'll cover that cap. 3mm I.T. 100 armour plate. I jest but I will cover it.

I need a regulator to go on the end of that.
I have a LM317AHVT 60V linear regulator and other components for a suitable basic circuit with diode protection I found to drive the 2N3055's. A single turn resistor may be a bit coarse so either a multi turn or 2 pots - coarse and fine

2 stage, 0V - 26V at 3.33A and 0V 0 52V at 3.33A (for 26V - 52V) 2N3055's with fan.
I wont be regulating more that 26V under normal circumstances (85W) but the fact that the 0V - 52V at 3.33A could be accidentally run at it's max output concerns me. That's 170W. The 4 x 2N3055's might handle that for a short period but I thought a thermal cut out or thermal warning device might be a good idea if overheating occurs. (I fancy I nice big red LED with Klaxon) The junction of the 2N3055 I have read is 1.5 times the case temperature. Would say a 100C thermal cut out be too high on the case?

The only concern I have is the emitter resistor value and Watts. I have been getting conflicting advice.
Anything from 1W - 0.1Ohm to 10W 0.47Ohm. Could I meet everyone about half way with a 0.2Ohm 4W. I don't mind dropping a few volts. I believe irregularities between the 2N3055's can cause imbalanced sharing of the load?

The volt/ammeter that came yesterday is a 0A-10A and 0V-100V 4 digit, 0.1% accuracy (maybe). It needs 4.5v DC to 30 VDC to power it. I thought I could use the 12V DC fan supply?
**broken link removed**

Should I ever need a 'clean' supply this could be it I believe - worth keeping I think. If I have anything 'dodgy' to power up I should be able to repair this power supply.

I intend to add on as necessary smaller regulators with reasonable meters.

The smallest to adjust down to 20mA at 2v for ie single small 3mm LED. I thought a LM338/LM317 for V with a LM338/LM317 Amp regulator. I have the circuit and parts for this. No meter yet.

I also have 3 small 5A x 1.3V - 32V DC converters. No Amps regulation but it says :- Load regulation: S(I)≤0.8%. I don't see any.
https://www.banggood.com/3Pcs-5A-XL...-Supply-Converter-p-1031856.html?rmmds=search

The largest will be the 400W 60V with V & A regulation with built in meters. This will bypass the linear regulator.
**broken link removed**

I'm going to be honest spec. I really am enjoying myself with your help and others. I was getting terribly bored.

Right now though I have been 'at it' for over 4 hours and my brain urts. Please do not hesitate to criticise any of the above. It's been quite an interesting couple of months to get this far.

When we have finalised or at least got a good idea where this thing is going I'll have a go at a circuit. There is quite a lot to consider having these little 'add on' regulators some plugging into 26V and others 52V.

At least you can have a few days off from my pestering you. I really do appreciate your help and the others of course.

 

[spec]

The only concern I have is the emitter resistor value and Watts. I have been getting conflicting advice.
Anything from 1W - 0.1Ohm to 10W 0.47Ohm. Could I meet everyone about half way with a 0.2Ohm 4W. I don't mind dropping a few volts. I believe irregularities between the 2N3055's can cause imbalanced sharing of the load?

The power dissipation in the emitter resistor is definitive, and and equals, I * I * R. It also equals, (V * V)/R, and it also equals V * I. There is no doubt whatsoever about that (all three equations will give the same answer).

One thing to know is that you would not normally run a resistor at full dissipation, more like half. Thus, if you wanted to dissipate say 5W, you would typically use a 10W resistor. The scientific way to determine the minimum power rating needed for a resistor would be to do a thermal calculation but, while that is not difficult, it may cloud the issue at the moment- just use the x2 rule of thumb.

About the value of emitter resistor, there are two conflicting requirements:
(1) High value for best current sharing
(2) Low value for maximum efficiency (emitter resistors waste power)
Unlike power dissipation there is no definitive answer to this question, but a good rule of thumb is that the voltage drop across the emitter resistor at the highest emitter current should be at least half the emitter/base voltage of the transistor at that emitter current.

