Servo requires "tapping"


 

AJones

TVWBB Member
I notice that my servo sometimes doesn't move. It seems to require a bit of tapping on the RD3 case to get it to wake up. Anyone else have this problem? I'm guessing it's because the servo is a cheap one, just wondering if there was another possibility. I have several more, so I'll probably just change it out.
 
if the servo had to be forced into the body it may be binding, if so, widen the slot a bit until you dont have to force the servo in.
 
With the servo powered down, you should be able to smoothly move the servo arm back and forth using the same force. if it requires more force at some points it's probably got some flat spots on the gearing.
 
plug the servo in and leave it loose and make it move, if it doesnt do what it is supposed you have a bad servo or a bad controller. You can get a servo tester for about 3 bucks on ebay if you want to identify if it is the controller (HM) or the servo at issue.
 
I'm trying another servo now and am having a different problem. This one is moving back and forth wildly and then returning to the same spot it was before. It also feels hot to the touch when not moving!

Edit: Getting the same thing from two different Tower Pro MG90S servos. My previous servo was a Tower Pro SG90.
 
Last edited:
Aww your HeaterMeter just misses me so it is going into withdraw shakes! If the servo is just goes nuts and all over the place, it is because the 5V power at the servo end is dipping to the point that it loses all logic control. You can fix this by putting a capacitor on the servo end of the cable (inside the RD3 box). Testing here, just a 100uF capacitor seems to work, although it never hurts to go larger if you have one. If you don't have one you can swing by my place and I can hook you up.
 
Last edited:
Also, you may have received some counterfeit MG90S servos. They're probably the most copied servo in existence. I always get mine from a reliable source such as Adafruit. I used to get them directly from TowerPro, but I believe they only sell them through distributors now.
 
Aww your HeaterMeter just misses me so it is going into withdraw shakes! If the servo is just goes nuts and all over the place, it is because the 5V power at the servo end is dipping to the point that it loses all logic control. You can fix this by putting a capacitor on the servo end of the cable (inside the RD3 box). Testing here, just a 100uF capacitor seems to work, although it never hurts to go larger if you have one. If you don't have one you can swing by my place and I can hook you up.

Thanks Bryan, I may take you up on that. Sorry to be so needy!

- - - Updated - - -

Also, you may have received some counterfeit MG90S servos. They're probably the most copied servo in existence. I always get mine from a reliable source such as Adafruit. I used to get them directly from TowerPro, but I believe they only sell them through distributors now.
That's quite possible. I ordered some from China that took forever to arrive and I thought the order wasn't coming. I ordered more and now have a bunch. Only problem is I can't find the envelope where I keep most of them! :(
 
Yeah when I pay like $1.50 per servo shipped, I know they may not be the best quality items. I had one in my drawer that always spazzed out so I was able to do my testing with it. I also bought one of those servo testers so I could look at the signal it puts out and compare it to HeaterMeter's to see if I could do something in software to fix it. What I did find was interesting: if I took the HeaterMeter output and plugged it into the servo tester as a power source and tested the bad servo, it worked perfectly!

I took the servo tester apart to see if there was any magic inside, or at least if they had a capacitor on the output. Nope, just a bunch of fets and resistors for the battery testing function. What I did notice with the servo on the scope was that the servo drive pulses were 4.2V (ours are 5V). To simulate this, I put a 1N4001 diode on my HeaterMeter cable and ran the servo on that to give it a little voltage drop to 4.3V. Guess what? The servo works perfectly!

So what I am wondering is if it has something to do with if the servo power voltage drops below the servo pulse voltage. This would explain why it works with the capacitor (the 5V line doesn't drop enough to cause the problem) and with the diode (the 5V line drops, but stays above the 4.3V PWM).
 
Honestly, I had better results with the current boosting servo booster circuit I put together than the voltage booster you put in the HM. Perhaps what you have discovered here about pulse voltage vs supply brown out voltage explains why. The voltage boost was an improvement over direct connect, but some Tower Pro MG90's were pretty twitchy still. IDK if you recall, but when you changed the servo control to "move and rest" mode I had a problem with the servo where I would find the damper open a bit when it should be closed. Frustrated I tried a HexTronics servo which was much less twitchy and I've never found my damper open since.
I guess the moral of the story is the Tower Pro MG90s is a picky little servo, or at least the bootlegs are, but who can tell if what you get is legit these days....
 
I also had a few issues with the TowerPro SG90 - in my pre-HM days I used them in my three fan/damper pods I had on my UDS. I had them hooked up to the 3 outputs of a very cheap servo tester with the OMRON PID controller's relay output to the tester's pot so the servos only ever opened and closed to its max positions as the OMRON did its thing. Out of a batch of 10 SG90's, I ended up replacing 7 of them over the span of about 6-8 months, the final 3 were good ones that lasted for more than a 18months until I replaced my entire control setup with the HM and my integrated damper design. These SG's were very jittery and noisy nonetheless.

When I got the HM, I ordered a TowerPro MG995 metal geared servo which was local stock in Aus and around $10 delivered; I was impressed with how powerful it was and well built & solid it appeared. But it only lasted 2 cooks on the HM, at the start of my 3rd cook it started to spazz out, and caused my HM to keep on shutting down and resetting. I thought something in the HM were toasted but then touched the servo which was extremely hot. Luckily the HM survived this ordeal unscathed!

My damper moves with minimal force so it did not work hard at all. As I needed to get the HM going on the spot, I recalled that my son's old 1/7th scale nitro HPI brand RC buggy is still somewhere in the shed from when he was a youngster so I pilfered it's steering servo which is the same dimensions than the TowerPro MG995 fitting my damper brackets perfectly. Was a pity to wreck the car fo it but then realised the engine is all seized up and rusted anyway from years of no use and not being prepared correctly for long storage. This HPI servo is a reeal beast and working a treat, unlike the TowerPro it is not chattering at all and it closes perfectly everytime.

sorry. long post but I just thought to share my experience fwiw.
 
Last edited:
Honestly, I had better results with the current boosting servo booster circuit I put together than the voltage booster you put in the HM.
Would you mind refreshing my memory as to what your circuit looks like? I would love to check it out now that I have a pretty reliably broken servo to play with. Another option could be to add a second resistor to the 5V pulse output to drop it to ~4.2V.

I did disassemble a couple of the servos a while back and found that the SG90 and the MG90 that I had used different driver chips on them which would explain why the the two often behave differently. Of course, it could just be that some knockoffs get one chip and some knockoffs get others. I might pop this one back open and see if I can find a datasheet on its chip and maybe it has information about what the pulse specs need to be.

EDIT: The servo driver is a AA51880 and the datasheet doesn't say anything about the VIN (PWM input) specs, just that the absolute minimum VCC (power input) is 4.8V. I wonder if it borrows from the VIN line if the VCC drops below it? The block diagram doesn't show any sort of equivalent circuit.
 
Last edited:
I also had a few issues with the TowerPro SG90 - in my pre-HM days I used them in my three fan/damper pods I had on my UDS. I had them hooked up to the 3 outputs of a very cheap servo tester with the OMRON PID controller's relay output to the tester's pot so the servos only ever opened and closed to its max positions as the OMRON did its thing. Out of a batch of 10 SG90's, I ended up replacing 7 of them over the span of about 6-8 months, the final 3 were good ones that lasted for more than a 18months until I replaced my entire control setup with the HM and my integrated damper design. These SG's were very jittery and noisy nonetheless.

When I got the HM, I ordered a TowerPro MG995 metal geared servo which was local stock in Aus and around $10 delivered; I was impressed with how powerful it was and well built & solid it appeared. But it only lasted 2 cooks on the HM, at the start of my 3rd cook it started to spazz out, and caused my HM to keep on shutting down and resetting. I thought something in the HM were toasted but then touched the servo which was extremely hot. Luckily the HM survived this ordeal unscathed!
The two MG90s that I recently tried would also get hot! Never had that problem with the SG90. I thought they were basically the same thing, but now I'm reading there are significant differences.
 
Would you mind refreshing my memory as to what your circuit looks like? I would love to check it out now that I have a pretty reliably broken servo to play with. Another option could be to add a second resistor to the 5V pulse output to drop it to ~4.2V.

I did disassemble a couple of the servos a while back and found that the SG90 and the MG90 that I had used different driver chips on them which would explain why the the two often behave differently. Of course, it could just be that some knockoffs get one chip and some knockoffs get others. I might pop this one back open and see if I can find a datasheet on its chip and maybe it has information about what the pulse specs need to be.

EDIT: The servo driver is a AA51880 and the datasheet doesn't say anything about the VIN (PWM input) specs, just that the absolute minimum VCC (power input) is 4.8V. I wonder if it borrows from the VIN line if the VCC drops below it? The block diagram doesn't show any sort of equivalent circuit.
Bryan, you may have already thought about this, but getting a known good servo when ordering the kit would be a great addition and worth whatever cost difference.
 

 

Back
Top