Hi.
I'm making a 5 to 15.5V/0.75A dual power supply, and have a couple of nagging doubts about the output/variable voltage stage, maybe some-one with far more experience than I have with circuits can see mistakes in the design, if anyone has the time to look over the schematic I'd appreciate feedback and suggestions. I've done a lot of reading, prototyping, struggling to do the important maths, and making small improvements to the design here and there, but I know my limitations so would appreciate a little help.
Information I can provide is: (Like everybody else, I have limited PCB space.) The voltage regulators and MOSFETS are all TO220. The transformer, capacitor filter, bleeder, and soft start parts are built and on a pcb, and working okay. So far I have only done short testing for a few minutes at a time, but understand that this part is fine. I am at the stage of bread-boarding the variable voltage stage, and between my poor maths skills, and a lot of reading, and observation, have doubts that the 7805 will hold up for long periods of time as I think I understand the junction temperature will rapidly rise to unacceptable levels with a load that could reach 0.75A, even at low current and with the MOSFET current bypass stage and with a fan to cool the whole thing. Due to not understanding that lower C/W is better than higher I bought a 23C/W heat sink, so I may as well just throw that in the bin, as from the calculations I did I saw it seemed it would make things worse rather than better... I think the 5V/1W fan I have lying around in a drawer is undersized for its intended purpose, but I'll worry about that aspect later.
The 24V part: 7824 puts out a fluctuating 23.8 to 23.9V - could be slightly better, but it'll have to do. Bleeder discharges in around 3 - 4 seconds to 0V, so that's fine. The bleeder MOSFETs and the 7824s on each supply only reach about 35 degrees centigrade, that is also fine.
The 5 to 15.5V bread-boarded part: The 7805 (at 5V output, with only 3 x SE555 and 1 x CD4060 as the load) reaches 50+ degrees centigrade in the space of a few minutes - not good, I suspect. This is in open air, not even in a small metal box.
Thinking of all the circuits I usually make, I really actually only need two supplies that can handle at most (and very rarely) about 350mA at 5/9/12V, but would like to have a robust PSU that can cope with 0.75A worst case scenario and can be operated from 5V up to 15.5V.
Hope not to have waffled too much, I just hope I have provided sufficient information for anybody who may answer, along with the schematic.
So... I suspect that I could house the 7805 in an igloo, but even so it will never work with this design as it will always overheat well beyond the maximum junction temperature. I understand the main problem is that the 19V change from 24 to 5V is asking ridiculously too much of the 7805, no matter what I throw at it to alleviate the heat created, due to any current flowing through it. Is this assumption correct or am I worrying too much?
Many thanks.

 

A few observations at first glance:

The 1kΩ resistor in series with the 5V regulator will limit the current to about 20mA.
If you reduce that to a 20Ω resistor, it will drop 15V at 0.75A and reduce the dissipation in the 7805.
It should be a 10W resistor.
The maximum dissipation in the 7805 would then be about 7W. That would require a heat sink with about a 10°C/W thermal resistance.

The inductor to ground at the output of the 7805 is a short. What's its purpose?

The diodes at the output of the adjustable supply will cause that output to have poor regulation with a change in current.
I see no purpose for them and they should be eliminated.
If you want reverse voltage protection then connect the diode across the output as you did for the 5V output.

 

Dont see the point of the fets, or a soft start, why dont you use a Lm317 regulator instead for variable voltage,the diodes on the outputs are not much use,

Is the inductor a cooling fan?

 

Hello,

Do you want to make a dual power supply or a dual TRACKING powersupply.
Both versions can be found on this page:
http://members.shaw.ca/roma/supply-2.html

Bertus

 

A few observations at first glance:

The 1kΩ resistor in series with the 5V regulator will limit the current to about 20mA.
If you reduce that to a 20Ω resistor, it will drop 15V at 0.75A and reduce the dissipation in the 7805.
It should be a 10W resistor.
The maximum dissipation in the 7805 would then be about 7W. That would require a heat sink with about a 10°C/W thermal resistance.

The inductor to ground at the output of the 7805 is a short. What's its purpose?

The diodes at the output of the adjustable supply will cause that output to have poor regulation with a change in current.
I see no purpose for them and they should be eliminated.
If you want reverse voltage protection then connect the diode across the output as you did for the 5V output.

