Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Voltage Reg Heat?

Status
Not open for further replies.

kentken

New Member
I have a circuit that has 6 dig led display, 6 74hc4511 drivers, and a pic 16F872.
From what I calculate this should draw less then 400ma. But it draws over 600ma.
A 500ma regulator gets to hot and shuts down.
I just put a 1A regulator on it, and the heat sink gets to hot to touch.

What Can cause this? How can I check it?

Thanks
Kent
 
Its hard to say what might be causing this without seeing your schematic.

If you get a Regulator with a package like a TO-220 you can attach a heat sink to keep the heat to a decent level. How much power is the regulator dissipating if its more than a Watt or so you will need a heat sink even if its operating correctly.

Brent
 
Here is the sch.
 

Attachments

  • countersch.jpg
    countersch.jpg
    307.8 KB · Views: 917
kentken said:
I have a circuit that has 6 dig led display, 6 74hc4511 drivers, and a pic 16F872.
From what I calculate this should draw less then 400ma. But it draws over 600ma.
A 500ma regulator gets to hot and shuts down.
I just put a 1A regulator on it, and the heat sink gets to hot to touch.

What Can cause this? How can I check it?

You don't mention what voltage you are feeding into the regulator?, dissipation in the regulator is the current (600mA) multiplied by the voltage across it. You also don't mention what size heatsink you have on the regulator, I would suggest you will need a pretty good sized one when drawing 600mA.
 
I am using 12v in.
the heatsink is 3/4 x 1 3/4 x3/8 ~~|_|~~ shape
Is the draw of these parts sound right? Or is it to high?

And Yes it does work, the 500ma would overheat and shutdown, the 1A just gets really Hot

Thanks
Kent
 
12v down to 5v is a pretty significant voltage drop. (By significant, I just mean a worthy drop, not "Hawly cow, dropping 7 volts!!").

As you know, drawing 600mA and dropping 7 volts will produce some heat, so I woundn't be alarmed if you have a nice and toasty regulator. 7(.600)= 4.2watts.

If you're not comfortable with the heat, shoot, put a larger heat sink on, or a heat sink with more fins. It sounds like the heat sink that you have is pretty small to start with.

If you're interested in figuring out the approximate thermal rating of a heatsink, based on the material, dimensions and physical structure...check this link out.

**broken link removed**

Good luck bud!
 
Johnson777717 said:
If you're not comfortable with the heat, shoot, put a larger heat sink on, or a heat sink with more fins. It sounds like the heat sink that you have is pretty small to start with.

Yes, I would suggest using a MUCH!! larger heatsink, 4.2W is a lot to dissipate in a 7805 regulator.

You could also fit a series resistor in front of the 7805, a 5.6 ohm wirewound would roughly half the dissipation in the 7805 - the other half dissipating as heat in the resistor - I would suggest using a 10W resistor, in order to stop the resistor getting too hot, but you could experiment here.

Certainly 600mA sounds quite high for your current consumption, what resistors are you using to feed the LED's?.

On a purely design point of view, using seperate 4511's is inelegant and expensive - normal practice would be to use a single 4511 and multiplex the six displays - using 6 simple transistors as drivers. It would also make your board considerably smaller.
 
In addition to what everybody said you can use 2 more methods:

A. Get a TO-3 7805 (You can see the ST microelectronics 7805's datasheet for the packaging information) it gives very good heat dissipation.

B. Put another TO-220 7805 in parallel (but not to close to each other) it will give you the best heat dissipation you can get with the little heatsinks.

You might want to check another things to ensure you did evrything right to prevent the regulator ho heat, did you put a lorge capacitor in between the input and the ground?
This capacitor intend to drop the ripple from the input causing the regulator to get unstable and gain heat.
Try to put something between 220uF-470uF/16V.
 
Nigel Goodwin said:
On a purely design point of view, using seperate 4511's is inelegant and expensive - normal practice would be to use a single 4511 and multiplex the six displays - using 6 simple transistors as drivers. It would also make your board considerably smaller.

Multiplexing them would also lower the current consumption a lot...
if you have program cycles to spare i would consider it
 
[/quote]Multiplexing them would also lower the current consumption a lot...
if you have program cycles to spare i would consider it[/quote]
??? Not if you want the same apparent brightness it wouldn't.
 
udi_hakim said:
In addition to what everybody said you can use 2 more methods:

<snip>
B. Put another TO-220 7805 in parallel (but not to close to each other) it will give you the best heat dissipation you can get with the little heatsinks.
<snip>
Parallel regulators don't share current equally. The one with the highest output voltage will hog most or all of the current, and it will be just as hot as before.
 
Thanks for all the great feedback. :)

I also have a 1N4001 diode before the regulator, and a 10uF 16v Tant cap before the regulator.

The reason that I went with 6 4511's is so I didnt have to multiplex the display, and so I could get the brightest display. But this does sound like another option that I could use, please explain how.

The heat sink that I am using is a Aaid Thermalloy 507002b00000 with a thermal resistance of 15.6.

I built three of these and this is the only one with problems, I will check the others to make sure.


If I lower the input voltage, will this lower the temp? What else can I do?

Thanks again
Kent
 
kentken said:
Thanks for all the great feedback. :)

I also have a 1N4001 diode before the regulator, and a 10uF 16v Tant cap before the regulator.

Have you got a capacitor on the output of the regulator as well?, it's important to have capacitors close to the inoput and output to prevent instability.

The reason that I went with 6 4511's is so I didnt have to multiplex the display, and so I could get the brightest display. But this does sound like another option that I could use, please explain how.

One of my PIC tutorials shows how to multiplex two 7-segment displays (without using any 4511's), you can easily expand it to cover more, and either use a 4511 or not, if you don't you would need drivers from the PIC, particularly as you want a bright display (and the PIC pins are only rated at 30mA).

You can run multiplexed displays just as bright, for your application you would run them at six times the current, as each one is only on 1/6th of the time.

The heat sink that I am using is a Aaid Thermalloy 507002b00000 with a thermal resistance of 15.6.

I built three of these and this is the only one with problems, I will check the others to make sure.

If they are all the same you may have a fault on it somewhere, check it carefully and compare it to the others - what current do the others take?.

If I lower the input voltage, will this lower the temp? What else can I do?

Yes lowering the input voltage will lower the temperature - W=VxI, for other ideas read all the previous suggestions - my resistor suggestion is probably the easiest - but I'm biased 8)
 
Suggestions

If I lower the input voltage, will this lower the temp?
Yes, an 8V supply would provide the 7805 regulator with 3V of headroom and lower the dissipation by (7V-3V)(0.6A) = 2.4W. Using a low-dropout (LDO) type regulator in place of the 7805 would allow you to reduce the input voltage even further and waste less power.

If you're stuck using a 12V supply, consider a simple step-down switching regulator. It requires a few more components than a 7805, but will have far greater efficiency than a linear regulator at this Vin - Vout difference. Linear Tech, National Semi, Maxim and Micrel are a few companies that offer suitable buck-type regulators. Here's a simple step-down switcher you can probably build from parts out of your junkbox:

**broken link removed**

And Ron's right, tying the outputs of multiple regulators together is not a good idea. If you're trying to spread out power dissipation among multiple regulators, split the IC's up and power separate sections of the circuit from each regulator, using a common ground return.
 
multiplexed displays

One of my PIC tutorials shows how to multiplex two 7-segment displays (without using any 4511's), you can easily expand it to cover more, and either use a 4511 or not, if you don't you would need drivers from the PIC, particularly as you want a bright display (and the PIC pins are only rated at 30mA).

You can run multiplexed displays just as bright, for your application you would run them at six times the current, as each one is only on 1/6th of the time.

Nigel makes a good point, why use 4511's at all? From the schematic you posted, the 16F872 has enough spare pins (3 more needed) to directly multiplex the 7-Segment displays (the 6 LE lines become the multiplex control lines). It's more effort on the firmware side, but does away with 6 ICs.

My data sheet for the CD4511 shows that it can sink up to 25mA. No spec is given for sourcing current, I assume it's in the same ballpark. The 16F872's pins can sink or source 25mA, so I don't know if you can run the displays at 6 times the current and 1/6 the duty cycle. It depends on your present current limit. Measure the "on"current for a single LED segment to find out.

Be careful during debug if you use this method, a firmware "mistake" which leaves the multiplex outputs turned on can easily fry the display. Suggest keeping the higher valued current limiting resistors in until you've fully tested the firmware (Dim displays are better than burned out ones). The higher current/lower duty cycle approach works well, just make sure not to dissipate more than the average rated maximum power of the LEDs.
 
Well, I tested one of the others, and it overheats to, it draws 750ma through a 78M05. :(

I have another project that runs off of 12v to that draws 350-400ma, and it dont heat up at all.

I think what I need to do is lower the draw of the componts, I gess multiplexing would be the best.

I was also looking at Max6951 or MAX6955 display drivers. Has anyone used these?
How well do they work?
Better then only using Pic and transistors?

Thanks
Kent
 
kentken said:
I was also looking at Max6951 or MAX6955 display drivers. Has anyone used these?
How well do they work?
Better then only using Pic and transistors?

It's all down to what you want (and the cost, both in money and board space). A PIC with simple driver transistors works perfectly, and gives you total control over multiplexing rate - you will need seven driver transistors for the segments, and 6 for the displays, plus the required resistors - however, these are just cheap NPN and PNP transistors. Using custom drivers chips is likely to be more expensive, take up more board space, and limit you to what the chip designers wanted.

But both should work fine.
 
That is what I came up with to, these chips cost about $8-$10.

What I decided to do was use the 4511 as a BCD to 7-seg converter, all display lines in perallel, and 6 CC transistors.
I heard of someone using a 74s138 in place of the transistors. Work? How?

I realy hope that this is the answer, I am tired of messing with this project.

Thanks again
Kent
 
kentken said:
That is what I came up with to, these chips cost about $8-$10.

What I decided to do was use the 4511 as a BCD to 7-seg converter, all display lines in perallel, and 6 CC transistors.
I heard of someone using a 74s138 in place of the transistors. Work? How?

Well 'laroche73' pointed out that the 4511 is only rated at 25mA per pin, so you could only feed each segment 4mA, which isn't a lot!, particularly when you mentioned you wanted the displays as bright as possible. As he also mentioned, you may as well use the PIC pins directly, they have a similar drive capacity anyway.

A 74s138 is an address decoder chip, I don't know what it's drive capabilities are - but probably no better than the 4511. As you have plenty of spare I/O there's not much point using either that, or a 4511.

What current are you wanting to feed the LED's with?.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top