Electronics > Beginners

DIY PSU part 2

(1/3) > >>

Pirateguy:
rather then dig up the old thread i decided to start a new one.

so i have drawn up a plan and with some help will be making a box to put it all in.
here's my schematic:


and the panel:

Pirateguy:
please let me know if any of the following is retarded:

ok so the plan right now is to have multiple poly switches to switch between,
so i can access somewhat higher currents without removing protection altogether
or limit the current to 100ma for sensitive projects.

these will be followed by an RGB LED so the colour will tell me which fuse i am set to.

i think i need those diodes to keep from lighting up all leds all the time.
i assume these diodes would need to be rated for at least the 6A of the heaviest poly fuse

the M box represents this: https://www.ebay.com/itm/Dual-Blue-Red-Red-Red-LED-Digital-Voltmeter-Amperemeter-Panel-Amp-Volt-Gauge-DC/322982162973?ssPageName=STRK%3AMEBIDX%3AIT&var=512120455709&_trksid=p2057872.m2749.l2649 meter display thingy
or this one:
https://www.ebay.com/itm/DC-0-30V-3-Wire-LED-Digital-Display-Panel-Volt-Meter-Voltage-Voltmeter-Car-Motor/182952047363?ssPageName=STRK%3AMEBIDX%3AIT&var=690269460764&_trksid=p2057872.m2749.l2649

The load resistor will be 4R7 10watt to run at 5.3 watt, just over the minimum 5 required.

a simple signal generator and relay timer will be included in the box.
i was thinking of having each of them drive a relay that connects a line to ground, and that
that way i could take any voltage and use that ground connection instead of the regular one
and i would have that voltage modulated.

the third relay is simple a pushbutton and a connector in parallel so i can connect a remote button or switch
or something like that. it could be used as a dead man's switch or sometimes i imagine i will want to conveniently
switch things on and off without having to reach for the unit, or give short bursts of power to test something or other.

is a relay the appropriate choice for this?

the thermometer would have a switch between an internal sensor and a set of binding post so i can use it externally.
i was thinking i need some thermal insulation to keep the temp inside the box from throwing off the thermometer.
maybe some of that expanding insulation foam?

ok so let me know if any of that doesn't make sense etc.
also: any tips on how to wire this in an organized fashion?

PointyOintment:

--- Quote from: Pirateguy on October 13, 2018, 11:59:32 am ---rather then dig up the old thread i decided to start a new one.

--- End quote ---
I had to dig up the old thread because I couldn't understand this fully without it. It seems that this machine will be based on an ATX power supply.


--- Quote ---[SCHEMATIC]

--- End quote ---
Your schematic is both messy and cramped. It's more cramped than it is messy (IMO), so it shouldn't be that hard to clean up. But you should also make more use of named nodes so you don't have lines going all over each other. Also, the components should be labeled with reference designators. They make the circuit much easier to talk about, so people would be more willing to help you.

In the schematic, you have USB type-B sockets, whereas on the panel you have type-A (which is what you want).

The third of the four panels with switchable polyfuses has only two current options, not three. I presume this is because the 7809 can't supply 6 A. In the old thread, there was much talk of pass transistors, so I know you already know that's an option if you want more current from that channel.

For the modules on the right, why not put switches on their power inputs if you want to be able to turn them off? The switches in series with the output relays seem strange.


--- Quote ---[PANEL DRAWING]

--- End quote ---
Let me see if I can understand your front panel. Labels would help. You might not need them yourself, but maybe a friend will use this power supply at some point, or maybe you'll put it away when you get a better one, and pull it out years later to use for something, and not remember what does what. I would understand none of the front panel if I couldn't look at the schematic simultaneously, and I would only understand 75% of it if I hadn't read through your old thread fully.

Anyway, here's how I interpret it. Everything I don't understand or get wrong in my interpretation is probably something you need to correct.

Column 1:
Power Good LED
master power switch
button for manual relay output
LED that isn't in the schematic
external control input for manual relay output (the "unspecified 2-pole connector")
manual relay output switched ground terminal—see below on using relays to switch ground
pointless LEDs that will always be on
USB ports

Column 2:
4 pairs of RGB LED and switch for the switchable current-limiting polyfuses for the 4 switchable-current-limit positive channels

Column 3:
4 voltage/current displays for the 4 switchable-current-limit positive channels

Column 4:
4 positive terminals for the 4 switchable-current-limit positive channels

Column 5:
4 ground terminals shared between the 4 switchable-current-limit positive channels and the three negative channels (all grounds common, but one per row for convenience)

Column 6:
2 yellow terminals together for the thermometer's external input
3 yellow terminals for the negative channels

Column 7:
LED that isn't in the schematic
Thermometer internal/external switch
Thermometer display
Three voltmeters for the negative channels
Three LEDs to indicate polyfuse states

Column 8:
DPS module
Relay timer module
4–20 mA semi-arbitrary waveform generator (a strange choice)

Column 9:
Switch that isn't in the schematic
Ground terminal—where's the positive output terminal of the DPS module as shown in the schematic?
Switch for the relay timer
Switched ground terminal from the relay timer—see below on using relays to switch ground
Switch for the function generator
Switched ground terminal from the semi-arbitrary waveform generator—see below on using relays to switch ground and on using a relay on the output of the function generator

Why do you have voltmeters when the voltages should not vary by a large enough amount to register on them? (Do you think you need them to know which channel has which voltage because you don't know the concept of labels?)



--- Quote from: Pirateguy on October 13, 2018, 12:33:07 pm ---ok so the plan right now is to have multiple poly switches to switch between,
so i can access somewhat higher currents without removing protection altogether
or limit the current to 100ma for sensitive projects.

--- End quote ---
Using switchable polyfuses for current limiting is strange, but it could work. (I would find this power supply inadequate for testing LEDs, though. For those, I usually set a current limit of 20 mA and a voltage around 4 V and hook up the LED in both directions until it lights up (because I can never remember the polarity indicators). When I figure out the correct polarity, it runs in CC mode at 20 mA (which is a safe current level for most LEDs). I set the voltage to 4 V so that when I hook up the LED in reverse, I don't exceed its reverse breakdown voltage and destroy it.)


--- Quote ---these will be followed by an RGB LED so the colour will tell me which fuse i am set to.

i think i need those diodes to keep from lighting up all leds all the time.
i assume these diodes would need to be rated for at least the 6A of the heaviest poly fuse

--- End quote ---
Yes and yes. On the 100 mA polyfuse, won't the LED take up a significant amount of that current?


--- Quote ---the third relay is simple a pushbutton and a connector in parallel so i can connect a remote button or switch
or something like that. it could be used as a dead man's switch or sometimes i imagine i will want to conveniently
switch things on and off without having to reach for the unit, or give short bursts of power to test something or other.

is a relay the appropriate choice for this?

--- End quote ---
A relay is fine for that, but you've connected its switch to ground. When you press the button, the relay output will be connected to ground. When you release the button, the relay output will be unconnected (floating). I initially assumed that this is not what you want. Then I realized it probably is. You're probably planning to connect your load between whatever positive voltage terminal you want and this switchable ground terminal. This seems unwise—it will leave your whole load floating at the positive voltage when the relay output is off. If a grounded part of the load touches a grounded object on the bench, it will still be powered when the relay output is "off".


--- Quote ---a simple signal generator […] will be included in the box.
i was thinking of having [it] drive a relay

--- End quote ---
A relay is not fine for that. In fact, that's probably the worst possible thing you could do there.

The function generator that you linked to in the other thread is actually an arbitrary waveform generator, which would seem like a good thing, but it's actually an extremely limited one. It's meant for testing PLCs by simulating a sensor that provides a 4–20 mA proportional output (which is an industrial de facto standard for that stuff). Also, it can only do 9 linear segments of a function (though that would be enough for you most of the time).

I can't tell whether it pushes the set amount of current through the current loop, or just allows the set amount of current through itself under an externally applied voltage (which would come from the PLC, in its intended application).

I can't figure out what the hookup diagram on the product page means, either. And the textual description of how you operate it makes it sound like it will be extremely frustrating to operate.

Anyway, assuming you can hook it up and get it to work, it will give a current output, not a voltage output. The current will vary between 4 and 20 mA. This is probably not suitable for driving a relay, but you haven't provided a part number or datasheet for the relay, so I don't know how much current its coil needs.

Also, are you aware that relays are either on or off, and cannot provide a proportional output? Knowing that, it makes no sense to use one as an output stage for a function generator. If the function generator could drive the relay, most of its purpose would be wasted because the relay could not reproduce the proportional output of the function.

