EEVblog #315 - Korad KA3005P Review/FAIL

[pomonabill221]

The version of the Korad that Dave destroyed is an old one and Korad has upgraded the supply (the output transistors).
You really need to do a little more research on the supply before you completely write it off and call it crap.
Yes it is a cheap supply, but for what it does, and for what it costs, it is ok for most.

I have two Lambda supplies (an LQ-531 ( 20v/8.6amp and LQ=533 60v/3.3amp) and they are tons better than the Korad, BUT are about 3x heavier and at least 2x larger.  For my testing, I use the Korad and knowing the limitations and faults, the Korad is just fine.

If you want to be a brand snob... so be it, so good luck finding a "good" supply for around 100 that is remotely programmable.

Did you even bother to look at the schematic and understand what is going on?  May be educational as well.

Seems strange that you are bashing the Korad but you didn't even know that the B&K was a programmable dummy load!

[GiskardReventlov]

If you want to be a brand snob... so be it, so good luck finding a "good" supply for around 100 that is remotely programmable.

First I think you should refrain from name calling it makes you look bad and it makes it look like you want to avoid discussing the reality.
You're throwing brand names around and you seem to be personally offended by completely deserved criticism. Do you have a stake in Korad?
As for wishing luck, I wouldn't pay $5 for that, it's a waste of my time.  You need to understand that crap is crap no matter what the price.

And I never saw him do any remote programming with it, we don't know if it works.

Seems strange that you are bashing the Korad but you didn't even know that the B&K was a programmable dummy load!

Again trying to deflect the real issue.  Explain how or why that's relevant to discernment of that Korad crap.
And that issue is that people should either find a good used supply for similar price point or save up some more and get good quality that's reliable, that has sound and logical thought behind it and that's safe to use with confidence.

I watched the teardown, like I said Dave was really generous.

Here's a previous quote which sums it up well:

I wasn't impressed with the teardown stage. Everything is 'there' to get a DC programmable PSU, but it seems rubbish built. Uncleaned PCB's, random holes drilled, poor soldering, weird R2R DAC (apparently PCB space and SMT parts placing is free for them) and overall shitty stuff

Well said Hans.

[pomonabill221]

You have your opinion... we all have ours...
Whatever blows your skirt up!

[Bored@Work]

@GiskardReventlov

You are running in the classic problem here. Those who bought it will defend their buying decision with teeth and claws. AKA choice-supportive bias http://en.wikipedia.org/wiki/Choice-supportive_bias You have the same here with people defending their shitty multimeters.

The Korad is the Yugo of programmable power supplies. Let them drive around in their Yugos and pretend they are driving Porsche, making some va-va voom voom sound with their lips.

[pomonabill221]

I never said it was a porch (that is what I think of overpriced VW's).
I know it is a cheapy, and not having a job right now, I have to watch what I spend.
If you would like to give me a high priced power supply, do so... I won't refuse a freebie.
I am NOT defending the supply... I know it has alot of problems/limitations, but for what I use it for, it fits my purpose.
I could drag out my Lambda's, but they are big, take up a lot of room, and I don't need that precision for what I am doing right now.
I don't know why I even responded... the complainers here are obviously just complaining about nothing... what do they have that is better?

[GiskardReventlov]

AKA choice-supportive bias

I know, sometimes it's so bad though that I can't refrain.  But mostly people have to realize it ain't personal. In fact it's painful to see and in some cases (e.g. PS and multimeters) there are real and significant safety issues.

If calling out a heap-O-crap was just complaining Korad would never have fixed it.

But it's also very entertaining to see what horrors Dave reveals. So for that we are grateful.

[jonathanpatate]

Hello from France. Here this PSU can be bought under the Velleman brand with the PS3005D reference.
As said before, it seems to be an upgraded version, so it bought one, and yes it is !

The fan is noisy and the buzzer stay weird too.

But the encoder for selecting voltage or current is now "stem" (with steps).
The 7-segment displays are well aligned/positioned (no more gap between the decoration window).
Some caps are Rubycon, others stays cheapo.

See pictures.

[pomonabill221]

That's a spittin' image of my Korad!  Not surprising though EXCEPT your front panel shows a M5 where mine shows LOCK (panel lock).

[jaxbird]

@GiskardReventlov

You are running in the classic problem here. Those who bought it will defend their buying decision with teeth and claws. AKA choice-supportive bias http://en.wikipedia.org/wiki/Choice-supportive_bias You have the same here with people defending their shitty multimeters.

...

It works both ways.

Frequently even stronger for those who spent a larger sum on a brand name, refusing to acknowledge that a cheaper product, can be adequate, perhaps even superior in some areas, and capable of performing the same duty.

One thing I do find interesting, are those, having no personal experience with a specific product, feeling the need to express strong (often 2nd hand) opinions about it's qualities.

[David_AVD]

I bought one of those Korad units.  It feels somewhat nasty to use (poor action of the buttons mainly) but I guess I got what I paid for (just).  I don't think I'd buy another one though.

[Electro Fan]

Everyone has their opinion but for what it's worth the KA3005P I've been using has been solid.  It's easy to operate (the buttons, display, and dial work fine) and it checks out reasonably close to my DMMs (including a Fluke) and voltagestandard.com devices.  For the price it's a good value.

Here are some reviews from other users:
http://www.amazon.com/KORAD-KA3005P-Programmable-Precision-Adjustable/product-reviews/B0085QLNFM/ref=cm_cr_pr_top_recent?ie=UTF8&showViewpoints=0&sortBy=bySubmissionDateDescending

[3w_049]

For USA folks.  SRA solder has these on sale for $95 shipped on ebay today.

http://www.ebay.com/itm/KORAD-PROGRAMMABLE-PRECISION-VARIABLE-30V-5A-DC-POWER-SUPPLY-DIGITAL-LAB-GRADE-/110888868115?pt=LH_DefaultDomain_0&hash=item19d17dd913

The ebay user id is randyruby - but it's SRA solder.

I couldn't resist - I order two.

The price has gone up significantly, hurry before it costs more than Agilent craps.

[madires]

Just saw the Korad at a local online shop, EUR 80 for the KA3005D and EUR 100 for the KA3005P (RS232 & USB).

[rr100]

Can somebody outline the programming capabilities over USB/serial? I assume you can control it from PC but can you also read the values on display? Can you also log them? Any programming capabilities/API/command line/etc?

[poorchava]

IMO programming PSU from PC is useful if you are trying to do some automated testing or stuff. I've built myself a PSU that can be connected to the PC and I find it that it's easier to just turn the knob or press 'recall from memory'. That's it.

[rr100]

Obviously dialing 5V for example is easier without going through the computer but I'm interested to either program it to use some specific algorithm or to log how much energy goes to something or how long it takes.

I looked at some screen shots and for now it looks it has the capability to set whatever (or most of what) the front panel buttons set and also some time-based "scripting"/change of some parameters.

However, I don't know for now:

- if it reads at all to the computer the actual values
- if you can do anything with them (like vary the voltage depending on the current, or maybe even current and time, etc)
- if there are some scripting capabilities so you can read/write whatever you want to the PS.

Other than that what would be a simple way to tell if you're getting the "good" version? I assume the decent shops by now should have the good one anyway but checking would be good.

[Bored@Work]

Obviously dialing 5V for example is easier without going through the computer but I'm interested to either program it to use some specific algorithm or to log how much energy goes to something or how long it takes.

I looked at some screen shots and for now it looks it has the capability to set whatever (or most of what) the front panel buttons set and also some time-based "scripting"/change of some parameters.

You will find one of the problems these things have is you can't precision-time the settings. That is because of the programming interface, but also because of the PSU. Especially reducing the output voltage is a problem, because these things have caps at the output and their discharge rate depends on the load. The lighter the load, the slower the voltage change. What they don't have is what is called active down programming, an instrument-controlled rapid discharge path over the output cap(s).

You also can not precision-time the readings. I.e. if you think of a sampling system, you have one with a lot of jitter in the sampling frequency. And this becomes even more apparent if you want to calculate things like power. This requires two separate readings I and V, which should be from the same time but are not. And each reading itself has a jitter with respect to the sampling frequency. The result is you can only measure very slow changing DUTs.

[rr100]

That's correct indeed, if there's any significant variation (even for what is considered low frequency, like 10Hz) the measurements might mean nothing and/or the CV (and even CC) might not answer fast enough.

Speaking of that I think it would be worth having some tests and explanations from Dave on the pitfalls of power measurement when you can's assume the waveform and the phase of the voltage/current. Let me try to find if there's some thread with ideas for future videos :-)

In any case for this application I don't care much of "fast change" is just some kind of glorified charger/test bank.

[rr100]

Ok, I think I can answer the questions myself: as somehow expected the protocol is easy enough and reasonably complete (I assume is the same as for the Velleman clone): http://sigrok.org/wiki/Velleman_PS3005D/Info

This starts to go from "fail" to "best buy" in the "cheap" sector, at least in Europe where the cheapest 30V 5A seems to be about 70eur. And of course is not programmable, just one of the basic fine/coarse volts/amps, with only 3 digits on both A and V, without memories, without the possibility to set the current/disconnect load/etc. And of course you REALLY don't know what you get inside the box and for sure they try to skimp as well on power/expensive components.

[rr100]

I've got one for myself for 99 eur from reichelt.de.
Not much to complain yet, except that the binding posts are now "internal" (multimeter-like) connectors. Of course, that makes them better safety-wise (they too take shrouded banana plugs) but still a bit of a bummer. Any (ebay?) recommendation for some cheap banana->binding post adapters that can take a couple amps?

[soren]

Reichelt had the KA3005D on sale for 55 EUR last Friday, so I bought one.

I was a little concerned to see that it has the silly old M5 memory slot instead of a LOCK indicator, but despite this it seems to be a recent build.

Some impressions:

The power transistors are a pair of Toshiba 2SC5198's, which should be easily up to the task. I wonder what type the old ones were.

The fan sounds like it is decent quality but it's quite loud because of chassis turbulence and being run unnecessarily fast. Maybe I'll just add a series resistor as the exiting air isn't too warm even when the power supply is at maximum thermal load.

General soldering quality is good, but the hand-inserted components are all just thrown in rather carelessly. In my case, the overload thermistor had its leads shorted and it also wasn't really pressed up against the heat sink.

The three large caps marked Rubycon may look okay in the picture, but they feel very rough to the touch and unlike real Rubycons, so I suspect that they are instead "con rubies".

The 7812 on the controller board has a tall heat sink that is only attached to the regulator and doesn't even rest against the PCB. Extremely wobbly to the point where I'm a bit surprised that it didn't break off during transport.

No overshoot! Discharging the output cap takes half a second with no load.

Calibration is easy and works fairly well. There's some drift, but that's okay. I didn't expect this to be a precision product.

Pressing the mechanical power switch to turn the power supply on causes a fairly strong burst of noise which doesn't appear if the switch is left on and the IEC cable inserted instead. Having read about previous runs of this PSU in the 120V version, which had lots of problems with the power switch, I wonder if they are still using that same weak switch in the 230V product and if it might be underspecified even there, causing arcing. It's also only a single pole switch.

[Andreas]

Hello soren,

Thanks for the teardown and the measurements.
I also couldnt resist on black friday for 55 EUR.
And some weeks ago the KA3005P went for 79 EUR as a computer controlled powerstage for my peltiers.

I do not think that the fan is unnecessarily fast.
In some special situations driving into 5A current limit from high voltage repeatedly
 (so that the relay switches happily) you are using up the SOA of the transistor at 25 degrees.
If you have higher temperatures you will be above SOA.

So I think I will order some 2SC5198 just in case....

With best regards

Andreas

[pomonabill221]

Has anyone been able to run the software for the KA3005P on windows 7 64 bit?
If so, what was the software and version?
I have not had any luck with mine... did work on win2000, but 7 is a no go.

[rr100]

Yes [it runs on win7-64], I'll let you know when I have access to that box, can't remember if it was Korad or Atten software.
Can't remember if it was over serial or over usb/virtual port - most likely usb.
BTW does the PS come with a serial cable? I can't remember and if it does I can't remember where I put it (if I ever had it) and is nagging me. Not that I don't have plenty of serial cables...

[Andreas]

Hello,

I have tried the software on a Win XP system. (of course with blocked internet access).
The function of the software was fine but
the XP system had unusual (unstable) behavior after installation.
So I uninstalled the software. And system behaved as before.

For me the software seems to be somewhat crappy or even dangerous.
There is one file left after deinstallation. (KA3005P.aliases)

content:
[ÎҵĵçÄÔ]
ÎҵĵçÄÔ = "localhost"

That what I do not understand:
so for what in the world does a simple control program
for a power supply need a network access?

With best regards

Andreas

edit:

BTW does the PS come with a serial cable?

Mine came only with the (light blue) usb cable for connections.