Midibox PSU

[Altitude]

i'm seeing the same thing. I have an alternate solution that I will post here later

[NorthernLightX]

Guys, please keep in mind I'm not an electonics guru (far from in fact), but you should at least know how a restistor works. Power, tension, current?

Basically, the resistor will only convert tension into heat if the current becomes high. This is exactly what you want, because else the reg can cope with the higher current. In theory, the 8 ohm resistor will eat 8 volts when the current is 1 ampere, so if your mains output is 10% below average, your 7805 will have to cope with about 7,8 volt. Perfect operating values.

So the design is allright, heatsink the rg properly, it doesn't matter if you build the C64PSU optimized thingy or this circuit, the 7805 will get hot anyway.

But I believe this has alredy been told in this very post.

modnaR: if your voltage drop over re resistor is about 1v, your load would have been about 120 mA on the 5V branch. If you oad it highr the resistor will drop more, so the heat of the reg wil stay the same.

Altitude PMed me with a similar question. Please post all your findings here, I hope it encourages people (or evenyually discourages people if it turns out this thing is not the way to go) to build this PSU.

I have been very busy lately, and have not had any time to work on the midibox or the psu. I hope I'll have any findings for you soon.

Cheers,  Alex.

[Mr modnaR]

i was told in another thread (about the pot that alters lcd backlight brightness), that the resistor acts as a current limiter, and does not drop the voltage significantly. i don't know who to believe.  ;D

[Altitude]

i think the problem is that the 7805 can only pull 1 amp max, so anything under that, there will be a ton of voltage to dissipate as heat and if your pulling 1 amp on the 5 volt rail, then you should be using a TO-3 size regulator.  I think Alex's design expects a huge draw from the regulator, and in practice, this does not really occur and as a result, the resistor does not really do much

it doesn't matter if you build the C64PSU optimized thingy or this circuit, the 7805 will get hot anyway.

Not that hot. Thats what i have been using until now and it does not even get half as hot, and thats with 4 sids and cores with high current LEDs

[Mr modnaR]

hmmm, so what's the solution? cascading voltage regulators? can't think properly at the moment, as i've just fried the fuse in a brand new 12VAC wall wart i got for this psu! i was testing this design on strip board, and had it layed copper tracks down on my work desk. little did i know that there was a renegade leg cut from a capacitor lurking around, and of course did not get noticed until i tried my 9VAC wall wart for my x0xb0x on it first! damn. darwin award here i come.....

[Mr modnaR]

ok, i think i've realised the problem. the resistor should have been used as a potential divider. that way the voltage across the 7805 would be dropped and should remain cool. does anyone think it could cause problems when the current draw on the regulator is high though?

[Altitude]

I found a schematic for a MOSFET based setup.  I'll post it when i get home

[Mr modnaR]

I found a schematic for a MOSFET based setup.  I'll post it when i get home

sounds interesting, aren't mosfets expensive? i only worry because of my clumsiness, and the likelyhood of draining my wallet dry replacing them....  ::);D

[NorthernLightX]

i think the problem is that the 7805 can only pull 1 amp max, so anything under that, there will be a ton of voltage to dissipate as heat and if your pulling 1 amp on the 5 volt rail, then you should be using a TO-3 size regulator.  I think Alex's design expects a huge draw from the regulator, and in practice, this does not really occur and as a result, the resistor does not really do much

Not that hot. Thats what i have been using until now and it does not even get half as hot, and thats with 4 sids and cores with high current LEDs

My design anticipates a draw of 1A on each rail. I've had the schamatic checked by several guys who DO know their electronics stuff (and I do not), and they approved of the design. You made your own schematic, are you sure you placed everything right?

I completed my PSU tonight, but I don't know when I'll be able to test it. As soon as I know anything I'll post it.

[Mr modnaR]

Not that hot. Thats what i have been using until now and it does not even get half as hot, and thats with 4 sids and cores with high current LEDs

My design anticipates a draw of 1A on each rail. I've had the schamatic checked by several guys who DO know their electronics stuff (and I do not), and they approved of the design. You made your own schematic, are you sure you placed everything right?

I completed my PSU tonight, but I don't know when I'll be able to test it. As soon as I know anything I'll post it.

[Altitude]

My schem is right, its based on Ray Wilsons with your added power resistor.  I spent a whole weekend reading up on this and your design is sound and fits everything I read.  I think the problem lies with the 7805 pulling a full amp, I will try to test mine on a pair of bench power supplies and see how much each rail is pulling but i seriously doubt mine was anything near 1 amp.  They do tend to get hot when they pull over 500 mA and that coupled with the high voltage results in a hot ass part.  I think going up to a to-3 size regulator with a big ass heat sink may be the way to go since a full C surface has tons of LED's and other stuff that will draw on that single little regulator and I can see it going over and amp easily.

I also found this in and issue of EDM, and it seems like a decent scheme to deal with the extra voltage issue

http://www.edn.com/index.asp?layout=article&articleid=CA6339248&spacedesc=designideas&industryid=44217#ref

[Mr modnaR]

this is how a potential divider works:

V2 = (R2 / (R1 + R2)) * V1

basically i think there needs to be two resistors to 'split' the voltage to give 5V (or probably 7V for the 7805). the only trouble with that is that the resistors would need to be able to disspate a fair amount of heat, and would need to be ones with high power ratings (=£$£$£$... :'(). i may be wrong though, i haven't been through in detail what resistances would be needed.

i think Alex had the right idea, but forgot the extra resistor.

[Mr modnaR]

ahh, ok, just read your post Altitude, yeah that's a good design, but how do you work out the values required?

[Altitude]

I have not done the math yet, but the only thing that needs to be adjusted is the resistance values for the voltage divider, everything else stays the same.  Heres a good site re: voltage dividers (especially the part about how they work under load)

http://www.play-hookey.com/dc_theory/voltage_divider.html

here is my preliminary schem:

[Mr modnaR]

looks good, let us know how you go with that one. i'm thinking now along the lines of getting a 2x6V trafo and using one 6VAC to make the 5V, then using the whole thing together to make the 12V. i only have cores and sids to worry about, so i don't need the -12V.

[Altitude]

More developments:

I tested the current draw on my rig (4 6581s and Cores + DIN + DOUT + 8 x Bankstick + LTC) and I am drawing 420 mA on the 5v and 130 mA on the 12v

[Mr modnaR]

really? that's crazy! what about when it's producing sound? also, you don't mention an LCD, did you use one?

[Altitude]

yep, using Tim's (smashTV) LCD.  It really does not go up when producing sound (even in 4x link mode)

[Mr modnaR]

is that the full control surface? you only list DIN and DOUT, perhaps the extra DOUTs needed (especially with all the mod matrix LEDs on) would increase the load?

[Mr modnaR]

another question: do you plan to have a transformer sitting in your midi box? if so, do you need to shield it in any way to prevent noise? i'm going to put a trafo in my 4xMBSID, and the layout i'm working with at the moment has the transformer right up close to the audio outputs (i'm hoping to put them all on the same pcb if possible). anyone seeing any problems?

[Altitude]

its the B surface so there is no matrix or other stuff.  I am using a shielded transformer and a filtered power entry module, both will be located inside the case.  I have not adjusted the LED resistors so each led pulls close to 100 mA which is really high (i am using bright blue ones) so there is no way that the remaining stuff should pull another 500 mA from the 5 volt line

[stryd_one]

transformer right up close to the audio outputs

Sounds like a bad idea to me.... I'd have it as far away as possible....Actually I'd have everything as far away from the audio as possible. Not a 'must', but I am sure it won't hurt!

[Mr modnaR]

yeah that'swhat i was thinking, o well, redesign no. 437 coming up.....

[Mr modnaR]

its the B surface so there is no matrix or other stuff.  I am using a shielded transformer and a filtered power entry module, both will be located inside the case.  I have not adjusted the LED resistors so each led pulls close to 100 mA which is really high (i am using bright blue ones) so there is no way that the remaining stuff should pull another 500 mA from the 5 volt line

there's what 7 LEDs in the b control surface? there's 99 in the C. surely that would draw significantly more??

[Altitude]

We will see. I dont think it will go up by a factor of two though