Also, the relay has hysteresis: it requires some level of coil current to close the switch, but then when the coil current decreases, the switch doesn't open again until the coil current reaches some much lower level. This will distort the timings of the function as well.

Also, switching a relay rapidly will wear it out quickly. However, this doesn't apply to your unwisely chosen function generator because its minimum time per segment is 1 second.

A much better output stage for this function generator would be a transistor. Either a BJT or a MOSFET could reasonably be made to work. However, if you want to do that, post the schematic for review before you build it!


--- Quote ---the thermometer would have a switch between an internal sensor and a set of binding post so i can use it externally.
i was thinking i need some thermal insulation to keep the temp inside the box from throwing off the thermometer.
maybe some of that expanding insulation foam?

--- End quote ---
Huh? What's the purpose of the internal thermometer? If it's to measure the internal case temperature of the power supply (to check whether it's overheating), just put it in there, probably in the exhaust air stream. If it's to measure the room temperature, put it in the intake air stream. Insulation would just make its readings meaningless.


--- Quote ---also: any tips on how to wire this in an organized fashion?

--- End quote ---
Maybe make a wiring diagram first. But before you can do that, you'll need a schematic that's correct and ideally also easy to read. Dave has a video on that:

Pirateguy:
lol yeah sorry about that, ADD got the better of me there.

All of your interpretations are correct.
some clarifications:


--- Quote ---Column 1:

LED that isn't in the schematic
pointless LEDs that will always be on

--- End quote ---

the led not in the schematic would be in series with the button, not sure if i will include the button on the unit itself.
the point of the LEDs is to confirm the usb ports are active.


--- Quote ---Column 7:
LED that isn't in the schematic
--- End quote ---

indicates thermometer switch state


--- Quote ---Column 8:
4–20 mA semi-arbitrary waveform generator (a strange choice)
--- End quote ---

it's the cheapest one that comes as a panel unit like that.


--- Quote ---Column 9:
Switch that isn't in the schematic
Ground terminal—where's the positive output terminal of the DPS module as shown in the schematic?
--- End quote ---

thats a booboo, the switch is supposed to be the missing positive terminal.


--- Quote ---Why do you have voltmeters when the voltages should not vary by a large enough amount to register on them? (Do you think you need them to know which channel has which voltage because you don't know the concept of labels?)
--- End quote ---

cheap ebay module, and it looks sexy :)



--- Quote ---Using switchable polyfuses for current limiting is strange, but it could work. (I would find this power supply inadequate for testing LEDs, though. For those, I usually set a current limit of 20 mA and a voltage around 4 V and hook up the LED in both directions until it lights up (because I can never remember the polarity indicators). When I figure out the correct polarity, it runs in CC mode at 20 mA (which is a safe current level for most LEDs). I set the voltage to 4 V so that when I hook up the LED in reverse, I don't exceed its reverse breakdown voltage and destroy it.)
--- End quote ---

right, i have gotten into the habit of using a 2032 coin cell for testing leds.
i made a little holder with breadboard leads and a female header for this purpose.

the poly switches aren't really for control, i have the DPS for specific needs, but rather to set a general safety level.
originally they were just 1 fuse, but i figured it might as well do it like this, in case i want to use more power then the fuse allows.
maybe a 1A, 6A and unfused connection might be more practical?


--- Quote ---Yes and yes. On the 100 mA polyfuse, won't the LED take up a significant amount of that current?
--- End quote ---
well since it is only an indicator led, i plan to run it at 2ma or so. certainly no more then 5.
would u recommend i run it through a transistor off the 5v rail?


--- Quote ---A relay is fine for that, but you've connected its switch to ground. When you press the button, the relay output will be connected to ground. When you release the button, the relay output will be unconnected (floating). I initially assumed that this is not what you want. Then I realized it probably is. You're probably planning to connect your load between whatever positive voltage terminal you want and this switchable ground terminal. This seems unwise—it will leave your whole load floating at the positive voltage when the relay output is off. If a grounded part of the load touches a grounded object on the bench, it will still be powered when the relay output is "off".
--- End quote ---

good point.
so i should do it the other way around, have an in and out terminal and use a jumper wire between that in and the voltage i want to use.


