LT1528 voltage drops too much

[integral]

I'm trying to get around 4V for a project from a 5V input, but while following the designs in http://www.telit.com/fileadmin/user...ads/3G/Telit_UL865_Hardware_User_Guide_r7.pdf as well as http://cds.linear.com/docs/en/datasheet/1528fb.pdf and none of the resistor combinations or even strict adherence to the low ESR caps, and still when ever there is a load put on it, the voltage drops at least half a volt.

Without a load, I get from the linear site's resistor configuration, 3.95V. I put a 20Ω resistor on there which should be just under a fifth of an amp, but then the voltage measured across the resistor lowers considerably more than what the datasheet indicates it should, and the actual load I measure matches the new voltage divided by the resistance.

I checked the power supply, which is just an ATX attached to a breakout, and there are no anomalies there.

I've tried varying the capacitance on both input and output, but the behavior is the same either way.

I've also measured the input pin to make sure it's getting sufficient voltage, and the sense pin is 3.27V which is probably around where it needs to be.

I've also tried swapping out the regulator with another one, but that doesn't really help.

This is frustrating.

Please anyone tell me if there are some pit falls with this particular linear regulator, or can anyone suggest a better circuit, preferably one where the regulator comes in through hole rather than surface mount?

 

[Alec_t]

Welcome to EP.
How much current do you need to draw from the regulator? Was the 200mA just for test purposes?

Edit: Perhaps a high-resistance contact in the breadboard is dropping the voltage?

 

[integral]

Welcome to EP.
How much current do you need to draw from the regulator? Was the 200mA just for test purposes?

Up to 0.735A. The 200mA was for test purposes since the modem I'm connecting to it uses around that or less during normal operation. The LT1528 is supposed to be able to handle up to 3A.

 

[davenn]

Up to 0.735A. The 200mA was for test purposes since the modem I'm connecting to it uses around that or less during normal operation. The LT1528 is supposed to be able to handle up to 3A.

99% probability is that the difference between input and output voltages isn't large enough
show us a schematic of the circuit you used as your construction is pretty bad and difficult to follow

Take note of @Alec_t 's comments

 

[integral]

@Alec_t do you then think I should solder it?

I uploaded a schematic of what I'm doing.

 

[davenn]

@Alec_t do you then think I should solder it?

that would be an extremely good idea even if a rough layout just for testing

I uploaded a schematic of what I'm doing.

schematic looks good .... NOTE the datasheet shows 4 x 47uF caps rather than a single cap
try 2 or 3 x 47uF in parallel

I'll take back my comment on the voltage difference, going by datasheet, that shouldn't be a problem


cheers
Dave

 

[integral]

I have tried 2 in parallel. It didn't make a difference.

 

[Alec_t]

It would be unwise to try and push 0.7A through breadboard connections. You really do need a pcb/stripboard with soldered connections.

 

[integral]

Well, it's not quite perfect, but soldering it did lower the drop, so now it's going from 3.93 with just the 10μf cap on the input, and the 100μf on the output, down to 3.76V when I put the 20Ω resistor on there.

That still seems kind of high, it should be closer to 3.8.

 

[integral]

Well for some reason, it seems to be giving me the full 4 or just within a couple of hundredths of it. Thank you guys so much for sharing your 1337 knowledge.

Also, as a side note, even though the datasheet calls for low ESR tantalum caps. it seems to be doing well non-tantalum electrolytic can caps.

 

[(*steve*)]

The dataset recommends components for a reason. It is possible that the regulator will oscillate with the incorrect type or value components.

This MAY be a reason why the output suddenly changed (in this case because it may have stopped oscillating).

 

[davenn]

The dataset recommends components for a reason. It is possible that the regulator will oscillate with the incorrect your or value components.

This MAY be a reason why the output suddenly changed (in this case because it may have stopped oscillating).

exactly
a number of posts ago I did say to use multiple caps on the output as per the datasheet, instead of a single one

@integral

 

[integral]

I think it was a loose connection, but I will try another cap if need-be. So far it's working for everything I need. Thanks for all your help.

 

[OBW0549]

The dataset recommends components for a reason. It is possible that the regulator will oscillate with the incorrect type or value components.

This MAY be a reason why the output suddenly changed (in this case because it may have stopped oscillating).

Yup. Regulators (and opamps too, for that matter) can also oscillate due to poor construction techniques such as the breadboard shown in the picture, with that tantalum capacitor connected to the rest of the circuit via a long wire. Wires have inductance, and a long wire has a LOT of inductance; together with the capacitor, that forms a resonant circuit and could easily cause oscillation.

Such oscillations are usually at RF frequencies and unless you have a good, wideband oscilloscope handy you'll never see them. But they do tend to cause serious DC voltage shifts through rectification, which is probably what's happening here.

Linear Technology has an excellent application note, AN-47 High Speed Amplifier Techniques which discusses (among various other issues) all the horrible, cruel things that can result from poor circuit layout; note especially Appendix F on breadboarding techniques, beginning on page 98. AN-47 applies just as much to voltage regulator design as it does to amplifiers; both can have aspects of high-speed design.

In any case, with a properly constructed LT1528 circuit you shouldn't see more than a few dozen millivolts change in output voltage going from no load to full load.

 

[cjdelphi]

It would be unwise to try and push 0.7A through breadboard connections. You really do nead a pcb/stripboard with soldered connections.

You beat me to it, breadboard losses are insane if you're playing with just volt difference

Eg, i created a switch mode (step up) power supply on a breadboard, calculated it to be around 10uf for the inductor, but because of the breadboard it failed

You might be ok with an ldo if you solder it on perfboard