Looking for 30V 5A schematic

[edalsan]

Hello everyone!

I'm trying to set up a small lab now that I ended my studies and I'm looking for a good power supply, so I wanted to ask you all first. I have two transformers here from an old never done project, a 15V-0-15V 5A and a 30V-0 5A transformer (both are same size and weight).

At this point I'm considerating some options to fit them in a good power supply/supplies:

• Just one big power supply, capable of handling 100V 10A, so by choosing parallel/series transformer I could get two ranges of 60V 5A or 30V 10A
• Use two separate power supplies 30V 5A, and series/parallel them when needed

So, I'm thinking about the connection in the diagram attached, which will allow to use two 30V 5A power supplies, or 30V 5A and 15-0-15 5A supplies (maybe the 15-0-15 could be used too while using the other two 30V 5A, but not sure about that because will need to be sure current will not get that high from center tapped transformer).

So far I've being looking at two main possibilities:

- Simple LM723 schematic with current limiting, for example http://electronics-diy.com/30v-10a-variable-bench-power-supply.php.

- But, as I've been hearing lot of bad things about poor 723 current regulation, I'm considering this one too with an LM317 without current limiting, LM317HVT will be needed as voltage will be more than 40V actually. An example of 10A 30V with LM317 http://www.cuteminds.com/en/elettronicagen/electonics-projects/55-alimentatore-variabile-1-30v-10a

At this point I would like to ask you all which option you think will be the best one to do the  best with those two big unused transformers I have, and if can recommend any good power design supply for my needs. I don't need ultra low ripple or so fine regulation, and I'm not even sure if I will need current limitation, so any help/schematic will be appreciated.

Thank you all!

[Kleinstein]

With normal rectification and filter capacitor a 5 A transformer is good for only about 3 A. From 30 V AC one would get something between about 37 and 48 V, depending on load and line voltage.  So the 30 V voltage is realistic and even a little more is possible. For a pure linear supply this would be about 130 W of worst case power dissipation - possible, but not that easy to handle.

The 2x15 V transformer could be used to reduce the power dissipation to a little more than half.  At this power level it is usually worth using that extra tap. There are a few option for this.

I would consider a single supply - much more than about 30-40 V gets more difficult anyway. Series connection of two separate supplies is simple, parallel connection a little more difficult.

For experimental use a current limit is definitely a good idea. Both to protect the supply and external parts. The LM723 current limiting is not accurate, but good enough for protection purpose. It needs a little extra effort to make the LM723 work down to 0 V, but it is possible. However the supply for the 723 is limited to 40 V - so it can't stand the full raw voltage it need an extra limit / regulation. If you don't need a super stable voltage and accurate adjustable current limit, the 723 is OK. However some (many) of the circuit one finds in the web are poor, with obvious flaws. A 0-30 V output range is possible.

The LM317 is not a good solution if one needs more than about 500 mA - it can't really stand a lot of power loss and extending the current with external transistors is problematic / not worth it.

There are circuits for a full lab supply with a precise adjustable current limit. Most circuits use an additional small transformer (e.g. 5-10 VA) for the control circuitry and often a display.

[hermit]

Someone posted this a while back and I found it interesting.  You can parallel the regulators.


Linear Tech PDF

[b_force]

Someone posted this a while back and I found it interesting.  You can parallel the regulators.


Linear Tech PDF

Interesting design!
Wonder if there is also a way to make a symmetric PSU?

[tautech]

The TO3 package LM338 makes it reasonable simple but be sure to heatsink them well and include the protection diodes.
The first decent PSU I built I used these and switched transformer taps to reduce dissipation.

[Zero999]

A lot can be done with the LM317.

The LM317 can work at higher differential voltages than 40V, by adding a couple of transistors, forming a pre-regulator, which can be made to track the output voltage. You might want to use a lower voltage zener and a Darlington transistor. It depends on the application.


A pass transistor can be added for higher currents than 1.5A. Unfortunately, this defeats the short circuit protection, which can be reinstated with more components.


It can be made to go down to 0V, using a negative power supply and an op-amp.


A centre tapped transformer and another transistor + diode can be used to switch to the centre tap, when the output voltage falls below a certain level. Unfortunately I think I undersized R2 in the schematic below. Try 470R.

[edalsan]

Thanks for your replies!

