595 issue

I am using two serial in parallel out shift registers. I am using 2 555 timers to clock in the latch and clock. Data is supplied from my microcontroller.
It was working fine on day one, but today (day 2) the shift registers ONLY work if I touch them (which is probably not a good idea).
I have been sniffin around the circuit with a multimeter. I am using a 7805 voltage reg. to power both 555s and both 595s. ALL grounds are tied (ground from micro, ground from 7805, and ground from 12v DC [used for high power lights that 595s are controlling and also used as input to 7805]).
I mention all this because when I look at the voltage from the 7805, the input only reads around 1 volt and the output is even lower (0.5 volts).
Why would this be? I thought maybe using a second power source just for the regulator would solve the issue, but w/ a seperate 9 volt still the same issue (but everything WAS working w/ this configuration yesterday).
I am guessing the logic levels are too low for the 595, especially because output of the timers is even lower then their input voltage.
So, it DOES work when i touch them, but that is not really feasible. I didn't use a socket for them, so i can't pop in new chips.

Any suggestions, similar experiences? I am I just adding current by touching? If so how can I get more juice into/out of the 7805. It tried replacing the 7805 twice and still same result.

thanks for the patience,
if i need to post a schematic thats fine...
Edit/Delete Message

Touching a chip to make it work suggests a breadboard with wonky tie points. Pressing the chip down helps to make better contact.

Why would the input to the 7805 be 1 volt??? That's something to investigate further. I thought you had it hooked up to a 12V supply.

Perhaps your power supply is having too much current drawn from it and is current-limiting?

The 595 circuit WAS working just fine so I can't imagine it is/was a case of cold solder joints.

Current limiting: That might be it. My adapter is 1 AMP and is power 12 150mA LEDs as well as x2 555s and x2 595s. Say I use a 2 amp adapter, would i run the risk of overheating the ICs?


I'd have to see the schematic. Something tells me you're not driving 12V right into the 595 outputs (at least not for very long).

Hey RC,
OK, some words about the schematic. Things get really messy when it gets to the outputs. All the outputs of the 595s (12, im not using 4 of them) are connected to base of 12 individual TIP120npn Transistors. The collector of each tip120 goes to one leg of a high power led, and the other leg is tied to 12v+.
Second, I can't get the 595 supply pins (8 and 16 i believe) to show up in Eagle. But trust me, those are connected to 5v+ and ground coming from the 7805 regulator.


I really hope this helps. I am totally self taught in the realm of circuit building, it being only a part of my artistic practice. Thus I need some help when it gets this complicated (for me at least)!
As always, thanks, thanks and thanks again.

Can you tell me more about the LED's? Are they meant to be driven directly by 12V?

You at least need resistors in series with each output of the 595's. I would recommend something like 2.2k-3.3k to start with.

Oh, and you can use the INVOKE tool in Eagle to show the power pins.

hi there,
thanks for fast reply. Yes, the LEDs are deafeningly bright six SMD arrays each. They run on 12v 150mA. Oh, damn I forgot to draw in the series resistors between 595 and base of TIP120. They are 1K each

OK, then it sounds like your 12V 1A supply is the problem...it's just not powerful enough. For worst-case 12 LED's on at the same time at 150mA each you need 1.8A minimum output current.

I don't think you will overheat anything if you go up to a more powerful power supply.

Thanks for the Help RC, but here is my next issue/confusion:
Say I plug in a 12v 2 Amp supply. Because the lights are "doing there thing" w/ the 595s, sometimes only 3 (or 1 or none) will be on at a time. Am I incorrect in thinking that if 9 of the Leds are not taking the load (which is over 1 amp for sure) won't all that current be forced into the IC's and overheat them?


Current is demanded by the "load" (LED's and various circuits). If they don't need it, the power supply doesn't provide it. The 555, for example, demands 1mA (just a guess). It doesn't matter how much current is flowing elsewhere in the circuit, it will still always demand 1mA.

In the extreme case, imagine your power supply connected to nothing at all. Then clearly there is 0A of current flowing. No load-->no demand.

OK, so IC's can't be burned out by a power supply that has tons of amperage. So then why do LEDs (the normal kind) burn out if they are connected to a large current? Are resistors in series with a power supply to limit voltage, not current (obviously w/ ohms law both are affected anway tho)

I'll stop the tirade, you have answered my question and I'm looking forward to seeing this work w/ a larger power supply. I still don't get the whole "this delicate CMOS chip won't burn up at all using a large amperage power supply but an led will???"
Thanks man,
I'll report back hopefully with some measure of success.


Power supplies supply a constant voltage, not a constant current. The "2 Amp" part of the spec of a power supply is the maximum amount of current that it can supply without overheating (or deciding to switch off to prevent overheating), but the actual current that goes through a given circuit depends on what the components draw from the power supply. So a 12V, 2A power supply always supplies 12 volts, but will provide from 0 to 2 amps, depending on what's connected to it (the lower the resistance of the device, the more current it will draw).

And you can blow out both LED's and IC's by connecting them to too high a voltage (which forces too much current through them), but most logic IC's are designed to run on 5V, while LED's often require less (1 to 3V will push enough current for most LED's, while higher voltages will force too much current through).

Also, FWIW, you could simplify your circuit by using one of these instead of the 595's and transistors:


Is like a 595, but can drive high voltages / bigger currents.


Thanks for the info paul!
The Micrel chip really looks great. Too bad i just ordered a ton Tip120s and 595s from Jameco :-/
even so, I think the Lights I am using would have too much draw for that chip anyhow (150mA * 8 = 1.2 A) don't think it could handle that much current for too long.

Another device to consider is the TPIC6B595 from TI which is basically a 595+2803. But be aware that its pin outs are not the same as the 595. :exclamation

too much draw for that chip anyhow (150mA * 8 = 1.2 A)

I think that also applies to the TPIC6B595, while it might be able to handle the current the power dissipation will rule it out if it's anything like the 2803 which is limited to switching 650mA at any one time.