-
Normally I avoid buying lemons by checking things out here first. But I guess this is too new. I decided to take a chance because Korad made things right on the KA3005 design that failed on Dave a while ago. I don't know that I'll ever have a need for it, but I bought the programmable version. It was only $20 more for that option, so what the heck.
It arrived today, and here it is in pictures:
It comes with a USB cable and one pair of really cheesy leads (also a driver disk, thin manual and power cord, not shown. The manual is better than many, but still Chiglish.)
Inside there's a fair bit of empty space. I assume this must be necessary for adequate cooling around that transformer?
The transformer core is about 4-1/2" x 3-3/4" x 3-1/4". Other than that I can't tell you anything about it -- there's no printing on it whatsoever on any side.
The power board is behind the transformer, dominated by a giant heat sink and fan. The little board above handles the programming communications, with both serial and USB ports.
Each side of the heatsink features a row of four large semiconductor devices well bedded in thermal paste. There's also 3 TO-220 devices with individual heatsinks on the right. They are not secured to the board, relying on the device itself for support. I'm not too happy about that, but other than that the boards seem well made.
I'm not sure what exactly these devices are, there's no ID on the chips. 5 of them have 3 terminals, 3 of them have 4 terminals. At a guess, I'd say IGBTs and bridge rectifiers, but I didn't try tracing anything out:
The fan blows down on the heatsink. The fins run fore-and-aft, up against the rear exhaust vents. The front of the heat sink has been blanked off with a metal plate to ensure all air goes out the exhaust. The fan is variable speed. It doesn't run at all when no current is flowing and ramps up when the PSU is delivering power. I don't know if I ever taxed things enough to hit full speed, but what I've heard so far is very quiet. (BTW: the beeps when a button is pushed are equally unobtrusive).
The brains of the thing are mounted to the back of the front panel. Mostly a whole lot of smd devices, but there's also an unoccupied 4-pin header. Jtag?
It comes with binding posts, which I like. The holes are a bit smaller than I'm used to though and they didn't bother to orient the holes in any particular direction on assembly! Personally, I think the holes should be aligned vertically rather than pointing at the adjacent post.
Fortunately, the posts are secured with screws to the control board (J6 to J17), so it was easy to align them all properly while things were apart. I'm still debating whether or not I should drill out the holes a bit larger though:
Looking at pictures on a website, I did not realize that the footprint was so damn big! It takes up a fair bit of real estate, which as you can see is at a premium in my little corner.
So, how does it work?
It seems to be fine with a dead short across the terminals. It survived the "welding test". So hopefully it has whatever improvements that Korad made to the one Dave tested. I don't have a fancy programmable load like Dave, so I tried to load it up with a carbon pile battery load tester. Now, that thing is built to measure hundreds of amps at 12v, not 5 or 10 amps at 30 or 60 volts. The adjustment is just too coarse for any semblance of precision, so my results may not be completely accurate.
That said, I could not get it to deliver more than 4 amps at 62 volts (serial mode*). If I tweaked the carbon pile a touch more, the current jumped to 5.1 amps, but the voltage dropped to 48 volts. A touch more and the current remained at 5.1a while the voltage dropped even further (sometimes down under 2v), so it looks like the current was being actively limited. I'm going to have to find a more suitable load for this test. I had similar difficulties in parallel mode. It will deliver 10.2 amps though.
The fixed 5v channel has an open circuit voltage of 4.997, which drops to 4.927 under a 1 amp load (5 ohm power resistor). The 5 volt channel is always on, unlike the two variable output channels. Both variable output channels are controlled by a single on/off switch. The display shows the set current and set voltage for each when the switch is "off" and the actual delivered values when the switch is "on", unless you are actively changing a value while powering something, in which case it temporarily reverts to displaying the set point.
In serial and parallel modes, channel 1 is slaved to channel 2, so changing the setting of channel 2 automatically changes channel 1 to the same value. Channel 1 cannot be changed independently in those modes. The display behaves a bit differently too: when actively changing the settings, the channel 1 display will continue to show the actual output while the channel 2 display shows the set points.
Changing the setpoints is not quite as simple as twirling a mult-turn pot. First you push the button for the channel you want to change. Then you push the voltage or current knob until the desired digit is flashing, then you turn the appropriate knob to change the digit. The knobs have a nice detent at each change of digit, but can revolve infinitely. As a digit passes 9, the next most significant digit is incremented. So if you are adjusting the left-most digit, the knob is a very coarse adjustment, but the adjustments are progressively finer if you choose to adjust a digit to the right. EG: Adjusting the leftmost voltage digit from 10.81 will increase to 20.81 then 30.81. But if you are adjusting the first digit to the right of the decimal, it will increase to 10.91, 11.01 etc up to 30.91. If you need the full 31 volts, you have to then switch to adjusting the right most digit. Clear as mud?