About the LM723, I really forgot about the 40V limit, so as Kleinstein said the power supply must use secondary voltage for the electronics control, and the beefy transformer for the ouptut power stage. Regarding that, I've found a circuit published in Elektor in 1983 (number 35, April on spanish edition) which I'm including as pdf attached. The power supply relies on 723 for control powered by 12V transformer, and just command the power transistors connected to the 30V 5A one.

I've been lucky, looking for this power supply and I've found this thread on a german forum https://www.mikrocontroller.net/topic/323755 in which an user has made new PCB design which allows to sepparate the control PCB with 723 and surrounding elements, with output connectors to drive power transistor bases. I'm including what seems to be the final version for this updated PCB and schematic in a zip, all credits to Luca E. (derlucae98) from that forum.

So far it seems a good option, well regulated power supply, and that can fill my needs of ~30V and ~3A. And as it has no power stage PCB you can make your own with bridge rectifier and adding any power transistors you want to use.

By usin a 3 pins regulator I know I will lose current limiting and I think I wanna have it to control and check any flaws in circuit designs and avoid burning them. LM317 built around center tapped transformer looks good, but I was looking maybe for something bit more "regulated".

[Kleinstein]

The linked in Elector regulator is using the LM723 only as a reference voltage. So one could replace it with an LM329 is well and still get better performance.

The circuit is kind of a standard low drop lab supply, somewhat similar to the circuits used in the cheap Chinese designs (e.g. HY1803, Korad 3005), but still without the transformer tap switching and no measurement.
Adding tap switching reduces the demand on the output transistors and heat sink.

There is no absolute need for a second transformer for the LM723. One could use a simple voltage limiter / simple regulator (e.g. LM317(HV) like, or just zener + transistor) for the supply of the regulator. It is only with transformer tap switching (if done AC) it really helps to have a separate transformer. The second transformer also reduces the voltage lost at the regulator. So the floating circuit could also deliver up to about 35 V (with a little less current).

One of those cheap Chinese smps modules from Dave's last video (#1030) might also be an option - however the up to nearly 50 V might be a problem.

[b_force]

This is an interesting transformer to use.
Price is decent
https://www.digikey.nl/product-detail/en/triad-magnetics/F-360U/F-360U-ND/7318113

Other idea is just simply buy a (second hand) Chinese PSU and use your own (better) schematic.
I have seen some for around 50 bucks, just a transformer and a case would be more expensive.

edit:
I mean, something like this. 43 bucks for 5A.
https://www.amazon.com/ZTDM-QW-MS305D-Stabilizer-Precision-Adjustable/dp/B06VXGLFWK/

[Kleinstein]

The circuit in the cheap supplies is often not that bad. The basic circuit is generally good, it is more like details that can be bad.  A common problem is that they allow two much current - so something like reducing the current limit from lets say 5 A to 3 A would already fix some of the reliability problems. 

There is a chance the relays used for tap switching will not last very long - so if used a lot they can wear out. This is especially a problem with highly variable load, that sometimes activates the current limit. Due to the limited capacity of the relays it is generally not a good idea to increase the filter caps.

The TO already has the transformers, though he still needs case, heat sink and similar.

[edalsan]

That's a good option to repurpose a second hand power supply updating it's circuit b_force, but as Kleinstein pointed out I'm already the happy owner of those two transformers, and as far as I can remember they were not so cheap.

For the power supply seen in last Dave's video, there's a version capable of delivering 50V 5A, called DPS5005, which costs around 25€ in aliexpress (there's even one version with USB connection to computer). So, as it's a DC-DC converter I will need to set up the bridge rectifier and filter capacitors before powering this little module. I've been checking the ripple and specs and people are pretty impressed with the performance of the power supply. I'm considering that option as a good one.

Have been checking also for 48V 5A power supplies (the switching metal enclosed ones) and they cost around 18-20€... so I'm not either sure about throwing my fellow transformers or go ahead with them. But so far provide them with a rectifier bridge + caps and add the 50V5A power supply, seems a good option. What do you think about series/paralleling two of those 50V 5A power supplies? I think that if each one uses it's own transformer there will be no problem.

[Zero999]

By usin a 3 pins regulator I know I will lose current limiting and I think I wanna have it to control and check any flaws in circuit designs and avoid burning them. LM317 built around center tapped transformer looks good, but I was looking maybe for something bit more "regulated".

