I'm designing a utility bus to supply power to various components on my GEM. The bus is powered by a single lead from a PSU 12v 12.5amps, which is more than adequate.
The bus will be mounted on the mount with short leads to each service - mount power, camera, dew heater, logic (various stuff), LED lighting, power ON indication and a spare 12v outlet.
Not being an EE, and with very basic skills in this area, I'm stumped as to how to maintain 12v across all services (which I have tried to represent with various resistances depending on what I think maximum power demands will be - manufacturers specs and my own measurements).
I think I'm making a fundamental error/s. Having tried a few different configurations, the diagram below is the closest I can get to achieving 12v across all services.
An in-line master power switch (not shown here) reduces voltages significantly, which is not ideal.
Is using a 13.8V 10A an option?
These are the normal power sources used to top-up/ replace the 12V battery.
If the capacity of the transformer is well above the actual peak current draw I haven't found any issues/ problems with voltage drop to all my cameras/ mounts/ spectroscopes etc etc.
Occasionally I use my power meter to check both voltage and actual current....
There is an error in your schematic. If you have a 12v rail as shown, you will have 12v on the bottom of all the components like the LED has.
Quote:
Originally Posted by rcheshire
I'm designing a utility bus to supply power to various components on my GEM. The bus is powered by a single lead from a PSU 12v 12.5amps, which is more than adequate.
The bus will be mounted on the mount with short leads to each service - mount power, camera, dew heater, logic (various stuff), LED lighting, power ON indication and a spare 12v outlet.
Not being an EE, and with very basic skills in this area, I'm stumped as to how to maintain 12v across all services (which I have tried to represent with various resistances depending on what I think maximum power demands will be - manufacturers specs and my own measurements).
I think I'm making a fundamental error/s. Having tried a few different configurations, the diagram below is the closest I can get to achieving 12v across all services.
An in-line master power switch (not shown here) reduces voltages significantly, which is not ideal.
I was hoping to avoid throwing more juice at it and will need to limit voltage to 12v for other services.
At 13.8v, the 3amp draw gets over the 10v lower limit for that device, just. I guess a 15v supply and some pots to tweak things is one way of doing it - but not very efficient.
Rowland you are trying to represent each rail with varying resistance loads via a series led, this is wrong. For example ....if you discarded the LEDs then each rail have the full 12V (Vcc) applied to each Rx. Put a LED in each rail by all means (parallel to the load) but have set resistor for each rail and try not to run all power thru each led as the leds have max Ic of ~30mA.
Lets say each led drops 2.2v, Vf = 2.2v. Then 12-2.2 = 9.8V (across Rx). If we set the led current, Iled @2mA this is R=V/I = 9.8/.002 = 4.9K ohms Rx required for each led (around about, resistors and LEDs in parallel to the load).
The important thing is to remember not to run all cct power thru each LED here, run the LEDs in parallel with a dropping Rx.
Oh and consider you voltage drop across your supply cable. If you draw lets say 8Amps max thru this cable then power loss (V drop) acorss the cable can be significant, P=I^2R. Industry std is a 3% V drop across supply cables.
Lest say you use 3m of a supply cable with .00252 ohms per m (a good quality cable, Jaycar 8 gauge). Thus power loss @ 8A 20degrees C is P = 8^2 x 3 x .00252 = .484v drop across the cable.
This is .484/12 x 100 = 4% V drop across supply cable! You electronics will only get ~11.5V @8A.
Ok all this is worst case scenario with I draw but consider that cable resistance I quoted is a good quality cable with low resistance per m (8 gauge). Thus I'd be looking at around at minimum an 8 gauge cable of no longer than 3m...preferably maybe even a lower gauge and maybe a power supply of 13.8V DC!.
I have thrown a cat amongst the pigeons for you and given you lots to think about but I'm happy to supply further info as I have recently built my own power distribution box mounted on my scope dovetail. Ok mine has a USB hub and +5V and +8V PSU in it but the principle is the same.
Oh one last thing..I wouldn't run dew heater control power (if you are using PWM dew controller) on the same cct as the camera etc power. It can get a little electrically noisy!
Rowland,
the 13.8V is equivalent to a fully charged 12V battery.
It would be worthwhile checking the actual voltages and current draws with a Powermeter - the Trurnigy unit I use is pretty good...measures up to 130A
I'm designing a utility bus to supply power to various components on my GEM. The bus is powered by a single lead from a PSU 12v 12.5amps, which is more than adequate.
The bus will be mounted on the mount with short leads to each service - mount power, camera, dew heater, logic (various stuff), LED lighting, power ON indication and a spare 12v outlet.
Not being an EE, and with very basic skills in this area, I'm stumped as to how to maintain 12v across all services (which I have tried to represent with various resistances depending on what I think maximum power demands will be - manufacturers specs and my own measurements).
I think I'm making a fundamental error/s. Having tried a few different configurations, the diagram below is the closest I can get to achieving 12v across all services.
An in-line master power switch (not shown here) reduces voltages significantly, which is not ideal.
Your design will not necessarily have a constant voltage at the loads because your design is basically a current source not voltage source because of the resistance of the line from the PSU. However modern electronic devices of the types you envisage using will have wide primary voltage tolerance and any critical internal circuitry will be supplied by internal dedicated regulators and the voltage drop from the line resistance will be negligible.
A simple two wire bus with intermediate sockets will do what you want if your bus cable is capable of carrying the maximum required current without voltage drop. If your 12v supply is some distance from the equipment your could consider running extra wires to supply the regulator sensor at the end of the line rather than the beginning to ensure that there is 12 volts available at the load rather than the source.
We use a simple 12v 15Amp PSU that I built 25 years ago and an array of sockets to suit the specific equipment on the G11. The real need for distribution is the USB line form the computer to the various devices and that is done with a hub fastened to the mount with Velcro strips. The cabling to each device is then made to size and bound into a harness that feeds each device neatly.
Barry
Last edited by Barrykgerdes; 02-02-2013 at 11:28 AM.
I will have to digest all the info and do a bit more design work by the looks of it. A two wire system was what I had in mind. The PSU and cable are capable of supplying 12.5 amps at 12v. PWM will have to go as it will play havoc with the camera.
This is take 2. A 12v rail with; power ON LED running at 10mw, next a switchable LED light source. The resistors are there to simulate the maximum rated load of equipment - total is well within the PSU rating.
The program Every Circuit introduces a fixed resistance value for switches and I have placed them to, ground the circuit ON - is this OK? Any combination of switch on/off has no effect on rail voltage or individual load resistance, which is what I was attempting to achieve - not sure what I was doing before, but this configuration is stable.
The rail is also scalable/expandable.
Critique - refinements - try again?
EDIT:
Quote:
have thrown a cat amongst the pigeons for you and given you lots to think about but I'm happy to supply further info as I have recently built my own power distribution box mounted on my scope dovetail. Ok mine has a USB hub and +5V and +8V PSU in it but the principle is the same.
Nice work Brendan. I like the way you have configured things there. I was going down the DIY PCB path, but have decided to stick to audio jacks, which will work with the type of case you have used.
Yep that seems better. Can I ask why you are using 2 LEDs?
You don't need to run the LEDs are 30mA indeed you have one running at 33mA...very high ILED indeed. You can lower the LED current quite a bit, the 10mA is probably ok but you can go lower. Most will run ok at 2mA (and less) and produce a tad less brightness.....just test them.
Just remember your power drop across the cable and use a high capacity cable...maybe add a fuse.
I agree with Brendan, it is better, although your current (amphere) values are out a bit.
As long as your power supply is capable of supplying more current than you draw AND you cables are large enough to keep voltage drop to a minimum, then you are on a winner.
Not sure why the LED's are there either.
Edit: One thing I would alter is having the switches in circuit before the load. Then when the switch is off, no power is at the load side.
I haven't given much thought to current protection. A fuse is probably the simplest method. - will need to read up on reverse polarity protection.
