Doubling up a regulator for more current ??

[madhippiescientist]

The stepper motor circuit (timer, dividers, shift register) is always on (connected to the 7810 is only the motor). The minimum draw is several hundred ma, spiking up to circa 1.8A when the servo motor plots activity (the servo shares the 7810).

If the current never drops below a certain level you could connect a resistor between the input and output of the LM7810.

Suppose the regulator can continuously supply 1A but you need 1.8A and the minimum load is 1A. If you connected a resistor that always passes 1A from the input to output , the regulation won't suffer as long as the minimm load current is drawn.

Could you explain how to determine the value of this resistor? Tell me most importantly, if the draw varies and the resistor contributes, will the voltage vary as well (I need it to STAY at a VERY specific voltage level somewhere within a few tenths of 10v).

 

[Mr RB]

Ok now we know a bit more about the application it might deserve it's own thread!

If you want a very accurate drum/pen recorder there are simple ways. In the '60's and '70's they used to make drum and chart recorders with synchronous motors that would run from the mains, these motor turn at an exact speed that is locked to the mains frequency and is regulated for accuracy by the government.

You can get a synchronous motor that already has a gearbox built in, from any old microwave oven, they use them as the turntable motor and usually run direct from mains at about 2 watts and about 1 rev every 5 seconds or so. Then just use some gearing on it.

A 555 timer is completely unsuitable for accurate drum rotation because of the large temperature drift.

Anyway this is an involved project for someone who is struggling with basic regulator stuff. It's probably best if you make a new thread with a name like "help with accurate drum/stylus recorder" or similar and break the project down into the main parts that you need to get done, then people who have experience with that type of thing may be able to help.

 

[mneary]

I didn't try a crystal because in time the gearbox/etc will develop more/less drag, and I didn't know how much drag I would have to begin with

If you're driving a stepper motor, the speed won't change until you stall and it begins to skip steps. These skipped steps won't be consistent and changing the clock cannot correct for it. Drag has no effect until you begin stalling. A constant clock is your only defense.

As Mr RB said, the best solution for accurate speed is a mains-driven synchronous motor. Second-best is a crystal clock divided to run a stepper or synchronous motor (your best backup when power is out). But even a 10ppm crystal can stray up to 6 seconds per week, so it should only be used as a backup when the mains fail.

Analog (555) timers are lucky to keep 2000ppm even with a beautiful design and perfect voltage and temperature. (That's 10 minutes per week).

 

[unclejed613]

https://www.electro-tech-online.com/custompdfs/2010/02/LB-51.pdf

the article describes a way to parallel 3-terminal regulators by using resistors to balance the current.

 

[madhippiescientist]

Guys and/or Gals,

In regards to keeping the drum accurate with an astable/shift register/stepper control, it actually does do very well. I thought of using a synchronous motor, but when I lose the power and go to battery I would be up spit's creek (unless I had an inverter). I would still have no control over timing the drum with original and/or long-time variations in drag (the original drag added over 800 pps). After the initial run, and it being terribly off on 1rph to the drum, for a few weeks now, I've been using the circuit to fine tune the drum (see next paragraph). Also be advised though... I can't simply run the drum from a motor because I also have a threaded rod which provides linear motion to the inking station that travels the length of the drum, and it rotates at a different speed than the drum. I have to be able to reverse the motor direction because after a week's worth of plotting, it needs to start the stylus back the other way (turn the threaded rod a different direction) or I would need to manually disengage the rod (which I tried with a coupling split in half, one welded to the stylus plenum and the other bolting to it, but it was mechanically hard to remove & slide and would also pose problems with how tight to bolt the removable half). The reversible motor allows me to simply flip a switch, after greasing and fine tightening of the threaded coupling to ensure no binding/noise throughout the entire length of the threaded rod. "This drum is not a cash register type paper roll being fed" plotter like you see in many of the science shows. It is if you can imagine, a drum the width of the seismometer, plotting left to right and right to left as the drum rotates forward and backward.
Timing. I have got it down to the drum rotating about 8 hours, and per my ledger sized plotter excel grid paper, it is only about 30 seconds off <4 seconds per hour> (just needs finer tuning). I have a 20 turn pot, that I keep turning about 15 degrees after each test, and I'm STILL closing in on more accuracy (though my 3.5 meter still reads 773x). The case is enclosed so it's not susceptible to drafts, and temperature co-efficiency is not a problem (it's in a residency) since I got the capacitor with a better tolerance. I know I'm going to get it VERY close, upon the use of a 4+ digit frequency counter, but simply have concerns (after "final" timing) of any VCC variance (to the 5v regulator for the astable timer), especially if the 10v wallwart / 12v battery has sudden shifts in draw/voltage (either due to power cut out or cut in, and the servo's sometimes ravid activity, like when my neighbor's run up and down the stairs!). Also remember everyone, I'm sooo close to finished already, and 90% has already been constructed... Please, no more on my design... Help me supply 10v2A, that will keep it's voltage (circa 10v) at a VERY stable level! My servo may burn if the voltage varies too high, and my 12v stepper may skip if the 10v (which DOES drive it well) drops who knows how far down...

 

[madhippiescientist]

the article describes a way to parallel 3-terminal regulators by using resistors to balance the current.

This article is utilizing the parallelization (is that a word?) of the $23 LM338 which is what I was trying to avoid (I only need 2A, not 10A or more). Try to think, the 7810 (in parallel? <or with a passQ>). I know I could find a circuit to simply do the 338 and set it to 10v, but it's very expensive (compared to the $1 7810), much much more space, needs an expensive heatsink (which I don't have), and is 250% of what I need at a infrequent maximum. Thanks for trying !!

 

[dougy83]

Just because your seismometer writes a continuous line doesn't mean you have to (the enter key works quite well to create spacing between paragraphs).

Just use the 7810 with a pass transistor and be done with it; it's a tried and true method. It doesn't require any minimum current draw (like the resistor paralleled with the regulator) to regulate. Not to mention it's straight-forward & simple.

 

[madhippiescientist]

dougy83...

Believe me, I'm going to try your method, with mneary's suggested value (.5 ohm) for an Rsource (as is described in the 78series spec sheets) and I'll do it at 15W or bigger. Then I'll see how stable the voltage remains with and without servo activity / during power cutout and re-instatement, etc. Your suggested method is so far the best that I have seen. Since it's finally my Friday, workbench, here I come!

Sorry I forgot to skip a line. The seismo writes a squiggly (continuous) line, and maybe just maybe it's affecting my brain!!

 

[mneary]

Did you say battery? As in lead-acid? Some of the suggestions intended to save your 10V regulator from overheating on 14.5V will fall out of regulation with only 12V input.

Oh, are heatsinks expensive? Not where I shop.... **broken link removed** You'll need two of these, 45 cents each. You have to get rid of 9W at (14.5V-10V)*2A. Really, it won't go away. A resistor will get as hot as anything else does, and a 10W resistor isn't cheaper than a 10W power transistor.

.....just saw your post. 0.5 ohm resistor will help keep the regulator cooler at 14.5 input, but the regulator WILL NOT work if the input is 12V, because 0.5 ohm will reduce the regulator input to 11V which is TOO LOW. My suggestion was for 14.5V NOT 12V.

I am sorry if I misled you. I've been working with the assumption that the input is 14.5V.

 

[dougy83]

My suggestion was for 14.5V NOT 12V.I am sorry if I misled you. I've been working with the assumption that the input is 14.5V.

I don't think any of the suggestions will work with 12V.

 

[madhippiescientist]

The input MAY be as high as 14.5v (from the sealed gel-cell, or 12v from the wallwart <which may be a half a volt or so more or less than 12>). As the battery drains when in back-up power mode I still need circa 10v (to stay at it's stable voltage level). I know when VCC gets to circa 11.5v I'll start having problems. The heatsink I was speaking 'bout was for a TO-3 (for a 5A adjustable regulator) which I wish to avoid. I have plenty of TO-220 heatsinks in my supplies already. OK, anybody else? Any other ideas?

 

[Hero999]

The stepper motor circuit (timer, dividers, shift register) is always on (connected to the 7810 is only the motor). The minimum draw is several hundred ma, spiking up to circa 1.8A when the servo motor plots activity (the servo shares the 7810).

Could you explain how to determine the value of this resistor? Tell me most importantly, if the draw varies and the resistor contributes, will the voltage vary as well (I need it to STAY at a VERY specific voltage level somewhere within a few tenths of 10v).

The minimum load you've specified is too low for my idea to work so there's no point in me posting it, unless you're really interested?

 

[mneary]

A 7810 won't give you a dependable 10V out with 11.5 in , even at low current. 7810 needs 2.0V overhead at 1A, which is 12.0V in. Never mind the overhead of a boosting circuit. When your original question was about putting 7810s in parallel, I mistakenly assumed that you had already verified that there was enough input voltage.

LT1258 will do the job ($4.35 ea, TO-220) if you have a good heat sink. If your 1.8A draw is brief enough (25% duty cycle max averaged over a two second period) and your average consumption is 1A or less, then the temperature should be manageable with a 10C/watt heat sink. It will get too hot to touch, but the LT1258 is good to a junction temp of 125C. Running this hot isn't a good idea, but your requirements are a bit demanding. If you need 1800mA continuous or for periods > 2 seconds, then you should shoot for a much bigger heat sink.

You can get a general idea of how big a 10C/W heat sink needs to be (and compare it to yours) by doing a parametric search on Digikey.

[edit] Bad mistake - LT1528 is the regulator I have in mind [/edit]

 

[unclejed613]

if you're running from battery power, then look up LDO regulators (Low Drop Out) that will operate down to Vout+0.6V. otherwise your "standard" 7810 will stop working as soon as the battery terminal voltage drops to about 12V. maybe look up a "buck-boost" or flyback regulator that will keep supplying 10V as long as the battery voltage is anywhere between 5-15V. really, modern power supply technology can be your friend....

 

[Ubergeek63]

The input MAY be as high as 14.5v (from the sealed gel-cell, or 12v from the wallwart <which may be a half a volt or so more or less than 12>). As the battery drains when in back-up power mode I still need circa 10v (to stay at it's stable voltage level). I know when VCC gets to circa 11.5v I'll start having problems. The heatsink I was speaking 'bout was for a TO-3 (for a 5A adjustable regulator) which I wish to avoid. I have plenty of TO-220 heatsinks in my supplies already. OK, anybody else? Any other ideas?

actually a crap 12V wall wart peaks at 20V at no load and voltage variation, even without surges (20% variation and 41% for AC peak)

you would be much better off leaving it unregulated and using a current sourcing stepper driver.

 

[madhippiescientist]

The 10V 1.0A wall wart is always very hot, and it's draw peaks at around 1.8A upon the servo's sudden activity (the servo/driver is always on, but it plots left to right at a very rapid rate when drawing lines back and forth as the pendulum magnet moves in the coil). After the servo starts, it (all of the 10V circuitry) only draws back and forth between 1.0A and about 1.3A, decreasing as the plotting slows down / decreases. I'm aiming at eliminating the wall wart, and making my own power supply, hence the need for 10V of regulation, and desirably a little more current than the wall wart claims to supply (on a steady basis). I won't need over 1.0A unless it's plotting activity, but simply wish to take it easy on the power supply's regulator, and sometimes when plotting, it does so for a minute or two (so the duty cycle can be for a long duration above 1A). I'm aware the wall wart supplies "excessive" voltage without draw (the meter's reading), but all of my readings are while the device is under operation (and it runs circa 10v, so I know the wall wart is "doing the job" <but gets hot!>, so I don't wish to utilize it, and even if I do, I surely don't want to have to replace it every six months or a year).

