Design new 0 to 30V power supply.

[Hero999]

Sorry it was a long day, It should have said is easy enough done. so it's a yes.

I'm still not sure if you know what I mean, sorry if you do I'm just boring you.

I'm interested to see if the tap switching works.

For output voltages under about 12V, M1 should be off, so the power being delivered to +V is coming from the centre tap (via D5): +V should measure about 21V, give or take a few volts.

When the output voltage is set above 12V, M1 should turn on, so the power to the circuit is supplied from the higher voltage end of the transformer: +V should measure about 42V give or take about 5V.

Did you test for any of the above and if so did it work as I expected?

Yes I substituted a few resitors, Maplins did not have thevalues I wanted plus rather exspensive 16P each, Rip off or what. I searched through my bins for the correct value and out of the thousands I got I gave up. but has this is just a test PCB I thought it would not matter has long they are the right values.

Yes, you can do that, I was just interested in which resistors you substituted.

Rather than buying single resistors you can buy a whole set of E12 values. Unfortunately the set I linked below is 5% and no good for the critical resistors in the circuit which need to be 1% otherwise the voltage and current settings will be way off.
http://maplin.co.uk/Module.aspx?ModuleNo=2163

For the moment becasue I've got a .1R resistor

Something doesn't seem right.

Before you said the current limit was 1.6A?

I think the current limit should be lower with R5 at 0.1Ohms.

The calculation for R5 = 15m Ohms:

Calculate the voltage reference
Vref = 2.495

[PicMaster]

I'm still not sure if you know what I mean, sorry if you do I'm just boring you.

Sorry what I meant is I shall carry out the test on M1 with a 1A load on it. Make sure M1 is off below 12V. Youre not boring me I'm finding this very interesting and learning lots more about power supplies. What I meant by it's been a long day at work (14hr shift) I love it when you get stuck into something that is really interesting. I just look forward to the finsh product and will be proud to show it off

Something doesn't seem right.

Before you said the current limit was 1.6A?

I think the current limit should be lower with R5 at 0.1Ohms.

R5 at the moment is a 0.1R 10W instead of the 15MR like in the shcematic, I'm sure that I was getting 1.6A. But may be I was carrying out the test wrong. How I did it was for the load testing I just used my multimeter in the current mode in line of the +&- which is using the shunt in the meter which is not very good. I shall pick a low ohm resistor and work it oiut like you have.

What I shall do is get armed with pen and paper start from the beging again measure and check all the Vref and the gain of of the  differential amplifier, M1, I will make sure i look at every angle of it, Then recored all data jot it onto the schematic then upload for you to have a ponder over. Then we can take it from there.

I've just remembered when I had a quick look at it last night and just looking at the schematic I did put in a 10K and not  a 100k, So I will get that changed silly me  :-[ ;D

[Hero999]

I've just remembered when I had a quick look at it last night and just looking at the schematic I did put in a 10K and not  a 100k, So I will get that changed silly me  :-[ ;D

That's an easy mistake to make.

Which resistor was it?

I'm still questioning this because there are no 100k resistors apart from the pot's and replacing a 100k pot with 10k should have this effect.

The only explanation is that you were intending to subsitute a value you couldn't get hold of. Were you intending to replace one of the 160k resistors with 100k, 33k and 27k in series but used 10k instead of 100k?

[PicMaster]

Which resistor was it?

I'm still questioning this because there are no 100k resistors apart from the pot's and replacing a 100k pot with 10k should have this effect.

It was R20 100K. I shall change it and start to carry out the tests above

[Hero999]

As I said before, that shouldn't make any difference to the current at which it starts to limit. R21 just changes the delay.

There must be something else which is wrong.

[PicMaster]

There must be ishall look into it and go over the rest of the pcb,I have just tested the v-ref 4.31v no load connected and both pots at 0 I turned the voltage to 10.v v-f ref went down to 2.8v I then placed a 1A load and with the current turned to max the v-ref semms to be jumping up and down,I then placed my a .1R resistor turned current pot to max and I only get 1.2A which mens v-ref is only 1.2v the v-ref components seem to be getting warm, I shall look into just in case it's me but do you think that the v-ref circuit is up to thr job ?

