Author Topic: DIY SCPI programmable dual channel bench PSU 0-50V/3A (now EEZ H24005)  (Read 451962 times)

0 Members and 1 Guest are viewing this topic.

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #275 on: April 04, 2016, 09:32:57 pm »
Hi Prasimix,

this is your chance. http://www.ebay.de/itm/Cosel-Power-Supply-PBA-1000F-36-Netzteil-Stromversorgung-/111957579751?hash=item1a113117e7:g:d5kAAOSwpDdVGqyN

New device with much power ;)
:) Thanks for that. I'm not sure that they are willing to ship to Croatia, but I can ask.

Thanks for your answer. At the moment I try to find a circuit where I can track the pre regulated voltage from the cosel power supply depend on what is entered for the uset.
I was seen, that you track the voltage depend on vout. But I think maybe it is better to track voltage a uset, cause than the cosel supply must not go up and down with the voltage if the vout drops down for example when current limiter kicks in.
I will do that by using an op amp which compares uset and pre out and generates a voltage between 1-12 volts for regulate the power supply.

What do you think ?
Sven

Think again. If for example post-regulator output is shorted or you entering for any reason "deep" CC mode, then pass mosfet will be under heavy stress. Let's say that you set your output to 20 V, and that you wants 2 V of difference i.e. 22 V on pre-regulator output. If you are in CV mode and source 3 A than 6 W is dissipated on the pass mosfet. But if your load is changed that CC is activated and starts to limit output voltage to e.g. 5 V you'll have 51 W to dissipate! Not to mention short circuit when output voltage goes below 100 mV.

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #276 on: April 04, 2016, 09:38:56 pm »
 
The following users thanked this post: HK

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
LTC3864 pre-regulator PCB arrived
« Reply #277 on: April 13, 2016, 01:10:59 pm »
LTC3864 pre-regulator (post #225) PCB ordered from OSHpark arrived yesterday. That's a little bit more then 3 weeks without paying any extra to speed it up. Not bad. In coming days it remain to see if the PCB layout is well done what is apart from electrical design a very important detail for the SMPS regulator.


Offline Electronix

  • Contributor
  • Posts: 39
  • Country: de
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #278 on: April 16, 2016, 10:57:18 am »
Hi prasimix,

The pcb's are looking very good and I hope to hear the new progress soon.
What I'm interested in is, how is the accuracy with your design ?
Means, are you able to set a voltage exactly 3 digits after the decimal point ?
Which means 1mV ? I don't know exactly if that is possible with the choosen op amps.

Cheers

Sven
 

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #279 on: April 16, 2016, 11:12:28 am »
Didn't tested throughly yet, but from what I can remember goes easily below 10 mV and 10 mA.
3-digit precision has never been a requirement for this project. But as you can see a standard SO-8 packages are used that gives you more freedom to experiment with different opamps and also remote sensing is exists that is one of prerequisite for precise solution (I actually wonder how is possible that someone advertise something as 1 mV precision with few Amperes on the output without remote sensing inputs).

If you have some suggestion for opamps please post it here for possible further discussion.

Offline Electronix

  • Contributor
  • Posts: 39
  • Country: de
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #280 on: April 16, 2016, 05:20:26 pm »
Hi Prasimix,

of course 10mv resolution is for most cases enough. But sometimes, if you experiment with rf designs it will be useful for can adjusting the voltage in 1 mV steps.
I plan to use in most cases the analog stage design from liv, which is also in most cases used in  your post regulator. For the op amp I have selected a max44246.
Very low drift, very low offset and 5MHz GBP.
As Dac I will use an AD5686 and for adc the same as you choose an ads1120.

Maybe you can have a look at this op amp.

Cheers

Sven
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #281 on: April 16, 2016, 06:32:22 pm »
The max44246 is an interesting OP for precision tasks - may be an alternative to the LTC2057 in some cases. But I don't think it is a good choice for an voltage supply.

Here the only OP that is a little critical is the one to amplify the voltage at the shunt. Here something like an LT1037 / OP37 should be a good choice (needs to be reasonable fast do to high gain and not to add extra phase shift). The other OPs don't need to be that fast of low drift. So even the TL072 might be good enough, though slightly better version are available (e.g. TL052).
 

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Latest update ...
« Reply #282 on: May 02, 2016, 08:22:33 am »
A lot of things have happened since my last post so I'll mention some that could be interesting for people who follows this project.

First, Electricar decided not to wait anymore and ordered his batch of PCBs based on preliminary design of the latest revision published on GitHub. He sent me some pictures of them:

Power board (SMPS pre-regulator, post-regulator)



Aux PS (+5 V for Arduino shield, soft-start/stand-by, AC input protection, 12 V fan control)



Arduino shield (Due/+3.3 V only, W5500 Ethernet, external digital trigger, remote programming)



On my side I spent some time thinking again about mechanical aspects of the power supply. For the revised design a new enclosure is proposed without exposed huge heatsink (mounted outside enclosure) since expected power dissipation should be much lower. My intention was to use extruded sides of the Galaxy enclosure as an heatsink. But I comes to another idea inspired by ELV's DPS5135 power supply. It comes with so-called cooling aggregate:



That gives me an idea that two sides of cooling aggregate could be used for mounting power boards. Of course if sides are flat not rounded as in case of LK-40. I found that Fischer elektronik offers many cooling aggregate and that e.g. LA 6 could be perfect candidate for such task. Actually not so perfect since it can manage much more power that we need and it's price is way too high at least in quantity of one.
I've checked what could be workable substitute for such professional solution and found that with two very cheap heatsinks (RAD-A5723/100) and 60 mm fan (Sunon MB60251V1-000U-G99) that could be done for under 10€:



Following that concept and inspired by Ian Johnston info about Varisom, a Portuguese enclosure manufacturer, I contacted them and ask if they can offer a customize enclosure. Surprisingly I got in almost no time 3D model of asked enclosure. See below a couple of screenshots, and PDF with embedded 3D model can be found in attachment (it seems that Adobe reader is required, some other PDF viewers cannot display such data).





The new enclosure size is 280 (W) x 80 (H) x 240 (D) mm and front panel printing could be something like this:



Both power outputs are now moved on the right side and if we still wants to plug power boards directly to the Arduino shield, that requires modification of the Arduino shield PCB layout. If already proposed shield will be used then some 26-pin flat cable has to be deployed.

Finally, I spent some time assembling SMPS pre-regulator what is a topic for separate post. Progress on the firmware side is also significant that will be covered in the separate thread.



« Last Edit: May 02, 2016, 08:29:44 am by prasimix »
 

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
SMPS pre-regulator is assembled and tested..
« Reply #283 on: May 03, 2016, 01:05:29 pm »
The SMPS pre-regulator is assembled and tested. I had some trouble with stability when tracking is added but thanks to LTC's support and Dave's video #329 I got nice results. It looks like this:



This design will be used on the new power board presented in post #245. the corrected schematic is shown on the picture below.



For testing an existing power supply is used. I just remove old pre-regulator and plug a new one:



Measurement is performed on two points: switching mosfet output and post-regulator output. Standard 10x Rigol probe is connected via 1K resistor to the first point. For second an hand-made 1x AC probe is used (see post #181).


No load


Output shorted, Iout=3.5 A


Iout=1 A, Load=8R2


Iout=2 A, Load=8R2


Iout=3.5 A, Load=8R2

Output voltage ripple and noise figure is pretty good: ~1.5 mVrms and ~10 mVpp. I didn't tested yet this configuration with 100% duty cycle switch. Also sync with another channel and on-board LM5574 has to be done.

You can also find in attachment LTspice simulation.
 
The following users thanked this post: Aigor

Offline void_error

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: ro
  • I can transistor...
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #284 on: May 04, 2016, 04:28:12 pm »
Nice! Now I want my PCBs purple too :'(. OSH Park?
Isn't that big filter capacitor going to be a bit too close to a heatsink?
Trust me, I'm NOT an engineer.
 

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #285 on: May 05, 2016, 01:19:12 pm »
Hi void_error, capacitor is close to the heatsink but thanks to efficiency of such pre-regulator temperature of the heatsink shown on the picture is pathetically low in comparison with previously phase-controlled solution. Anyway this PCB was more proof of concept to proceed with "integrated" power board presented in previous post (still green not OSHpark's purple :)) on which both pre- and post-regulator are located.

I can continue with testing modified arduino shield soon (Elektricar told me that he sent PCBs to me!) and after that we are step away from organizing group buy for all people on this and some other forums that shows interest in this project.

Offline void_error

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: ro
  • I can transistor...
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #286 on: May 05, 2016, 01:45:35 pm »
Looked at the last schematic you posted and realized the diode bridge was replaced with schottky diodes, so the board can be mounted away from any heat source, the only things which might heat up would be the diodes.

This might have been previously asked but what is the reason behind choosing the LTC3864 and not the LTC3824?

Just out of curiosity, how many people are working on this project?
Trust me, I'm NOT an engineer.
 
The following users thanked this post: Aigor

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #287 on: May 09, 2016, 08:18:59 am »
Looked at the last schematic you posted and realized the diode bridge was replaced with schottky diodes, so the board can be mounted away from any heat source, the only things which might heat up would be the diodes.

Not just diodes from rectifying bridge but also mosfet switch and diode. Due to that I'm planning to put a "thermal bridge" between PCB and heatsink plate that will assist heat transfer. Also a MCU controlled fan will be deplyed.

This might have been previously asked but what is the reason behind choosing the LTC3864 and not the LTC3824?

Because it's better, more capable? See table in attachment.

Just out of curiosity, how many people are working on this project?

Still two, my colleague is fully dedicated to firmware and I'm trying to do the rest. Due to that everything lasts so much :).

Offline electricar

  • Regular Contributor
  • *
  • Posts: 89
  • Country: ch
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #288 on: May 09, 2016, 05:46:03 pm »
Hi folks,

I wanted to share some of my impressions, because I was so impatient and already ordered a batch of the preliminary PCBs.
Yesterday I finished with the AUX and Arduino PCB, although some parts are missing, the basic functionality is given.
First I have to say, that prasimix makes a great job in helping and fixing problems (if there are any)!
He also gave me some hints how to start testing the different sections of the PCBs.

Here are some points concerning the AUX PCB:
  • You have to trim the NC pins of the transformer like shown on the attached picture to fit the PCB.
  • I ordered most of my components from TME. Because the power resistor listed in the BOM “CRL7W-22R” was not available on TME I bought “CRL7W-15R”. Unfortunately this one is axial and the footprint on the PCB is radial, so it won’t fit. So find a radial one on TME (although I don’t think they have such a resistor) or buy it directly from Farnell (1687978).

Arduino PCB:
Nothing special, everything went fine, although I was very lucky to have a 6-pin IDC cable from an Atmel programmer to connect the AUX and Arduino PCB (forgot to order one).

After putting all together prasimix sent me a touchscreen demo sketch and luckily everything worked from the scratch! :D (see attached image)
Be aware, that my power resistor on the AUX PCB is quite dangerous, so I put some duck tape around it until I get the right one :scared:

The power board PCB is still missing in my setup and I hope that I can finish it by the end of the week and finally start some real testing!
Because I have the preliminary PCBs and the software is until now only for the “released” PCBs, there will be some issues e.g. the rotation of the touchscreen would be wrong. But for me that’s not a big deal. Did I get this right prasimix?

Until now I’m really happy with this project and impressed of the support by prasimix and can’t wait to go on further testing and finally have a great power supply on my bench!  :clap:
 
The following users thanked this post: prasimix, Aigor

Offline void_error

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: ro
  • I can transistor...
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #289 on: May 09, 2016, 05:55:18 pm »
This might have been previously asked but what is the reason behind choosing the LTC3864 and not the LTC3824?

Because it's better, more capable? See table in attachment.
Double the gate driver current and an almost twice as accurate internal reference as well as the wider range of switching frequencies seem to be the things that make the LTC3864 a better choice by quite a margin.

Just out of curiosity, how many people are working on this project?

Still two, my colleague is fully dedicated to firmware and I'm trying to do the rest. Due to that everything lasts so much :).

It's always nice to have more than one person working on something like this. However it doesn't seem to take so long, been watching this evolve over many iterations since the beginning and it seems to go quite fast. At least faster than I could do it but then again my day job eats up most of my time and my mountain bike almost half of my spare time :(

I'm curious about how the touch screen interface is going to look like when it's finished. I'll be stuck with a Nokia 3310 - like UI for my PSU (and a few more things that share the UI module) and it's going to take almost forever to finish anything as I'm writing the firmware myself :scared:.
Trust me, I'm NOT an engineer.
 

Offline Aigor

  • Contributor
  • Posts: 29
  • Country: it
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #290 on: May 09, 2016, 07:31:23 pm »

The power board PCB is still missing in my setup and I hope that I can finish it by the end of the week and finally start some real testing!
Because I have the preliminary PCBs and the software is until now only for the “released” PCBs, there will be some issues e.g. the rotation of the touchscreen would be wrong. But for me that’s not a big deal. Did I get this right prasimix?

Until now I’m really happy with this project and impressed of the support by prasimix and can’t wait to go on further testing and finally have a great power supply on my bench!  :clap:
Hi,  and thanks for the hints, do you mean that actual software from github doesn't work on a preliminary rig?
I'm interested on building this amazing PSU but i'm afraid i have to wait Parsimix :D
 

Offline electricar

  • Regular Contributor
  • *
  • Posts: 89
  • Country: ch
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #291 on: May 09, 2016, 08:17:45 pm »
Quote
Hi,  and thanks for the hints, do you mean that actual software from github doesn't work on a preliminary rig?
I'm interested on building this amazing PSU but i'm afraid i have to wait Parsimix :D

Hi aigor,

no, I meant that e.g. the display rotation is different because it changed between the revisions.
Because I don't finished the power board yet, I cannot say much more about the software differences.
I think only prasimix could clear things up :) 
 

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #292 on: May 09, 2016, 09:05:58 pm »
I'm curious about how the touch screen interface is going to look like when it's finished.

I made some intro video recently about using touch screen for programming output values. Your input is welcome.

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #293 on: May 09, 2016, 09:25:26 pm »
The power board PCB is still missing in my setup and I hope that I can finish it by the end of the week and finally start some real testing!
Because I have the preliminary PCBs and the software is until now only for the “released” PCBs, there will be some issues e.g. the rotation of the touchscreen would be wrong. But for me that’s not a big deal. Did I get this right prasimix?

Thanks Electricar for sharing this with us. You're right it is not a big deal. That maybe sounds suspiciously enthusiastic since contents of all screen has to be adopted to landscape format. But, idea is to provide screen editor that anybody can modify existing screens or even introduce their own. Screen editor works with various screen elements, can be used to define touch screen interaction, has font editor and can be used for image manipulations (e.g. logo, icons, etc.). Finaly it offer embeding SCPI commands for interaction with screens and build a complete code ready for uploading and execution on Arduino board.
The first step will be to rotate and adjust existing portrait screens to landscape and continue with support for both version. And yes, this is not wishful thinking but something that will be ready very soon for deployment. Such approach can slow progress at the beginning but will starts to "pay-off" very quickly. The same situation we had with firmware software simulator. It already spares us hundreds of development hours since you can instantly compile, execute and debug code without using Arduino board.

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #294 on: May 09, 2016, 09:38:43 pm »
Hi,  and thanks for the hints, do you mean that actual software from github doesn't work on a preliminary rig?
I'm interested on building this amazing PSU but i'm afraid i have to wait Parsimix :D

Hi Aigor, depends how fast are you with building everything. Support for landscape screen format is not ready now, but I can bet that we'll do our job before you get your PCB :). Please note that even in this moment you can count with nice feature set accomplished using SCPI commands and you can even test that without having physical device. If you'd like to explore what is already done use software simulator in version M1. If you wants more feel free to get latest public build of M2 firmware where you can even start to test local screen functionality.

Offline void_error

  • Frequent Contributor
  • **
  • Posts: 673
  • Country: ro
  • I can transistor...
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #295 on: May 09, 2016, 09:55:53 pm »
I'm curious about how the touch screen interface is going to look like when it's finished.

I made some intro video recently about using touch screen for programming output values. Your input is welcome.

I remember watching that...
I have a few suggestions though, nothing too fancy or complicated. I'll make a list of things I think would make the UI more intuitive some time this week, need to watch the video a few more times carefully, and make a post in the thread where that video is. In my opinion you shouldn't need a manual for using the basic features of a bench PSU.
Trust me, I'm NOT an engineer.
 
The following users thanked this post: Aigor

Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #296 on: May 09, 2016, 09:59:39 pm »
I remember watching that...
I have a few suggestions though, nothing too fancy or complicated. I'll make a list of things I think would make the UI more intuitive some time this week, need to watch the video a few more times carefully, and make a post in the thread where that video is.

Excellent, looking forward to see that.

In my opinion you shouldn't need a manual for using the basic features of a bench PSU.

You're right. There is many rooms for improvements. Hopefully we can catch with all suggestions easily when "screen editor" will be ready.


Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
New power board assembled ...
« Reply #297 on: May 14, 2016, 10:14:07 pm »
Electricar who decided to make PCBs using preliminary new revision layouts sent one PCB set to me. I spent some time yesterday and today to assemble power board which include SMPS pre-regulator. I didn't find any surprise yet. I borrowed one channel from existing PSU for digital control. Output enable, voltage and current programming works fine. Only correction on the PCB was one capacitor in pre-regulator's tracking circuit. I used one THT and solder it over SMD component.
I didn't test it yet with real load. For that I need to mount heatisink for post-regulator's pass element and also some kind of thermal bridge to improve cooling of pre-regulator's mosfet and diode.


Offline prasimixTopic starter

  • Supporter
  • ****
  • Posts: 2022
  • Country: hr
    • EEZ
Power board testing...
« Reply #298 on: May 16, 2016, 02:14:57 pm »
I did some new testing trying to load it with up to 5 A. That requires some cooling and I temporately mount post-regulator's power mosfet on external heatsink and fan for blowing air over PCB for cooling pre-regulator's Q1, D5 and L1, L2 and L3.





For testing such load I had to made few changes. Referring to schematic from the post #285 R5 is increased from 36K to 47K that increase switching frequency to over 300 kHz. That is required if 33uH power inductor L1 is used. Higher output current require lower sense resistor therefore R2 is decreased from R015 (that is enough for up to 4.5 A) to R012 or could be R010.

Switching frequency synchronization between LM5574 and LTC3864 is now establised. LM5574 is assign as master and its frequency is set about 330 kHz (Rt on pin 7 is decreased to 18K, and Cramp is now 3n9). in operation it looks like this (magenta trace is LTC3864, cyan is LM5574).



Two new functions: 100% duty cycle operation and remote programming has to be tested but that requires additional code in firmware.

Offline Electronix

  • Contributor
  • Posts: 39
  • Country: de
Re: DIY SCPI programmable dual channel bench PSU 0-50V/3A
« Reply #299 on: May 21, 2016, 04:17:23 am »
Hi Prasimix,

I'm also build now the first prototype based most on liv's analog stage.
I have problems with the prototype to get down to 0 volts.
I used fixed bias voltage at 38volts and -15v for dp.
I would like to ask you, if you can go down to 0 volts ?

Cheers

Sven
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf