My laptop adapter is too noisy for LC! Can anybody recommend me a way out?

[SounDuke]

Hi, I've tried to power a MBLC with a toshiba adapter switching PSU which delivers 15V5A. I put the 15V through a 2200mF cap then to a 7809 then to a 100mF cap. on the motorfaders I notice jitter, which obviously is not present when I use a linear adapter 9V 500mA. I tried to use the toshiba adapter to regulate 12V and 5V for MF module and Core respectively, but this solution revealed to be a waste of time. Lots of people are complaining because the fortron PSU is no longer available at pollin. I'm about to try an ATX pc psu but I don't trust really about this idea. Could anybody recommend me another psu or a place to find a fortron FSA55P31? It is a pity to see that everything is ok on the LC and not being able to power it properly :(

[NorthernLightX]

All computer power supplies are switching, or so have I been told. Those are useless for midibox. You need 12v for your MF module, so the C64 PSU is not an option for you. I got my trafo from a local "dump-shop", where you can buy second hand electronics and obsolete army stuff. Mine is still not working  either, but for different reasons. If I were you I'd check if such a shop is in your vicinity. Good luck!

[manvsmachine]

It is possible with some older ATX PSU to modify them. If you do that it acts like a normal power supply without a switch. We use them at school as power supply's  ;D

[NorthernLightX]

It is possible with some older ATX PSU to modify them. If you do that it acts like a normal power supply without a switch. We use them at school as power supply's  ;D

You can, but keep in mind that the trafo's used on those are often very small, so if your box draws some current they'll not be able to cope...

[Jidis]

There's also been talk here (I think) of the issue with not using the exact load they expect, and the possibility of ruining them. In ATX supply testing stuff I've read on the web, they're usually using some "dummy" load with them (lightbulbs,etc.).

George

[manvsmachine]

We use them only to power a 8051 microcontroller and to power some leds. That is not a problem. We use a philips pm2811 for testing.

[SounDuke]

I've finally drew out my own design for a power supply, using a 6+6 5A toroidal transformer (linear and noiseless!) . In serial, I'll obtain 12V 2,5A. I can work out the 5V ine by means of a Zener diode (5V) a resistor and a 3055 transistor. I'll post my schematic I'd be very happy to know the opinion of more skilled people since I never designed psu's. Toroidal transformers are cheap and I think it should be the time to solve ther lc psu trouble.

[manvsmachine]

I don't know how much current you need but you can use an 7805 too. 

[SounDuke]

I need more than 1A of current (2 cores with 4LCDS and over 128 20mA leds) so I must use the transistor. My doubt is: using toroidals, a 2x06 outputs in serial doubles the voltage but the current is the same right? (in my previous post I think I made a mistake by saying that the maximum current will be 2,5A). Thanks for support, I think an 80W toroidal transformer will satisfy my power needs

[moebius]

My doubt is: using toroidals, a 2x06 outputs in serial doubles the voltage but the current is the same right? (in my previous post I think I made a mistake by saying that the maximum current will be 2,5A). Thanks for support, I think an 80W toroidal transformer will satisfy my power needs

Yes, then it's 12V 5A - and for the regulation: I'd say - Go for LM317 with a heatsink. Was it a couple of added resistors (over 78** series) and this should give You precision supply with max 3A output.

Moebius

[manvsmachine]

Yes, then it's 12V 5A - and for the regulation: I'd say - Go for LM317 with a heatsink. Was it a couple of added resistors (over 78** series) and this should give You precision supply with max 3A output.

Moebius

an lm317 works fine too. a normal 7805 can not handle a current of 5A

I have to think about a power supply too. My 2 displays for the sequencer use together 900 mA and that is too much for 7805 on the core module. 

[SounDuke]

well an lm317 scares me a bit since the max current is 1,5A. with a 3055 npn transistor, I'll be sure I'll have no heat problems. The con is that protection in a 3055 is inexistent and a short circuit will definitely break it out. Let's do a comparison between 3055 and lm317 :)

[monokinetic]

My 2 displays for the sequencer use together 900 mA and that is too much for 7805 on the core module.

I was just about to start a thread on this and what do I find? Great coincidence, thats what I love about this place :)

I did a bit of googling and found this:

http://www.electronics-lab.com/projects/power/010/index.html

Which is a schematic for a variable power supply based around the LM317. They also have a calculator for working out the resistor values:

http://www.electronics-lab.com/articles/LM317/index.html

which I tweaked with a bit and the final value for R2 seems to be 720. I presume this is in Ohms?

