DIY bench power supply PSL-3604

[Liv]

Features:

• output voltage: 0…36 V
• output current: 0…4 A
• voltage resolution: 10 mV
• current resolution: 1 mA
• operation modes: CV, CC
• down programmer circuit
• remote voltage sense amplifier
• pure linear design
• multilevel output stage supply
• synchronous rectifiers
• digital calibration
• interface: USB (optocoupled)
• case type: Z-2A (Kradex, Poland)

To minimize dissipation power output stage uses 4-level supply. It provides high dynamics, while conventional circuits with transformer tap switching or preregulator need time for bulk capacitors charging. This power supply also uses MOSFET synchronous rectifiers. Fast discharging of load capacitance provided using down programmer circuit. High dynamics and very small output capacitor prevents current spikes at CV-CC transition. Power supply circuit ready for remote voltage sensing. Control system based on ARM-microcontroller STM32F100C8T6B. User control elements – incremental encoder and 4 pushbuttons. Settings -via local menu or via USB interface.

1-ampere and 2-ampere versions here: PSL-2401/2

[Liv]

PCB project for manufacturing:

[DmitryL]

Nice try!
Good to see you here
I've made several devices by your design and I'm completely happy with them (though minimalistic amount of controls was discussed elsewhere

[firewalker]

I seems to be a nice unit. Is there a place fore someone to buy it?

Alexander.

[dom0]

Nice project! Clean mechanical build, too.

The toroid is custom manufactured I guess?

[BloodyCactus]

that looks pretty sweet mate.

[prasimix]

Thanks Liv for sharing all details with us. Schematics looks awesome, I have some questions if you're are willing to address it. Some details is far beyond my comprehension.

[Liv]

Nice try!
(though minimalistic amount of controls was discussed elsewhere

Thank you!
Minimalistic control - this is real drawback. Mainly concerns of my sinal generators projects. The power supply has enough balanced controls: separated buttons for V, I settings, and encoder with adaptive speed and "Fine" mode allows edit setpoint quickly.

Is there a place fore someone to buy it?

I don't know. Project is open, somebody probable produce it.

The toroid is custom manufactured I guess?

Toroid based on 105W transformer for halogen lamps. Secondary winding (12V) was removed and replaced by 4 windings 2 x 8.5V and  2 x 10V.

Schematics looks awesome, I have some questions if you're are willing to address it. Some details is far beyond my comprehension.

Schematic has been designed to meet the high dynamics and precision (if you apply more precision opams and resistors with low TC). Very low output capacitance provides very low current spikes. I can set 36V, 0.005A and connect LED to the output. There in no even visible light flash, of course, no damage of LED. ADC uses oversampling to increase resolution. DAC uses addition sigma-delta part using DMA to increase resolution to 16 bits. For example, Agilent U8002 uses 12 bit DAC, voltage steps is not equal. This power supply does not have such a drawback because of increased DAC resolution. Any questions can be sent to e-mail wubblick@yahoo.com, just hard for me to speak in english.

[deephaven]

That is a pretty amazing power supply, congratulations! Could you tell me how you made the front panel?

[DmitryL]

That is a pretty amazing power supply, congratulations! Could you tell me how you made the front panel?

Take a printer, print on paper (I use "silvery" looking paper), laminate. "Cold laminate" film seems to be the best option.
Use double-sided scotch tape to glue it.
As for the palstic carving - alsk LiV

[Liv]

I used the blank panel of Z-2A case, that represents the polystyrene panel thickness of about 3 mm. Holes done using desktop milling machine, although it can be done using a drill and needle files. For display window I used a dark plexiglass. The sticker is printed on the film Oracal, cut on the plotter and laminated in the firm, engaged in advertising.

[deephaven]

Thanks. Very nice finish.

[BravoV]

Cool ! Thanks for sharing.

Really appreciate if you can share few captured scope's shots at it's dynamic performances like :

- Load transient response and recovery
- Over load/short circuit performance and it's recovery
- Over/under-shoot at CV->CC transition and vice versa.

[Liv]

I have only analog scope, it's hard to make screenshot. I have about 150$/month, I can't buy digital oscilloscope. But I observe good compliance between Spice model and reality. Spice diagrams show the following: 1 - step from 2A to 4 A and back (C_load  =100 uF); 2 - short circuit (0.01 Ohm, C_load  =1 uF); 3 - CV->CC transition and vice versa (I = 3A, R_load transition 18 Ohm->6Ohm, C_load  =1 uF).

For example (last image), in debug process (tuning frequency correction for CC-CC step) I compare real scope image with Spice diagram and found good compliance.

[nickds1]

Lovely work.

One question - you use a discrete synchronous FET bridge rectifier - did you consider using one of the monolithic ideal diode bridge controllers like the LT4320(-1) ? - cuts out a load of components and real-estate, plus its cheap, good to 72V and inputs from DC-600Hz.

[Liv]

I know about LT4320, but it's two different universes - price LM393 and price LT4320. My version of synchronous rectifier is very inexpensive.

[BravoV]

I have only analog scope, it's hard to make screenshot. I have about 150$/month, I can't buy digital oscilloscope. .... <snip>

Ah .. don't worry, just build the circuit below and it should be cheap & easy, especially with your building skill there  . 

Designed by one of the well respected member here, and this circuit should be handy to verify DC power supply whether it has a decent "dynamic performance" or not.

Click here -> Dynamic Electronic Load Project by Jay_Diddy_B

[nickds1]

I know about LT4320, but it's two different universes - price LM393 and price LT4320. My version of synchronous rectifier is very inexpensive.

LT4320 is about USD 4 here - you save a whole bunch of components by using something like that, plus a load of board real-estate (under the FETs) - I can't believe that its more expensive to use the 4320 than the discrete solution... over twenty components replaced by each single driver.

Not in anyway a criticism of your design, just you use other integrated solutions in the design, not least the uP... just seems a bit odd...

Cheers

[BravoV]

I can't believe that its more expensive to use the 4320 than the discrete solution... over twenty components replaced by each single driver.

You should believe it, at certain part of the world, for example mine  , say on a few bucks purchase on few chips, the shipping & handling charge sometimes can be as high as $70 to $100 like from well known Digikey/Mouser etc, or purchase it directly at the manufacturer e-shop like Linear Technology which sometimes even worst. 

Oh yeah, we're not discussing about the local custom import tax yet and bla..bla... 

[Liv]

just build the circuit below

I'm just designs electronic load (see attachment), a project to link too primitive.

LT4320 is about USD 4 here

Very expensive. But you can not do without a microcontroller. We have completely different conditions in terms of money.

[PeterHouse]

The unit looks very professional and clean.

My $400 Agilent power supply will let the smoke out of an LED if connected at 36 volts and set to .005A!

I would like to hear more about how you used the Oracal film - did you get a small piece and use and inkjet printer - how did you do the front panel?

Good Work.

[Liv]

how did you do the front panel

First, make all the necessary holes in the blank panel. Then the glass is made from smoked acrylic. You can use a transparent acrylic and film for window tinting for cars. Then stick label. Label ordered the firm engaged in advertising. They are printed on the Oracal film, laminated and cut on the plotter. One label is worth about $ 0.2. Install a circuit board through the glued plastic spacers - panel ready.

[ivan747]

Do you happen to have a bill of materials for the components (specially the mechanical things like that heatsink)? I'm interested in this, looks very nice. I have been looking to build a power supply for years, but I am never satisfied with the projects out there, many times they are too simple, others are too much hassle for the outcome and the rest never have what I need but yours seems to be so professional that I might put the effort into it. Also, it would be great to share the Gerber files, since the design is open 

[JSnyder]

Do you happen to have a bill of materials for the components (specially the mechanical things like that heatsink)? I'm interested in this, looks very nice. I have been looking to build a power supply for years, but I am never satisfied with the projects out there, many times they are too simple, others are too much hassle for the outcome and the rest never have what I need but yours seems to be so professional that I might put the effort into it. Also, it would be great to share the Gerber files, since the design is open 

+1 on the BOM, I've already managed to convert the files to gerber  I also agree, this looks like a great addition to the lab.

[Liv]

Do you happen to have a bill of materials for the components (specially the mechanical things like that heatsink)?

Mechanical things are custom. For heatsink I used aluminum profiles with the treatment on the milling machine.
Attach several documents and control software.