The maximum current that any of the 2N3055 will be able to conduct in your power supply will be 1A (due to junction temperature and SOA limitations). At 1A emitter current, the VBE of a 2N3055 transistor will be about 1V (if I remember correctly from the 2N3055 data sheet), so the minimum voltage drop across the emitter resistor should be 0.5V. This means that the emitter resistor should be 0.5V/1A = 0.5 Ohms (0.47 Ohms is the closest standard resistor value, and will be fine). I know from experience, though, that any emitter resistor value from 0.22 Ohms upwards will be fine, from a current sharing point of view.

The power dissipation in the resistor will thus be, 1A * 1A * 0.47 Ohms = 0.47 watts so, using the x2 rule, a 1W minimum resistor would be required.

And that is all there is to it.

spec

 

[spec]

I really am enjoying myself with your help and others.

That is good. You may find this hard to believe, but I also enjoy the posts on ETO and every day I learn something new. Also, it is surprising how explaining something focuses your view.

spec

 

[spec]

2 stage, 0V - 26V at 3.33A and 0V 0 52V at 3.33A (for 26V - 52V) 2N3055's with fan.
I wont be regulating more that 26V under normal circumstances (85W) but the fact that the 0V - 52V at 3.33A could be accidentally run at it's max output concerns me. That's 170W. The 4 x 2N3055's might handle that for a short period but I thought a thermal cut out or thermal warning device might be a good idea if overheating occurs. (I fancy I nice big red LED with Klaxon) The junction of the 2N3055 I have read is 1.5 times the case temperature. Would say a 100C thermal cut out be too high on the case?

You will need to provide current limiting on your power supply, or the 2N3055s will not last long.

A simple circuit can ensure that the 2N3055s never have more than 26V across them.

A thermal cut-out is a good precaution, but is not normally used, and will not be needed.

spec

 

[spec]

I need a regulator to go on the end of that.
I have a LM317AHVT 60V linear regulator and other components for a suitable basic circuit with diode protection I found to drive the 2N3055's. A single turn resistor may be a bit coarse so either a multi turn or 2 pots - coarse and fine

I didn't realise that you were basing your PSU on an LM317AHVT. That will simplify the design greatly in some respects.

spec

 

[John Potter]

The power dissipation in the emitter resistor is definitive, and and equals, I * I * R. It also equals, (V * V)/R, and it also equals V * I. There is no doubt whatsoever about that (all three equations will give the same answer).

One thing to know is that you would not normally run a resistor at full dissipation, more like half. Thus, if you wanted to dissipate say 5W, you would typically use a 10W resistor. The scientific way to determine the minimum power of resistor you would need is to do a thermal calculation but, while that is not difficult, it may cloud the issue at the moment- just use the x2 rule of thumb.

About the value of emitter resistor, there are two conflicting requirements:
(1) High value for best current sharing
(2) Low value for maximum efficiency (emitter resistors waste power)
Unlike power dissipation there is no definitive answer to this question, but a good rule of thumb is that the voltage drop across the emitter resistor at the highest emitter current should be at least half the emitter/base voltage of the transistor at that emitter current.

The maximum current that any of the 2N3055 will be able to conduct in your power supply will be 1A (due to junction temperature and SOA limitations). At 1A emitter current, the VBE of a 2N3055 transistor will be about 1V (if I remember correctly from the 2n3055 data sheet), so the voltage drop across the emitter resistor should be 0.5V. This means that the emitter resistor should be 0.5V/1A = 0.5 Ohms (0.47 Ohms is the closest standard resistor value, and will be fine). I know from experience, though, that any emitter resistor value from 0.22 Ohms upwards will be fine, from a current sharing point of view.

The power dissipation in the resistor will thus be, 1A * 1A * 0.47 Ohms = 0.47 watts so, using the x2 rule, a 1W minimum resistor would be required.

And that is all there is to it.

spec

Hi spec Another explanation that will be copy pasted into my notes. Thanks. Strange how I got such varied answers though. I shall treat yours as :- The Gospel According to spec, Verse 12, line33.

 

[John Potter]

You will need to provide current limiting on your power supply, or the 2N3055s will not last long.

A simple circuit can ensure that the 2N3055s never have more than 26V across them.

A thermal cut-out is a good precaution, but is not normally used, and will not be needed.

spec

I do appreciate that the data sheet on the 2N3055 is almost a fairy story from any practical point of view.
I thought that splitting the transformer to give 26V at 3.33A between 4 x 2N3055's (22W each - 0.86A ) would be within their capabilities. The heat sinks and fan provides just under 1W /C at a reasonable sound level.
There is of course the possibility that I will forget to only use 26V to 52V with both windings connected - that doubles everything.
To be safe where might I find the circuit you mention.
I was looking forward to my big red LED and Klaxon. I'll meet you half way and just have an LED. I have a thermal 60C switch.

 

[John Potter]

I didn't realise that you were basing your PSU on an LM317AHVT. That will simplify the design greatly in some respects.

spec

I was almost about to give up when I had leaned enough to realise both the LM338 and LM317 were around the 30 odd volt level. I then noticed a circuit where the LN317HV was mentioned. That does not exist now but the LN317AHVT does and I have a few. I gave up on CPC as they didn't stock the LN317HV so I sent away to the US for them and got LN317AHVT's back. I then found LN317AHVT on CPC. What fun we have spec. £17 postage from the US wasn't much fun though.

 

[John Potter]

Just for you spec I found the correct non standard resistor :
https://cpc.farnell.com/multicomp/mcknp01wj050ka10/resistor-wirewound-0r5-5-1w/dp/RE07259
0.5Ohm 1W. Only because CPC don't stock a 0.47Ohm 1W.
They do do a nice 0.47Ohm 200W one though.

Does the watt rating matter for this resistor? Is it just for aesthetic / cost / that's all it needs to be reasons?
Just for future reference - nothing to do with this circuit. Maybe sometime in the future I have a 5W one with the correct resistance and that's all I've got when a 1W is fine.
In other words providing it's enough watts it's going to work ?
Please just answer 'yes' or I shall need to take more medication.

Don't panic I am not fitting 200W ones. Not at £19.43 each I'm not.

 

[spec]

Just for you spec I found the correct non standard resistor :
https://cpc.farnell.com/multicomp/mcknp01wj050ka10/resistor-wirewound-0r5-5-1w/dp/RE07259
0.5Ohm 1W. Only because CPC don't stock a 0.47Ohm 1W.
They do do a nice 0.47Ohm 200W one though.

Does the watt rating matter for this resistor? Is it just for aesthetic / cost / that's all it needs to be reasons?
Just for future reference - nothing to do with this circuit. Maybe sometime in the future I have a 5W one with the correct resistance and that's all I've got when a 1W is fine.
In other words providing it's enough watts it's going to work ?
Please just answer 'yes' or I shall need to take more medication.

Don't panic I am not fitting 200W ones. Not at £19.43 each I'm not.

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

 

[spec]

I was almost about to give up when I had leaned enough to realise both the LM338 and LM317 were around the 30 odd volt level. I then noticed a circuit where the LN317HV was mentioned. That does not exist now but the LN317AHVT does and I have a few. I gave up on CPC as they didn't stock the LN317HV so I sent away to the US for them and got LN317AHVT's back. I then found LN317AHVT on CPC. What fun we have spec. £17 postage from the US wasn't much fun though.

The TL783 will take 125V between its input and output.

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

 

[John Potter]

The TL783 will take 125V between its input and output.

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.

 

[spec]

I was almost about to give up when I had leaned enough to realise both the LM338 and LM317 were around the 30 odd volt level. I then noticed a circuit where the LN317HV was mentioned. That does not exist now but the LN317AHVT does and I have a few. I gave up on CPC as they didn't stock the LN317HV so I sent away to the US for them and got LN317AHVT's back. I then found LN317AHVT on CPC. What fun we have spec. £17 postage from the US wasn't much fun though.

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