--- Quote ---A relay is not fine for that. In fact, that's probably the worst possible thing you could do there.
--- End quote ---
yeah the func gen and relay timer thing are mostly there to fill the space. ill have to experiment some before i decide if i actually include them.
anyway the idea is to provide a PWM signal to drive motors and that kinda thing.

this whole right end of the thing i plan to work on over time with upgrades n stuff.
i shouldn't have added that to the circuit, as i plan to work that as a separate project after finishing the rest of it,
but i wanted to know about the relay n if this idea made sense outside of my own head :P

i think i will start another thread for this part once i get around to it.


--- Quote ---Huh? What's the purpose of the internal thermometer? If it's to measure the internal case temperature of the power supply (to check whether it's overheating), just put it in there, probably in the exhaust air stream. If it's to measure the room temperature, put it in the intake air stream. Insulation would just make its readings meaningless.
--- End quote ---

the idea is to display the internal temp of the unit. mostly just to have something relevant on the display when i'm not using the external.
the isolation would be around the unit itself so that the sensor picks up the temperature, but when disconnected the unit itself, and the remaining wire
from the sensor do not pick up temp from the psu when i am using the external.
the external of course is for measuring temperature in projects.
not sure if ill ever have a need for that, but if there is no practical issue with having a switch like this, i figured why not.

PointyOintment:

--- Quote from: Pirateguy on October 16, 2018, 04:00:52 am ---
--- Quote ---Using switchable polyfuses for current limiting is strange, but it could work. (I would find this power supply inadequate for testing LEDs, though. For those, I usually set a current limit of 20 mA and a voltage around 4 V and hook up the LED in both directions until it lights up (because I can never remember the polarity indicators). When I figure out the correct polarity, it runs in CC mode at 20 mA (which is a safe current level for most LEDs). I set the voltage to 4 V so that when I hook up the LED in reverse, I don't exceed its reverse breakdown voltage and destroy it.)
--- End quote ---

right, i have gotten into the habit of using a 2032 coin cell for testing leds.
i made a little holder with breadboard leads and a female header for this purpose.

the poly switches aren't really for control, i have the DPS for specific needs, but rather to set a general safety level.
originally they were just 1 fuse, but i figured it might as well do it like this, in case i want to use more power then the fuse allows.

--- End quote ---
I realized after I'd posted and put away my computer that you could just use the DPS for LED testing and other low-current stuff, so I guess having only the three polyfuse levels is fine.


--- Quote ---
--- Quote ---Yes and yes. On the 100 mA polyfuse, won't the LED take up a significant amount of that current?
--- End quote ---
well since it is only an indicator led, i plan to run it at 2ma or so. certainly no more then 5.
would u recommend i run it through a transistor off the 5v rail?

--- End quote ---
No, that's probably fine. You're not going to get any good trip current precision out of the polyfuses anyway.


--- Quote ---
--- Quote ---A relay is fine for that, but you've connected its switch to ground. When you press the button, the relay output will be connected to ground. When you release the button, the relay output will be unconnected (floating). I initially assumed that this is not what you want. Then I realized it probably is. You're probably planning to connect your load between whatever positive voltage terminal you want and this switchable ground terminal. This seems unwise—it will leave your whole load floating at the positive voltage when the relay output is off. If a grounded part of the load touches a grounded object on the bench, it will still be powered when the relay output is "off".
--- End quote ---

good point.
so i should do it the other way around, have an in and out terminal and use a jumper wire between that in and the voltage i want to use.

--- End quote ---
That's probably how I'd do it.


--- Quote ---
--- Quote ---Huh? What's the purpose of the internal thermometer? If it's to measure the internal case temperature of the power supply (to check whether it's overheating), just put it in there, probably in the exhaust air stream. If it's to measure the room temperature, put it in the intake air stream. Insulation would just make its readings meaningless.
--- End quote ---

the idea is to display the internal temp of the unit. mostly just to have something relevant on the display when i'm not using the external.
the isolation would be around the unit itself so that the sensor picks up the temperature, but when disconnected the unit itself, and the remaining wire
from the sensor do not pick up temp from the psu when i am using the external.
the external of course is for measuring temperature in projects.
not sure if ill ever have a need for that, but if there is no practical issue with having a switch like this, i figured why not.

--- End quote ---
The internal thermometer doesn't get disconnected when you use the external one? That seems like it would prevent accurate reading of the external thermometer except when the two happen to be at the same temperature.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod