EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: d4n13l on July 27, 2019, 04:18:12 am
-
Hi,
I'm working on my first power supply 0-30V 0-4A, and I'd appreciate some feedback, specially in two points:
1) Right now the slightest turn of the potentiometer VR1 allows for a lot of current, I'd like to make it more sensible, but I'm not having much luck
2) I want to add a voltage/current meter, I tried a L7805 connected to the second transformer to feed 5V to the meter but the moment I connect the terminals to the output I get an extra volts in the output instead of zero, not sure why.
Also I'd appreciate any comments on the design. Thanks
-
IC1B is operating with far too much gain. Put a resistor say 1K from pin 6 to ground and 100nF from pin 7 to pin 6. These values are just starting points and may be way away from optimum. This will roll of the high frequency response of the op amp and hopefully make it more stable. It is most probably oscillating like crazy.
-
Hi,
I tried a bunch of resistors values and didn't see any aparent effects on the behaviour. Same for the capacitor.
-
Hi,
I'm working on my first power supply 0-30V 0-4A, and I'd appreciate some feedback, specially in two points:
1) Right now the slightest turn of the potentiometer VR1 allows for a lot of current, I'd like to make it more sensible, but I'm not having much luck
2) I want to add a voltage/current meter, I tried a L7805 connected to the second transformer to feed 5V to the meter but the moment I connect the terminals to the output I get an extra volts in the output instead of zero, not sure why.
Also I'd appreciate any comments on the design. Thanks
How does the voltage regulation work?
-
It works as supposed to, 0-25V with current resistor values. I think current regulation is working also, just as I mentioned very light changes in the potentioner go from 0 to too much since ground is receiving about 0.9V/0.22ohms allows about 4A. Which is a wide range to cover.
-
With the PSU unloaded, is the wiper of VR1 able to be smoothly adjusted from 0V upward?
Although likely not related to the problem, I don't understand the function of IC1A.
Put a small load on the output and confirm that there is a corresponding voltage drop across R2. Check that the voltage drop is being applied properly to IC1B.
-
IC1A might be fighting IC1B, not allowing the ADJ pin to be pulled down by enough.
-
Hi xavier60,
So U2 is adjusted to have -3V in the lower rail, R5 and R6 make a voltage divider with -1.5V give or take that is seen by the non inverting input of IC1A, so the op amp puts the same voltage in the inverting input lowering the output by that much, and since the min voltage of the LM338 is 1.25V it comes down to zero. R11 is used to adjust it properly. If you remove IC1B you have a 0-30V 4A power supply.
For your comment on R2, there won't be voltage drop due to IC1B that keeps the voltage at the value specified by VR1 and so we have constant current.. Not sure about the op amps fighting thought, but I don't think that would be the case, maybe you can elaborate on that.
-
Hi xavier60,
So U2 is adjusted to have -3V in the lower rail, R5 and R6 make a voltage divider with -1.5V give or take that is seen by the non inverting input of IC1A, so the op amp puts the same voltage in the inverting input lowering the output by that much, and since the min voltage of the LM338 is 1.25V it comes down to zero. R11 is used to adjust it properly. If you remove IC1B you have a 0-30V 4A power supply.
For your comment on R2, there won't be voltage drop due to IC1B that keeps the voltage at the value specified by VR1 and so we have constant current.. Not sure about the op amps fighting thought, but I don't think that would be the case, maybe you can elaborate on that.
The R5/R6 divider would make more sense to me if it went to GND rather than the ADJ pin.
Because of the way the circuit looks to me, it seems that IC1A also has control of the ADJ pin?
There must be a voltage drop across R2 corresponding to load current. Like 220mV for a 1A load. This should be checked first if not done so already.
-
I'm sort of new to electronics so correct me if I'm wrong but you have the high potential (the output of the LM338) and the low potential (-3V rail), why do you need ground? when you connect two batteries in series you use the extremes of the batteries you don't care about the middle.
The current flowing throught the divider is very small and compensated by R11, as to not affect the adj pin. IC1A input connects to the divider but as I undertand it no current is flowing through it.