So I will play around with laying this out on stripboard but 1st I have a few questiosn for those more experienced -

I have already bought a power supply for my Sequencer which is capable of outputing DC 5V-15V at up to 1000mA. From reading the above links and the national.com datasheet it seems that the input supply should be 1.5v more than the planned output, which is obviously 5v for the core. So if I input 6.5v that should be sufficient shouldn't it? I presume if I input more the Lm317 will convert this extra voltage into heat......

All the information I have found about designing this type of supply say using a heatsink is essential. I plan on following this advice but I havent ever used heatsinks before. Do I need to use heatsink compound between the LM317-TO220 and the heatsink? Or will it be sufficient to just clamp it on?

BTW the stripboard layout  found is here:

http://www.pinkfishmedia.net/forum/showpost.php?p=32069&postcount=19

The image has 2 possibilities, the left one seems to be "standard" whilst the right is more for audio applications. I'm going to use the left one because the core only needs a stable supply. What do you think?

I look forwrad to having my MBSeq on for longer than 5 minutes (ie before the 7805 is hot enough to fry and egg with all those lights and LCD's attached!)

Dave M

[SounDuke]

sorry but my aim in this topic was to introduce a different technique to handle larger currents. Since I have to power up 2 MBLC, I do not trust the lm317 which I think will not be able to handle a current of 5A even with the largest heatsink. That's why I asked  more skilled people if a proper alternative could be a 3055 npn and a 5V zener diode to handle the 3A I need for 2 cores (no prob for the motorfader modules which requires 1A)

[smashtv]

sorry but my aim in this topic was to introduce a different technique to handle larger currents. Since I have to power up 2 MBLC, I do not trust the lm317 which I think will not be able to handle a current of 5A even with the largest heatsink. That's why I asked  more skilled people if a proper alternative could be a 3055 npn and a 5V zener diode to handle the 3A I need for 2 cores (no prob for the motorfader modules which requires 1A)

I have seen the 3055 scheme work 24/7 for years on many of the Atari coinop games, but they always used a massive (talking huge) heatsink.  Every Centipede or Dig Dug game you ever put a coin in has 3055 kickin' one of the rails inside.  ;)

Check out this thread/post for yet another path to your goal: http://www.midibox.org/forum/index.php?topic=3244.msg21308#msg21308

Back to the order packing, smoke break over....  ;)

Best

Smash

[SounDuke]

Hi smash thanks for the advice. I began building the power supply connecting in parallel 3 7805's like described in the PDF (http://www.reed-electronics.com/ednmag/contents/images/72204di.pdf) page 3 you smartly showed me. I built it but because of the diode drop I obtain 4.7V from the last regulator when COM line is connected to ground. connecting com line to ground with a 1N4007 I can raise the voltage to 5.4V. Is it ok to power the pic with such voltage? The datasheet for the pic18f452 states that the supply range of cmos permits it (I hope i'm not wrong). If not, how could I compensate the voltage drop?

[SounDuke]

ahem... Sorry maybe I understood.. I forgot the the resistor ladder between ground and +5.4V ^_^ I can now decrease the voltage like japanese men do! RESISTOR-cap-RESISTOR-cap--------  ^^

[smashtv]

Hi smash thanks for the advice. I began building the power supply connecting in parallel 3 7805's like described in the PDF (http://www.reed-electronics.com/ednmag/contents/images/72204di.pdf) page 3 you smartly showed me. I built it but because of the diode drop I obtain 4.7V from the last regulator when COM line is connected to ground. connecting com line to ground with a 1N4007 I can raise the voltage to 5.4V. Is it ok to power the pic with such voltage? The datasheet for the pic18f452 states that the supply range of cmos permits it (I hope i'm not wrong). If not, how could I compensate the voltage drop?

Heya SD!

5.4v is definitely not right....looking at Figure 1 on the pdf, D1 and D2 are there to keep the regulators' outputs from backfeeding into each other, but due to the drop to 4.7 you have to float all of the regulators ground/com through the same type diode to compensate.....

So: the only thing the regulator coms should connect to is each other, and the anode side of  a like diode ("D" in the pdf) with that diode's cathode tying to ground.....If you connect the regulator com to ground anywhere but through "D", it won't work correctly.

Also to function as expected, all regulator outs have to hit the anode side of like diodes "D1, D2" and have the cathodes all tied together for the final output, and depending on the internals of the your regulators (they do vary from manufacturer to manufacturer) it might need a load present to stabilize at the correct voltage.

Simply put: Since the Vcom diode's drop exactly matches the drop from the Vout diodes, the only way you can get over or under 5v is if your regulator grounds/coms are tied directly to your circuit ground, or a diode is missing, etc.  Keep in mind that -most- of these regs have a grounded tab (pin2 connected to tab internally), that might be connected to your circuit ground through a heatsink or something when it should be isolated.

This circuit won't work for people who want to split the power circuits up, etc....all of the regulator output pins must tie together (via the respective diodes) to make up one power rail, otherwise the loading will be off balance and probably cause some noise.

Lemme know what ya find!  :)

Smash

[smashtv]

ahem... Sorry maybe I understood.. I forgot the the resistor ladder between ground and +5.4V ^_^ I can now decrease the voltage like japanese men do! RESISTOR-cap-RESISTOR-cap--------  ^^

No!

That resistor symbol there is your load......The designer is trying to tell you the circuit might test high or low without a bit of load there (depends on the regulator).

Just to be clear: all diodes must be same, and prefer from the same batch....

Strapping a resistor across the output rails would not be good, and it would have to be a large one, then that will eat lots of your power as heat....etc. 

Built exactly as shown the wrong voltage should turn back into 5v.... :)

Best

Smash

[SounDuke]

Hi smash thanks i was about to make a mess over here! Well, I followed the schematic for a 3A psu so I used 3x7805. I have the 5.4V exactly when I connect the diode between com and ground. I have not connected any load, The heatsink connects COM pin of each regulator but ground is absolutely not connected to com line. The only connection between heatsink (so com line) and ground is anode on common line and cathode on ground line, used the same diode type as specified (1N4007). Everything is like specified in the schematic, except the load. It is possible that if a load is connected it will stabilize. I'll try to connect a lamp to see if this happens ^_^

[SounDuke]

I put a load of a 1kohm resistor+a led. The reg's begin the regulation and the voltage drops to 5.1 .  Could you confirm it's okay? putting a load which brings the regs to 5V and then connecting the core should not harm the pic right?

[SounDuke]

sorry for the numerous posts, hey it was nearly 3 months ago that I last posted!  Smash, putting a 220ohm resistor between the 2 rails solved the problem: exactly 5V. I'm supplying the core with no problem now but still I must test it at full power.

[smashtv]

sorry for the numerous posts, hey it was nearly 3 months ago that I last posted!  Smash, putting a 220ohm resistor between the 2 rails solved the problem: exactly 5V. I'm supplying the core with no problem now but still I must test it at full power.

hehe no worries!  ;)

I'm guessing the 220 is providing the load that the regs want to see since it fired up as expected.  I'm also betting that the resistor is getting hot....not a permanent solution, and it's wasting some of that available 3amps as heat.  :-\

I would try an expendable device/circuit as a load, without the 220 resistor, and see if the voltage stays at 5.  My guess is the resistor/led combo did not present enough load (current limited to 25ma or less by the resistor), but a real-world load of an amp or so would show you exactly 5v.

This reminds me of an old Nintendo coinop/arcade game design, the main PCB did not draw enough load to stabilize the switching power supply, so they strapped a 15w low-ohm resistor complete with heatsink across the rails.  And boy did those run hot.  (was a common failure cause too, since the Nintendo engineers opted for connectors on the big resistor vs. solder, which would bake from the heat 24/7 and fail)

Let me know how it goes! ;)

Smash

[moebius]

Hmm..

And I would trust core /(midibox - would that be "an expendable device/circuit" Smash refers to??) to load PSU enough to stabilize PSU at a proper voltage.. 5.4V would not kill PIC/Logic anyway.

Moebius

p.s. Smashy also gave a nice tech tip there for those techy types: If you have a device that behaves strangly and it uses load resistors/ Power resistors at PSU - resolder those first.

[SounDuke]

I can't think of any device which consumes less than a 220 ohm resistor and gives the right load... maybe a 5V lamp? I think it consumes a lot more...