Author Topic: Inside the new Korad KA3305P linear PSU  (Read 58992 times)

0 Members and 1 Guest are viewing this topic.

Offline Pinkus

  • Frequent Contributor
  • **
  • Posts: 596
Re: Inside the new Korad KA3305P linear PSU
« Reply #50 on: September 20, 2015, 01:40:27 am »
you may ask them to send you a new main board with a new software release installed instead. This would be the least expensive way for everybody.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #51 on: September 20, 2015, 04:02:01 am »
Just paste the youtube link.

I don't do youtube anymore. I closed down my account when Google tried to force me into signing up for a bunch of stuff I didn't want. The video is hosted by me. But thanks for the thought.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #52 on: September 20, 2015, 04:03:34 am »
you may ask them to send you a new main board with a new software release installed instead. This would be the least expensive way for everybody.

I'll ask.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #53 on: September 21, 2015, 04:04:07 am »
I asked, and was surprised to receive a reply on a Sunday. Sounds like they are willing to do a board swap. The guy who is handling this is on vacation this week, but they will have him contact me to hammer out the details when he gets back next week.

I have to say that Korad (and their distributor) seem to be standing behind the product.

I also asked about the maximum output rating. Their response was
"It depends on the fan and heatsink inside the unit. As long as that fan keeps going you could theoretically run all three channels at max indefinitely, but the fan is the weak point. "

This would seem to confirm Klienstein's earlier comment:

The real drawbacks of these cheap chinese PSUs is the thermal design: they are made to deliver much power at low cost. Here less safty marging saves costs and thus things get hot right to the limit.
 

Offline mos6502

  • Frequent Contributor
  • **
  • Posts: 537
  • Country: aq
Re: Inside the new Korad KA3305P linear PSU
« Reply #54 on: September 21, 2015, 04:53:38 am »
I also asked about the maximum output rating. Their response was
"It depends on the fan and heatsink inside the unit. As long as that fan keeps going you could theoretically run all three channels at max indefinitely, but the fan is the weak point. "

Hmm. Usually the weak point is the transformer. Because copper is expensive. So if manufacturers want to skimp, they just use a smaller transformer and rely on airflow to keep it cool.

Another trick is to use a very small filter cap and higher transformer output voltage and rely on the pass element to dissipate the extra heat. The small filter cap keeps the RMS current low and allows the transfomer to deliver more power. This is also why you should never "upgrade" the filter caps in a linear supply with larger ones.
for(;;);
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #55 on: September 21, 2015, 05:25:27 am »
Another trick is to use a very small filter cap and higher transformer output voltage and rely on the pass element to dissipate the extra heat.

The filter caps are 6800µF each. If I did the math right, that should work out to 3% ripple with a load of 10A at 30v. From what I've read, that's a reasonable figure for a linear supply. Or would you consider that "small" for the application?
 

Offline mos6502

  • Frequent Contributor
  • **
  • Posts: 537
  • Country: aq
Re: Inside the new Korad KA3305P linear PSU
« Reply #56 on: September 21, 2015, 06:06:06 am »
Another trick is to use a very small filter cap and higher transformer output voltage and rely on the pass element to dissipate the extra heat.

The filter caps are 6800µF each. If I did the math right, that should work out to 3% ripple with a load of 10A at 30v. From what I've read, that's a reasonable figure for a linear supply. Or would you consider that "small" for the application?

6800uF for 5A is fine. I was merely explaining what some Wun Hung Lo manufacturers do. The Korads actually seem to be well designed.  :-+
for(;;);
 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 7695
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Inside the new Korad KA3305P linear PSU
« Reply #57 on: September 21, 2015, 01:42:41 pm »
I remember seeing this new model on pre-order for a while. Thanks for the teardown photos. What all changed from the prior triple-channel model?
I TEA.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #58 on: September 21, 2015, 08:06:00 pm »
I remember seeing this new model on pre-order for a while. Thanks for the teardown photos. What all changed from the prior triple-channel model?

Not a lot, as far as ripple/regulation specifications go. The 3-channel 3amp predecessor was called KA3003D-3S. It looked just like they took two single-channel supplies and stuck them in a double-wide case. They even retained separate on/off buttons and separate keyboard lock buttons for channel 1 and channel 2. There was no obvious serial/parallel operation, (although one review I read suggested it was possible, so it may have been added mid-production). It had separate fans and reviews I read mentioned they are running all the time and a bit on the noisy side.