Hi, thanks for the useful advice. I appreciate the input, it helps a lot. As you can see, I have a lot to learn, so I'm a bit wary about frying components for the fun of it. I realised halfway through - a bit too late to abandon the project - that my power supply is really badly designed, ...you live and learn. I was finding the calculations and online calculators for calculating heat sinks confusing or those I think I understood were giving not optimistic results...
The inductor is supposed to represent a little 5V fan, for cooling purposes, I may have drawn the wiring incorrectly - the regulator to power it goes to 0V and the black cable of the fan goes to 0V. No I may be wrong about that as well, I hope not, I calculated the power dissipation for the tinny fan, which only uses about 50mA, so I had thought the regulator feeding it should be okay heat-wise.
I saw supplies with one diode across the output, as you say for reverse protection, I thought it was okay to do it the other way, so thanks very much - really glad you said that, it saves the supply from even worse regulation.
Can I place two 10 Ohm 7 Watt resistors in series, instead of one 20 Ohm 10 Watt resistor? I was reading about that this morning, and wonder if that will help. The main reason I ask is that I'm a bit loathe to fork out even more money on this circuit, except for an appropriate heat sink, as it has gone over budget in my opinion, to make a joke of a bad situation: I almost have enough 3 and 7 Watt resistors to build a log cabin by now.
Thanks for your help.

 

Dont see the point of the fets, or a soft start, why dont you use a Lm317 regulator instead for variable voltage,the diodes on the outputs are not much use,

Is the inductor a cooling fan?

Hi, thanks for answering. Any suggestion that vaguely sounds like spending a penny more on this thing by changing components sends a distressed shiver down my wallet's spine by now (!), and the L78 datasheet has a few variable voltage schematics so I'll stick to/ I'm stuck with that for this circuit. Yes, the inductor represents a little cooling fan, the schematic design program I use doesn't appear to have fans among the components from what I saw after looking for one.
The MOSFETS are so the supply is more robust, and for this circuit I prefer how they turn on and off compared to the transistors I have, also, I had a couple of that type here to test something, and as they are high current types I decided to include them where I could have put transistors or lower value FETs.
Why don't you see the point of a soft start? I see that its quite normal for power supplies, to limit the current inrush at power up.
Thanks for the recommendation about the diodes, I'll do what crutschow suggests with the outputs.
Thanks for your input.

 

Hello,

Do you want to make a dual power supply or a dual TRACKING powersupply.
Both versions can be found on this page:
http://members.shaw.ca/roma/supply-2.html

Bertus

Hi, thanks. If I ever finish it successfully (!) (without having a nervous breakdown in the process - if this were a battle of wits, the power supply wins, hands down.) it'll be a dual supply, the original idea was so that when I want to do frequency counting or pulse timing between circuits they'll both run off (in so far as is possible) the same supply references, for slightly better (in)accuracy.
Thanks for directing me to the Laurier Dandy web for power supplies, much appreciated - he has nice circuits, I learnt how to make toggle latching pushbuttons with a 4017 after finding his web a few months ago, so it's nice to be reminded of his helpful web.
Thanks for your suggestion.

 

......................
Can I place two 10 Ohm 7 Watt resistors in series, instead of one 20 Ohm 10 Watt resistor? I was reading about that this morning, and wonder if that will help. The main reason I ask is that I'm a bit loathe to fork out even more money on this circuit, except for an appropriate heat sink, as it has gone over budget in my opinion, to make a joke of a bad situation: I almost have enough 3 and 7 Watt resistors to build a log cabin by now.

Certainly you can put resistors in series.
No different than one big resistor as far as the circuit is concerned.

 

Certainly you can put resistors in series.
No different than one big resistor as far as the circuit is concerned.

Thanks.

 

I don't see any need for the soft start stage, the inrush current at power up is caused by the large electrolytic capacitors after the bridge rectifier. Your soft start is after the capacitors so it really doesn't do anything. Also I don't see any point in the bleeder stage either. If you are using a selector switch to adjust the voltage the output is going to spike to maximum voltage when the switch moves between contacts.

