Korad KA3005P improve efficiency

[80+platinum]

The Korad KA3005P has an lineair power supply with 50 Hz transformer and power electronics to reduce the voltage (and dissipate heat). I think the Korad KA3005P can be much more efficient. Below are some measurement with different power outputs and different gridvoltages:

Gridvoltage   Power in   Vout   Iout   Power out   Power loss   Efficiency
230 V   111 W   9,22 V   3 A   28 W   83 W   25%
160 V   75 W   9,22 V   3 A   28 W   47 W   37%
230 V   207 W   15,41 V   5 A   77 W   130 W   37%
160 V   138 W   15,41 V   5 A   77 W   61 W   56%

The Korad KA3005P has multiple taps on the transformer that are switched with relays. The tapsetting is only depended on the actual output voltage. It does not take the gridvoltage in account.

I want the tapsettings to be more efficient. The powerloss is pretty high now. How can i change this tapsettings without rewinding the transformer For longer high power loads i am using an autotransformer to reduce the gridvoltage to 160V...

[coromonadalix]

To gain what in the end ?   more design complexity ?  lower temp losses  etc ...

It is designed like this,  it is normal in a sense you have thermal losses etc ...  LINEAR says it all,   even having tons of supply taps,  you have the load variations,  if the supply get a little bit low on the input power section  you may loose efficiency and stability ...

This is a pretty basic psu ... and lucky you you have xformer taps in it, other dont  loll 

Not my 52 pourd   Lambda 80vdc - 8 amp psu,  when i push it to the max,  i have an 864w heater  loll

Get a switchmode ??   or some  psu  like Kikusui, they are in some models  pseudo phase controlled with a mix of buck and linear mode,  they dont heat at all

If not,   get some sorensen or other psu based on phase controlled ac main inputs  ...  they get bulkier because of the big coil input they have, since they almost work lik buck converters


my 2 cents

 

[80+platinum]

I understand the principle of a lineair power supply. But the heatloss is unnecessary high.

An switchmode was better. But I already bought this one a few years ago. Can I fix or improve this bug 

It seems my device is designed for 160 Volt, but in my country I have at least 207 Volt.

[Kleinstein]

The supply needs some reserves for different effects. One is a possible drop in mains (e.g. 220 instead of nominal 230). Another is the ripple in the filter capacitor, that can be larger when the current is high. A 3rd point is a reserve for the short interruption of the power when switching the relay. The relay is on the AC side an apon switching on may just miss much the a half cycle. So for this transition the ripple may be about twice the normal size.

Because of the stress to the relay contacts is would be problematic to make the filter capacitor much larger.

A few more voltage taps could allow for a slightly higher efficiency, but with more effort.
One could take the current (at least the current limit setting) into account and in some case allow for a slightly lower voltage for tap switching. Still this in a relatively large effort for an often small gain.
A rather crazy idea could be synchronizing the relay switching with the mains phase and this way reduce the dopt due to the break before make time a little.

A much better efficiency would need a more drastic change, like a switched mode preregulation or may be electronic cross over between 2 supply levels possibly in addition to the relay switching.

[coromonadalix]

any switchmode reputable brand should do at least minimum 85-100vac  up to 240vac,  if not search for goodbrands,  not china crap stuff

you main losse is  before the main power stage and after it is versus the consumed current as you may guess,  not necesarly on the primary xformer

for the xformer you choose an toroid one, better efficeiency, less thermal loss for a start

if the main rectified filtered voltage is 40 volts, and after the pass transistors, you ask 20 volts, you have 20v of loss multiplied by the consumed current, 1 amp is already 20 watts of loss ...

you have to find better design psu if it is a real consern to you


scr controlled bridge rectifiers  etc ....

in the end you redesign almost everything  loll and you have no garantie it will work reliably with and without load ...

[tunk]

No expert, but a few questions:
- how do you measure input power, does it include power factor?
- are the selected voltages just above the relay switching?
- at 160V, are you sure the output voltages are constant?

[coromonadalix]

thats why on linear psu  sometimes a minimum of 10vdc or more before the pass transistors is needed

It play a role in the load and line regulation, energy reserve  etc ..  if the main dc power get too low  it could become instable

Thats why some psu designs have a floating regulator for the low power side, and use a main power section,  the floating section will help stability


BUT    i never saw the schematics for this psu, is it an switching mode and a have a linear section  like the korad 3005d ??     


And for the OP   you have threads here about  some 3005 models

for 160v    tests have to be made with a variac ?

[radiolistener]

The Korad KA3005P has an lineair power supply with 50 Hz transformer and power electronics to reduce the voltage (and dissipate heat). I think the Korad KA3005P can be much more efficient. Below are some measurement with different power outputs and different gridvoltages

this is a lab power supply, it doesn't needs to be efficient, it's goal is different - to provide clean, stable and regulated Voltage/Current for your needs. If you want good efficiency, you're needs to see for a specific PSU which fits your device.

Also, if you're thinking about efficiency, probably you have device with a high power consumption, such device usually doesn't require clean linear PSU, so you can look for a switching mode PSU, it will be more efficient, but with noise.

[80+platinum]

The power measurement is including Power Factor (displacement en distortion power factor). So the Watt power
I don't know. But with charging an 24V battery this has been tested. The exact switching points I don't know.
At 160V the output voltage is stable at stable load. I haven't done any testing with load changes.


A lot of people in this topic say the powerloss should be this high because for different reasons. Can it be possible that this is not necessary and it is just an design fault? Because I have seen SMPS with lineair after regulator. I don't think they waste for example 20V of heat

Buying an new device is not the best solution. I like to improve products that are designed wrong.

[Kleinstein]

The swtiching points may not be optimized very much. So one may get a little less heat loss with adjusted switching points. However this comes with the danger of getting glitches of a drop out under difficult conditions. I think the most difficult point is when the mains voltage is low, the current high and the set voltage is raised so that the relay will switch to the next high voltage tap.

Depending on the filter capacitor one could expect a voltage ripple of some 5 V during switching this may about double to some 10 V from a half wave missed, depending on the relay.
The output stage and shunt could need some 3 V. So 13 V of reserve may be actually needed. Under light load the transformer voltage may rise by some 10% which would be another 3 V and with a slightly higher than normal mains votlage one could get another 10% extra and this way get close to 20 V to just avoid a drop out under unfavorable conditions. The 20 V of drop at a more random set voltage, not just may even be already on the low side to make sure not to get a glitch even under worst case conditions. Just on the edge of switching to a lower tap, 20 V of reseve for the linear stage sounds like a bit too low. 

One may reduce the drop a little with a lower drop output stage (maybe save 1 V or even 2), larger capacitors (but this comes at the price of a reduced relay life) or maybe mains sychronized tap switching to avoid the extra drop. The cheap version is accepting a chance for glitches -  not sure if that is an improvement.

They may do a bit better with making the switching levels dependent on the actual voltage in the filter capacitors and thus the mains voltage and loading effect. This could still fail if there is a rise in the load current that causes the need for tap switching.

Supplies with a SMPS and than a linear post regulator can get away with less voltage drop in the linear stage: the SPMS can react faster and does not have the extra drop out from switching. With a SMPS it is also common practive to let he SMPS follow the output voltage, so that mains variations do not matter. With a SMPS pre-regulator I would more like expect 3-6 V of drop for the linear stage.