Supposedly there was a KA3005P-3S version providing 5 amps and programming capability, however that may have been vapourware as I never did find one actually for sale anywhere.

This one seems better integrated (to me) and perhaps more importantly, available.
« Last Edit: September 21, 2015, 08:08:29 pm by torch »
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5963
  • Country: de
Re: Inside the new Korad KA3305P linear PSU
« Reply #59 on: September 22, 2015, 05:49:32 am »
As transistors a cheap today, the real weak spot is often the transformer - but sometimes even a few cents saved on the output transistors are still tempting.

Having a rather small filter cap helps to reduce the current peaks, but you than also need more reserve in voltage. So you can not gain that much by reducing the cap size, unless you risk that ripple might be apear at the output at high voltage and high current. This might only happen if the caps are at the lower end (tolerances are quite large) or mains voltage is less than nominal.
Using a relay to switch transformer taps also calls for the cap not to be very large, as this gives extra current peaks when swithing to a higher voltage.

Transformers can stand more than nominal load for quite some time. They get hotter however and thus age faster. Having the fan might also help to ensure cooling of the transformer.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #60 on: September 22, 2015, 06:29:35 am »
I wonder what the effect would be to stick a nice big aluminium CPU heatsink right on the transformer core? Maybe even with it's own fan? Would it help keep the transformer cooler? Would it somehow interfere with the magnetic flux of the core?
 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 7695
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Inside the new Korad KA3305P linear PSU
« Reply #61 on: September 22, 2015, 05:46:51 pm »
It might help, just depends how quickly the heat can be dissipated. For a large transformer, the surface may not overheat, but the interior still might.
I TEA.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 5963
  • Country: de
Re: Inside the new Korad KA3305P linear PSU
« Reply #62 on: September 23, 2015, 01:47:24 am »
The heat in the transformer comes mainly from the windings. So cooling the core will help a littel, but not that much.
At least it will not interferre with the normal magentic flux. However still some care may be needed, not to cause eddy currents because of eletrical contact to more than one sheet of the core.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #63 on: September 29, 2015, 01:31:40 am »
Update: it's first thing Monday morning and the guy who is handling this must be back from his vacation because he sent me a message that they are packaging up a replacement board with instructions on swapping it out and it will be in the mail in the next day or two.

 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 7695
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Inside the new Korad KA3305P linear PSU
« Reply #64 on: September 29, 2015, 02:51:00 pm »
Although issues have come up with Korad power supplies, the manufacturer is quite fast with fixes. Korad and SRA are pretty responsive and flexible with rectification too, which is good to see. If I didn't already have plenty of vintage power supplies, I'd probably get one of these triple-channel ones.
I TEA.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #65 on: October 15, 2015, 04:22:02 pm »
The replacement board arrived in the mail today (but postmarked October 6).

The new one is on the right, beside the original. They look identical, right down to the funny twist to the crystals.



The M4 bug is fixed, but I found another, minor bug. I'm not sure if it was always there and I just missed it, or if it was introduced with this latest firmware. It happens in Serial mode only. When Serial is selected, the current limit is displayed as 0.000 (both channels). When setting the current limit (ie: press CH2 twice, so a digit flashes) the current limit is displayed temporarily. When setting either OVP or OCP (ie: press and hold OCP for 2 seconds) the OCP set point is displayed. When the output is turned on, the actual current is displayed.

The current setpoint is properly displayed when the output is off in both Parallel mode and Independent mode.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #66 on: October 20, 2015, 12:58:53 pm »
I still haven't figured out the serial port connection but either it's dead or maybe requires a null modem cable (manual shows straight pass-through).

Ok, I was wrong and the serial port works with a pass-through cable as described in the manual.

However, I still can't get a terminal program like hyperterm or MTTTY to work, as they don't send the command until "Enter" is pushed and pushing "Enter" adds a CR character. The Korad ignores any command with an appended CR. The only way I can make this work is through the Send tab of Realterm, which sends the command without a CR or LF when you hit the Send ASCII button. 

