Servos - Jaycar YM2763 & YM2765

[westaust55]

I had purchased a couple of 6V Servo's from Jaycar a while back and after some research to ascertain the correct connections (which lead to my separate "tutorial") today I conencted them up and did some simple tests.

The findings so far:

1. Rated at 6Vdc and on no-load operate happily from 4.5Vdc.

2. Did a test using a single 4.5Vdc supply for the PICAXE (08M) and the Servo motor without experiencing any resetting problems that I have read about. However in general I will be running from a separate supply.

3. These Jaycar Servos (both the YM2763 and YM2765 - made by Hobby Tech) operate exactly the same. Jaycar have no datasheets for them and when I asked what the output shaft rotational range was, the Jaycar (Perth) sales reps first stated they were 270 degree but then turned the output shaft by hand and found they had just over 180 degree of mechanical travel and thus stated they were 180 degrees operation.

4. While most information on the internet suggests that the shorter pusle duration moves the shaft anticlockwise, these Jaycar Servos move clockwise with shorter pulses.

5. My checks by gradually increasng the range of the control pulses found that:
(a) the "nominal" angular range might likely be 120 degrees
(b) they both can operate through 142 degrees
(c) the minimum pulse duration is 0.64 ms (SERVOPOS setting = 64)
(d) the maximum pulse duration is 1.99 ms (SERVOPOS setting = 199)
(e) relative to a nominal "neutral position at 1.5ms, they move roughly 90 degrees clockwise and 52 degrees anticlockwise.

Not exactly what I was expecting in terms of balanced rotation either side of the "neutral" (1.5ms) position.


Being new to using Servos, any thoughts or comments from those more experienced in Servo operation in relation to the above would be appreciated.

 

[Andrew Cowan]

All servos come up different. You can reverse the direction of the servo by opening it and reversing the wires to the motor and to the pot.

A

 

[Wrenow]

"Normal," operation such as it may be (as I understand it) is for 1.5ms to be relative center, with normal operating range being about 1.25ms to about 1.75ms, resulting in somewhere between a 45 degree and 60 degree total range of travel. Sounds like your servos fit that bill of goods. That it's operational/relative center is different from a true center is irrelevant until you try to push it to the stops (which the manufacturers do not really recommend). In its normal operating range, 1.5ms will be center, 1.25 should be 22-30 degrees in one direction and 1.75 should be 22-30 degrees in the other direction. Operationally, it is "balanced."

If that makes sense.

Cheers,

Wreno

 

[goom]

In the course of playing around with my sevo stretcher, I did find that, for several Futaba servos, the neutral position (1.5ms) was considerably biassed towards the +ive end of the travel (as defined by the mechanical end stops). I suppose that it could be brought nearer the center by taking the gear-train apart and moving the pot drive shaft round a few cogs.
Even with this offset, I could get full 180° rotation with a pulse of about 0.5ms to 220ms.
All the servos that I have used rotate clockwise with a shorter pulse.

 

[westaust55]

Thanks Wrenow and Goom.

The operating range for the Servo's I have is short of the mechanical stops so clearly the electronics is preventing further travel.

obviously some variance between manufacturers both in terms of pulse timing and direction.


From the Hitec Servo Manual:

Pulse Data: All Hitec servos require 3-5V peak to peak square wave pulse. Pulse duration is from 0.9mS to 2.1mS with 1.5mS as center. The pulse refreshes at 50Hz (20mS).

and

Direction of Rotation: All Hitec servos turn Clockwise direction ( CW ).

And from the HiTec datasheets:

DIRECTION: :CLOCK WISE/PULSE TRAVELING 1500 TO 1900usec

but the bit I really like in the FAQ:

Servo gets hot: Get out the fire extinguisher!

someone has a sense of humour

 

[BeanieBots]

One word of caution.
Although 1.5mS is almost universely regarded as the 'neutral' position. There are some servos out there which use 1.3mS for neutral. These are rare but do exist, so anyone buying from e-bay may encounter this type. It was a standard used many decades ago but seems to have crept back into use on cheaper two channel toys. Probably by some engineer who reverse engineered his grandad's old toy but they are now out there again!

 

[MPep]

I also remember hearing about 1.7ms being neutral. Again, on older equipment only.

 

[westaust55]

One word of caution.
Although 1.5mS is almost universely regarded as the 'neutral' position. There are some servos out there which use 1.3mS for neutral. These are rare but do exist, so anyone buying from e-bay may encounter this type. It was a standard used many decades ago but seems to have crept back into use on cheaper two channel toys. Probably by some engineer who reverse engineered his grandad's old toy but they are now out there again!

1.3ms would about be the neutral/centre of travel for those I have from Jaycar.

 

[westaust55]

Extending the rotational range of Jaycar YM2763 & YM2765 Servos

Bit the bullet, so to speak, and decided to see what I could modify to increase the rotational angle to 180 degrees for the Jaycar YM2763 & YM2765 servos I have.

Opened one servo, disconnected the feedback pot and measured the resistance = 4.53 kOhm.

Also measured the voltages when in operation:
1. Voltage across the pot = 2.95 V
2. relative to more negative leg of the pot,
a. feedback voltage at min pulse width (0.64ms) = 0.48V
b. feedback voltage at max pulsewidth (1.99ms) = 2.64V

Calculated that I need to add 64 Ohms to the min side of the pot and 1200 Ohms to the max side of the pot.

Closest that I had to try tonight was 75 Ohms and 1500 Ohms.

results predicatable in that with pulse of 0.64 ms, did not move as close to the clockwise mechanical limit, and by pulse length of 1.82ms reached the anticlockwise mechanical limit.

