power supply / circuit gurus, please help

[mindburner]

hi I am building a power supply which used a board that I already had at it's heart. It was based on a LM338 regulator

It was getting quite hot so I decided to swap it for a LT1083 7.5 a regulator

Pin outs were the same and sample circuits similar.

Upon swapping to the LT1083CP, the PSU powers up ok and the stepped down outputs I made up work great

but I am no longer able to adjust via the pot, and get a fluctuating 14-16v at the output

This question shows up my ignorance of electronics big time but is there a reason why the pot has no effect?

can I change a component to get it to work?

Here is the original circuit for the LM338 and below a sample circuit for the LT1083CP

[AstroTux]

I'm not familiar with the LT1083CP but in passing noticed that the pot for the 338 in 10k whilst the pot in the example is 1k. It could be that the bias current required for the LT1083CP requires a 1k pot. Too high a value and the regulation will not work.

Hope this helps,

Alan

[woodbob123]

Looks as if the bottom trimmer is doing nothing. try putting centre tap of trimmer to ground as shown in top diagram.

[martin_h]

I'm not familiar with the LT1083CP but in passing noticed that the pot for the 338 in 10k whilst the pot in the example is 1k. It could be that the bias current required for the LT1083CP requires a 1k pot. Too high a value and the regulation will not work.

Hope this helps,

Alan

Also the resistor from the pot to the output is a different value.(R2 in the top diagram, R1 in the bottom)

[AstroTux]

Also the resistor from the pot to the output is a different value.(R2 in the top diagram, R1 in the bottom)

Yeah, didn't spot that, but it would need to be to give the same ratios available in the voltage divider.

[mindburner]

Looks as if the bottom trimmer is doing nothing. try putting centre tap of trimmer to ground as shown in top diagram.

hi it's already at ground, just not clearly shown in the LT1083 example

[mindburner]

Also the resistor from the pot to the output is a different value.(R2 in the top diagram, R1 in the bottom)

hi yes I suspected this, but though that the pot should have some effect on the output, even if the range as not great due to the differences in resistor values

[mindburner]

I'm not familiar with the LT1083CP but in passing noticed that the pot for the 338 in 10k whilst the pot in the example is 1k. It could be that the bias current required for the LT1083CP requires a 1k pot. Too high a value and the regulation will not work.

Hope this helps,

Alan

hi yes thats great and something I can experiment with

[woodbob123]

Do you get a change in voltage on adj pin as you alter the trimmer

[simonmc]

Its probably got something to do with the regulators bias current as already mentioned, or the load your pulling from it is not enough to make the regulator, er regulate. I did quikly look at the datasheets and the formula for working out the Vout for both circuits is the same.

Simon

[mindburner]

hi will see if these a change on the ADJ pin. I have also ordered a 1k pot to see if that helps. Will post back.

Thanks guys

[mindburner]

Do you get a change in voltage on adj pin as you alter the trimmer

hi yes from 1.5 up to 18v

so that seem to be ok, just no change at the output of the LT1083 reg

[woodbob123]

Have you wired it in correctly always the thought that it's faulty? As a thought how hot did the previous reg get could you touch it and did you have a heatsink, Also don't forget that metal body of device will be connected to one of the pins. Strange value of resistance 90.9 ohms!

Without seeing it not sure what else to suggest, sorry

[woodbob123]

After some more head srratching lm338 diagram shows pin 1 as adj data sheet for lt1083 shows pin 3 as adjust or am I just confusing things, Check pin outs for both devices to be sure you have them correct. Bob

[mindburner]

hi yes I thin kit's correct and the pinouts are the same as far as I can tell

The LM338 gets very hot. I have a decent heatsink on it but even then it would burn ones skin, so not good. Hence the desire for a larger regulator

I could go switched mode but I have already spent a fair bit of time on the case design etc - so wanted to avoid this if at all possible

[woodbob123]

I'm now out of ideas sorry

[mindburner]

I'm now out of ideas sorry

yep, Im convinced the regulator is faulty, as it well does not regulate. I have ordered another. If it works great, if not i'll get a 7A PWM power supply and just stick it in the existing box I have made

[woodbob123]

Another thought struck me replacing the regulator may not be the answer, you will still be dropping the same voltage at the same current so the heat generated will be the same, can you put a fan on heatsink or split the load between 2 regs. Your original reg would have switched it's self off or blown had it become overheated. Another way would be reduce the input Voltage if that's possible. Bob

[mindburner]

Another thought struck me replacing the regulator may not be the answer, you will still be dropping the same voltage at the same current so the heat generated will be the same, can you put a fan on heatsink or split the load between 2 regs. Your original reg would have switched it's self off or blown had it become overheated. Another way would be reduce the input Voltage if that's possible. Bob

hi yes that makes sense I guess. The bigger reg will just take longer to heat up.

The input is 19v from one of the windings of a big transformer. It's regulated down to 12v in the circuit. I don;t actually need the pot really but it's nice to be able to alter the voltage.

One of the biggest hassles was using poxy rocker switches from ebay. Three out of 4 failed or the LED would not light. A real pain

I have a fan taking hot air from the inside of the case. A fan grill and filter is ordered.

I could mount a small fan on the external heatsink as well, it just gets a bit more convoluted. The heatsink does keep the hands warm on a cold nights session:)

The regulator idea is a good one. Do I just take the output of one into the input of the other

Thanks for all your help with this

A pic of the cooled DSLR power supply

[woodbob123]

How many 12V outlets do you have best would be to use 12V fixed regs for each one. Or as you say drop 19V- 15V then 15V -12V. are there any lower voltage taps on transformer? If the load does not vary much a high wattage resistor would do it.

Bob

[mindburner]

it's one variable 10- 14v (could be fixed at 12) This is for the peltier cooler and fan.The variability allowed me to vary the amount of cooling a tad.

One at 16v. I use a little DC to DC converter to get the voltage up. This is for the resistor heater in the cooled camera sensor chamber.

One at 7.4v for the Canon DLSR power

The load seems to be about 4.3A and is pretty constant.

No lower taps ont he tranny unfortunately. It was one I had from the spares box and seemed to provide the amps needed.

I did try a 7A switch mode PSU but the little DC to DC converter for the 7.4 volt camera supply refused to work from it. May be as the input is already PWM. This is one of the reasons i'm still flogging this old nag:) Also I actually like the fun of messing about with it (most of the ttime)

If I changed the resistor in series with the pot to a higher wattage version it should handle some of the load going to the reg?

cheers

[woodbob123]

Why not split the load between 2 regs, one feeding the peltier which is taking most of the amps, 2nd for all the rest.

Bob

[Lonestar70]

Hi All,

I am new here, having just joined, but have been lurking on the forum for some time.

I have been following this thread with interest (as a retired electronics engineer) and I thought a couple of things may have been overlooked in the change over to the LF1083.

Firstly: - have you made sure that the TAB of the new regulator is isolated properly from the heatsink?... especially if the heatsink is attached to the project case, which may be either at 0v and/or earth.... this would short out the output pin and prevent the regulator from turning on.

(note... you may not get a big bang in such a case as the internal short circuit limiting would prevent the output from passing current.

Secondly: - The LF1083 requires a more strict bias regime than the LM338 in that it must have at the very minimum 5mA of bias current (10mA would be best).

It also requires a minimum load of 10mA.

The 300 Ohm resistor in your existing circuit would only provide 4.16ma of bias current.

If you change this resistor to 100 Ohms this would provide 13.75mA of bias current and also provide the minimum load current.

The control pot (10K at present) is far to large to be used with a 100 Ohm bias resistor and would make setting the output voltage very touchy (sensitive)... A value of 1k Ohms - 2K Ohms would be more appropriate and would make setting easier.

Hope this helps.

Best regards.

Sandy.

[woodbob123]

Thanks Sandy starting to get beyond my comfort zone Bob

[mindburner]

Why not split the load between 2 regs, one feeding the peltier which is taking most of the amps, 2nd for all the rest.

Bob

hi yes thats a great idea. I also found this circuit which looks interesting