The point of IC1B is to not allow the voltage to change at R2 no matter the load. IC1B see in the non inverting input 0.9V tops from the divider. The same voltage is put in the inverting input and so you would allow 0.9V/0.22=~4A max, if you set VR1 to supply 0.1V the max current would be 0.45A.. which is my problem, even thought 0.1 is the low end of the voltage it still supplies high current for small electronic projects, which make it hard to regulate in this circuit.
-
I'm sort of new to electronics so correct me if I'm wrong but you have the high potential (the output of the LM338) and the low potential (-3V rail), why do you need ground? when you connect two batteries in series you use the extremes of the batteries you don't care about the middle.
The current flowing throught the divider is very small and compensated by R11, as to not affect the adj pin. IC1A input connects to the divider but as I undertand it no current is flowing through it.
The point of IC1B is to not allow the voltage to change at R2 no matter the load. IC1B see in the non inverting input 0.9V tops from the divider. The same voltage is put in the inverting input and so you would allow 0.9V/0.22=~4A max, if you set VR1 to supply 0.1V the max current would be 0.45A.. which is my problem, even thought 0.1 is the low end of the voltage it still supplies high current for small electronic projects, which make it hard to regulate in this circuit.
I don't think I'll ever make sense of the voltage regulation side. I'll just accept that it works for now.
It sounds like your current regulation is working but not all the way down to 0V output such as when the output is short circuited.
You need to do some testing. Set the current to some low setting like 1A and apply a load with gradually decreasing resistance.
I use a 6Ω 300W rheostat.
After the 1A limit is reached, IC1B should be capable of pulling the ADJ pin down to -1.2V as the load approaches 0Ω.
-
Sorry man, but I think you don't get it. The supply works great as is. It goes from 0 to 30V and 0 to 4A. The problem is that I have a coarse current regulation when I want a fine current regulation. If I adjust R10 to output 0.1V VR1 works great now because the output range it was to work with goes from 0 to ~450mA whereas if R10 outputs 1V it has a range of 0 to 4A. What I'm looking for I guess is a way to allow this fine/coarse tunning for VR1 through the whole 0-4A range.
Also, I have problems adapting the meter display, if I power it from yet another supply, it works great, but if I feed power it it from B3 with a L7805 or something like that it rises the output voltage a few volts. I would like a suggestion on how to power it without having to add yet another supply.
-
You could add a Fine Pot. One way of doing it is to wire the outer pins in parallel and then sum the wipers with 2 resistors say a 10K for the course and a 100K for the fine.
I used 10 turn Pots. They were cheap enough on ebay but I had trouble. I'm not certain if was due to poor quality or bad luck.
I explained what happened here.
https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2368746/#msg2368746 (https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2368746/#msg2368746)
https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2388873/#msg2388873 (https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2388873/#msg2388873)
-
Also, I have problems adapting the meter display, if I power it from yet another supply, it works great, but if I feed power it it from B3 with a L7805 or something like that it rises the output voltage a few volts. I would like a suggestion on how to power it without having to add yet another supply.
How were you powering the meter from B3 as well as U2? The meter would need a + supply?
What voltage is there across C2 at the moment?
-
I just connected the a L7805 in parallele with C2.. I didn't realize before that this causes the voltage to come down.. guess I'll try to connected it somehow to B2
-
I just connected the a L7805 in parallele with C2.. I didn't realize before that this causes the voltage to come down.. guess I'll try to connected it somehow to B2
You still have options with B3 by using a full bridge split rail rectifier or a 2 diode doubler to give + and - rails depending on how much voltage is needed.
Another option with the CC adjustment is to use a Logarithmic potentiometer which will allow fine adjustment for the low end of the range and coarse for the high end. Most audio Pots are this type.
To find out, turn the Pot to the middle of its range and measure the resistance on each side of the wiper. There will be a large difference with a log Pot.
-
One more option for the current control is to add a range switch.
-
For those not familiar with this design, a simplified example is shown below.
The idea is that the operational amplifier, configured as an instrumentation amplifier (notice the 4 equal valued resistors), copies the voltage across R7 as shown below which is always 1.25 volts as maintained by the regulator, inverts it, and then applies it to the bottom of the resistive divider which sets the output voltage. So with -1.25 volts applied to the bottom of the divider, the regulator's output can reach zero volts. R8 helps the operational amplifier sink the current through the resistive divider.
d4n13l, your idea is sound but the implementation of that current control loop has at least one problem.
1. The voltage reference for control of the current is unregulated so the output current is going to follow the ripple voltage at the input of the regulator. Maybe that is acceptable though.
2. 50 kilohms worst case impedance at the non-inverting input of IC1B may be a problem. Pin 5 should be bypassed to the bottom of the potentiometer (the node where VR1 and R2 connect) and *not* ground as drawn in the schematic.
3. I think the current regulation loop will be stable but if it is not, Circlotron is wrong about the gain but right about how to add external frequency compensation of IC1B if necessary. C5 severely limits the bandwidth of IC1B which should make it stable albeit slow.
4. IC1B should be able to to sink the roughly 5.7 milliamps coming from R4.
5. The bottom of R6 should be connected to ground as shown below and not the negative supply.
-
Hi David,
Thanks for the reply. About your simplified schematic, you're right R7 voltage is always 1.25V however after some thought I realized that the voltage copied by the op amp is not this 1.25V but the output voltage minus 1.25V, thus allowing the voltage to go to zero. Great! now I can delete the voltage regulator from the second transformer in my schematic.
As for your other comments, from
1. You're right it's unregulated, I don't know a good and efficient way to get 1V regulated from 34V input, I'd love suggestions.
2. If I understood correctly, and you meant to simply delete the connection from the bottom of the pot VR1 to ground, that wouldn't work because the pin 5 would see the voltage from the voltage divider R1-R10 (0-1V) , and I wouldn't be able to regulate the current with VR1.
The current regulation part works and is stable but I get coarse instead fine tunning (I'm not happy about it).
Also the voltage/current meter that I have works with 3-30 input voltage, but I get 34V from B1, I fixed it with a transistor as in shown in the schematic, but I'm not sure how good of a solution it is.
-
It's likely that the meter should be fully referenced to GND. The brand and model of the meter will be helpful.
-
I guess is a short of cheap one, the brand is radox, 535-177.
I just discovered that the IC1B is blown :(, probably why I'm having problem with the current regulation, but I'm not sure how that happened, if it was with the current circuit or previously when I was making tests. According to the datasheet the supply voltage for the LM358 is 32V max. which may be other cause :(
-
I wonder if the LM358 will work properly if powered from the PSU's output which can be allowed a defined maximum of 27V.
The negative pin would have to be supplied with -5V to satisfy the LM358's minimum supply voltage of 5V.
There might be other tricks to increase the voltage range.
-
Working with 35V from B1 is turning to be a pain, I'll switch to a 12V transformer and see if everything works as expected. Also I could directly connect the display, it's gonna be much simpler.
-
1. You're right it's unregulated, I don't know a good and efficient way to get 1V regulated from 34V input, I'd love suggestions.
A zener diode shunt regulator would be a big but simple improvement. Or 1.25 volt bandgap shunt references are inexpensive. If you do not have anything better, than even an LM317 or LM317LZ with the adjustment pin tied to ground and a 120 ohm or whatever resistor shunting the output makes an instant 1.25 volt reference.
-
The divided output from a 5V regulator could be used for the CC reference. The 5V could also power the meter.
-
Hi,
I updated the schematic for a 0-15V 0-0.9A PSU. I added a second potentiometer so that I can tune the low currents, and I connected in the meter in parallel with C1. The only thing I notice is that there is about 10 mV with the lower for voltage instead of a perfect zero. I don't know if it's a limitation of the op amp but close enough.
It's easy enough to scale to 30V and 4A output as I originally planned but the voltage is outside specifications for the LM358 and I don't have a appropiate replacement.
Maybe it'll be helpful to someone.
-
Some bench power supplies add a deliberate offset to the error amplifier so the output goes negative by a couple of millivolts and zero can be reached under all conditions.
