EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: mitrynicolae on November 30, 2018, 09:10:06 pm

Title: Supply design
Post by: mitrynicolae on November 30, 2018, 09:10:06 pm
Hello all,

I have an issue with the power supply of a regulator board (that is using multiple op amps). The board is powered from a 20V transformer which is rectified and filtered.
On the first attempt the board was using 2 linear regulators 7812 on the input and 7805 on the output. The issue with these regulators is that if the difference between the input and the output voltage is high so is the power dissipation (heat). Then I changed the input regulator with a switch mode power supply like this one (https://www.tme.eu/ro/details/pololu-2103/convertoare/pololu/ (https://www.tme.eu/ro/details/pololu-2103/convertoare/pololu/)). Everything cooled down and "seemed" to work fine, except for the ripple that is not filtered by the output regulator as Dave Jones describes in this video: https://www.youtube.com/watch?v=wopmEyZKnYo. (https://www.youtube.com/watch?v=wopmEyZKnYo.) I say "seemed" because that ripple get to the op amps which don't like it very much and as a result it is reflected on the output.
Since the design is pretty easy to change (only by adding a resistor and a transistor) I decided to do so. The only concern that I have at this moment is that the transistor will overheat. From my calculations and based on Dave information the transistor should not dissipate more than 200mW (200 mA x 1V (drop voltage) = 200mW). I have decided to use this transistor : https://www.tme.eu/ro/details/zxtn25020dfh/tranzistori-smd-npn/diodes-incorporated/. (https://www.tme.eu/ro/details/zxtn25020dfh/tranzistori-smd-npn/diodes-incorporated/.)

The question is: Do you belive that the transistor temperature will exceed 60C at an ambient temperature of 25C?

P.S. Please ignore some components (the first regulator or the capacitor multiplier transistor) since the schematic is not using the actual parts.
P.S.2 Using a lower voltage transformer is out of the question.

Thank you!
Title: Re: Supply design
Post by: SiliconWizard on December 01, 2018, 02:57:36 am
You can refer to the transistor's datasheet for thermal management using thermal resistance figures.
At 200 mW, it won't "overheat" whatsoever (unless it's very poorly mounted on the PCB), but it will probably get warm to the touch. Probably not a problem at all unless it's in a very tight and closed enclosure.

That said, I haven't thought about your schematic much, but one thing that comes to mind is that if the transistor is only dropping about 1V, The 5V regulator will still have 11V at its input, still a lot to dissipate, so it wouldn't change much. I may have missed something though - just took a very quick look.
Title: Re: Supply design
Post by: ArthurDent on December 01, 2018, 04:24:38 am
If you are using the 7812 as shown in the schematic, it doesn't have the recommended input and output bypass capacitors. The 7805 is also lacking an input bypass capacitor that is recommended by the datasheets. This can be a real problem.

If you're only using the 5 volts and the 7812 is only used to drop the input voltage to the 7805, then I'd eliminate the 7812 and use a resistor to dissipate the heat, lowering the parts counts, and lowering the chance of noise generated by the first regulator. If the load is fairly constant why not dissipate the heat that has to be dissipated either way by a resistor?

It isn't clear from your drawing why you appear to be driving an LED with a zener. The drawing seems to be missing something.
Title: Re: Supply design
Post by: spec on December 01, 2018, 11:29:15 am
Title: Re: Supply design
Post by: mitrynicolae on December 01, 2018, 03:26:57 pm
Thank you all for your answers.
@SiliconWizard. As mentioned in the post the LM7812 was replaced with a switch mode power supply that will output 7.5v. If the capacitance multiplier circuit drops 1 volt then I will have 6.5 volts at the input of LM7805. This is just above the lowest allowed voltage.
@spec. As mentioned in the first post I am using this transistor : https://www.tme.eu/ro/details/zxtn25020dfh/tranzistori-smd-npn/diodes-incorporated/. (https://www.tme.eu/ro/details/zxtn25020dfh/tranzistori-smd-npn/diodes-incorporated/.) The schematic is used more for the PCB layout and some parts are placed on the schematic only for their footprint. On the PCB different components might be mounted (having the same footprint) than the ones on the schematic. This is because the application that I am using is too old and it does not know about newer components. Changing the CAD application at this stage of the design is out of the question. At least for this project. In the future we will see.
Using a large resistor to dissipate the excess power again is not an option for two reasons: efficiency (in this case it is better to replace the input transformer with a lower one) and PCB layout (which will be a hell of a work to redesign it as the current space does not allow to add another 1206 SMD package resistor, not to talk about THT power resistors)
Also @spec, you mentioned that there are now better regulators on the same TO 220 package. Can you provide some examples and some pros and cons between LM7805 and the ones you know about?
Also regarding the large capacitors with low ESR, the schematic is divided in multiple sections and the input section is not available in the provided photo. On that section I have the bridge rectifier followed by some large capacitors.
And as a final question, (and conclusion from my side), I supposed that putting larger capacitors on the input and output side of a regulator should not affect in a bad way its functionality. I supposed that those were the minimum allowed capacitors in order to obtain the datasheet values. Please correct me if I am wrong.
Title: Re: Supply design
Post by: spec on December 02, 2018, 12:32:20 am
Also @spec, you mentioned that there are now better regulators on the same TO 220 package. Can you provide some examples and some pros and cons between LM7805 and the ones you know about?
The 7805 was the first low-cost, easy-to-use three terminal regulator, and is still popular, especially for digital circuits. But there are new regulators which have tighter absolute accuracy, better regulation, better ripple rejection, and lower noise. Also, the ground current is less and better controlled, which means you get fewer problems with voltage drops, noise, etc on the ground line.
The two regulators that are used for high-end amplifiers are the positive TI TPS7A4700_TPS7A4701 and the negative TI  TPS7A33. To get the best out of these regulators you need very strict layout and wiring/PCB routing, as mentioned previously.

http://www.ti.com/lit/ds/symlink/tps7a47.pdf (http://www.ti.com/lit/ds/symlink/tps7a47.pdf)
http://www.ti.com/lit/ds/symlink/tps7a33.pdf (http://www.ti.com/lit/ds/symlink/tps7a33.pdf)
http://www.ti.com/lit/an/slyt504/slyt504.pdf (http://www.ti.com/lit/an/slyt504/slyt504.pdf)

Not necessarily in the same class as the two regulator families above, but there is a whole raft of other high performance regulators including: LP38690_LP38692, LT1086, LT1763, LT3080, LT3081

Also regarding the large capacitors with low ESR, the schematic is divided in multiple sections and the input section is not available in the provided photo. On that section I have the bridge rectifier followed by some large capacitors.
And as a final question, (and conclusion from my side), I supposed that putting larger capacitors on the input and output side of a regulator should not affect in a bad way its functionality. I supposed that those were the minimum allowed capacitors in order to obtain the datasheet values. Please correct me if I am wrong.
Putting a large value electrolytic capacitor (with a low ESR) on the input side of a regulator will reduce the ripple voltage and is a good thing. As far as capacitors on the output go, you definitely need to refer to the regulator manufacturer's datasheet but, in general, a large electrolytic capacitor is beneficial. The way it works is that the voltage regulator supplies the current to the output capacitors, both low frequency and high frequency. Then the low frequency capacitor (electrolytic) provides the large low frequency current peaks and the high frequency capacitor (ceramic or film) provides the high frequency current peaks. Of course, this is greatly simplified.

This action is illustrated by the old dodge of putting a 100uf capacitor across the coin cell in solenoid operated wall clocks to extend battery life.