-
Thanks Technix for your input. This project is started with possibly different ground that yours. I'd like to list few of them:
- Modularity, the PSU is broken into few smaller modules that helps anyone to reach easier the final goal: a fully functional device. That also mean that one could stop for a while in "the middle" like using manual control without MCU assistance and add it in the future using here offered solution or (re)using something else.
- Open hardware and software, that offer anyone to freely adopt existing solution to personal ideas. That also means that if someone found attractive only one part of the whole like pre-regulator and wants to deploy it with his/her own post-regulator it is able to do it without extra effort to "extract" it from the bigger PCB.
- Arduino "compatibility", there is a huge group of people who already own and have some experience with Arduino boards, many of them is still 8-bit (5V) and 32-bit (3.3V) is currently available with the same pinouts only in one model the Arduino Due (Zero has a size of Uno which is not capable enough for this project and therefore is not a candidate for 5V version).
I think that possibility to simply take existing Arduino board and plug it to Arduino shield for this PSU is quite attractive (e.g. forget hassle with uploading bootloader, etc.). Of course anyone who wants to go with 3.3V (Due) that can be very easily done by redesigning the PCB and remove not just level shifters but also binding posts section and make the PCB much smaller. - Functional software, I already spent many hours with hardware part of the project. I'm still not tired and software part is not taken lightly, like write few lines of codes, set and display measured voltages and currents and that's it. Thats in general works for the author but making something usable and intuitive enough that (almost) anyone can use it without thick manual or exchanging lots of support mails is quite another thing. Such kind of device cannot rely on presumption that operator (that could be uninformed user) will use interface without making any error. My existing experience with software tells me that significant part of any software that has an user interface belongs to the user's error handling.
So we (I cannot talk here in singular anymore because two of us is now involved in software design, writing and debugging) would like to make a intuitive (and secure) user interface when we talk about local access (front panel with touch screen TFT display) and standardized communication when the PSU is accessed remotely. Due to that we decided to restructure software development plans in a way to take into account immediately all aspects for local and remote control. Accessing any device remotely could be done in myriad ways that can be at the end divided in at least two categories: "proprietary" and "standardized". First is a great sandbox for making many quick-and-dirty solution but at the end it easily become a dead end (otherwise you a not a mega-corporation which can push it as de facto or de jure standard). Taking that into account we decided to use for the remote control the SCPI that will simplify connection with existing "controllers" for instrumentation and also will offer a quick terminal/text console access (thanks to "serial over USB" feature).
Adding invisible SCPI is boring in comparison with painting in different colors messages on the TFT display but taking into account both local and remote access from the beginning could help us to structure a core in a better way (you can add here if you wish a few marketing buzzwords such as "future-proof", "scalability/flexibility", etc.
).
Again, if someone found that local access is not required since PSU will always be close to some PC then great, it can forget TFT display and take USB or even Ethernet as the only way of controlling the PSU.
-
New PCB comes few days ago and I took time to assembly first Arduino shield that replaces three PCB used before. Here is how bare board looks:
Top side of the PCB with all parts inserted without Arduino board:
... or with Arduino (Mega2560) inserted:
Bottom side that will be visible on PSU front panel carry TFT touch screen display, binding posts, LED indicators and connectors for battery's NTC. I didn't use spacer sleeve for LEDs for better mechanical strength, but I'll order possibly this one.
I tested all parts and can say "so far so good" - buzzer, Ethernet, RTC, external EEPROM and LED/relay driver works. Communication with TFT display also works. Optional V/F converters for monitoring battery temperature while charging also works. No error on PCB layout or logic is found.
For selecting Arduino Mega2560 or Due only two jumpers has to be set on the PCB.
In parallel a good progress is made on software side. First set of SCPI commands are implemented and I'll post few sample session soon. I'm continue to works on new AUX PS that will supply Arduino shield and which also has MCU controlled soft-start for main toroidal transformer.
-
It looks really nice!... Besides color matching,
but any board color except green or yellow would have looked better with the Arduino board.
If I remember correctly you used a Czech PCB house.
Looking forward to see how it all comes together... -
Looks amazing! Posting mostly so I can keep tabs on your awesome project
-
It looks really nice!... Besides color matching,
but any board color except green or yellow would have looked better with the Arduino board.
If I remember correctly you used a Czech PCB house.
Looking forward to see how it all comes together... Yes, I'm still with printed.cz where you can order many other colors. The green one is the cheapest and I'd like to spare some money in this phase. If GB is ever happen we can choose any other color.
Anyway I forgot to add a picture of how it looks on the other side:
-
Very nice project! I can tell a lot of work (and money) went into this. Kudos to you.
Hey, what kind/brand of binding posts are those? Where can I get them? -
sorry if its already asked,
is this project can be available for a "kit" form
-
Very nice project! I can tell a lot of work (and money) went into this. Kudos to you.
Hey, what kind/brand of binding posts are those? Where can I get them?
Here is an extract from BOM (it will be available online when ready):
I'm using TME but Farnell has something similar (and more expensive). -
sorry if its already asked,
is this project can be available for a "kit" form
My intention is to organize GB for PCB and possibly enclosure with customized front panel. For all PCB detailed BOM will be published and everything can be ordered from just two supplier: TME.eu and Farnell (TME is cheaper but cannot offer all ICs). -
The latest version of AUX PS PCB that will be used for powering Arduino shield is now also assembled and tested. Bare board looks like this:
Assembled with all components:
-
When Arduino shield is connected to the AUX PS from the previous post and everything is turned on (all LEDs and three relays) the +5V output (yellow) and SW signal (cyan) looks like this:
Since this revision of AUX PS PCB also include main transformer soft-start I also made some testing of that functionality. For current measurement (magenta) a small current transformer is used to sense what's going on on AC mains.