-
Just a thought but I was wondering why theirs no feed back loop from pin 7 to pin 6 .In the first schematic your putting a -33V reference directly to the inverting pin.And on the second schematic theirs a -15 volt going to the inverting pin .All the current controls I've seen and built have a feed back from the out put to the inverting input.Wouldn't that stabilize your control more? I'm learning so if someone could explain it would be appreciated.
Also some digital volt and amp meters must have a separate power supply to operate and some don't .Some 4 wire meters don't use a common ground.Found that out the hard way.But it's difficult to find proper wiring diagrams for some particular meters.
-
As I understand it, the negative feedback loop provides a path for the output of the op amp to regulate the inverting pin. In the schematic U1G2 can do the same by controlling U2.
I think you're right about the meter, I tried the 5V regulator as xavier60 suggested and it messes with the regulation, guess I'll have yet another small transformer.
---
David,
for the offset you mean something like the trimmer (R11) I had in the first schematic?
-
As I understand it, the negative feedback loop provides a path for the output of the op amp to regulate the inverting pin. In the schematic U1G2 can do the same by controlling U2.
I think you're right about the meter, I tried the 5V regulator as xavier60 suggested and it messes with the regulation, guess I'll have yet another small transformer.
---
David,
for the offset you mean something like the trimmer (R11) I had in the first schematic?
I'm surprised about that. I couldn't find wiring instructions specific to your panel meter. Is there an instruction sheet that came with it?
https://www.youtube.com/watch?v=RnWrsT-QBG0 (https://www.youtube.com/watch?v=RnWrsT-QBG0)
-
As I understand it, the negative feedback loop provides a path for the output of the op amp to regulate the inverting pin. In the schematic U1G2 can do the same by controlling U2.
I think you're right about the meter, I tried the 5V regulator as xavier60 suggested and it messes with the regulation, guess I'll have yet another small transformer.
---
David,
for the offset you mean something like the trimmer (R11) I had in the first schematic?
Do you have a diagram showing how the 5V regulator and meter were wired?
-
will charge pump 7660 work here, instead a second transformer and all around ?
-
It's even possible to obtain the negative supply from the main transformer and use a voltage regulator (like a 79L05) if consumption is not too high.
(http://)
-
Thanks on the tips... I'll try them out, so... I tested everything on the protoboard and worked great.. then tried it once soldered to a stripboard, the results were not that good.. I guess it's expected from my first project u.u
maybe someone can give me some tips, like do you tested the circuit each time you solder a component? also something I hadn't considered, how much current do these boards can handle?I don't know if it should have an external cable to handle the current or simething like that..
-
not good, means, not a stable output when apply a load ? Need a symptoms
-
yep, I'm sure I screwed up somewhere in the wiring, I got an output range of about 3V on the upper end. I'm also a bit concern about the amount of current the stripboard can handle, apparently it may be somewhat higher that the ~1A I need but I'd like some confirmation.
-
yep, I'm sure I screwed up somewhere in the wiring, I got an output range of about 3V on the upper end. I'm also a bit concern about the amount of current the stripboard can handle, apparently it may be somewhat higher that the ~1A I need but I'd like some confirmation.
I don't think you should have any problem with just 1 amp .But you can always double up the wire where most of the current travels. That would be the main ground plain and main positive plain.from rectifier to output.
I usual just check continuity between connections.This finds bad solder joints and solder shorts.If you tin the entire solder strip that also helps.A little extra flux won't hurt either.Just make sure that you clean the board well after soldering is complete.Sometimes just a tiny piece of dirt or carbon can cause problems..Another tip is to make a hard copy of the schematic so you can check off each connection as you weld it.
-
The biggest capacitor seems to be on the right side and not next to the bridge rectifier. Is it normal ?
May be you swapped some components ? It's easy to make mistakes :(
-
Hi ggchab,
What happened there is that I have a bridge rectifier that doesn't fit the holes in the stripboard as you can see in the picture.
---
Anyway.. I finilly had time to try again and works great, the only problem is with the volt/amp meter that apparently I can't connect anywhere without messing up the current regulation in the circuit, I guess I'll just power it from a different source.
Thanks for the help