Tomorrow will try to buy some resistors of right value (or trim pots) and will then have 180 degree (+/- 90) of shaft rotation. Mechanical limits are greater than 180 degrees = ~200 degrees.

As both Jaycar servos I have (two different models of same make) exhibit the same rotation ability, think I can presume with some confidence that the 1200 Ohm and 64 Ohm values will be close for all Jaycar YM2763 & YM2765 servos.

Certainly room inside the servo for a couple of resistors and maybe even some 25 turn mini trimpots between the feedback pot and the control board.

Obviously this will VOID warranty, so be warned if you follow my lead.

 

[Andrew Cowan]

Certainly room inside the servo for a couple of resistors and maybe even some 25 turn mini trimpots between the feedback pot and the control board.

I've replaced the main pot with a 25 turn pot on a Futaba servo, to make it multi turn.

It gives about 19 turns, however, the centre point is extreamly poorly designed, and wanders by up to 120 degrees. Generally works, however...

A

 

[goom]

See my "Servo Range Extender and Failsafe" in the miscellaneous section of finished projects. With this you can set the neutral position, both end points and the failsafe position.

 

[Marcwolf]

I've replaced the main pot with a 25 turn pot on a Futaba servo, to make it multi turn.

It gives about 19 turns, however, the centre point is extreamly poorly designed, and wanders by up to 120 degrees. Generally works, however...

A

I have done likewise with a 10 turn pot on a 9g Servo. Very fiddly but it works

 

[westaust55]

See my "Servo Range Extender and Failsafe" in the miscellaneous section of finished projects. With this you can set the neutral position, both end points and the failsafe position.

Unless I am missing something, I do not think that your project will work for me without changing the pot or adding resistors.

1. The electronics board in both my servos only accepts a pulse in the range 0.6 to 1.99 ms. Shorter or longer signals have no effect.
This is standard on the servos I have and gives a range of 142 degrees. Physically this is from 90 degree clockwise relative to the case to 52 degrees anticlockwise relative to the case.

So I am trying to get a further 38 degrees anticlockwise.

The way to achieve this is add some resistance to the anticlockwise side so the feedback voltage at 90 anticlockwise is what the original has at 52 degrees clockwise. Then a small resistor on the other side to compensate for reduced volatge at the clockwise limit.

I have now done this and it does work with the original pot fitted.

 

[goom]

I guess that not all servos are created equal. I have only tried 2 different brands (Futaba and Bluebird), and both could be streched to 180° rotation, albeit with pulse widths way outside of specifications. I guess that the actual hardware has not read these specs.
Congratulations on the courage to play with the feedback, and the knowledge to make it work.

 

[westaust55]

Modifications to Jaycar YM2763 & YM2765

A few more details for those who may be interested.

1. Of the two Jaycar YM2763 and YM2765 servos I have, the electronics react to a very slightly different pulse range.
- The first works with a pulse from 0.64ms to 1.99ms
- The second works with a pusle from 0.65ms to 2.00ms.

2. I finally used a couple of external pots for a while to optimise the actual range of shaft rotation to determine the optimal values better than my first calcs.

End results for the first Servo are:
- feedback pot resistance 4350 Ohms.
- 943 Ohms added to the high side of the Pot
- 676 Ohms added to the low side of the Pot
(high = side to which pot wiper moves with longer pulses.

by measuring actual values of various fixed values I had in my storage box, I managed to get some resistors very close using
- 523 Ohm (nominal 510 Ohm) + 422 (nominal 410 Ohm) = 945 Ohms
- 335 Ohm + 334 Ohm (both nominal 330 Ohms) = 669 Ohms

Further tests and I am getting exactly 180 degree movement.

Now about to modify the second Servo.

Attached are a couple of photos.

First shows the the difference in the output shaft rotation range.
Left side shows both at min pulse length position and right side shows after (top) and before (bottom) at max pulse length position.

Second shows the added fixed resistors to extend the range.

 

[westaust55]

More on getting 180 degree rotation from Servos

End results for the second Servo are:
- feedback pot resistance 4940 Ohms.
- 848 Ohms added to the high side of the Pot
- 665 Ohms added to the low side of the Pot
(high = side to which pot wiper moves with longer pulses.

by measuring actual values of various fixed values I had in my storage box, I managed to get some resistors very close using
- 514 Ohm (nominal 510 Ohm) + 332 (nominal 330 Ohm) = 846 Ohms
- 333 Ohm + 331 Ohm (both nominal 330 Ohms) = 664 Ohms

Attached as a couple more pics.


First shows what these Servos look like internally.

Second is with a couple of 25 turn 2 kOhm trim pots to determine the best values to add to get 180 degree rotation.

 

[BeanieBots]

What you are discovering is what many RC hobbiests know all too well. There is a vast range of characteristics from one servo to another. That's why most 'respectable' Tx controllers have a mechanism by which those characteristics can be trimmed out.

You will also notice (but not necessarily with the ones you have) that position is also a function of supply voltage. At first, that might appear to be a major design flaw but 'experienced' flyers will tell you that it can actually be a handy feature that lets you know the state of your receiver battery.

Then of course, there are digital servos. These are far more precise but as with any other product, there are good ones and bad ones. A "bad" one will have poor response time with quantized positions and will probably fail completely at anything other than optimum input voltage.
A good one will blow your mind. It will also allow you to program just about every single charactersitic you can think of.

You get what you pay for and with a price difference in the region of a factor of one hundred, there is also a factor of around one hundred difference in performance and features. Performance wise, it's like comparing an 08 to a 28X2.