Hi there! Thanks for your knowledge input, I'm still learning simple things, so everything helps. I found this circuit, and copied the schematic as I understood it slowly increases output voltage, reducing the inrush current:
http://powersupply88.com/single-power-supply-with-soft-start.html
http://www.zen22142.zen.co.uk/Circuits/Power/softstart.htm
But based on your remarks, obviously not...
If I have understood, I imagine that from what you say, if I turn a rotary switch from 5 to 9V there will be a 24V spike at the output? Oh dear, in that case I won't be able to change the voltage and have a load connected to the supply without turning it off first, so thank you. Silly question maybe: Does it make a difference if the switch is make-before-break?
The bleeder is because there was 27V lingering for minutes after unplugging the supply, which I was uncomfortable with, I prefer something that is definitely off and without current/voltage present when I switch it off.
Thanks.

 

The first link you provided shows a soft start to power halogen bulbs this would extend their life by slowly ramping up the power. This is fine if you are planning to power bulbs, but most electronic circuits don't like a slow power up. Some IC's will latch up if the voltage rises too slowly.

A make before break switch will probably work for you.

The 27 Volts should discharge through the regulator circuitry after the power is switched off, but if it worries you use the bleeder. Myself I wouldn't worry about it.

 

The first link you provided shows a soft start to power halogen bulbs this would extend their life by slowly ramping up the power. This is fine if you are planning to power bulbs, but most electronic circuits don't like a slow power up. Some IC's will latch up if the voltage rises too slowly.

A make before break switch will probably work for you.

The 27 Volts should discharge through the regulator circuitry after the power is switched off, but if it worries you use the bleeder. Myself I wouldn't worry about it.

Thanks for the additional info., that's really interesting to know about ICs at power up.
I wasn't too fussy when I got the switches, and on checking the part number to its datasheet yesterday, the rotary switches are actually break-before-make, which I hope is better for the load after reading (amongst other things) a tutorial page on switches here on AAC which explains that make-before-break create a moment of two voltages at the output - not keen on that bearing in mind that, for example, 9 + 12 = exceeds the maximum ratings for ICs... Anyway, either way, not to worry, I'll have a look with the tester when I reach that stage and if there's a voltage surge on selecting another position on the rotary switch, I'll just turn off the circuits that are the load before changing the voltage - that's another use for on/off switches.
Can't remember who said "Every answer is another question.", which sometimes with circuits could have been "Every solution leads to another pleasant little problem to solve or get around."

Thanks.

 

Not to take anything away from your efforts, because I'm all for making your own circuits for all sorts of stuff, but... If you want a second way to look at frequencies, pulses, etc (and don't want to spend serious money on an oscilloscope) there are multimeters that read frequency and duty cycle (I really like my Extech EX330) that might help with projects like this.

 

Not to take anything away from your efforts, because I'm all for making your own circuits for all sorts of stuff, but... If you want a second way to look at frequencies, pulses, etc (and don't want to spend serious money on an oscilloscope) there are multimeters that read frequency and duty cycle (I really like my Extech EX330) that might help with projects like this.

Hi. Do you work as a salesperson for that company/do sponsored content? It looks like a nice multimeter, and the user manual has useful information you can apply to using other 'meters, especially like how hot the temperature sensor cable can get before it'll melt. I have one that reads frequency and so on, and a cheap oscilloscope - it would be hard to calibrate a frequency counter otherwise, but in my case the point is making the effort to learn how things work and being able to make them or basic versions.
Anyway, after only a week of signing up to these forums, between the prattish shouting style infantile criticism from one guy who does exactly the same thing with other people, and I'm surprised the moderators haven't picked him up on his mean-spirited, unhelpful dopey remarks like "The capacitor's the wrong way round." to some-one else, offering no help whatsoever, like a sneering oaf, and on top of it giving dodgy advice that some-one else had to correct - what's the point of a member like that in an allegedly "friendly, helpful geeks" website..., and by now my feeling like the butt of a snyde joke around here after genuinely seeking friendly - neither hostile nor snotty - help with a power supply that has been pretty hard to research and put the pieces together before starting to make it, all I can say is that I left primary school a long time ago. if it makes you feel good about yourselves, best of luck with that attitude.
AAC has felt like a rather unwelcoming place, shame because it's a nice web. Shame, some of the other posts I read have genuinely helpful and polite interaction from decent looking members. Whatever. So long...

 

Ouch!

No, not working for anyone or getting kickbacks. Just didn't know how well equipped you were and thought I'd mention the in between product - until recently I thought it was oscilloscope or nothing at all for these issues.

As for your other comments, I've seen a lot of the negativity you referred to as well. It's a shame you saw so much, so soon, because in my experience it's usually quite rare here and has only flared up recently. Hopefully it will die down, and hopefully you'll give aac another shot.