Adding current limiting is fairly easy. This one effectively drives the pass transistors with a multiple of the LM317's current, it will reduce the current, when the IC current limits. The voltage across R3 is dependant on the current flowing into the LM317. It sets the voltage on Tr1's base, which acts as a common emitter amplifier. The voltage across R5 will be equal to the voltage across R3 minus the base-emitter drop of Tr1. The voltage across R4 is roughly double, the voltage across R5 and sets the voltage on Tr2's base which acts as a current source. The higher the base voltage on Tr2, the higher the voltage across R_E, thus the current. In short the current through R3, controls the current through R_E.



The centre tapped transformer design is regulated. Only the input voltage to the LM317 is changed, depending on the output voltage.
Does the current limit need to be adjustable? If so, how accurate does it need to be? The LM723 only has rough current limiting.

[edalsan]

Does the current limit need to be adjustable? If so, how accurate does it need to be? The LM723 only has rough current limiting.

Yes, I would like to adjust the current limit, the schematic looks great, but I'm not sure if that adjust limit could be done in your schematic as both resistors are dependant.

Regarding to accuracy I'll be happy moving between 0.1A steps, but if 0.01A is possible it would definetely be much better. The LM723 seems capable of providing that, that's why I thought about it at the beginning, but now looking into those 50V 5A modules it makes me wonder if wouldn't be better to just get one of those off the shelf solutions + rectifier bridge + filter cap.

[Kleinstein]

A linear lab supply and those cheap  SMPS modules will have different properties. So it depends on what you need it for.

The SMPS modules have noticeable higher frequency ripple, though it does not seem to be bad for a SMPS it is still much higher than with a linear regulator (like 50 mV (at a few 100 kHZ) compared to maybe 50 µV at 50/100 Hz).

The regulation of a SMPS is usually slower than with a linear regulator. So things like the voltage drop on transients will be larger ( longer. Still those SMPS modules look like they are reasonably well tuned - so only moderate overshoot for the CC-CV transition. With a self build linear supply the tricky part can be finding the right balance between fast reaction and not to much overshoot / ringing.

The SMPS modules can deliver more current. At the full voltage there will be not much difference - so the 5A (Ac) transformer would set a limit at about 3 A (DC), but at lower voltage SMPS version could deliver more current as the input side current will be lower. So one could get the 5 A from about 20 V down. The linear supply is always limited by the transformer current. It's only a version with tab switching that might deliver a little more current (e.g. 5A) at lower voltage, like below about 15 V, depending on how it is made.

The Linear regulator will need a large heat sink and likely a fan. So it is not a small solution.

The linear regulator is usually a good idea for low power and sensitive circuits. The SMPS version is more something for high power loads, that are not sensitive (e.g. motor, lights, computer that has a switched more regulator anyway).

It can still make sense to use a transformer with one of these SMPS modules. The switched mode supplies (e.g. for 48 V , or maybe better a little lower as Chinese specs on such low cost parts are often stretching the envelope to the corners) have extra common mode noise. They are either grounded (thus the supply will not be isolated from ground) or have common mode caps (class Y) and thus have a voltage relative to ground. The conventional transformer provides a better isolation and suppression of common mode noise.

The current limit in the basic 723 circuit is only crude and usually not very temperature stable. Its adjustable, but depending on how it is done it gets less useful below something like 1-10% of the full range.

The LM317 based circuits have a few additional problems: the stability with a difficult load can be tricky. The drop out is rather high (thus getting 30 V out might need larger filter caps and thus lower power factor and more load to the transformer). When combining with external power transistors the 723 is usually the better solution than the LM317, at least it is made for this purpose.

[Zero999]

Does the current limit need to be adjustable? If so, how accurate does it need to be? The LM723 only has rough current limiting.

Yes, I would like to adjust the current limit, the schematic looks great, but I'm not sure if that adjust limit could be done in your schematic as both resistors are dependant.

It's possible to have adjustable current limiting by adding a low value (10R to 50R) potentiometer across R3 and connecting the base to the wiper, but the minimum current limit will be determined by the LM317's internal circuitry, which will be more than 1.5A. The LM317L could be used to take it down to 100mA, but it could still be easily double that. Still no good, if you want to limit the current to under 200mA.

Regarding to accuracy I'll be happy moving between 0.1A steps, but if 0.01A is possible it would definetely be much better. The LM723 seems capable of providing that, that's why I thought about it at the beginning, but now looking into those 50V 5A modules it makes me wonder if wouldn't be better to just get one of those off the shelf solutions + rectifier bridge + filter cap.

The LM723 and LM317 are analogue solutions, so they have a theoretically indefinite number of steps. The trouble is, the current limit is not accurately controlled and is purely there to protect the driver transistor from overheating. It will depend other factors, such as the temperature and input-output voltage differential.

