Servo current spike

Well the difference between different batteries is the internal resistance of the battery, and it may vary under load, and that certainly can make a difference in the situation you describe.

Let us know how you get on with the diode.

That sounds like a creative fix for the problem. Thanks for the description.

Yeah, I can't argue that I suppose. Obviously the battery has a lower series resistance than the 9 volt power supply, but it would have to drop well below 7 volts to affect the output of the 5 volt regulator. Unless the issue is the 5 volt regulator having too high a drop out voltage. Do you have tools like a scope and a variable power supply? I'd like to see the 9 volt rail when the servo comes on.

You also said a carbon-zinc battery didn't work, so it clearly is a power supply issue, but not necessarily that the PSU isn't "good".

Actually, the PSU units I looked at like this one were only rated for 5 volts and they ran the 3 volt regulator off the 5 volt rail loading it even more.

No, I don't think you can clearly say where the fault is until you get better data on the 9 volt rail as well as the 5 volt rail.

The diode and cap would be a band-aid, but it would still solve the problem no matter where the source of the problem is. You do understand how it works, right?

I'm not following this at all. What does the power switch have to do with it? Isn't the CPU turning the servo on and off? Or are you saying the servo draws power when power is turned on preventing the CPU from booting? I had the impression the CPU boots, but then the servo comes on and the CPU reboots with this cycle repeating.

Even 120 mA is a bit high for a CPU board. The receiver should only be drawing a few mA and the CPU should be a few 10's of mA. What else is drawing current?

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

If the power mdule LED goes dark that's a sign that it's been overloaded and gone into shutdown.

some powersupplies just go out or regulation when overloaded (the voltage droops) others go into temporary shutdown. it seems you have this type. it's going to be hard to run your servo from that power supply

If it's only 700mA that's needed then a larger powersupply seems like the obvious solution,

--
This email has not been checked by half-arsed antivirus software

When I put only the servo and the power module on a breadboard, and power that with the alkaline battery, and turn on the power, the power module LED comes on and stays on. If I use the other 9V supply, that lED cycles on and off. That's with no PSU load at all.

I do understand how the diode/cap works, but I don't think it works if the 5V supply continues to cycle off and on, and that's what's happening.

Yeah, I've tried to explain the setup but just haven't done that adequately I guess. The CPU is NOT turning the servo on and off. Both the CPU and the servo are tied to the same

5V rail. However, if I disconnect the jumper tieing the servo V+ to 5V, and let the PSU boot up normally and start sending PWM, THEN stick the servo jumper into the 5V rail, the whole circuit starts power cycling.

I don't know about carbon zinc batteries, but the 9V wall wart I'm trying to use is a switching supply rated at 650 ma. And, you know, that's the problem. I thought it would work fine at 120 ma, and it does, but not at 700.

That's the total draw of the entire circuit when the servo is being moved. The servo is about 75 ma of that, and the rest is the PSU and other stuff. That's using the alkaline battery of course. That's why I thought the wall wart would work. But I didn't know about the powerup spike then.

Or, it's a sign that the 9V supply feeding the module has been overloaded and has gone into shutdown. And since the module works fine when fed by a 9V alkaline battery, I think it's preety clear that the module IS adequate if the

9V feed is adequate.

Forget about the 9V and 5V supplies dipping. Let it dip for 22ms or whatever. Add the diode and big cap to power the CPU through the dip. Be sure to power the servo directly from the supply before the diode.

--
 Thanks, 
    - Win

I'm not sure I understand your circuit. If the servo is not connected to the 5 volt PSU, you are saying it is directly on the 9 volt power line? What does the voltage look like? It would seem to me your 9 volt power source is not up to the job which is what you are saying.

Ok, so the servo is on the 5 volt output of the PSU which means the PSU is

*not* unloaded in your previous test. So it may be the 9 volt supply or it may be the 5 volt regulator. You need to find out what the voltage on the 9 volt rail looks like when the servo comes on.

Duh!

Ok, but if you know the servo uses 700 mA (not counting the rest of the circuit) and your wall wart is only rated for 650 mA, why are we still discussing this? I don't recall you saying before the wall wart is only good for 650 mA.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

That won't help him at powerup when the servo zaps the power rail. In fact, the circuit never comes up because the servo surge causes the 5 volt PSU to shut down and when it comes back up it repeats the process. That's why he kept saying the capacitor would be an even *bigger* drain on the PSU.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

Yes. The problem is the 9V supply.

Here is the circuit:

If it was the 5V regulator, the circuit wouldn't work with an alkaline battery. But it does. Therefore the problem must be the 9V supply when anything other than alkaline is used.

I did mention in the beginning that the 9V was rated at 650 ma, but I didn't know then the size of the powerup spike. I started the thread in an attempt to find out whether this powerup servo spike was typical of all servos, or whether anyone knew of ones that didn't do that. Reliably. I guess we're still discussing it because I'm the only one here who doesn't think a diode will make it work with the

650 ma 9V supply. As it turns out, I can't find a schottky in my junque box, so I'll have to try it with a 1N914. I don't know how the Nano will feel about the voltage drop, but we'll see.

I will test the diode because I said I would. But I will be shocked if it works. Because, let me say again:

If the servo alone makes the power module cycle, I would be hard pressed to explain how adding the diode/cap would prevent that cycling. But I WILL test it. In fact, I will let the PSU circuit, diode and all, power up without the servo powered at all, then connect the servo V+ to the 5V rail. That will give it the best chance of working.

This helps a lot. I can see the servo does have a control line from the nano. Does the servo surge on power up regardless of the state of this control line?

Not if the 5 volt supply is not LDO. Normally a linear regulator only needs about a volt of head room to operate. The PSUs I found like this say it works with inputs down to 6.5 volts. But if this is not true it may need higher voltages. When the surge from the servo pulls the 9 volt rail to maybe 8 volts the 5 volt regulator cuts off. The battery doesn't droop as much, but then you are pulling more current than the 9 volt supply is rated for so that's understandable.

Since you have not responded to my requests for a scope trace I assume you don't have one?

The nano may well be a 3 volt MCU with an LDO regulator on board anyway. Or it may be a 5 volt MCU which will work ok at 4.3 volts.

Everyone was thinking the problem had to do with the nano resetting because of the power droop. But you are saying the 5 volt rail is cut off because the 9 volt rail is drooping too much. So, no, the diode and cap fix won't fix anything. It will keep the nano from rebooting, but it won't matter because the PSU will keep cycling.

I'd still like to see the voltage on the 9 volt rail with a scope. Or power the circuit from a variable supply and see just where the cutout on the 5 volt PSU is.

Also, I think JL is right in that you could probably replace the servo with an LED. Motion detectors don't actually see motion. They see variations in intensity caused by motion through a bug eye lens. Shine a bright, point source light on them and they trigger. But a pipe cleaner waiving in front of the lens has a certain "geek" appeal.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

I've tried tying it low, or high, or waiting until the Nano is sending PWM before powering the servo, but none of that makes any difference.

The regulator is the 5V version of the AMS1117.

The problem is that the 9V power wall wart I'm using is the one I normally power my scope with. I guess I could temporarily power the scope with a 9V battery, but I think it's pretty clear what's going on.

Indeed. I just finished the diode test with a 330 ufd cap. With the 9V battery it all works fine. With the 9V wall wart both the Nano and the power module LEDs begin cycling as soon as I connect the servo. So the diode/cap doesn't even keep the Nano from rebooting. The power module must be off longer than the 330 ufd will bridge.

That part of the project was another guy's responsibility. I just did the opamp stuff. I will ask him if he ever considered using an LED. But I have to say, waving the pipe cleaner is pretty cool. I mean, what's the fun in just turning on an LED? We really should have several more Rube Goldberg steps before getting to the pipe cleaner. :-)

None of this is unexpected now that I understand everything you are saying. The PSU likely is holding off power a lot longer than the 22 ms you see in your servo surge. Otherwise the nano would stay up.

The other thing is you can power the servo with a separate PSU. That's not a bad idea regardless. I'm not even sure why you are using the 5 volt PSU when there are tons of 5 volt wall warts in the world. I guess you wanted to be power independent just in case.

--

Rick C 

Viewed the eclipse at Wintercrest Farms, 
on the centerline of totality since 1998

yeah, that's what I meant. get a bigger "feed" like 2A perhaps.

--
This email has not been checked by half-arsed antivirus software

As the 9v drops down to 6v and its impedance rises greatly, you may run into much trouble.

NT

Couple of things... maybe the wall wart PSU has over current protection and just switches off momentarily ? What about incresing the CPU reset time at power on ?

--
This email has been checked for viruses by Avast antivirus software. 
https://www.avast.com/antivirus