Strange psu/radio issue

[adibrook]

Im having some strange problems with a project.

Basically, the project itself, to give it a brief backround, is a tiny radio controlled vehicle with an infra red camera mounted on it. Its designed to be lowered into an abandonned wwii tunnel system, but since the only hole into it is rather small, this has been abit of a challenge to build.

The system consists of the video stuff, which is a camera, mic, illumination controller and IRled's, and also transmitter, in a metal case, and also the radio control module, which was scavanged from a radio controlled toy.

Now...the video module absolutely must run on at least 9.2v, otherwise the automatic illumination system dont like it and you get erratic illumination.

The radio module must run on less then 6v, because otherwise the dinky little power transistors on it will burn out. I cant change the power transitors to bigger ones as unfortunatly i stupidly encased the whole thing in epoxy alredy.

The battery i chose to power it all is a 9.6v unit, as its small enough and light enough to power the vehicle, yet powerful enough to power the illumination system, which is pretty beefy.

So the whole thing looks like this

http://img.photobucket.com/albums/v4...rook/psu-1.gif

I decided to use a LM317 regulator, as a last minute attempt to get it running last night, to power the RC module.

It all seems to work. The LM317 is putting out about 5.something V, which is good enough, the RC module is supposed to run on 4.5v but will run on anything up to 6v ok.

Exept...for some reason, there is a massive drop in radio range.

The RC module is basically a single chip 45MHz receiver, which has almost no external components, just a little preset tuning coil and thats about it it seems, one of those ''all in one'' chips, and also power transistors to allow the chip to control motors.

When the RC module is powered by a 6v dry battery, or a 4.5v battery pack, it works absolutely fine. But when powered by the voltage regulator from the main 9.6v battery, it goes wierd. It seems like the chips on-board radio signal gain drops dramatically. It still demodulates OK, so it powers the motors on and off cleanly when it gets the appropriate signal, but it only works liturally with the transmitter within 30cm of the antenna, any further and it just stops dead. Connect to the 4.5v pack, and it works fine again! Or the 6v battery.

I dont know how it can be effected like this...perhaps there is some kind of wierd radio feedback going on. I tried putting a RFC and decoupling cap between the regulator and the RC module, but it made no difference.

The regulators voltage output is nice and even, its waaaaaay below its maximum current even with all motors runnins, and i cant see anything wierd on the power rails with the scope.

i cant figure it out. Why would the receiver chip in the RC module not like being powered through a voltage regulator?

[Skywave]

Hi Adi -
Sounds like instability to me - which, when the scope is connected, will mask that instability - and may even give the necessary decoupling .

If you haven't already done so, try this:

Make sure that the 780R has fairly short leads and close to the LM317.
Wire the 0.1 uF and 1 uF decouplers directly from ther respective connections on the LM317 to the 780R. Now wire this common point to the -ve terminal of the 9v battery.
Run a second, seperate wire from the -ve connection of the RC module to the -ve terminal of the 9v battery.
Run a third, seperate wire from the -ve connection of the camera to the -ve terminal of the 9v. battery.

That should eliminate all earth loops.

Al / Skywave.

[KeithsTV]

Agree with everything Al has mentioned. It may be worth changing the 1uF cap to a higher value as some of these regulators need a minimum value in that position to keep them stable.

Keith

[XTC]

I agree with Al and Keith.

Something I'd do which should be easy and quick, is to connect a 0.1µF or 0.47µF ceramic across the electrolytic. Electrolytics have a resonant frequency. I'd also try putting a small electrolytic on the supply side of the reg.


Years ago, when IC regs first came out, they were notorious for instability and going into oscillation. Usually in the datasheet there's some mention of connecting a 0.1µF cap between a couple of the pins as close to the chip as possible. It's not a thing I've noticed and I think that as time went by the problem was lessened with later production chips, but it's possible you've drawn the short straw, or that the LM317 is still susceptible to this.

I suppose another possibility is to try it with a fixed 5V reg.

Pete.

[Skywave]

Having read the two Posts above, a foggy memory has come back to me - I'm sure I've experienced this type of trouble with an LM317 (or was it a 323K ?) in the past.