The reason I could not get the serial port to respond before was a classic PEBKAC: I set the port speed to 9600 in the port tab in Realterm, but did not realize it doesn't take effect until you hit "Change" -- despite the big green checkmark on the Change button. Doh!

I did finally successfully communicate through the USB cable to the virtual Serial-USB adapter installed in the PSU itself.

It's too bad the supplied software is so buggy. I can see where the programming and logging features could be real handy. What I need now is a Mechatrommer: someone with the skills to create a decent software interface!
 

Offline Messtechniker

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: de
  • Old analog audio hand - No voodoo.
Re: Inside the new Korad KA3305P linear PSU
« Reply #67 on: October 20, 2015, 04:34:42 pm »
It's too bad the supplied software is so buggy. I can see where the programming and logging features could be real handy. What I need now is a Mechatrommer: someone with the skills to create a decent software interface!

Why not roll your own with this:
http://www.abacom-online.de/uk/html/profilab-expert.html.

For an example I programmed, see here:
https://www.eevblog.com/forum/projects/building-your-own-voltage-reference-the-jvr/?action=dlattach;attach=177041;image
Agilent 34465A, Hameg HMO1022, R&S HMC 8043, Voltcraft VC 940 and M-Audio Audiophile 192
 

Offline Gall

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: ru
Re: Inside the new Korad KA3305P linear PSU
« Reply #68 on: October 21, 2015, 12:15:21 am »
Ok, regarding the bugs, it is possible to replace the MCUs with something else with our own program? I bet these funny chips with "Korad" marking are nothing else as ATmega or MCS-51, so it shouldn't be hard to find a pin-compatible replacement.
The difficult we do today; the impossible takes a little longer.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #69 on: October 21, 2015, 12:42:12 am »
Why not roll your own with this:
http://www.abacom-online.de/uk/html/profilab-expert.html.

That looks pretty nice. Unfortunately Korad is not one of the supported devices listed on their website. Reading between the lines, the software uses some sort of driver for each device. Does it include the capability of writing a custom driver for an unknown device? Can it send data alone to the serial port without a CR or LF?

Also unfortunately: that software costs about as much as the Korad. Mind you, it appears to have Spice-like capabilities as well?

Ok, regarding the bugs, it is possible to replace the MCUs with something else with our own program? I bet these funny chips with "Korad" marking are nothing else as ATmega or MCS-51, so it shouldn't be hard to find a pin-compatible replacement.

Two are marked Korad1 and one is marked Korad4. Possibly identical HW with different firmwares. They seem to be associated with the individual channels by location and there are also a j-tag looking header near each. Beyond that, I couldn't say what the heck they might be. I think Korad is trying to protect itself from competitors reverse-engineering and copying the product. Quite possibly they programmed in something to make ID more difficult too.

The one remaining bug in the firmware (showing current limit of 0 in Serial mode) is livable. The remaining bugs are in the PC software for running a script, not in the PSU itself. So for me personally, it's not worth the effort of trying to ID and reprogram the chips.
 

Offline Gall

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: ru
Re: Inside the new Korad KA3305P linear PSU
« Reply #70 on: October 21, 2015, 05:01:07 am »
I'd like to reprogram just to make the interface more usable. I personally don't like the way the voltage is set by individual digits.

Also, is it true that the display shows the preset voltage/current, not the actual one?

I don't want to reflash the firmware, I want to replace the whole MCU chip. Anything that is pin-compatible would fit, and there are not so many possibilities. Could you please identify the power, XTAL and SPI (to the DAC) pins?

I'm I right that the controller board is not the same as display board? That is, could one replace the whole board leaving the original display and controls in place?
The difficult we do today; the impossible takes a little longer.
 

Offline Messtechniker

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: de
  • Old analog audio hand - No voodoo.
Re: Inside the new Korad KA3305P linear PSU
« Reply #71 on: October 21, 2015, 05:39:22 am »
Does it include the capability of writing a custom driver for an unknown device? Can it send data alone to the serial port without a CR or LF?

As to custom drivers one would have to write a dll. Something which is beyond my capabilities.
But as long as the target unit is responsive through its serial interface then you have many options with "Profilab Expert".
Provided the protocol is known resp. given in the manual. As to doing without CR and LF: I would simply have to find out.

