DIY SCPI programmable dual channel bench PSU 0-50V/3A (now EEZ H24005)

[HAL-42b]

Are the counterbores for the banana posts necessary for some mechanical constraint?

They are likely to up the costs and are a bit finicky on thin material. It would be nice if a prospective builder could skip them in order to save cost.

Of course it won't be possible if screen thickness or something like that makes the counterbore necessary.

[prasimix]

Are the counterbores for the banana posts necessary for some mechanical constraint?

They are likely to up the costs and are a bit finicky on thin material. It would be nice if a prospective builder could skip them in order to save cost.

Of course it won't be possible if screen thickness or something like that makes the counterbore necessary.

No, the counterbore are not necessary for biding post that I'm going to use. I just added it for better appearance.

[void_error]

Nice front panel there! Although I'm not a fan of touchscreens for test equipment, it does indeed look really clean.

[prasimix]

Nice front panel there! Although I'm not a fan of touchscreens for test equipment, it does indeed look really clean.

Thanks void_error, do you have some experience with touchscreen on the test equipment that you can share with us? What kind of obstacles did you find?

[kolesar007]

Hello, for my opinion just one rotary encoder with push button would be usefull for set voltage and current, so you don't need to type on the screen for every little change.

[prasimix]

We are currently working with touchscreen that can be used for output values setting in three (actually 5) different mode:

• Slider (interactive and non-interactive)
• Step (interactive and non-interactive)
• Numeric keypad (non-interactive only)

The first one emulate linear potentiometer that are popular on analog mixer devices. But we are going beyond that because touchscreen is 2D device and we are using both X and Z axis for setting desired value: X defines precision of the movement, Y changes the value. Thanks to that you can quickly change e.g. output value from 10 to 30 V and then fine tunned it to let's say 30.25 V
Second one is similar to the first one but here we have predefined steps (e.g. 100 mV, 1 V, 2 V or 5 V).
The third one is "boring" keypad that can be find on so many digitally programmable instruments: you can simply enter and confirm new value.

The difference between "interactive" and "non-interactive" mode is that in later you need to confirm new value, when in former the value will be set immediately (will follow movement of your finger).

Now please let me know if we didn't cover some scenario that is possible with multi-turn pot, keypad or encoder.

[kolesar007]

I think that rotary knob is most ininuitive and user frandly aproach. In my project I use push to select step and rotary for value.
Few month ago I try to make velocity control so that step is interactively define with rotary speed and I made it. Then I reprogram all my other projects to velocity control .

I like that you think about that . Interesting modes and it is worth to test this new aproach.

[prasimix]

So far I can say that touchscreen is an valid substitution (and even "addictive" to some extent, but my opinion should not count since I'm too much bonded to this project). Hopefully some video that I'm planning to make when everything is ready will help other to get an idea about it.

[prasimix]

Thanks to few people from the North America region with 115VAC mains that decided to join me in the next "experimental PCB group buy" I spent some time to check which PCB transformer could be suitable for the AUX PS PCB. It was much harder then I initially expected but at the end I found Triad Magnetics' VPP28-180 (Newark: 37B9209, Digikey: 237-1086-ND, Mouser: 553-VPP28-180). It comes with pinout that can coexist with  230VAC transformer using the same real-estate.
Since the AUX PS PCB could also carry few AC input protection components my logic tells me that all selected values for 230VAC should be simply divided by two. In that case we should have the following changes:

SAR1: 2027-23-BLF (230VAC) instead of 2027-47-BLF (470VAC)
VAR1, VAR2: S20K140 (140VAC) instead of S20K275 (275VAC) and
ZD1, ZD2: 1.5KE200CA (200VAC) instead of 1.5KE400CA (400VAC)

Please let me know if I miscalculated something.

[void_error]

Thanks void_error, do you have some experience with touchscreen on the test equipment that you can share with us? What kind of obstacles did you find?

My only reason to not use a touchscreen is that it can get smudgy (full of fingerprints - looks at tablet) therefore it would require more maintenance (cleaning), other than that I guess it's fine as long as the software for it is not poorly written (same goes for regular mechanical buttons/knobs). Other than that, I haven't used any test equipment with a touchscreen but I did 'play' with something that has both buttons, a joystick and a touchscreen and for that particular piece of equipment I can say the interface was horrible (although it was running embedded linux and the response was instant).

I think that rotary knob is most ininuitive and user frandly aproach. In my project I use push to select step and rotary for value.
Few month ago I try to make velocity control so that step is interactively define with rotary speed and I made it. Then I reprogram all my other projects to velocity control .

I like that you think about that . Interesting modes and it is worth to test this new aproach.

Velocity control is a nice feature in some cases. I might add that to the user interface I'm going to build for my DIY test equipment projects, although I'm more inclined towards manually selecting the increment value using another button.

