0
\$\begingroup\$

I've designed a circuit and I needed to use low noise regulators. The regulator that I chose for my positive rail is the LT1761-5, the 5V fixed output version of the regulator. I put 6V in and I haven't been able to get the 5V out. I generally get somewhere between 4.2 and 4.6 V. I've tried increasing my Vin in case it was a headroom issue but that hasn't changed anything. I milled a board just to test the regulator (in case my circuit was causing the issues) and I'm still having the same issues. I'm attaching the schematic and the PCB here. If anyone has any ideas of what this could be please let me know. LT1761_Test_Circuit_Schematic LT1761_Test_Circuit_PCB

Thanks!!

\$\endgroup\$
4
  • 2
    \$\begingroup\$ Do you have any load connected to the output, except the 1 kOhm resistor? Can you probe the pins and see if some oscillation is going on? \$\endgroup\$ Commented Sep 14, 2018 at 19:04
  • \$\begingroup\$ Your decoupling is a bit far away from the IC and layout isn't exactly optimal, though I'm not sure I'd expect this to be so sensitive. \$\endgroup\$
    – Shamtam
    Commented Sep 14, 2018 at 19:35
  • \$\begingroup\$ @VladimirCravero The 1 k resistor is the only load that I have on there. I just tried probing the pins and there's no oscillation. \$\endgroup\$
    – Bex.1233
    Commented Sep 14, 2018 at 19:50
  • \$\begingroup\$ Can you post a picture of your actual PCB? This is quite an odd problem, my bet is that there is something you deem unimportant that you are not mentioning. \$\endgroup\$ Commented Sep 15, 2018 at 8:57

2 Answers 2

1
\$\begingroup\$

If there are RF sources nearby, shielding will help. However, filtering is generally less expensive.

The most critical place for filtering is the output, as regulation failures are often caused by RF falsing the sense circuit. However, the input should be protected as well. Keep the components as small as possible and as close as possible to the pins they're protecting (any distance creates inductance, whether it's on the protected pin or the ground pin).

A pi filter with a ferrite and two caps is a good start. Choose the ferrite and caps based on the band(s) that EMI is expected to be encountered. Different values of caps can be used to address different bands, but if they are in parallel antiresonances can occur; having ferrites in between decouples them enough to avoid this.

Simsurfing is a good tool for capacitor selection. Other manufacturers' caps of the same value and size will behave similarly. http://ds.murata.co.jp/simsurfing/index.html?lcid=en-us

\$\endgroup\$
0
\$\begingroup\$

After checking your schematic, layout and the datasheet, the only reason I can think of for this circuit not to work is the current limiter of your power supply.

\$\endgroup\$
5
  • 2
    \$\begingroup\$ Linear regulators can be susceptible to RF, and EMI has a tendency to reduce the output voltage. Are there any RF sources near your test setup? \$\endgroup\$ Commented Sep 14, 2018 at 19:49
  • \$\begingroup\$ @stefanwyss I have my current limit set to 100 mA so I don't think that could be it. \$\endgroup\$
    – Bex.1233
    Commented Sep 14, 2018 at 19:51
  • \$\begingroup\$ @CristobolPolychronopolis yes there are definitely RF sources nearby. Is there anything that I can do to minimize their effects on the circuit? \$\endgroup\$
    – Bex.1233
    Commented Sep 14, 2018 at 19:53
  • \$\begingroup\$ To test, cover your circuit with something metallic (a mint tin lid, aluminum foil, etc) and ground it. That will block RF from getting to the circuit. Unless either a really strong or really close RF source is nearby, this probably won't do much. But it's easy to check. \$\endgroup\$
    – rdtsc
    Commented Sep 14, 2018 at 20:56
  • \$\begingroup\$ What about ESD damage of your regulator? Did you exchange it once? \$\endgroup\$ Commented Sep 15, 2018 at 7:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.