Yours Messtechniker
Agilent 34465A, Hameg HMO1022, R&S HMC 8043, Voltcraft VC 940 and M-Audio Audiophile 192
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #72 on: October 21, 2015, 06:19:53 am »
I'd like to reprogram just to make the interface more usable. I personally don't like the way the voltage is set by individual digits.

I'm not sure you understand that part correctly. The granularity is variable by digit, however, it increments/decrements the next leftmost digit after 9. So you can twirl the knob from 0v to 31 volts by single volts, tenths of a volt or hundredths of a volt depending on what you need to do. Or you can go from 0 to 30 volts in 3 steps if you set the tens digit to flashing.

That said, it won't let you go past 31 volts, so if you started at 1.01v and adjusted by tens of volts, the maximum would be 21.01v. You would have to use a different granularity to go higher. If you then switched to single volts, you could go up to 30.01 volts. From there you could switch to tenths and go up to 30.91v.

I can do a little video if the above is still confusing.

For what I'm doing, I don't generally need to adjust hundredths of a volt. I just leave the hundredths digit at 0 and set the granularity at tenths (the granularity remains set unless changed). Changing the voltage is a matter of tapping the channel button once and twirling the volts dial. Similarly, tap the channel button twice and twirl the current dial.

Quote
Also, is it true that the display shows the preset voltage/current, not the actual one?
Only true until the output is turned On. Once the output is turned on, the display changes from showing the presets to showing the actual output (EXCEPT in serial mode, when the display shows 00.00 when outputs are Off and not being actively adjusted -- that's the minor bug I mentioned earlier.)

Quote
I don't want to reflash the firmware, I want to replace the whole MCU chip. Anything that is pin-compatible would fit, and there are not so many possibilities. Could you please identify the power, XTAL and SPI (to the DAC) pins?

Speak slowly and use small words! I will attempt to follow directions, but I'm an old dog trying to learn new tricks. My last quarter century was spent riding a firetruck, not sitting at an EE bench: "Me strong like bull. And smart -- Like freight train."

Quote
I'm I right that the controller board is not the same as display board? That is, could one replace the whole board leaving the original display and controls in place?

Yes, that is correct. There is a 20(?) pin header between the boards as well as a ribbon cable. The outputs are on the control board and the display/knob board sits above those in front of the control board.
[/quote]
 

Offline Gall

  • Frequent Contributor
  • **
  • Posts: 311
  • Country: ru
Re: Inside the new Korad KA3305P linear PSU
« Reply #73 on: October 21, 2015, 07:12:34 am »
I'm not sure you understand that part correctly. The granularity is variable by digit, however, it increments/decrements the next leftmost digit after 9. So you can twirl the knob from 0v to 31 volts by single volts, tenths of a volt or hundredths of a volt depending on what you need to do. Or you can go from 0 to 30 volts in 3 steps if you set the tens digit to flashing.
Yes, that's exactly how I understood that. I think the smaller KD3005P has exactly the same control. I'd like to change the granularity by rotating the knob while pressing it.

And I'd like to make it blink faster.

Quote
Quote
I'm I right that the controller board is not the same as display board? That is, could one replace the whole board leaving the original display and controls in place?

Yes, that is correct. There is a 20(?) pin header between the boards as well as a ribbon cable. The outputs are on the control board and the display/knob board sits above those in front of the control board.
Fine! Thank you! In the worst case I'll just replace the whole board.

I wonder if this ribbon cable should have been shielded...

The difficult we do today; the impossible takes a little longer.
 

Offline torch

  • Frequent Contributor
  • **
  • Posts: 339
Re: Inside the new Korad KA3305P linear PSU
« Reply #74 on: October 21, 2015, 10:43:32 am »
I'd like to change the granularity by rotating the knob while pressing it.

Well you do press and release the knob to change the granularity. Each press moves the granularity one digit to the right. However, if you press and hold the knob then you lock/unlock the keyboard (volts) or turn the beep on/off (amps).

Quote
I'm I right that the controller board is not the same as display board? That is, could one replace the whole board leaving the original display and controls in place?

Quote
I wonder if this ribbon cable should have been shielded...

Why? I'm not saying you are wrong, I just don't understand what makes you suspect that. I don't think there is a lot of HF radiation from either board, it's a linear supply and the regulation seems to be at the other end of the PSU.
« Last Edit: December 08, 2015, 10:13:39 am by torch »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf