LM1117 3.3V with almost 4V at output...

I continued experimenting with this and I was able to use same part in a breadboard and output voltage is 3.3V which is Ok. I also use this output to power the board with the bad regulator (without powering this one...) and everything looks good.

However, the regulator on board is still giving ~4V... I don't understand why... same part, same caps... it's a very simple circuit...

BTW, this is not just one part, in fact, I have 4 boards and all show same problem. Regulator's output goes from 3.6V to 4V... It seems when regulator is on the PCB board, the over voltage happens. This is very weird...

I appreciate any help on this.

Thanks

Gus

Hello Gus,

I am looking into the issue and will be trying to figure out what is happening. I am curious to know what the LM1117 device is connected to or what is on the load of the LM1117. Also, do you know the approximate ESR of the output capacitors? Too low of ESR at the output could effect device stability. Lastly, when you measure the 4V on the output, are you using a voltmeter reading or is it a stable DC 4V measurement as shown on an Oscilloscope? It might be possible that the output voltage measurement is not stable but can appear stable if reading off a voltmeter. I will do my best to try to figure out the problem! Thanks.

-Michael

Hello Michael:

Thank you very much for taking the time. We appreciate it.

We (my partner and I) prepared a document with the issue, which is attached.

I hope it has enough information for you to figure it out what is going on...

Please let us know if you need other tests or have some suggestions. We are running out of ideas now...

Regards;

GusLPM1117 3V3 issue.pdf

Gus,

Thank you for the report, that provided a lot of useful troubleshooting information. I believe that the issue is with the PCB since the circuit works in the breadboard even when the output cap is removed. I also noticed that the "normal" Aluminum Electrolytic output cap was not tested on the PCB which might be a valid test since the PCB layout will introduce different parasitic than what is seen in the breadboard. The closest to getting the PCB circuit working is when you did "Other Tests" in which I think you had the breadboard completely set up but simply connected the output of the LM1117 device to the PCB to see if the load and everything else on the PCB is working correctly IF the LM1117 is correctly regulating? Is this understanding correct? Was the LM1117 device on the breadboard or the PCB for this test? Could you please provide the layout of the board? If you would like to email me directly, my email is michaeldesando@ti.com I will continue to look into this issue and hopefully help you get it solved!

Another thing you might want to investigate is the GND of the PCB and verifying that the GND the input supply uses is at the same potential as the GND the LM1117 device uses. If the LM1117-3.3 is connected to the PCB, the output should never be above 3.365V unless the GND is at a slightly higher potential than 0V which will cause the output to be shifted by that much. Unless the device is damaged but it works on the breadboard so I have to believe the GND of the PCB is causing the issue.

-Michael

Hello Michael:

I just sent you an email with the layout.
I just checked again the ground in the PCB and all grounds including the modules, caps, chips and the LM1117 3.3 are all connected together, so it's good.
There is some detail I didn't mention before but the Tab pad (Vout in SOT223) is not connected to anything in the PCB. But I think this shouldn't be a problem since I'm using pin OUT (middle pin OUT) as an output and I verified they are both connected internally at the device).

Regarding the "Other Tests" scenario, since I have the breadboard circuit working fine and the PCB circuit working wrong, I wanted to check the behavior of the PCB "injecting" the power from a good supply (regulator in the breadboard). So my thinking was if the PCB was affecting the regulator on board, I wanted to check if powering the PCB would also affect the regulator in the breadboard. In other words, in this scenario the regulator in the PCB was not powered. The power came from the breadboard. And the results showed the regulator in the breadboard continued working fine, even when powering the PCB with the problem...

Please let me know if there is something else I could try.
Thanks
Gus

Gus,

I have reviewed your layout and everything seems fine. It would be nice if the output cap was a little closer to the device but I don't think that is the issue. I believe it's possible for there to be interference coming from the surrounding components. Could you do one test for me? I would like to you completely populate the PCB. I would like you to then remove all power from the PCB and apply the supply power and GND directly to the LM1117 device on the PCB. By putting the input supply directly to the LM1117 device while on PCB with all other components off, this will hopefully isolate the issue. If the LM1117 device regulates in this condition, then I believe the issue is interference from the surrounding devices. This is a solvable issue but may require some isolation techniques. If the LM1117 still does not regulate properly with that device being the only device powered, then I will need to continue to investigate. Please let me know! You might need to remove the U7 device and maybe also U4 to power only the LM1117 device since it appears those devices share the same input supply. This might also explain why it works on the breadboard when devices are spaced out further and not on the PCB when the close proximity creates stronger interference.

-Michael

Hello Michael:

Yes, I agree the caps should be closer to the regulator in the layout. However, I soldered additional caps on the regulator pins, but didn't help.

BTW, U6 (GSM Modem) and U7 (4.1V regulator) are not populated.

I'm not sure if I'm following the test you suggest. You mean all board components populated, including the LM1117, and apply external power (3.3V) to the on board regulator (pins Vout and Gnd)? Because this is exactly what I did in the "Other tests" scenario described in the document.

In fact, something I did not mention before, I run a test where I populated the regulator, the caps, LEDs with resistors, the buffer U3 and relay U4, and everything worked fine (3.3V). Then, I soldered the radio module U1, and then the regulator started supplying ~4V. So the radio module was interfering the regulator. HOWEVER, I proceeded to apply external 3.3V to the PCB (that is, not powering the on board regulator, and connecting the Vout and gnd from external regulator in breadboard to the Vout and gnd on regulator in PCB) and the supply still measure 3.3V. So if the radio module affects the on PCB regulator, why it doesn't affect the regulator in breadboard???

Please let me know.

Thanks

Gus

I did one more test;
since the regulator on the breadboard was supplying power for both, the gateway circuit assembled in the breadboard and the one in the PCB with the issue, I decided to disconnect the power from the circuit in the breadboard, so I wanted to be sure this one was not "helping" or "positively interfering" with the PCB....
So basically I run the circuit in the PCB powered from the regulator in the breadboard (this one supplying the PCB only) and with the regulator on PCB soldered but not powered (Vin not powered), and everything worked fine.
So the failure happens when PCB circuit is powered from the regulator on board, but not from regulator on breadboard. This means that if the radio module (or any other component) is causing the failure, it only affects it when regulator is on PCB.

Thanks
Gus

Gus,

I thought the "Other Test" you did has the LM1117 on the breadboard. In my test, everything is on the PCB and there is no breadboard being used at all. In my test, the power supply is being directly applied to the LM1117 on the PCB and preventing any power to the other surrounding modules. It sounds like you have done this and it worked (3.3V regulator output). It seems like the issue is the interference caused from the Radio Module U1. The reason it works on the breadboard and not on the PCB may be because of the distance between the devices is closer on the PCB which creates stronger interference and the interference can travel through the traces on the PCB.

You might want to try some shielding techniques or see if you can reduce the EMI/RF interference coming from the radio module. Your options are limited after the layout has already been created. Here is some info from Digikey. It might be worth trying a shield if interference is in fact the issue:

www.digikey.com/.../rf-shielding-the-art-and-science-of-eliminating-interference