If the above suggestions fail - or, even, in addition to - replace the I/P and O/P caps. on the regulator with 10uF units and wire 0.01 uF (polyester will do) in parallel with each electrolytic. Try to keep the wiring to and from the regulator fairly short and direct; avoid using thin wire for this.

As for a choke in the +ve feed, I wouldn't fit this: it's likely to undo the decoupling that you have then achieved, or worse still, convert the whole reg. cct. into a low power transmitter !

HTH

Al / Skywave.

[KeithsTV]

I wouldn't fit an inductor in the supply to the regulator. This can make the situation worse if the regulator is oscillating. I once had a regulator with an inductor in the supply and a tantalum cap on the regulator input. The regulator was oscillating and the resulting current throught the cap was too much for it and it blew releasing the all important smoke. I changed the cap to a normal electrolytic and all was well.

This type of regulator needs a minimum load current to maintain stability, usually between 3 & 10mA. This can be achieved by selecting the voltage setting resistors to provide this load. The values in the original post would give a load of about 5mA. It may be worth checking if an additional load resistor across the output would help.

Keith

[adibrook]

Hey.

Thanks for all the help guys.

I gave up and found a fixed 5V positive regulator from a computer PSU.

Fitted it with a couple of capacitors, as per the datasheet. Also changed the earths around to how you described.

And....its STILL doing it! Same symptoms as before.

UUUGH. I just cant figure it out. A fixed regulator would be much more stable then a variable one...right? But its still doing it.

Just as well that i seemed to have fried the output transistors on the radio control module too! they are little TO-92 devices, and thers about 4 or 5 of them. When thers a logic 1 on one pin of the chip, they connect the motor to the power rails. When thers a logic 1 on another pin, they connect the motor back to front to run it in reverse. there are two of these groups, one for each motor. One of the transitors suddenly overheated. i dont know why, but thesame thing happend to the other prototype too.

I think the extra strain of pushing around a much heavier machine then the original toy is taking its toll, the motor needs too much current moving from a stand still.

I think ill just replace the transistors with bigger ones.

But that still leaves the strange psu behavior. Its rather annoying. This thing is heavy enough as it is, and on board space is very very limited. The machine hardly moves as it is from being much heavier then originally anticipated.

Because of this, it would be a massive advantage to have only one battery, so carrying around two batteries and making it EVEN more overloaded isn't necessary.

[dominicbeesley]

Does the RC module work when the camera's not connected (i.e. through the regulator on its own?) Could it be RF pickup of the camera output overloading the RC module? If that's the case you might need some filtering in the aerial circuits?

AFAIK the 5V fixed regulators are pretty much the same sort of thing, only with the resistors preset, so they may well mis-behave the same.

In the past I've sometimes improved things by fitting a larger than spec capacitor at the output. i.e. 10u then go down in 100s until you get to pfs i.e. 10u, 100n, 1n, 10pf all soldered together then make a star earth.

Another thing to try would be a string of diodes the forward drop will do the same job as the regulator, though it won't be stabalised. But that won't matter if the RC thing works with 4.5 to 6V. 6 diodes 0.7V drop each will give about 5.4V out or 5 will give 6.1V with plenty of room to run down! You might want to make a string of diodes and an LED such that it goes out when the battery is getting low, and put it somewhere ahead of the camera - then you'll know when its time to come home!

Good luck getting the pictures, do tell us where they are when you've got em!

Dom

Dom

[kalee20]

Some regulators are rather noisy, and it could be that there is enough RF noise to interfere with proper reception. I had big trouble using an LM317 to regulate the 2V supply to a battery radio's filaments! So:

Disconnect the input and output of the regulator, and try running the RC module from a 6V battery (even if impractical) just to re-confirm the range. Then, reconnect the regulator but have it feeding into a dummy load leaving the RC operating from the battery, and see if the range reduces.

[Antlong]

Have you tried screening the regulator?

You could also try an opamp-based regulator which shouldn't generate hash.

Regards Ant

[XTC]

Originally Posted by adibrook

I think the extra strain of pushing around a much heavier machine then the original toy is taking its toll, the motor needs too much current moving from a stand still.

I think ill just replace the transistors with bigger ones.

But that still leaves the strange psu behavior. Its rather annoying. This thing is heavy enough as it is, and on board space is very very limited. The machine hardly moves as it is from being much heavier then originally anticipated.

Because of this, it would be a massive advantage to have only one battery, so carrying around two batteries and making it EVEN more overloaded isn't necessary.

Adi,

Sorry to hear that this irritating problem hasn't gone away. A solution you've found does work is to use a separate battery for the RC module, but you say that the extra weight is too much for the toy on which this was based. The weight of 4 AA NiMH cells isn't all that much. This device is to be lowered into tunnels and inevitably there'll be dirt and debris to plough through or go over.

From what you say, the mechanical basis for this vehicle is distinctly marginal and I think it would rapidly be stuck and lost.


I think you'd be better going back to square one and finding a better chassis.

Anyway, as for the problem with the regulator. You've tried another regulator, altered the earthing and fitted extra caps and the problem remains. I'd be inclined to think that the RC module was picking up hash from something else, maybe the camera unit. That doesn't really explain why the RC module works with its own battery but not with the regulator.

Have you tried investigating the RF churned out by the camera unit with a radio, preferably a scanner tuned to the frequency the RC unit works on? Of course, the hash could be coming through the supply line and through the reg. If you power up the camera unit and put a scope on the supply line to it, is there a lot of noise?



Pete.

[adibrook]

Hey, thanks for the responces.

Without wishing to go too far off topic, but i think its worth mentioning just to clear up the point above. The machine itself is suitable for the kind of terrain that would be found in there. Designing it has been a real challenge as it has to fit into a 200mm x 200mm vent, and is one giant compromise. But its safe to say that if it ACTUALLY worked, it would be adequate for the job.

This is machine No 4. So far, all attempts have failed. Stuff just doesnt work. Transistors blow for no reason, or something brakes, or i accidently push the only working prototype into a drill which catches its wiring loom and rips it appart, etc. Its very spooky...EVERYTHING goes wrong with this project.

But aaaaaanyway, back on topic. The video gear is in a screened enclosure, and is very 'clean' electrically. It has a camera, a mic, the illumination controller and IRled's and the 2.4GHz transmitter, and also filtering for its power supply.

And also the symptoms persist even with the video module entirely removed!

Will have a look at the fried power transistor today...or tomorrow...i have like the king of all flu's/colds at the moment and also i ran out of solder.

By the way, the project itself is pretty interasting, if anyone does happen to want to know all the detiels, PM me.

[XTC]

So, you've eliminated the camera unit and it's down to the battery, the reg and the RC module. You've tried another reg - it was possible that the problem was with the LM317 specifically. The problem still occurs, but you don't see it with a battery powering the RC module.

You've adjusted the earthing arrangements and the problem persists.

Have you put a scope on the power line? If so, do you see any hash or does the problem just go away when you attach the probe and return when you remove it?

Have you put electrolytic caps on either side of the reg and have you also put 0.1µF caps across them? Have you tried bigger electrolytics, maybe 100µF?

If you've done that and the problem still persists, I'd be inclined to make a regulator from discrete components, choosing a pass transistor without too high an Ft and put electrolytics on either side of it. Normally, if you want a couple of amps or less, this approach makes no sense as three terminal regs are cheap, compact and perform well, but in this case, you don't need very good regulation and for some reason the IC regs don't work with the battery and the RC unit.

I can't think what else to suggest.

Pete.

[adibrook]

Hey....well...i got more progress.

Sorry iv been busy bashign my head against the wall with frustration, i havnt replied to pm's and to this thread for a bit.

I fixed the problem...in a completely different approach.

The RC modules chip has a maximum supply of 14v i think, so that wasnt the problem. The problem with running the rc module on 9.6v is that the power transistors, which were MPS8050's and MPS8550's, were just no big enough to dissipate the heat from the current of starting the motor on 9.6v. They WERE to92 cased devices, and since the motor is under so much more load pushing the buggy around, and the increased voltage, they just counldn take it.

So i just replaced them with some NECD882's and NECB772's which i scavanged out of a radio control hovercraft power control board.

These are to126 cases, and look much more rugged. Although the datassheet says they only dissipate 1W as opposed to 650mW for the original ones, with a heatsink on them i think they can survive much more.

So i mounted these on a remote circuit board with heatsinks.

problem solved. Now i can power the whole device from a 9.6v battery, and it seems to even work ok.

[OErjan_S]

would be nice to see some clips later. always interesting.