Actual servo perfomance, tested

[kcaldwel]

I'm not an indoor pattern guy, but I've been looking through your forum and it's great info on lightweight servos for use in a Discus Launch Glider.

I noticed a lot of comparisons of servos based on the seller's published specs. There is a very good French program that allows easy comparison of the actual servo performance (torque, speed, weight, centering, drive train stiffness, etc.). There are some huge differences between the claimed numbers and the actual performance, especially for the small servos.

http://www.teaser.fr/~osegouin/aeromode/servos.phtml

Discussion and Babelfish translation of the manual:

http://www.rcgroups.com/forums/showthread.php?t=406059&pp=15

They don't have all the servos you might be interested in, but there is good data on the Dymond D47, Hitec HS-50, a lot of Waypoints, etc.

One thing to note is the Dymond servos are no where near their specs on torque. The D47 published spec is 15 to 20 oz-in, actual is 448 gm-cm (6.3 oz-in). Hitecs seem to be near their specs, and JR (Graupner in Europe, unfortunately with different designations for all their servos) seem to be about 20% low on torque.

Anyway, a very interesting tool that is worth the time to figure out the French!

[Jurgen Heilig]

...
One thing to note is the Dymond servos are no where near their specs on torque. The D47 published spec is 15 to 20 oz-in, actual is 448 gm-cm (6.3 oz-in). Hitecs seem to be near their specs, and JR (Graupner in Europe, unfortunately with different designations for all their servos) seem to be about 20% low on torque.

Anyway, a very interesting tool that is worth the time to figure out the French!

Well, unfortunately there is also some very odd data in this tool, for example:

Do you really think a standard sized C577 provides only 1.690 gcm of torque (Graupner/JR claims 3.500 gcm) and that the same servo with a ball bearing provides 2.761 gcm?

Or that the well proven mid-class C4041 (2.130 gcm) has less torque than the micro C351 (2.240 gcm)?

So please take the data with "a pinch of salt".

:) Jürgen

[Trisquire]

I was under the impression that the DLG guys discovered the D47 before we did.

Tom

[kcaldwel]

Well, unfortunately there is also some very odd data in this tool, for example:

Do you really think a standard sized C577 provides only 1.690 gcm of torque (Graupner/JR claims 3.500 gcm) and that the same servo with a ball bearing provides 2.761 gcm?

Or that the well proven mid-class C4041 (2.130 gcm) has less torque than the micro C351 (2.240 gcm)?

So please take the data with "a pinch of salt".

:) Jürgen

I think their data is quite good. They have an excellent looking test rig, and I believe they did 3 samples of each.

There is usually a reason for "odd" data, if you look at the details. For example, the C4061 is twice as fast as the C351 under load, so it may be sacrificing torque for speed. Not sure what the ball bearing equivalent of the C577 is...

I'll believe their numbers over the seller's fantasies most days!

Kevin

[Jurgen Heilig]

...
There is usually a reason for "odd" data, if you look at the details. For example, the C4061 is twice as fast as the C351 under load, so it may be sacrificing torque for speed. Not sure what the ball bearing equivalent of the C577 is...

I'll believe their numbers over the seller's fantasies most days!

Kevin

Hi Kevin,

if you read the documenation, you will see that they only test one servo - not three. They even show you the picture of the servo, and the C4041 actually looks quite battered (with a broken lug).

The ball bearing version of the C577 is the C5077.

According to the test documentation, the torque test stops when either the servo speed gets too slow, or when the servo current reaches 80% of the stall current. Unfortunately in many cases the tests are stopped before any of the above criteria are met. That explains some off the odd data, as you might compare apples with oranges.

:) Jürgen

[kcaldwel]

Thanks Jurgen. Your French must be better than mine, I'm struggling with some of it. I'll have to get my French speaking wife to translate some more!

I am very surprised how little independent testing of servos is done. Maybe there is more in Europe that doesn't turn up on English searches? Basic torque testing doesn't seem like it should be that difficult to do. The only other testing I can find is quite limited, old, and doesn't include any of the small servos:

http://www.servotestsite.com/index.html

[Jurgen Heilig]

Thanks Jurgen. Your French must be better than mine, I'm struggling with some of it. I'll have to get my French speaking wife to translate some more!

I am very surprised how little independent testing of servos is done. ...

Hi Kevin,

there was a time when my French was much better than my English (had a French girlfriend at university ;)).

The problem with independent servo testing is simple. Who will sponsor such tests? How many are interested in real data and how many can actually make proper use of them? Most importantly, those tests say nothing about reliability, resistance against shock and vibration etc. which is often more important.

:) Jürgen

[Harri Pihl]

Generally I rate precision, good centering and reliability much higher than absolute values of torque and speed. So IMHO it's best to just browse forums and talk with other people and sometimes when something interesting hits market, I'll buy couple examples and test them myself (with my own, not so precise methods) and share my experience with other people.