[Hero999]

That can't be right, Vref shouldn't change at all.

U5 shouldn't be getting warm at all, R18 might get very slightly warm because it's dissipating just over 0.1W but it shouldn't be very hot, I doubt you'd notice it.

I'm sure the circuit is right, it's the same as the datasheet.
http://www.fairchildsemi.com/ds/TL/TL431A.pdf

[PicMaster]

Right I have done a bit more testing. I will try and explain the best I can. I found out where I went wrong, It was on the connections from the transformer(my fault).
V-Ref I can only get 2.6V out of it. and the max output voltage was 23V. Also if I connected a load it went lower 2.1V.
So here is what I did, Removed R18 and R14 connect my other PSU to V-Ref(on the top end of the pot) which was set to 4.1V, This then enables the PSU to goto 0 - 35V, To limit the voltage I turned V-Ref PSU to 3.63V which give a max Voltage ouput of 32V. but what I did notice is that when the PSU gets to 7V my meter displays OL (and the high voltage symbol) only for a split second well enough for my eyes to see it, Then volatge contiues to rise upto 32V
So something is telling me that the swap over is happening at 7V not 12V,

My second test was I measured the current draw on V-ref line and it read 180MA. the current pot has to be turn up slighty to get the full 32V, the center of the pot read 333mv to obtain the full voltage with out no load. I can veary the current from 0-2.6amps with a 0.1R(R5) shunt and a 0.1R load but the PSU was set for 5V I did not want to risk higer voltage yet.
Is the 2.6amps due to the act that the shunt resistor R5 is 0.1R ?

This was the way I set my  velleman power supply up at 5V using a 0.1R has the load and can vart the full 5amps (8A max), So I think that the mosfet PSU should work in the same manner. The velleman PSU was a good desing it makes you wonder why they stoped selling it. 

Hope you can understand what I'm trying to say, ask any questions and I will try and explain in more detail a bit by bit.

I might be early in saying this or even thinking about it, I'm beging to have doubts  about the stability and the smoothness of the operation of the voltage control/current, (really I should not think this yet but can't help it)
BTW, I have double checked my PCB for any errors and shorts everything checks out oK  

[Hero999]

So here is what I did, Removed R18 and R14 connect my other PSU to V-Ref(on the top end of the pot) which was set to 4.1V

Did you remove U5?

So you remove R8 and R14 without removing U5?

If the current from the supply isn't limited it will destroy U5.

So something is telling me that the swap over is happening at 7V not 12V,

The tap change occurs when the voltage across R17  goes above the voltage across R13.

If Vref is changed, then the switch over point will change because the voltage across R13 is dependant on Vref.

I can veary the current from 0-2.6amps with a 0.1R(R5) shunt and a 0.1R load but the PSU was set for 5V I did not want to risk higer voltage yet.

So you still get 0 to 2.6V even though Vref has been lowered?

That doesn't sound right because if Vref is reduced the current limit should also be reduced.

[PicMaster]

Did you remove U5?

So you remove R8 and R14 without removing U5?

If the current from the supply isn't limited it will destroy U5.

I removed the 2 resistors and I did forget to say U5 as well

The tap change occurs when the voltage across R17  goes above the voltage across R13

It was the same if it running from U5 or the exteranl PSU set for 4.1V with common GND

So you still get 0 to 2.6V even though Vref has been lowered?

No when running off U5 The goes up and down due to the V-Ref voltage not been stable from the external PSU set for 4.1V I only get 2.6amps max

I will still carry on in looking and more testing I'm not giving up yet, I class this has a minor set back, But the PSU seems to work well in the sim program  ???

[Alex Tsekenis]

My second test was I measured the current draw on V-ref line and it read 180MA.

Is that a lot or is it just me?

Have you tried replacing U5?

Sorry, for some reason the forum removed my post notification from this topic.

[PicMaster]

Sorry, for some reason the forum removed my post notification from this topic.