11.5v or 12v, whatever the cutout I can live with, because the battery will have it's own regulation (ideally), and the two of them (battery / AC=>DC Supply) will be switched (in a power outage) after their individual regulators. I just designed the circuit to run as long as possible under battery backup because of "cutout" (if I ran it at 12V it wouldn't last long due to this "circa 2V cutout"). I vaguely noted 12V for the power supply, because I plan to eliminate the wall wart and create a 13.8V ac/dc converter (and then down to 12v and then 10v) for the power supply. Hero999, I'm interested in anything you think might work !! mneary, I will check into your LT1258 suggestion !! Ubergeek63, are you speaking of a "current sourcing stepper driver" for the drum rotation motor or the plotting stylus servo? Or is this some sort of power supply I've never heard of?? If it's the stepper motor for the drum, I am using pairs of darlington drivers on each phase (if that's what you're talking about). Or, are you suggesting I eliminate the stepper motor's regulation, and if I do so, won't that inflict the motor's possiblity of proportionately missing phases as the battery drains? unclejed613, I'll also look into some of the terminology you wrote about, though I have no idea what you're actually speaking of (the product, not the terminology).

 

[madhippiescientist]

The "Flyback Regulator" will cost more money, and take more space than an adjustable. I checked into the LT1258 and couldn't find anything until I tried LTC1258, but even then they note very small amperages. Enough. I know I've been toooo specific with all of my requests and restrictions, so I'm going to seek ending this all now. I will make either an LM317 (1.5A) or LM338 (5.0A) adjustable regulator for my power management, including feedback blocking diodes and the whole nine yards. But I do seek one more yard...

If I want to utilize a Zener Diode reverse-biased onto the output (to stabilize the adjustable regulator even further), at 2A draw / 10V, I need to know how much wattage I MAY draw (I figure it to be 20W). But the largest Z I find is the 1N5347 (5W). Any suggestions?

I appreciate everybody's input, but this has drawn out too long, and it isn't worth the $40 it may cost me to just do the LM338 thing. Who knows, maybe I'll try it with the LM317. Anybody with any feedback on my choices (re maxing out at 1.8A VERY shortly but drawing around 1.3A when the stylus is active, meaning, would it be wise to try the LM317??).

 

[unclejed613]

build a 14V 3A supply, use the regulator for the driver electronics and run the stepper motor off the unregulated supply. that's the way it's done most often. the stepper motor skipping and stalling is because it's running off of a severely current limited supply. and have a large filter cap across the supply too (10,000uf or more) to supply higher peak currents without the supply itself seeing much variation in current. don't put a raw zener across the output of the supply, it tends to let the smoke out quite rapidly.

 

[mneary]

I checked into the LT1258 and couldn't find anything until I tried LTC1258, but even then they note very small amperages.

My bad - I meant LT1528. I just changed my message above so it wouldn't confuse people in the future.

 

[madhippiescientist]

I already have a large filter capacitor. The motor isn't skipping steps with the 10V 1A wall wart, I just want to build a power supply to eliminate the wall wart, since the seismometer is also going to have a battery charger, I want just one power cord to it. In re a zener for regulating small voltage variations, I always thought you did this with reverse biasing (I'm not knowledgeable about the use of a Z), so, how can I use one to regulate small variations (if it's possible), and where can I find one for my 10v load (I still need the stepper to run as the backup battery drains, which is why I'm not running it at 12V or higher due to regulator drop out and most importantly battery drainage). Ideally a maximum draw up to 2A, even if the available supply is bigger).
LT1528. I finally found an app sheet that would load and also showed me how to calculate the desired voltage output. It also states no protection diodes needed, and what appears to be a minute load regulation variance and drop out V, that seems pretty stable. I think I'm going to have to try this one! Their formula (see attachment) lists a formula on page 7, that indicates Vout=(3.3v*(1+(R2/R1))+(Isense*R2), and with their resistance examples on page 1, I figured the needed R2 as 680 ohms utilizing the maximum suggested value of R1 at 330 ohms. Tell me guys if I'm wrong, and what value R2 should be. I'm a little offset by the "Isense" thing, but it seems to be such a small value that it really won't matter all that much. How about a zener on this circuit as well?

https://www.sparkfun.com/datasheets/Components/SMD/LT1528.pdf