First I check what's going on when transformer is simply connected to the AC mains (a 500VA Talema 55186-P1S2 transformer is used for testing):
Then if the same transformer is connected with 22R/7W used for soft-start functionality then I got the following in-rush current:
Then I tried with power-up sequence controlled by Arduino shield when first soft-start triac (cyan) is activated then after 700ms, direct triac (blue) is switched on, and finally soft-start is switched off after 100ms:
-
The optocoupler part looks dangerrously close. SMT opto-couplers are already a problem when is comes to creapage disdance. There usually is not room for an extra line under the OK. One might even need a cutout under the OK to get sufficient clearance.
Also under the transformer the creapage distance looks rather small - at least not suitable for a dirty / humid enviroment.
Usually the fuse should go before the MOVs and similar overvoltage protection parts. -
Well I have finalized my design of M130v1 (25V max at 1.5A, 7.5A max at 5V), and I used a hybrid design that drew concepts from both switch-mode and linear power supplies. I used a mains frequency transformer to step the voltage down to 35V full wave rectified (transformer not on board but rectifiers are), and used three step-down converters to provide the required power rails, +5V (MCU), +30V (op amps and digipot) and the adjustable output rail. The adjustable step-down is then cascaded with a NMOS linear stage to smooth out most the remaining ripple.
-
The optocoupler part looks dangerrously close. SMT opto-couplers are already a problem when is comes to creapage disdance. There usually is not room for an extra line under the OK. One might even need a cutout under the OK to get sufficient clearance.
Also under the transformer the creapage distance looks rather small - at least not suitable for a dirty / humid enviroment.
Usually the fuse should go before the MOVs and similar overvoltage protection parts.
Thanks once again Kleinstein for your valuable comments. I forgot to mention that schematic for this PCB is presented in post #138.
Do you referring to horizontal or vertical improper clearance? I think that MOC205x and MOC206x are designed to be used with AC mains (EU 230VAC included) and they have clearance of ~7mm. I'm using SMD parts and Cu traces are isolated at least "dirty" factor should not be an issue. Anyway I made a correction and one possible layout could be like this:
Regarding fuse, this one is not main fuse but only 80mA for PCB transformer. Therefore I believe that is on the right side.
Could you please let me know what traces under transformer look problematic? On the left side is AC mains and on the right 12VAC. -
Die optocoupler should be OK, but this is about the minimum distance one should have if the secondary side is not PE connected.
The distance from the 12 V AC part to the mains connected part on the right now looks a little too small. Like with the optocoupler this should be something like 7 mm minimum, better 8 mm. How much clearance is needed depends on the enviroment. Humid or dusty enviroment might need even more distance and thus a cutout under the optocoupler.
If there already is another fuse, than it's Ok to have an extra fuse just for the transformer. -
The distance from the 12 V AC part to the mains connected part on the right now looks a little too small. Like with the optocoupler this should be something like 7 mm minimum, better 8 mm. How much clearance is needed depends on the enviroment. Humid or dusty enviroment might need even more distance and thus a cutout under the optocoupler.
This is a corrected version where ~8mm spacing around secondary traces was added:
-
I am hoping you intend to place an insulator under the heat sink, on the aux board? Around the mains tracks on the top or red layer?
-
I am hoping you intend to place an insulator under the heat sink, on the aux board? Around the mains tracks on the top or red layer?
Thanks for your comment. It can be done easily without hassle of insulating heatsink. Here it is:
-
Wow that was quick. Don't want you to touch a mains hot heat sink :-)I am hoping you intend to place an insulator under the heat sink, on the aux board? Around the mains tracks on the top or red layer?
Thanks for your comment. It can be done easily without hassle of insulating heatsink. Here it is: -
The new mosfet pre-regulator board with SMD components is now also finished. In bias pre-regulator sections LM317 is replaced with LP2951. Bias voltage for op-amps on the post-regulator board is decreased from +/-15V to +/5V. The +5V for supplying digital section is still separated (using another LP2951). V/F converter is added for temperature monitoring and PCB also can carry now three smaller power capacitor that can provide higher max. ripple current figure then the single one. The bare board looks like this:
With all components (except sync isolator for syncing with LM5574 on another channel's pre-regulator):
-
The new post-regulator board with SMD components is also finished. Connection with pre-regulator is now accomplished using single connector. This board just as previously presented pre-regulator is 60x100mm that is within limits of Eagle freeware edition (that means modifications is also possible on top of opening and viewing). No errors is found in layout and no additional cuting and wiring was required this time.
The bare board looks like this:
Assembled (everything is hand soldered, connectors for manual OE and DP control and CC/CV LEDs were not mounted):
Mosfet pre-regulator and post-regulator when connected occupy 122x100mm:
Mounted on SK85/50/SA heatsink that is enough that without fan temperature (Tamb=25oC) goes to 60oC with 3A load:
With this a hardware part of the adventure is mostly finished. Actually what is still missing is manufacturing of the front panel that do not cost a fortune and can be easily obtainable to anyone. That topic is already discussed in this thread but I'm still not sure what to do.
Anyway in the time that comes focus will be on the software part. First with "boring" SCPI that will offer remote control using serial (via USB on Arduino board) or Ethernet (on Arduino PSU shield) communication. Support for local control using TFT touch-screen will follow.
Of course detailed BOM with parts number for TME and Farnell are in preparation and will be published together with Eagle "source" files (PCB schematics and routing). -
Really good job!!!
I love that smd work, looks incredibly close to branded manufacture. Keep it up. -
Looking forward to completed project.
-
Again, lots of work went into this. Great job!
That heatsink looks awesome... nice thick one!
Part#/Brand? -
That heatsink looks awesome... nice thick one!
Part#/Brand?
Please click the link in the sentence above the picture ...