I wondered where you went to. With it been Mothers day I think that it would be better left for today  ;D

[PicMaster]

I've add another hour on it and Finlay getting some where, I found out where I went wrong.
I checked my schematic against the original very thing check out OK. I then studied the values and foot prints, I got incorrect values on R12 68K instead of 33K and R14 15K instead of 56K so the swap over voltage is working correct now  ;D
Then onto the foot prints I selected the wrong foot print for U5 which explains the V-Ref voltage to be all over the place. So I now get 4.0V at V-Ref all the time  ;D, The output voltage is smoother and more accurate readings.

I started to look at the current side now, It's still not quite there I used a 0.1R for the load and can vary the current I get to 1.25A then it jumps to 3.25amps turn it up a bit more then back down 1.25A I can get to 3amps then when R21 is at it max the current is back down to 1.25A

Now I think is caused by me having a incorrect value pot 10L instead of 100K and R5 .1R instead of 15mr ?

Silly me I've read about the calculations you posted. I shall look into that next time, Then once I'm happy that the current side is working correct I shall see about carrying out the other stuff you talked about.

[Hero999]

Good, that seems to be much better.

The current limiting still isn't behaving as it should.

As I said before, changing the value of the pot won't alter the actual current limit, it'll just alter the length of time for the currrent limit to kick in.

Changing R5 to 0.1 will lower the current limit but 1.25A is still far too high, I worked out how to calculate the current limit a few posts ago.

Note how none of the pot. values are used in the calculation?
http://www.electronics-lab.com/forum/index.php?topic=19605.msg91375#msg91375

[PicMaster]

I realised that after I typed it. The reason why the current was reading higher and funny readings was down to my meter(not my fluke one). I brought home my spare meter from work(another fluke meter) I used that to measure the current and the other fluke for the voltage readings.
Your right I get right I was only getting .699ma. So I now know not to trust cheap meters well it was not that cheap 40quid from maplin (that may be why ;D)

I shall work out the calculations now so I get the right resistors to obtain the right current until next time I place an order so I can get the 15mr resistor. So now I know that I'm getting some where

Because I have to wait for the wife to get tired and go to bed it's a bit on the late side and you know the brain does not function as well  ;D,

EDIT: Right I have worked out R7 = 22k R6 160K a gain of gain of 7.27 = 5.63A. I dropped in some 22K for R7 & R8 but this seems to throw the voltage readings out and the current pot to max to get 20V with the voltage pot on max.

Do I have to just change R7 to 22k and leave R8 3K ?

You bare with me if I seem a little thick or slow analogue is not my strongest point

[Hero999]

Good, I'm glad it's performing as expected.

You probably can use a 0.1R resistor, in fact may be the circuit should be modified so 0.1R can be used?

The trouble is the resistor will need to be rated to 12.5W, even higher if the current is slightly over 5A, I'd probably suggest 15W.

For 5A I think you need a gain of 8.2 which would be 82k and 10k

I'll check your calculations tomorrow.

[PicMaster]

What is the lowest current you cna set it to ?

My V-Ref is 4.05V and the maxium voltage is 35V.
The lowest current I can go to is 123MA, The center tap of the curent pot reads 103MV This enables the ouput to go to 35V, If I set it the center of the pot reads 33MV the max out I can only obtain 13.60V

[Hero999]

I'm not meaning to be picky but don't put MA when you mean mA or MV when you mean mV. The upper case M means mega which is a million, m means a thousandth. I obviously know what you mean but not everyone will.

In theory you should be able to set the current all the way down to zero but I accept that the voltage ofset of the op-amp and the minimum voltage from the pot will be limiting factors.

The voltage when the pot is set to the centre position should be about 2V.

The maximum voltage should not go down unless you set the output current to less than 12mA when the output current through the 3k resistor will be enough to cause the current limiter to kick-in.

I've made a minor amendment to be BoM, don't worry this doesn't affect the circuit or what you've done. I've just corrected a typo and I've suggested, that if you can't get hold of 3k and 160k, you can replace 3k with 3k3 as long as you replace 160k resistor with 120k and 56k in series which eliminates a resistor. I've also put the ratios in so it should be obvious that the resistors can be replaced with other values, as long as the correct ratio is preserved.
http://www.electronics-lab.com/forum/index.php?topic=19605.msg90275#msg90275

[PicMaster]

I did have a 3k & 160k but with a .1R restor the current has you was low now I have stll got the 160k and a 22k but judging what you said a 82k &10k should be better with the .1R
so if The voltage was set to 5v darwing say 10mA I still should be able to alter the voltages before the voltage drop sorry it's a quick reply but got to go out

[PicMaster]

I changed the resitor values to 82K&10K this give me a max current of 4.36A

Not sure if it is me me, Put when I alter R20 the current does not seem stable and I get higher amps out put. I have got R20 set about half way current pot set near minimum voltge set to 15V then when I switxh on a .1R load the output current goes upto 16amps I turned if off quickly, I set R20 back to min and tired it again and I thought it was ok, But when I turned off the power supply then turned it back on and the same thing 16amps. It only does this once after that it does not seem to do it again until I turn off/on the Power.

I opened up the sim version on the PC and I noticed that is does the same in there but because it's a sim prog it's goes to 66amps but it does drop off down to it's normal current draw, I do not have the guts to let it run long enough in real life

Any ideas Why this should be happening

[Hero999]

R20 is not supposed to alter the current limit, that's R21's job. R20 controls the delay before the current limit kicks in so when it's set to the highest resistance setting there will be more of a surge before the current is limited.

There will be a surge anyway due to C7 rapidly discharging but the delay will make the surge last for longer.

I didn't experiment with changing the gain and sense resistor much with the simulator. It could also be possible that it's not so stable with the gain set lower but I doubt it, I imagine increasing the gain will make it worse and as you've reduced it, there's no need to worry.

[PicMaster]

R20 is not supposed to alter the current limit, that's R21's job. R20 controls the delay before the current limit kicks in so when it's set to the highest resistance setting there will be more of a surge before the current is limited.

This was before I changed the resistors to 82K&10K with a 0.1R shunt, The max current is 4.36Amps and altering tR20 does not effect the current readings now but I thought the output voltage was limited to 32V but I get 35V. Is this correct ?
I will investigate some more but can't understand why I get 16amps(to scared to leave it on for more than 2 seconds to see if it drops). Could this be-caused by me changing the shunt resistor I know you said it should have no effect, But really i can't go any further with all the other tests until I find out why or what is causing high amps

[PicMaster]

Hero
Can you explain a little how the current limit side of thigs work and about the role of R20 just has so I cna get a better understanding of how the circuit works

cheers

[Hero999]

R20 and R23 are in series so  for the sake of this discussion I mean R23 as well as R20 when I mention R20.

R20, C5 and U1C form an integrator or low pass filter. The cut-off frequency is equal to 1/(2πRC), in this case R20 and C5 are R and C respectively.

The interesting thing about an integrator is when the input is suddenly changed from one voltage to another, the output takes awhile to reach the same voltage as the input. An integrator is just an RC circuit, the time for the output voltage to reach within 38.7% of the input voltage is equal to one RC time constant, see the graph on Wikipedia.
http://en.wikipedia.org/wiki/File:Series_RC_resistor_voltage.svg

Note that the output voltage never, ever reaches the same voltage as the input but after five time constants it's within 0.7% so for practical purposes, it's equal to the input voltage.

The value of R20 makes no difference to the current or voltage setting of the power supply. It just alters how long it takes for the current limiting to kick-in. The integrator is connected to the current amplifier so there will be a delay from when the current limit is exceeded to when the output transistors are turned off.

[PicMaster]

I have done a bit more playing and something is not right, When I adjust R20 the amps do adjust I know you said it should not. The other thing I noticed instead of using a resistor has a load I used a 12V fan I set the voltage to 12.56V the fans runs OK then if I turn down the current pot to the Min the fans starts to slow down then stop on turning the current back up the fan spins mega fast, So I tried it again this time with a multimeter set up and I measured the voltage and it went up to 22volts ??? and that even without the voltage pot adjusted only the current
Now I'm not sure but it could be my fault I used IFRP3710 Mosfets has these are the only one I add at the time. could this be causing the problem ? to much gate capacitance ?

I shall have a look through my bigger junk box for different types  ;D