# PSU Question - (Newbie)

#### CCinBR200

Joined Oct 4, 2020
80
Hi, folks.

I'm working on a bench top power supply. I'm combining an instructible and the basic voltage design from Forst Mims' GSiE. So far having lots of fun and general success (burnt up a couple of small parts.)

Stuck on one issue. I want to power 2 small 12v DC fans. They come with no specs. I don't yet know how to draw circuits so I'll describe mine.

wall plug -> 10A fuse -> switch -> 25.2V 2.0A transformer -> bridge rectifier -> (2) 2200uF caps in parallel -> SH-PWS1 buck converter -> .1uF cap -> output posts

First, I used a L7812 in parallel with the last 2200uF cap. Fan didn't run until I powered off and then it ran for 1 sec (I guess as the discharging caps reached just the right point?) When I measure voltage after the L7812 without the fan it measures 12.0V. (Yay!) When I attach the fan (which won't run), I get 1.2V. Is the fan acting as a resistor?

To make sure I had a good fan, I hooked up the fan to the output posts and set voltage between 9.0V and 12.0V. It ran fine and sped up / slowed with the changes in voltage.

What am I doing wrong? Any help appreciated.

Carlo

#### dl324

Joined Mar 30, 2015
12,792
Welcome to AAC!
What am I doing wrong?
It sounds like the regulator is going into self protection mode either due to over dissipation and/or current draw. You should look up the specs for the fans to see how much current they require.

I take your L7812 to be the same as LM78L12, i.e. a low current regulator.

I don't yet know how to draw circuits so I'll describe mine.

wall plug -> 10A fuse -> switch -> 25.2V 2.0A transformer -> bridge rectifier -> (2) 2200uF caps in parallel -> SH-PWS1 buck converter -> .1uF cap -> output posts
Surely you've seen enough to draw a schematic for this. Even a hand drawn block diagram would be better.

#### MrChips

Joined Oct 2, 2009
23,239
You need a small capacitor (about 0.1μF) on the output side of the L7812 regulator.
Check the device datasheet for more details.

#### dl324

Joined Mar 30, 2015
12,792
Here's a schematic of your circuit:

You have too much filter cap and not enough capacitance on the output. An larger output cap would help the regulator from going into self protection when you put a large load on it (e.g. a motor starting).

The absolute maximum input voltage for 78L12 is 35V and you're very near that. The unloaded secondary voltage probably exceeds 35V after rectification.

The regulator will enter self protection mode at a current of about 34mA (maximum dissipation is 750mW). Does your fan require more than that? I suspect so...

Your fuse is too large to give any meaningful protection.

What is the voltage rating on the caps? I think the nearest suitable standard voltage will be 50V; 35V would be too low because your unloaded secondary could provide more than that.

Last edited:

#### AlbertHall

Joined Jun 4, 2014
11,222
Check what voltage you get across the two capacitors.
With a 25V transformer I suspect it will be very close to the 35V maximum input voltage of the 7812.

#### MrChips

Joined Oct 2, 2009
23,239
78L12 voltage regulator is not the same as L7812CV. The latter can handle 1.5A.

#### CCinBR200

Joined Oct 4, 2020
80
Welcome to AAC!
Thank you! This is an amazing resource!

Surely you've seen enough to draw a schematic for this. Even a hand drawn block diagram would be better.
My apologies. A newbie error. I actually had a hand-drawn circuit but thought it might be frowned-upon. Any suggestions for a schematic-drawing program that doesn't have Mount Everest sized learning curve? I have tried to get started on Eagle and LTSpice but it takes me forever to just draw a battery, resistor and an LED.

You should look up the specs for the fans to see how much current they require.
The regulator will enter self protection mode at a current of about 34mA (maximum dissipation is 750mW). Does your fan require more than that? I suspect so...
Unfortunately, I bought these from Aliexpress and I don't know their specs. l will try to get the specs. The fans are small (1.5" square faces). (see pic). Maybe I shouldn't buy everything from Aliexpress..............

You need a small capacitor (about 0.1μF) on the output side of the L7812 regulator.
Check the device datasheet for more details.
Thanks! I saw that Mims put one after the main power circuit but I didn't think that one would be necessary on this side fan circuit. This is the challenge for me doing copy-cat construction while I'm still studying. I don't have enough theory yet to generalize in all situations. (But I am studying Practical Electronics for Inventors so that should improve).

Your fuse is too large to give any meaningful protection.
Thanks for this. Mims doesn't have one but, from my solar work, I learned to fuse everything. I figured that 10A would be the largest current likely. How do I calculate a smart fuse amount?

What is the voltage rating on the caps?
Both the 2200uF and .1uF caps are 50V. From your input, I think this should be ok, right?

Check what voltage you get across the two capacitors.
With a 25V transformer I suspect it will be very close to the 35V maximum input voltage of the 7812.
I will check. thanks!

Thank you to all for taking the time to help a newbie!!

Carlo

#### MrChips

Joined Oct 2, 2009
23,239
My apologies. A newbie error. I actually had a hand-drawn circuit but thought it might be frowned-upon. Any suggestions for a schematic-drawing program that doesn't have Mount Everest sized learning curve? I have tried to get started on Eagle and LTSpice but it takes me forever to just draw a battery, resistor and an LED.
Carlo
Try Scheme-it on Digi-Key website

#### CCinBR200

Joined Oct 4, 2020
80

#### dl324

Joined Mar 30, 2015
12,792
Any suggestions for a schematic-drawing program that doesn't have Mount Everest sized learning curve? I have tried to get started on Eagle and LTSpice but it takes me forever to just draw a battery, resistor and an LED.
I use Eagle. I wouldn't recommend using LTspice to make readable schematics.
l will try to get the specs. The fans are small (1.5" square faces).
You can insert a small value resistor and monitor the voltage. The startup current will be highest, but you may not be able to capture it with a DVM.
I figured that 10A would be the largest current likely. How do I calculate a smart fuse amount?
You're using a transformer with a 5:1 step down rated for 2A, so a 0.5A fuse would be sufficient.
Both the 2200uF and .1uF caps are 50V. From your input, I think this should be ok, right?
Yes.

#### CCinBR200

Joined Oct 4, 2020
80
I use Eagle. I wouldn't recommend using LTspice to make readable schematics.
Thank you.

You can insert a small value resistor and monitor the voltage. The startup current will be highest, but you may not be able to capture it with a DVM.
I have an oscilloscope (but learning how to use it as you can guess by now). I'm guessing I look at Vmax for this spot on the circuit?

You're using a transformer with a 5:1 step down rated for 2A, so a 0.5A fuse would be sufficient.
I'm sorry. I'm not understanding how the math works here.

Thanks again for your time and knowledge.

Carlo

#### dl324

Joined Mar 30, 2015
12,792
I have an oscilloscope (but learning how to use it as you can guess by now). I'm guessing I look at Vmax for this spot on the circuit?
If you put the current sense resistor on the ground side of the fan, you just measure the voltage across the resistor.

You can't do this if you put the resistor on the high side. That will only work if you're using a battery operated scope or you float a line powered scope. The latter isn't safe for novices, so recommending that is discouraged on this site.
EDIT: It will work if the supply is floating.

You can set the horizontal sweep to some slow setting and watch the vertical deflection. You can set sensitivity based on a guess at how much current you expect. Or you can set it to a faster sweep and see the whole trace move. This should work for analog or digital scopes.
I'm not understanding how the math works here.
If you assume 100% efficiency for the transformer, power in will equal power out.
$$I_{primary}*V_{primary}=I_{secondary}*V_{secondary}$$
The secondary is 25.2VAC@2A and the input is 125VAC. So simple arithmetic will tell you that the maximum input current will be:
$$I_{primary}=\frac{2A*25.2V}{125V}=0.4A$$

Last edited:

#### CCinBR200

Joined Oct 4, 2020
80
If you put the current sense resistor on the ground side of the fan, you just measure the voltage across the resistor.

You can't do this if you put the resistor on the high side. That will only work if you're using a battery operated scope or you float a line powered scope. The latter isn't safe for novices, so recommending that is discouraged on this site.
EDIT: It will work if the supply is floating.

You can set the horizontal sweep to some slow setting and watch the vertical deflection. You can set sensitivity based on a guess at how much current you expect. Or you can set it to a faster sweep and see the whole trace move. This should work for analog or digital scopes.

If you assume 100% efficiency for the transformer, power in will equal power out.
$$I_{primary}*V_{primary}=I_{secondary}*V_{secondary}$$
The secondary is 25.2VAC@2A and the input is 125VAC. So simple arithmetic will tell you that the maximum input current will be:
$$I_{primary}=\frac{2A*25.2V}{125V}=0.4A$$
Now I get it (the math)! Thanks for all your help!

Carlo

#### CCinBR200

Joined Oct 4, 2020
80
I kept studying and decided to try a voltage divider. The voltage coming out of C2 is 40V. I have no specs on my fan but found a similar fan which states .96A current draw (I only have one fan - the other burned up - but I will be using 2 so I multiplied .96 x 2 and rounded to 2A total draw). I found a voltage divider calculator and came up with the 18ohm and 10ohm values to drop voltage to 10v. I used 1W resistors. Smoke at R1.

I reasoned: 1) the voltage divider would give 10V going into the L7809CV, within it's range. But, 10V * .96A = 9.6W. Do I really need 10W (or even 20W for 2 fans) resisters for such a tiny fan? It's face is 1.5" square.

Thank you in advance for any help.

Carlo

P.S. The Scheme-It recommendation helped!
c

Last edited:

#### dl324

Joined Mar 30, 2015
12,792
I reasoned: 1) the voltage divider would give 10V going into the L7809CV, within it's range. But, 10V * .96A = 9.6W. Do I really need 10W (or even 20W for 2 fans) resisters for such a tiny fan? It's face is 1.5" square.
The typical dropout voltage for L7809C is 2V (the ST datasheet doesn't give more detailed information). So a 10V input isn't a conservative design.

When you use a voltage divider, the current in the divider needs to be 10 times the load current. 1W resistors will be over stressed.

#### CCinBR200

Joined Oct 4, 2020
80
The typical dropout voltage for L7809C is 2V (the ST datasheet doesn't give more detailed information). So a 10V input isn't a conservative design.

When you use a voltage divider, the current in the divider needs to be 10 times the load current. 1W resistors will be over stressed.
Thank you for the reply. Can you point me in the direction of a functional design? The main power supply (a buck power unit - not shown) is testing out great. I'm just not yet figuring out how to put in a side circuit to power this little fan.

#### dl324

Joined Mar 30, 2015
12,792
If you don't care about efficiency, you can use a power zener diode to drop the input voltage to something acceptable for the voltage regulator. It's a hack, but better than your voltage divider idea.

#### CCinBR200

Joined Oct 4, 2020
80
If you don't care about efficiency, you can use a power zener diode to drop the input voltage to something acceptable for the voltage regulator. It's a hack, but better than your voltage divider idea.
Thank you, Dennis. I have read about a lot of different aspects of this after reading your response. I have been really enjoying learning. It seems like I can lower the voltage from 39 to 35 (the max voltage for the L7809CV) with 7 power Zener diodes in series.

I rethought the whole project. Originally, I bought the 25.2V 2A transformer thinking "bigger is better" (e.g. I could work with as much as 5V @ 10A if needed). The more I try to solve the fan circuit issue, I realize that the high voltage (39V unloaded after the caps) is a pain. So I've had fun learning some of the principles of power supplies but I'm going to take an easier route. I have an extra computer power adapter. It's output is 19.5V 4.62A. I think I'm going to substitute that for the start of my circuit. Then, the voltage should be within the range of the L7809CV. It feels like I'm cheating but, like I said, I learned some things I won't use here but will help later.

I really appreciate you taking time and patience with me.

Carlo

#### dl324

Joined Mar 30, 2015
12,792
It seems like I can lower the voltage from 39 to 35 (the max voltage for the L7809CV) with 7 power Zener diodes in series.
You can turn a regular 1/2W or 1W zener diode in to a power zener by using it with a power transistor.

From a NatSemi Voltage Regulator Handbook:

R should be sized to give 5-10mA in the zener. The transistor will dissipate (Vz+1V) * loadCurrent watts.

The more I try to solve the fan circuit issue, I realize that the high voltage (39V unloaded after the caps) is a pain.
The transformer secondary voltage is specified at the maximum secondary current. Unloaded it will be around 10% higher (about 38V).