One other complication: the adjustment mode times out fairly quickly when actively delivering power. Once it times out, you have to go through the whole rigamarole again to continue any adjustments. A good idea in principle, but in practice it would be nice if the delay was a bit longer.
The system is cute, but a bit complicated. It would be easy to accidentally crank up the voltage or current too high for the circuit under test. However, there are both over-current protection and over-voltage protection features. It took me a bit to understand how they work, but they can be thought of as a kind of a customized fuse. You can set a limit for the maximum currents and maximum voltages that each channel can deliver, separately from the normal adjustment. For example, one might set the OVP to 14.5 volts when working on an automotive device so that you cannot accidentally crank the voltage up to a damaging level. If you subsequently try to exceed 14.5 volts, the PSU will shut down. Similarly, you can set current limits. Each channel can have different limits provided you are not in Serial or Parallel mode, in which case the Channel 2 limits apply. (There is no OVP or OCP for channel 3 though).
There is also a "lock" feature that locks out all control panel inputs. It's not necessary to prevent accidentally spinning an adjustment knob, but it could prevent disaster if you bumped a memory button. There are 5 preset memories. Pressing M1 through M5 switches the outputs to the values that were used when that memory was last used. Any subsequent changes while that memory position is active are automatically saved.
The displays are reasonably accurate, to within 0.006v according to my meter. Which has not be calibrated by a lab or anything. It does not go down to 0, despite the display. My uncalibrated meter still showed a few hundredths of a volt which disappears when the On/Off button is pressed. I confirmed voltage was still present by checking for current across the leads.
So in conclusion then,
Pros:
Reasonably well made for the price.
Linear.
3 channels
Lots of available power.
OVP and OCP for those times when you don't want all that power available.
Can withstand short circuits.
Quiet fan. Much quieter than any other fan in the room.
Good cooling path
Nice clear easy-to-read displays
Good tactile feel to the controls.
Multiple memories.
No noticeable heat to the case when loaded.
Binding posts (as opposed to recessed jacks)
Unobtrusive beeps.
Cons:
Very large footprint.
A single set of cheesy leads with a 3 channel PSU?
Complicated adjustment procedure.
Short time-out when adjusting.
Channel 3 not switched
Unknown components when it comes time to repair things.
Some heat sinks not secured to a fixed point.
Did I mention how big it is?
To Do:
Find a more suitable load to test the extremes.
Try out the software and programming capabilities.
Figure out a better place to put it.
*A note for the uninitiated: Serial mode is simply connecting the two channels in series, internally connecting Channel 1 negative to Channel 2 positive. Available voltage doubles. Alternatively, either Channel 1 negative or Channel 2 positive can be connected to the circuit ground, turning the Channel 1 positive into a + voltage supply and the Channel 2 negative into a - voltage supply for things like op-amp circuits. Parallel mode connects the positives together and the negatives together, making a single channel with twice the available current.
-
You also want to check with a scope if it behaves on power up and power down (little, ideally no, overshoot).
-
Tell em the price son.
-
I'm sorry, I'm new at this
There is one North American vendor, and his price is $200. However, he offers free shipping if purchased via eBay, unlike purchases through his web store or Amazon presence. Prices may vary elsewhere in the world.
There is no overshoot. There is a momentary pause as it homes in on the final setting.
This is serial mode, 31v setting (so, 62v), no load, power on:
(note: by "power on", I mean via the On/Off key to turn power on to Channels 1 and 2, not powering on the main power for the whole unit)
Here is channel 1, 31v, no load, power on:
Channel 1, 1v, no load, power on:
Channel 1, 1v, across 5 ohm power resistor, power on:
Channel 1, 1v, across 5 ohm power resistor, power off:
Etc. etc. They're all the same so I didn't bother with all the nonsense of saving and uploading. And there's no momentary spike even setting the time base to 2nS. What you see above is a fair representation.
(I did, however, fix the glitch in my first post that prevented the last image from loading. Now you can see the footprint next to some other equipment for comparison.)
-
Oh, and I'm still digging around for that perfect load. SWMBO's hair dryer gets throttled back to 42v on high and only draws 3.6 amps at 62v on low...
-
The transformer does not seem to be large enough to deliver the full power for an extendet time. This would require something like a 500-600 VA transformer. This one looks more like 250 VA, maybe 350 VA. Also the elcos look rather small for 5 A. So the full current may not be available at the full current. Likely it would be better to have a maximum current level in the 3 A range to be save.