The second LED is a work light - free floating cable.
The switches are 1 ohm, as it turns out. It occurred to me that all my calculations were 1ohm in error - will relocate switches to the other side of the load.
Hopefully, the PSU will regulate voltage adequately at 12v.
This is the latest with protection using a P Channel power mosfet as per Brendan's post. The FET shown doesn't have an internal diode - representative only.
I'm chasing up dew proof switches and connectors and a suitable box.
Hi Rowland yep that would work. keep up the good work.
The reason a Gate source zener is used is to ensure the Vgs does not exceed the FET parameters. Many TO220 type FETs have a +/-15V maximum Vgs...some higher some lower. So if your PSU is 14V and your FET can take 15V Vgs or higher your really don't need this zener and resistor. That said it won't hurt either.
The other thing is the FET on Rds (on resistance). This changes with current, heat and other things and OEMs give a graph of on Rds and a multiplication factor Vs gate junction temperature you must use to calculate actual on Rds. Choosing as low as you can get Rds on as you can source is the best option but do the sums.
Lets say you use an IRF4905, TO220 P channel mosfet. This has Vgs +/-20V so you don't need zener and resistor, Rds on @20 degrees C (junction temp) = .02ohms.
So if you pass 5A, power dissipated by the FET = I^2R = 5x5x.02 = .5watts.
Now most these TO220 FETS packages can dissipate 2.5W max with good heat-sink. So it looks ok, a touch warm but ok...but read on.
The junction temp will rise with current and ambient. Taking a WAG lets say junction temp = 80C (you can find this by de-rating and taking measurements of the case temp, look at the pdf specs). Thus we multiply the new on Rds by a factor (see spec sheets on device) by 1.25@80C. Rds on actual @ 80C = .025ohms.
Thus @5 Amps, pwr dissipation = 5 x 5x x .025 = .625watts. Your getting hotter you may need a heat-sink on the FET.
I'm just trying to give you an outline on FETs and some simple hurdles you need to watch for. A lot depends on Rds and how it behaves with temperature. Your max I draw is around 6A at present do you'll need a good low on Rds FET...I suspect lower than .02ohms...lets say look for one around .015ohms or less.
Of course you can parallel 2 of the P channel FETs to share power dissipation & I flow.
Now here's a trick...most P channel mosfet have higher on Rds than their N channel cousins, the P FET I quoted above as very good on Rds characteristics but you can go lower with an N channel FET. It is much easier to obtain a low on Rds N channel FET.
This explains it (opens/downloads an on semiconductor pdf on the subject). Basically use a N channel FET on the -ve rail with the gate connected to the +ve rail. Acts just like the P channel except stops reverse current flow thru -ve rail.
Thanks Brendan. I have a very low rds on N channel FET that I used for the DSLR cooling system. I think it meets all of the specs you mentioned and didnt need a heat sink. Given your figures, I suspect that it will do the job.
I needed a 750R to GND to pull down the gate on the cooling system. But I think we are working in reverse and setting a trip point with resistance for this application? visualising the set up.
Thanks for the scematics and links too. Much appreciated. It's a fascinating subject.
I've progressed a little further with this, with low current switching through power transistors, and it's a whole lot smaller.
Two ways to do this - mechanical switching as shown, or digital with a microprocessor. Either way I need switches of some kind. The whole thing could be automated, but at this stage, I see no practical advantage.
I notice that the switch voltage determines the load pulled on the 12v circuit. I think this might depend on the transistor properties.
More development. Replaced the power transistors with logic level MOSFET's and removed switching to the Arduino - 12v on, Arduino on, so I don't forget it. The switches can be hardware or digital output from the Arduino, but this means more debounce programming for any pushbutton logic - so I think, hardware?
The big and small caps on the 12v supply are actually fixed to the PSU terminals. I put them in the schematic to see if there was any effect. 5v supply will be an inbuilt PSU for the switches.
Thoughts on this setup appreciated. Practical, not practical, too complex for the application?