I did look into using some big LCD with a touchscreen for the UI but the amount of programming (almost everything from scratch) doesn't make it worthwhile given the little time I have to work on it. I might however try to use one of those 'fancy' color LCDs if I don't have to use a really powerful/expensive PIC micro to drive it just because they're quite cheap and big... but I'm still almost clueless about where to start.

[prasimix]

My only reason to not use a touchscreen is that it can get smudgy (full of fingerprints - looks at tablet) therefore it would require more maintenance (cleaning), other than that I guess it's fine as long as the software for it is not poorly written (same goes for regular mechanical buttons/knobs). Other than that, I haven't used any test equipment with a touchscreen but I did 'play' with something that has both buttons, a joystick and a touchscreen and for that particular piece of equipment I can say the interface was horrible (although it was running embedded linux and the response was instant).

Sounds fair, so it's on my side to present a decent peace of software for controlling touchscreen . We already spent lots of time just to realize that initially selected touch library is useless for level of precision and usability that we wants to achieve. That ends up with rewriting a whole stuff based on more mature approach the you can find on Daniel Beer's pages (since simple averaging of multiple ADC readings is far from usable, it's more like a bad joke).

[kolesar007]

I also doubt the precision of this kind of settings but lets see in realitiy.
And what void_error is noticed is fingerprints. I do not bother few fingerprints on the side of display for some soft touch buttons but don't like fingerprints over the entire surface.

[prasimix]

Precision goes down to the smallest defined step or 10 mV (mA) in our case. But, seeing is believing (actually it's opposite but never mind) so I'll present that soon.
Regarding fingerprints I didn't notice that greasy or clean surface makes any difference . If something could make a real mess it's a "messy" software when pressing e.g. lightly and/or sloooowly on one side of the screen could be wrongly interpreted resulting in action on the opposite side.

[kolesar007]

I know what step do you offer by SW.
I mean how precisely and what is the feeling when you slide with finger to set the value ..and is it strict or jumping up and down when you slide .. and it is possible that jump lets say 1V when you lift finger from surface?

[prasimix]

Lifting finger from the surface should change new set value for exactly ZERO! Anything more/or less then that should not be tolerated.

[Electronix]

Hello Prasimix,

at first, I'm very impressed of all your efforts you put in that project and can't wait to see the finished psu.
I'm also planning a dual power supply with similar features you have at 0-30 volts and 0-10A.
I'm often test circuits with vacuum tubes so i need a little bit more current  .
My plan and I hope it will work is to use directly a switching supply from Cosel.  It is a pba 300 with 36 volts and current of 9 ampere but it could handle 10 ampere  .
This I will also pre regulate via a circuit so that the pre voltage which is feed into the post regulator is 4-5 volts higher than the output.
We will see, if that is working. Advantage of that is, that I not to have a big transformer with bridge rectifier etc.  so that will save money and place in the enclosure.
I saw in one of the post, that you are still search for how to save io's on the arduino.
And would give you the hint maybe to use a different display.
My plan is to use two 4" inch displays with touchscreen from nextion. These displays are controlled via a serial interface. The screen pictures keyboards etc. are all made with an editor. Also a scope function is existing, cause I will use that to show the traces current, voltage and power at a time line like a scope on that display. Maybe it is interesting for you too.
Here is the link :http://wiki.iteadstudio.com/Nextion_HMI_Solution

Hopefully we will hear from you soon, with some progress on that project 

Cheers

Sven

[prasimix]

Hi Electronix, and welcome.

Auch, Cosel PBA300-36 cost a fortune (more then €1 per watt) and I'm not sure is it possible to push it beyond specified output voltage range (25.2 to 39.6 V) but idea is valid and you succeed there is no need main transformer and pre-regulator anymore.

I'm aware of Nextion but one thing is attractive demo where everyone are happy and everything is simple and another one is trying to build a real stuff with it. It's price and capability together with serial interface looks appealing and possibly in a future I'll get one and check how it works.
In the meantime we'll continue development on dirty cheap and dumb 3.2" TFT which if properly managed could be really usable for controlling at least 2 channel so no need for dual display what you can see in video that I posted today here.

[Electronix]

Thanks 
Yes you're right normally they cost a lot of money but you can get them sometimes cheap on eBay.
So my two devices arrived today and a payed 15€ per device, not bad I think .
I have the 36 v with 9 ampere version .
And they have already an remote input where you can control the device from 0 -36 volts with a control voltage of up to 12 volt so very easy in use. The only thing which annoying me a little bit are the two smal fans which running in highspeed mode when the output voltage is over 2,5 volts. But I think that will work, I will spend a temperature control for the fans which should solve the problem than .
Regarding the display, I will buy one and will make some tests. I programmed already a test gui with the designer and it looks attractive. Of course I have to check how that will work in real operation but I expect that I can cover all the things I need for the power supply. It is also possible to enter custom code in the gui designer for some things if needed.

In the meanwhile I'm simulate your post regulator and try to get work him stable at max 10 ampere.
I replaced the tl and also the op27 against a max 44246 which have really good specs and he is not so expensive.
I think I will use two mos in parallel to handle that amount of current.
Let's see. The first simulations are not bad. Current control is working with 1ma resulation also voltage control works with 1mv resolution in simulation.
How far you are with the progress of your supply ?
Is there also a video available?
I'm very interested in that.


Cheers

Sven

[prasimix]

That's interesting. I'd like to see a link to 15€ ebay version since genuine one is over 300€/$ (check here).

Please note that if you are going to use post-regulator in combination with pre-regulation module that you are mentioned you'll need additional bias power supply for +/-5V (op amps), -9 V (DP) and output mosfet bias that is at least 5 V (better 10 V) higher then output voltage. If you are using e.g. +/-9 V for opamps then no needs for additional -9 V for down-programmer. Output mosfet bias could be floating or fixed above max. input voltage (e.g. 40 V and something in case of 2-32 V input).

I finished two power supplies but thinking about another more compact version. Software part is still in development. In the first milestone (M1) remote control using SCPI commands are accomplished, and now we are working on local control (I posted link to first video in my previous answer to you).

[Electronix]

Hi Prasimix,

I have had a look but the seller has only this types of devices :

http://www.ebay.de/itm/COSEL-PBA150F-24-mit-100-240V-Eingang-Netzteil-24V-6-50A-Spannungsversorgung-/111925331364?hash=item1a0f4505a4:g:dhgAAOSwmmxW2HTy

I think I have had many luck with that. But when you keep your eyes open you will also have luck I think.

Cheers

Sven

[prasimix]

OK, half the power and used/gebraucht makes more sense for 15€ price tag. Anyway, thanks for the link.

[prasimix]

I decided to postpone manufacturing of the latest set of PCBs (revision 5) since it include many changes of which a new SMPS based pre-regulator is the most significant one. Therefore I ordered this time from OSH park only the SMPS pre-regulator PCB that could be connected with existing post-regulator to see how it will work.

The GitHub hardware section of this project is recently updated with various changes. The Arduino shield (digital control) board is now shorter for 25 mm because a new type of connectors are used for remote sensing and load temperature sensors. Also an external digital trigger input is added that will offer more flexibility in controlling the PSU from the outside.
A new SMPS pre-regulator based on already presented LTC3864 controller is also now available (BOM is still missing). Since it allows 100% duty cycle mode of operation, small correction is added on existing post-regulator board that contains digital control of the channel power section.
Finally a complete new revision 5 is added in PRELIMINARY folder where you can find the following designs:

• Redesigned Arduino shield that can be now installed on 2U front panel since the TFT display is rotated from portrait to landscape
• New power board that include previously used post-regulator and new SMPS pre-regulator. Connection between this board and new Arduino shield is accomplished with the single 26-pin 0.1” connector.
• Redesigned AUX PS board has fan control, USB and Ethernet socket and
• Mechanical drawing for new enclosure such as Galaxy Maggiorato GX283 or Galaxy Maggiorato GX288 (or any other with inner width of 210 mm).

[Electronix]

Hello Prasimix,

thanks for sharing your progress on this project.
Did you or your colleague also working on a Windows program to enter a predefined voltage current list for automatically running on that power supply ?
I had a look at the software power view from liv and it is looking like that I have in my mind. The only thing I missed is a function where you can import predefined voltages currents on a time line or do it directly in the voltage current and power chart, by adding points.

Cheers

Sven

[prasimix]

Hello Prasimix,

thanks for sharing your progress on this project.
Did you or your colleague also working on a Windows program to enter a predefined voltage current list for automatically running on that power supply ?
I had a look at the software power view from liv and it is looking like that I have in my mind. The only thing I missed is a function where you can import predefined voltages currents on a time line or do it directly in the voltage current and power chart, by adding points.

Cheers

Sven

Hi Sven, we decided in the early stage to implement remote control using some standard or de-facto standard protocol that is used with such kind of device. The SCPI was an obvious choice because it's used by all industry heavy weights (Keysight at al). The SCPI is a plain text protocol therefore one can use telnet session for sending to and acquiring data from a device. Or you can write application in any programming language and OS that will use SCPI commands. More about our implementation you can find in the separate thread or in the SCPI reference guide.
Please also note that we created a software simulator that can be used for both development and evaluation of the firmware without having real device on hand. Our plan is to extent existing SCPI command set and also to create an platform independent remote controller application. The software simulator is already platform agnostic and it's tested on Linux, OS X and Windows and one of the possibility is to use that "front-end" not just for simulation but also to communicate with the real device.
Finally people who are in position to work with SCPI controllers such as NI LabView, Keysight VEE or similar could start writing application for this PSU like for many other laboratory devices and use Serial or Ethernet communication. I wish that MyOpenLab is more active because as an open-source project it looks like a logical 3rd party companion for this PSU project.

[Electronix]

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 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