I have no personal experience with those modules and haven't seen any videos. If cost is a consideration, then it might be better to buy a switched mode power supply and power it off that.

[not1xor1]

This is an interesting transformer to use.
Price is decent
https://www.digikey.nl/product-detail/en/triad-magnetics/F-360U/F-360U-ND/7318113

Other idea is just simply buy a (second hand) Chinese PSU and use your own (better) schematic.
I have seen some for around 50 bucks, just a transformer and a case would be more expensive.

edit:
I mean, something like this. 43 bucks for 5A.
https://www.amazon.com/ZTDM-QW-MS305D-Stabilizer-Precision-Adjustable/dp/B06VXGLFWK/

be careful
those cheap PSUs are not linear PSUs
if you check one of the pictures you can see, through the ventilation slots, that the case is almost empty

no 50/60 Hz multi tapped transformer, no large heatsink

I think linear PSUs are more expensive nowadays
even second hand ones you can grab on ebay, at least here in Europe, cost more than 50-60 euros (including shipping expenses)

[Kleinstein]

One can still get linear and SMPS based cheap lab supplies. The higher power ones (e.g. > 200 W) are usually switched mode. With the linear type the small ones (like 50 W) are still cheap, by the higher power one get more expensive.

One can usually tell from the weight if it is a linear or SMPS type. It also gives a clue on the size of the transformer and might be a warning if a unit is rather low weight for the claimed power, it might use excessive load to the transformer (and often other part too). However toroid shaped transformers are actually lower weight - so you can't buy by the kg only.

[b_force]

This is an interesting transformer to use.
Price is decent
https://www.digikey.nl/product-detail/en/triad-magnetics/F-360U/F-360U-ND/7318113

Other idea is just simply buy a (second hand) Chinese PSU and use your own (better) schematic.
I have seen some for around 50 bucks, just a transformer and a case would be more expensive.

edit:
I mean, something like this. 43 bucks for 5A.
https://www.amazon.com/ZTDM-QW-MS305D-Stabilizer-Precision-Adjustable/dp/B06VXGLFWK/

be careful
those cheap PSUs are not linear PSUs
if you check one of the pictures you can see, through the ventilation slots, that the case is almost empty

no 50/60 Hz multi tapped transformer, no large heatsink

I think linear PSUs are more expensive nowadays
even second hand ones you can grab on ebay, at least here in Europe, cost more than 50-60 euros (including shipping expenses)

I actually thought that most linear PSUs were the cheap ones, lol my bad.
I have seen linear PSUs for cheap as well, somewhere

[Kleinstein]

There are cheap small linear power supplies, like 0-18 V and 2 A. I found one in Germany for just under 30 EUR. However they get more expensive when they get larger. So the 30 V 3 A Version is more like $100 in a linear version. SMPS based versions can be cheaper in this power range. So the cheap one in the 100-200 W range are usually SMPS based ones. It is usually easy to tell by the weight.

A pure switched mode regulator will give quite some output ripple and the speed of the regulation and thus transient response is limited. Filtering the Ripple will slow down the transient response - so there are limits one how much can be done.

To reduce these two principle weak points, there are also supplies with a SMPS regulator followed by a linear stage. This way ripple can be low, regulation fast (least least most of the time, if the step is not to large up in power) without a massive heat sink and noisy fan. However these supplies are usually not the very cheap ones.

[b_force]

It would be interesting to find out if you can make these switching PSUs like a tracking psu.
Than it's just simply adding a decent good performing regulator to suppress the ripple and noise and you're done.

[Kleinstein]

The combined SMPS + linear supplies usually have the SMPS tracking the output plus a few volts so the linear stage has some reserve to work with and also to allow some drop in a filtering stage between the two.

[derlucae98]

all credits to Luca E. (derlucae98) from that forum.

Hi! Hope you don't mind it, when I pull this thread up again.

Just in case you're still searching for a power supply with tracking SMPS preregulator you might be interested in this design: https://www.mikrocontroller.net/topic/396405
In particular the final plans: https://www.mikrocontroller.net/attachment/298951/Labornetzteil-Schaltplaene-final.pdf
PM me if you want to have the EAGLE Files (PCB and Schematic)

Although it is designed for 24V and 2A it should be possible to upgrade it to 30V 5A.
I use it nearly every day and I'm quite happy with it.

Regards
Luca E.