116 Controller 6 FET Input Regulator Mod

[12p3phPMDC]

http://focus.ti.com/lit/ds/symlink/tl783.pdf

I'm going to order a handful of these regulators in a TO220 pack to try and reduce the amount current that R1 draws on the 6 fet board.
IT will do so by eliminating R1 because there is no need to drop any voltage through a resistor.

To short R1 see Jeremy's low voltage mod thread:
http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=16651

It is a high voltage regulator that accepts up to 125V as is without requiring R1.
Perfect for a 100V controller!

It should make high voltage mods easier because its a drop in...

edit. As Jeremy, pointed out in so many words, this is basically moving the dissipation into the regulator. Since it's still a linear reg, (Vin - Vout) * Iout = Power Dissipated.

 

[12p3phPMDC]

It occured to me this morning, that it's much easier to heat sink a to-220 regulator package than a through hole carbon
resistor.

 

[TMaster]

I have a few of those regulators. They are adjustable; meaning you will need to do a small circuit with some caps and resistors to get the right output voltage. They are not plug in play like the standard linear regulators, if you don't already know.

Its way easier to just slap in one or two 3W resistors, from calculating the voltage you are putting in. How much voltage will you be running?

 

[12p3phPMDC]

TMaster,

I hear you, but I think that the topology or circuit of the LM317 is nearly identical to the tl783 if not identical....so yea, pretty much just like a standard linear regulator.
Am I missing something? I looked at the datasheets, and well, now I'm gonna go double check.....

I'm planning on 88V or 4 x 6 cell packs in series for 24s1p setup.
That should give about 12V margin on the caps and fets...

So it would be a matter of just changing R2 and R3 to set the voltage to 12V.

Admittedly, Neither mod is very hard.... I just want to do it this way. :wink: I think it's cool that this part will withstand 125V across the input and output...
that's pretty damn good for a little regulator...

 

[TMaster]

The benefit of the regulator is so you can have pretty much any input voltage and it will put out the same output no matter what. That regulator outputs 1.25V stand alone, meaning no resistors in the circuit. You will have to use one of the regulator circuits from the data sheet and calculate the resistances to get the right output voltage.

What is the minimum input voltage the controller can handle without damaging it? And what regulator is the first one in line taking the input voltage?

 

[Jeremy Harris]

TMaster,

If you take a look at the schematic for these controllers (see my post referred to at the start of this thread) you'll see that they use an LM317 variable voltage regulator already, with resistors (and a protection diode) that sets the output voltage to about 12V.

The 783 regulator uses the same pin out and the same 1.25V reference voltage, so should work as a direct drop in replacement, but with a much higher allowable input voltage.

The problem with the LM317 is that it's limited to 40V maximum at the input. To get around this limitation, the designers of these controllers added a big power resistor to lose a few volts in front of the regulator. It isn't elegant, but it does work, after a fashion. Replacing the LM317 with a 783 removes the need for the power resistor; no other changes should be necessary as far as I can see.

Jeremy

 

[TMaster]

Ok if it has the same vref then the resistors should be right and you wont have to change them. You may want to change the caps though. The 783 shows a 1uf for the input side and a 10uf for the output side. I see the data sheet on the 317 has smaller ones. ALso be sure the 1uf cap is 100V minimum capable. I used ceramic ones for a circuit I did with this regulator.

 

[12p3phPMDC]

Thanks for providing the explanation Jeremy....saved me the trouble. :wink:
5W dumped into the heat sink will raise the Rdson of the fets, but for fets that can handle 500W of dissipation,
it seems to be in the noise, like 1%. But, I'm starting to see switchers in three terminal packages... Maybe someone will market
a nice little three terminal module with 125V input, 12V out . On the heat sink extrusion/box that I bought for this controller, it shouldn't matter much due
to the thermal mass and the huge surface area. :wink:

I just got some of these bad boys too...WSMS5515 from Vishay...
http://www.vishay.com/company/press/releases/2008/080123resistor/
My supplier got me some samples....ahh....the joys of big ol' passive parts.

Tmaster,

There is a 100V large cap on the input, I don't have it in front of me.
If you read the posts, I said I should have a 12V margin on the caps and fets, therefore, 12+88 = 100V rating.
The caps on the 6 fet board are much larger than those suggested in the datasheet.

The capacitance values in the datasheet are not hard and fast values....They are suggestions for a regulator that will work.. Read the datasheet again.
Thanks for the suggestions, you have good intentions... Don't worry it will all be covered design wise ...keep reading.

 

[TMaster]

Post some pics when you get it running What are the amp capabilies going to be of the 6 fet controller?

 

[fechter]

Nice. I haven't seen that one before. That should work great as long as the total dissipation is not too high.

Do you know how much current the controller draws off the regulator? I think most of the ones I've measured run around 50ma or so.

Any idea how much those cost?

I've been trying to make a switching regulator out of an off-line SMPS chip. I have wall warts that run as low as 19v on the input and can go up to 230v, so I know it's possible. I also seem to recall there is a way to add an inductor and diode to a LM317 to make it run like a switcher too. I'll have to look for that schematic.

 

[Knuckles]

You can also shunt it with a power resistor. That way it won't get that hot at higher V drops.

Great Find. I will study this. 8)

12p3phPMDC, Please post a high res pic of the 6-Fet voltage regulator circuit.
Heck ... we can solve this like easy peasy ... ASAP. :lol:

OMG! I had it for years and never knew I had it ...
http://98.131.176.65/endless-sphere/tl783-2005.pdf
http://98.131.176.65/endless-sphere/tl783-2008.pdf

Anyway ... Let's make it work.

Heck ... This is the 6-Fet WITH the LM317 on the heat sink. Dang, just swap LM317 with the 783 and you should be golden!
You can still add a 5W resistor across the 783 Vin - Vout to take the heat from the fet sink if you want.
All depends on how much ... maybe R=V/I = 100/.03 = 3300 ohm.
(This creates an "any voltage" mod. NICE! )

PS The shorted R1 location is a nice place for a VR input cap ... say 160v ... plenty of room for it.
(or swap out the input cap that is there now ... hard to tell from the pics on the other thread)
http://www.endless-sphere.com/forums/viewtopic.php?f=2&t=16651&start=0

Anyway,
Very very cool. 8)
Let's take a nice CLOSE UP LOOK.
pics pics pics

-K

 

[12p3phPMDC]

I'm trying to eliminate the power resistor. If we can get the heat out of the enclosure, it should help the overall temperature
rating of the controller in general. By moving the heat into the regulator ( we all would like a switcher)....you can
sink the heat into the case. The temp inside the case will drop some. It's not much difference really.... But you are eliminating a component, but not cost unfortunately.

LM317 $0.60
TL783 $2.00

pick your poison.... no free rides until we come up with an expensive switcher :wink: :lol:

Jeremy posted some pics of the 6 fet in his low voltage mod. Same mods almost, bypass R1. Insert TL783..
Adjust R2 and R3 if necessary. Use high voltage caps.

The shunt is good for 3W....200 uOhm at 99 cents....Shunts for meters like this are typically much more expensive
althought this probably more expensive than the existing shunt.

No offense, but I think it's solved... We just have to install it and run it!

The help I need has to do with finding the minimum set of connections on the 6 fet controller to make it run.
I have hall sensors (need to install), 2 6 fet boards, regulator on the way...finish connecting the new fets.
I have some switches for on off, a pot throttle, batteries, enclosure, power cabling....
ITs all the other stuff that I need to pull low, high or whatever, unconnected...I need to compile all this stuff buried in threads.
I bought a turnigy watt meter too for diagnostic help.

I'll take some pics when the parts arrive.

 

[Knuckles]

markcycle has dirt cheap 72v to 24v external switching regulators. He uses them on his Kelly controllers.
They run ice cold and outside the case.

maybe this helps somewhat.

 

[12p3phPMDC]

Knuckles,

re: external switcher

That's a good idea...thanks!

I'm trying to keep the case small, but since I'm approaching
this using a different case, we may be able to accomodate a switcher like the one Fechter brought up....
and just mount in there nicely.

Fechter, do you have a model number you care to share?

 

[fechter]

I tried a Viper22, but no luck.
Another one is a KA5M0365RN
I sort of got this thing I had a work to go, but I can't remember the part number on that one.

 

[pa.hioficr]

TOPIC RESURRECTION

I just wanted to thank all the contributors to this topic, special thanks to 12p3phPMDC and Jeremy Harris!

To finish two versions of a master controller for optically isolating and controlling up to 4x CellLog8 (of any type) I needed an efficient, low cost, low parts count way to produce a 5v regulated supply from any pack voltage between 6S and 24S. I had been considering a switching supply but for low current consumption (MCU Atmega2560 or RPI with 3 x 25ma draw USB TTL Adapters so 100-300 ma roughly max when active mode) the expense and complexity just wasn't worth it. I thought perhaps someone made a drop in LM7812 or 7805 but with higher limit than 40v and since I found a TIP122 based any voltage mod thread I figured I'd search here a bit.

Needless to say, this is exactly what I need to complete my schematic and PCB design so thank you very much!

William