Also I can't see any fuses and no thermal sensor at the transformer. At least a fuse on the primary sinde could be added relatively easy.
Ohterwise it least looks ok. The ramps on power on look a little strange, but not bad. -
The fan should blow directly from outside, as you can see on your home PC.
And the total output power is not specified. http://koradtechnology.com/en/cp-26.html -
There is a 20 amp fuse in the mains socket. It is possible that there is a fuse inside the transformer itself, I've seen that before, but there is no outward sign of it and there certainly is no external thermal sensor attached to the transformer.
Elcos? Electrical connections?
It's funny you mention 3 amps. They also make a 3 amp version that is outwardly very similar. And we all know how companies like to use the same components in similar models to save on inventory expenses. I think it probable that the 3 amp version is differentiated from the 5 amp via software and a Jtag interface. Likely the only difference between the programmable and non-programmable versions too. (Well, that and the little communications board tacked on this one.) It wouldn't really surprise me if an engineer figured out he could squeeze 5 amps from the transformer by shaving all safety factors to the bone. Let's face it: this thing sells for $200.
It is replacing my little home-made 1 amp supply. Most of my projects to date are at half the voltage or less. I don't foresee needing a constant 5a 30v supply for any extended length of time, I just figured it was better to have the capacity and not need it than to need it and not have it. That said, there is nothing mentioned in the manual about a duty cycle or other restrictions and it comes with a 1 year warranty. So if it burns up running it at full power and 100% duty cycle in that time, it's their problem.
-
Elcos? Electrical connections?
electrolytic capacitors -
And the total output power is not specified. http://koradtechnology.com/en/cp-26.html
No total as such, but the specification sheet says "0-30v, 5a * 2"
Of course, that's the same specification sheet that says it's only 260mm deep, and you all can see how that worked out for me. They only missed being accurate by 100mm or so. (claimed 220 x 156 x 260 vs actual 250 x 140 x 360) -
Hey Torch,
Thanks for being an Early Adopter on this one.
I've been waffling on this one for the last two days, but your review has pushed me over the edge. Ordered one just now.
The quirks you pointed out are duly noted (again, Thanks!) but all things considered, the Korad KA3305P certainly seems to offer a lot for the $'s.
Personal 'clincher' for me? I couldn't find any other supplies that offered the USB (and RS232) control and logging options, without going to crazy dollars. I need these features for some upcoming test/optimization/validation tasks, which ideally can be automated. At the price, I see ways to save myself many, many hours of hands-on testing and logging.
Back to your review: I do hope they give us a firmware update option to allow longer timeouts on output adjustments. As you described the 'flow' and timeout, that could be a bit annoying (or potentially "Magic-Smoke-releasing"!).
Thanks again! -
Personal 'clincher' for me? I couldn't find any other supplies that offered the USB (and RS232) control and logging options, without going to crazy dollars. I need these features for some upcoming test/optimization/validation tasks, which ideally can be automated. At the price, I see ways to save myself many, many hours of hands-on testing and logging.
I played very briefly with the software this evening. There's two modes, one a virtual control panel with separate knobs for voltage and current for each channel (ie: 4 knobs) that can just be "turned" with the mouse at any time. Plus a virtual version of all the other buttons. The second mode executes a programmed sequence of voltage/current settings for both channels. IIRC, you can change the start and stop, duration and also the number of loops. It all seemed to work correctly -- I didn't have a load on things but the relays clicked and clacked and the displays changed to reflect the virtual display.
There's a graph of each channel running alongside the interface window, a logging on/off "switch", and a file-folder icon to set the file path. The graphs jumped around each time the settings were changed -- either manually or as the sample default program executed.
For some reason, the software doesn't like the resolution I'm running on that old laptop. Or maybe it doesn't like Win XP. Some of the programming related text boxes are crowded and hard to use, and I can't resize the windows. I need to do some tinkering there, maybe try it on my Win 7 laptop.
I was actually kind of surprised it installed on XP, considering this is a "new" product. I'm guessing it's left-over software. Don't hold your breath waiting for firmware upgrades if that is the case!
Oh, and I don't think it's a native USB port. I think it's a serial port with a USB to serial adapter built-in, based on the drivers I saw it installing. On the plus side, there were no glitches or problems with the installation.
That's as far as I got. I haven't even looked at the PDF guide on the cd yet -- I have 3 very active grandsons sleeping over tonight, so Poppa's focus has been temporarily redirected. But hopefully it will do what you need.
-
Update: the supplied software is not by Korad. Apparently they licensed it from "National Instruments" and it's called "LabVIEW 2014". www.ni.com/labview. And it seems to run fine on this Win7 laptop.
There's also a PDF list commands and syntax in case you want to write your own scripting. No idea how to do that, through a terminal emulator perhaps? -
Good value for money new Korad PSU package. Unnecessarily bigger than its contents as you say though. I wonder, would it keep a straight enough rail line on the scope at 200mV per division resolution AC coupled when powering a high current dynamic load? Not necessarily a punishing pulse test. More life like things like a loaded audio power amplifier pulling strong amperage on a few kHZ sinewave for instance. We haven't seen such testing even on Dave's usually more expensive PSU models videos I think.
-
Good value for money new Korad PSU package. Unnecessarily bigger than its contents as you say though. I wonder, would it keep a straight enough rail line on the scope at 200mV per division resolution AC coupled when powering a high current dynamic load? Not necessarily a punishing pulse test. More life like things like a loaded audio power amplifier pulling strong amperage on a few kHZ sinewave for instance. We haven't seen such testing even on Dave's usually more expensive PSU models videos I think.
Sounds like a great test. I'm certainly willing to try but unfortunately my days of loud music are long past.
Hmmm. I do have an ARB function generator (hantek 3x25) and a few big IGBTs (g4pc40u). And some old transformers for a load. It should be possible to construct a simple amplifier around those, right? I am not an electrical engineer so I would need some help designing said simple circuit. Or some alternative ideas. -
Probably doable, but to avoid the hassle maybe there is some old receiver around? Connecting on one channel's rails fuses after disconnecting its own PSU, no big reservoir capacitors aid to the bench PSU, and some cement resistors combo 5-10 Ohm 50W dummy load hanging off its speaker output posts should run the test without needing a loudspeaker driven to some earsplitting SPL.
-
The old receiver is a tube powered floor model radio. Most of the energy goes into heating the tubes -- not really very dynamic. Everything else is pretty low-powered around here.
I'll see if I can figure something out. -
A bunch of old car headlight bulbs should do the trick. Wire them in parallel/serial as needed. Just keep in mind they're non-linear and will draw about 10 times the normal current for a few milliseconds while they heat up.
-
The car headlights won't be very dynamic to do the ripple test that Salas asked for.
Funny you should mention light bulbs though. I had a half-dozen new-in-box H3 bulbs that I bought on clearance for $1 or two each a few years ago. On my first attempt to load test I hooked them up to the Korad and they drew a heck of a current, but no light output. What the heck? I checked and rechecked connections before I discovered that every one of them was shorted out. I think the wires on each are pinched by the metal housings. No wonder they were on clearance. -
The car headlights won't be very dynamic to do the ripple test that Salas asked for.
Of course, you have to switch them on and off using your IGBTs. The good old NE555 should do the trick. 50Hz would be a good frequency. This will keep the bulbs hot so their resistance should remain fairly constant.Funny you should mention light bulbs though. I had a half-dozen new-in-box H3 bulbs that I bought on clearance for $1 or two each a few years ago. On my first attempt to load test I hooked them up to the Korad and they drew a heck of a current, but no light output. What the heck? I checked and rechecked connections before I discovered that every one of them was shorted out. I think the wires on each are pinched by the metal housings. No wonder they were on clearance.
You're better off collecting a bunch at a junkyard. You can usually get them for free. And you'll know most of them will be working. Oh, and also take the bulb sockets while you're there. -
You think 50HZ will be enough to show if the loop speed of the lab PSU is slow or decent enough though?
-
I'll see if I can figure something out.
http://gpete.blogs.keysight.com/2012/03/building-electronic-load-with-general.html
-
That looks similar to the 'Tenma' unit I bought from Element14 a couple of months back. Mine looks like two 'single' supplies, side by side. I don't think it has a USB interface. It works OK, and can deliver the rated load (60V at 5A or 30V at 10A) for a reasonable time. I should pull it apart for the giggles and look inside.
-
I found that five 100w "rough service" drop light bulbs in parallel draws 2.64A @ 62v indicated (PSU in series mode, 61.95v measured at the output terminals). I didn't have any luck using the IGBT as a switch (got the bulb output to oscillate slightly @ 1 hz) though.
No chance of overshoot with this load:
The ripple looks pretty bad in that shot, but that's just the screen resolution. Here's the ripple at the same load:
Those two bigger blips almost 200mV P-P about 8uS apart are regular and recurring.
-
Nevermind. Those blips must be EMF from something. They're everywhere, on both channels:
Even with everything but the scope shut off: