TBM Servo Comparison List

Discussion of the servo testing by TBM.

1. TBM performed its own side by side testing to compare servos under similar conditions.
2. When we indicate torque at 7.2v, the voltage is actually higher. We are using a new, unregulated 4-cell Li Ion battery. The no load voltage is over 8v, and the voltage under the load of 1 servo is over 7.2v. The voltage that the servo sees in actual use is dependent upon lots of factors, and will probably be lower than is stated here.
3. This is for one servo only, this is not the average of several servos.
4. The values we posted in the servo comparison table are the highest values we saw in the test.
5. Every servo has a higher torque the first time power is applied to the servo. Power drops off in torque and amps immediately and continues to drop. Servo manufacturers publish the highest value at the first instant of use because this is the highest attainable value.
6. The voltage that the servos were tested at were less than the starting voltage. For instance, we set the regulator for 6v, though when the servo was stalled, the voltage dropped below 6v, like 5.7v in some cases. Servo manufacturers will maintain a constant voltage of 6v for testing purposes, so their values will be higher.
7. We are measuring stalled torque, which is the torque available to move a surface, not hold the surface in place. If we were to first stall the servo, and then pull on the meter to try to move the servo arm, the load would go way up! It can be triple the stalled torque. For instance, if the servo read 10 lbs stalled, if you pulled on the meter to move the servo arm, the meter might read as much as 30 lbs, and the amps would increase similarly. Thus holding a control surface in position is much easier than getting the control surface into position in the first place. Some manufacturers post values somewhere in between the stalled torque and the holding torque. How they determine the numbers they post is vague, and I’m sure the methodology is different from manufacturer to manufacturer. Our method compares values attained from a similar method. It is not intended to duplicate values posted by servo manufacturers.
8. Amps is directly proportional to torque. However, keep in mind that gear ratios are changed so that speed is traded for torque. The motor will pull the same amps in a speed versus a torque version of the same servo (like the HS625 and HS645), however the torque is less with the faster servo. However, for the most part, if you want more torque, you will pull more amps.
9. Amps is directly proportional to voltage. The higher the voltage, the higher the amps, and the higher the torque and speed.
10. Some servos can dissipate heat better than others, so some will maintain a higher torque after 30 seconds (like on a knife edge) than others.
11. Voltage and amps is directly proportional to the gauge wire and plugs used. If you use HV plugs on the battery and use a power expander type unit, the power to the servos goes up tremendously on high powered servos. Just one JR8711 can’t be operated at its fullest capability if a normal JR type plug is used on the battery, switch, regulator, etc. the voltage drop across those little plugs is tremendous. Imagine what the voltage drop to the servo is if a bunch of servos are trying to draw current simultaneously! Using a JR8711 without a power expander is a waste of money, because the amps can’t get to the servo. It’s not the wire as much as it is the plugs.
12. Within experimental error, losses after time/heat, and voltage drops, there are several servos which are all about the same in torque. These are the HD-DS-120M, HD-1501MG, HD-9150, HS-645, HS-5645, HS-7985.  The JR-8411, and JR-8611A are a little stronger. The HS-7955 is quite a bit stronger. And the JR8711 and JR8711HV are best by far, with the JR8711HV being the biggest, baddest servo out there. It is king of the hill (with a price tag to match).
13. The "pole" motors aren't as fast as they show. True, that's the speed in one direction, but it takes a longer time to stop and turn around and then go the other direction. The motors are heavy so they take longer to stop and reverse direction than a coreless motor which is lighter. So if you want speed, then you want a coreless motor. The coreless motors are faster than a cored motor with the same speed rating. (I hope that makes sense!!)
14. We did not test the Seiko servo for fear that it would break our testing equipment.

JOHN DURANT'S AEROWORKS 42% ULTIMATE WITH SEIKO PS-050 ON RUDDER

John DuRant's Aeroworks 42% Ultimate Biplane Installation

I have been in RC for 30 years and am not usually easily impressed, but when I first took the PS-50 out of the box I got instant goose bumps. This is a real beast! Once my blood pressure came back down, I got to work putting it on the rudder of my Aeroworks 42% Ultimate.

The first step was to pick a location. I chose dead center in the fuselage and just in front of the fuselage former for support. The servo is pretty deep, so a couple sticks of 1/2 X 1" balsa were needed to raise the servo and stiffen the light ply fuselage floor. I considered using the supplied sheet metal screws to mount the servo, but quickly scrapped that idea. Instead, I used 4-40 machine bolts with blind nuts under the fuselage floor to hold the servo in place. At this point the aft servo mount was braced by the fuselage former, but the front mount flexed badly the first time I tried running the servo. I then used 1/2" square balsa braces from the fuselage former to the forward servo mount. For pull-pull set ups, the front of the servo attempts to rock up, so that is the axis that needs bracing. The servo was now rock solid.

The next step was the electrics. The servo has two leads. The first is a standard JR style connector which fits right into the receiver. The second is two plain wires (red and black) to be used as the power input. I was fortunate, as I am using 10.8 volt Powerflight batteries. They produce 12.2 volts when fully charged, and have voltage regulators to reduce this to 6 volts for the receivers. All I had to do was tap directly into the batteries prior to the voltage regulators. A few minutes soldering connectors was all that was required to power up the big servo. At 12 volts the servo is rated at 1800 inch ounces. If you do the math, that means it will pull 64 pounds with the arm Troy Built Models offers. Here come the goose bumps again! You can run the servo off a 6 volt power supply, just at a reduced torque rating, which shouldn't be a problem for any plane under full size.

Now for the rudder hook up. Troy Built Models offers an aluminum pull-pull servo arm that is perfectly set up for ball links. Use a drop of Locktite on the arm mounting screw. The post is nylon and tends to squeeze the screw out from the loads we are dealing with. My screw backed out every flight until I finally used Locktite, and now it stays put. I used Nelson Hobbies ball links, cable connectors and .030" pull-pull cables. A quick note here is warranted. In the directions, Jerry Nelson says soldering of the cables into the end links is not necessary. He was right - until the first time I selected high rate rudder travel. The cable connectors let go of the cables and my rudder was left flapping. I used silver solder and a torch to permanently fix the cables to the connectors and have had no trouble since then.

The last item is the rudder horn. My previous set up used a Nelson Super horn, which uses an 8-32 rod threaded through the rudder with ball links on either side. I knew this would not hold up to the full torque of the BIG servo, but I decided to give it a try. I decided the 8-32 horn would be a good "weak link" in the setup to keep from destroying something more expensive. After all, when this servo pulls, something is going to move. A key point with the horn; the servo arm offered by Troy Built Models is not a straight arm. In other words, if you hold a straight edge on the arm's end points, they line up behind the servo post. The rudder horn end points have to be that same distance behind the hinge line of the rudder to get the geometry right, and with 64 pulling pounds - you want to get it right! It amounts to about 3/8 inch behind the rudder post. On the AW Ultimate, this is exactly where the hard point on the rudder is located, so all was well. The servo turns about 85 degrees in each direction, so a quick calculation said I could use a 4 inch horn (2 inches per side) to get 45+ degrees of rudder travel. I still had to use the travel volume function of the radio to reduce the rudder deflection some more. I would have gone to an even longer horn for more precision, but that would have interfered with the elevators.

Now for the flying. The goal of a good control setup is to have it perform flawlessly without your worrying about it. This servo is just that. The surface moves faster than the 4-JR 8411's used to turn it. I was concerned about centering, but I cannot see any drift in rudder centering at all. The "dead spot" at neutral is about 1/16" at the very tip of the rudder, which is as good as I got with the 8411's. Many maneuvers I perform, like snaps and 50 degree nose high knife edges, cleaned up immediately with the big servo. You don't know how underpowered your rudder is until you throw this PS-50 in and see what is possible! With the exceptions I mentioned above, I have had no trouble with the servo in more than 100 flights. Recently, I found the limit of the 8-32 rudder horn. A high rate, high speed snap roll into a hover bent it. I put a new horn in and the same maneuver bent that one like rubber too. Imagine having enough power on a control to bend the horn at will! I'm planning to switch to a 10-32 rod and will see how that goes. The 8-32 horn does work fine for all "normal" maneuvers, including everything Chip Hyde did with the plane at the 2002 TOC.

In conclusion, this servo would be my first choice on any giant scale I build in the future. I am trying to figure out how to build it into the elevators and ailerons now!

PERRY GENOVESE'S 47% ONE DESIGN WITH SEIKO PS-050 ON RUDDER

Thought I would share with the group that I have finally found THE servo for rudder on large planes. I have been struggling for months with the rudder setup on my 47% DR109 propotype. I started with 4 x 5945's......not enough. Added one to get to 5 x 5945's. Still not enough. This is a large plane with probably the largest rudder on any model out there. I've been compensating and leanring how to fly it without enough rudder. So I've been waiting for the Hitec 5995 robot servos and planned on replacing all 5 5945's with the 5995's. In the mean time I discovered the Seiko P050. On 12V its starting torque is rated @ 1800 in-oz!. Being analog I was skeptical but for $190 I figured what the hell.

All I can say is I'm totally blown away. What was not enough rudder for a roller and basically 55% low rates for pattern flying using 5 x 5945's is now a new plane. I had to cut my low rates in half and add a bunch of expo. In fact I had to add expo and remix on all rates! And on 12V it's faster than any standard size servo out there! Knife edge pop ups:) ?

Kris Welter and I have been conversing and he just added one to his plane and we were joking if there is such a thing as too much rudder? Ha ha, never! This is one serious servo. I added a switch and a stand-alone 12V pack and it weighed out identical to 5 5945's, linkages, extensions, etc......15 oz out and 15 oz back in. The Seiko weighs in at 10 oz and Gene sells a custom truss arm made by Sam Monteleone. Very nice stuff. This servo is a freaking monster....both is size and power. Die-cast aluminum case and all so no case distortion! My 47%, 47lb bird is now a very happy camper with unlimited vertical and a silly "foamy" style rudder. Gotta love it! Now if only work didn't interfere so much with flying :)

Ken Moore's 42% Ultimate WITH SEIKO PS-050 ON RUDDER

Knife edge is out of sight with the Seiko! It has so much power that I can perform knife edge loops with ease! I plug the power lead for the servo into the Powerbox. The rudder sees 6.2v from the Powerbox. This voltage gives me plenty of power and speed for any maneuver that I want to do!

Ron Lindsey's 46% Ultimate WITH SEIKO PS-050 ON RUDDER

A couple of photos of my Seiko install in my Hangar 9 46% Ultimate. This is definately the way to go, lots of power and the hook-up couldn't be easier. I would highly recommend this servo to anyone setting up a large gas plane.

SERVO	INITIAL TORQUE @ 6V	WORST CASE TORQUE
Hitec HS-5955	248 oz-in	152 oz-in
JR 8711	358 oz-in	280 oz-in

SERVO	INITIAL TORQUE @ 6V	WORST CASE TORQUE
Hitec HS-645	96 oz-in	NA
HD Power 2550A	165 oz-in	NA

WHY ARE THESE VALUES DIFFERENT FROM THE MANUFACTURER'S? Even though we name this servo wars, we don't want to start a war! If you read this entire page you will understand the reasons, and there are many. However, I believe that the main reasons are:

1) We used a 6v regulator which is commonly used in giant scale airplanes. While the output voltage may read 6v unloaded, under load, the voltage is nowhere near 6v, especially when a servo is pulling 4.5 amps. The voltage is down around 5v. If a large power supply is used which maintains 6v under load, the values will be significantly higher.

2) The servo power drops off significantly with temperature. The values above are averaged over several tests with the same servo. The very first reading when the servo was cold was significantly higher. We averaged in lower numbers because the servo was heating up.

3) The test equipment I used took 1 second to stabilize the reading. Better test equipment will definitely show higher readings because those readings can be taken at the very instant that the load is applied, not 1 second later. Also, who says my fish scale is accurate? I did do some testing to verify that it looked pretty close, but was it off by 1%, how about 5% or 10%?

4) System losses. My system will cause some losses inherently. The servo must overcome some friction in the system and such which will give me lower values than a better system.

For these reasons, I can see that the values that are published by the manufacturers are correct because the manufacturers have perfect conditions. They are the values that are the maximum possible that the servo can possibly put out! It would be really bad marketing for a servo company to publish values which are any less than the best! No company in the world is going to publish advertising data which is under the very worst of conditions if their competitors are publishing data under the best conditions!

These values were derived using the test method and equipment shown above. The worst case torque figures are after a minute of hard operation which causes the servo to heat up substantially. These torque values will only be seen if the servo is stalled and allowed to heat up. It is not recommended to set up your plane so that you are stalling servos for long periods of time! You should set up your plane so that the servos are pushed to their maximum. If you suspect that you are stalling servos, you must put in more powerful servos! You should never experience the worst case torque values if your plane is set up properly. The only way to determine this is to test it yourself, or follow the lead of others you trust. We offer information below on which servos we feel work best in various situations, however your situation may be different. If you are losing power in knife edge or knife edge loops or sustained rolls, then you need servos with more power. You are pushing the servos past where they are intended to operate. In some cases you will need more amps to the servos which requires larger wires, better connectors, high output batteries and more. And this is not just for the best pilots! Any pilot can tell the difference between a plane which has weak servos and strong servos.

I have never installed a higher power servo, and not seen an improvement in performance. If this happens to you, then you don't have enough amps getting to the servo and you must do something about it. Planes up to 35% in size are easy. Just follow the recommendations below and you won't have any problems. It's planes of 150cc size engines and above where problems can arise. The biggest issue is the rudder, and right now there is no perfect solution. Seiko servos have tons of power and are great for 3D but lack precision and centering for IMAC. 4 or 5 HS-5955 or 2-4 JR 8711 servos might be required, and that's expensive and time consuming to set up properly.

We really need to do more testing and refine the test methods. The set up we used did not move the servo much. The servo moved a few degrees and then stalled. Stalling a servo should never happen in flight. It would be best to use some test equipment on board each and every plane available and determine the actual loads imparted on the servos and thus the appropriate servo can be selected.

JR-8711 vs Hitec HS-5955 in actual use. This test started with using one HS-5955 servo in my 35% Extra 260. I had just enough rudder power to perform a knife edge loop. However I noticed that the rudder lost power later on in the flight. Apparently this was due to the servo heating up and losing power. I added a second HS-5955 servo and the difference was tremendous. Rudder authority was like a foamy plane. I then switched to one JR-8711, and there was no detectable change to the rudder authority compared to the two Hitec HS-5955 servos. This confirms that the above test results of 152 oz-in vs 280 oz in is realistic and usable.

TBM SERVO TESTING: TBM conducted its own servo torque testing using the pictured equipment. We monitored the voltage at the Rx, the voltage at the battery, the amp draw at the battery, and the torque of the servo. We used a TBM 2-cell LiIon battery with a MPI Miracle switch (which has a built in 6v regulator). We tested other batteries and other regulators from Fromeco and Smart-Fly and there was no difference in the performance of the 1 servo. We were concerned with moving torque, not holding torque. We are interested in the power available to move a control surface into position. The most powerful servo, and the most power hungry was the JR-8711. It drew as high as 5 amps initially though the load dropped to 4.2 amps in less than a second. Having 13 of these on a very large plane like a 50% plane leads you to conclude that there could be a momentary current draw of 65 amps if all the servos were stalled. To put this in perspective, most of the circuit breakers in your home are 15 amps, and the best 4-cell LiIon has a burst output capacity of 18 amps. IN REALITY a plane with 13 JR-8711 servos will never require 65 amps. We estimate based on some testing that 50% planes will typically pull a peak of 40 amps, 40% planes pull a peak of 20 amps, 35% planes pull a peak of 10 amps, 33% planes pull a peak of 8 amps and 30% planes pull a peak of 5 amps on a routine basis. NOTE: I am not concerned too much about the torque values I obtained being lower than the stated torque values of the manufacturers because they are publishing holding torque, not moving torque, and their testing methods are different. I do feel very confident that the relative amount of torque difference from servo to servo is accurate and should be used to determine the servo you should use in your aircraft.

TBM SERVO ENDURANCE TESTING: TBM conducted its own servo endurance testing using the equipment shown above. We found that the first 1/10 of a second that the servo is used that the torque is at its maximum. The performance then degraded substantially over the next 60 seconds (of intermittent use), and stayed at that lower value for the duration of the testing. For instance the JR-8711 has an initial torque of 358 oz-in, and after 60 seconds (of intermittent use) the case heats up to 120 deg F and the torque drops off to only 280 oz-in! This is a decrease of over 20%. The Hitec HS-5955 lost 40% of its torque by dropping from 248 oz-in to just 152 oz-in in about 60 seconds (of hard but intermittent use). To define intermittent use: I did not simply stall the servo for 60 seconds to see what would happen. I operated the servo as if I was performing knife edge passes along the flightline and knife edge loops, so the servo was cycled on an off and was stalled for no more than 8 seconds at a time.

WHAT IS SERVO TORQUE? Servo torque in the US is in oz-in. That means that if you multiply the inches of the servo arm (where the hole is that you are using from the center of the output shaft) times the the number of ounces the servo can move, you get the torque. If you have a servo with 300 oz-in of torque, it can move 300 oz if the servo arm is 1". If the servo arm is 2" it can move only 150 oz because 150x2=300. If you have a 1/2" arm it can move 600oz.

HOW DO MANUFACTURERS TEST FOR SERVO TORQUE? The published values of servo torques from manufacturers are all done in different ways. There is no standard method for determining servo torque. This means that information from a reliable third party is necessary to actually compare the servo torque using the exact same conditions. This has not been done extensively that I know of. However, manufacturer's data is a very good starting point because the data from manufacturer to manufacturer is fairly close.

HOW GOOD WAS TBM'S TEST METHOD? TBM performed a test that we felt could be duplicated repeatable, though we only tested a few servos. The major problem we had with repeatability was that the servos heated up and lost torque very quickly when they were tested at their maximum. So, the maximum torque is interesting to know, but this may not be what is actually available during a flight. We don't know the actual torque that is available when the servo is in a plane and it has been used for several minutes, and of course this depends on the situation. So is the maximum torque rating the best rating to use to compare servos or should it be the maximum torque which is available after the servo has been run at 20% of it's maximum for 5 minutes? If we let the servo heat up for 5 minutes by applying 20% of the torque, this would simulate the control surface being used in flight. But is 20% the correct amount? Some planes may be 2% and some 80%. It's just difficult to test the correct way. Therefore the best test method is to use the recommendations of people that you trust, and if you feel that you are losing power in the flight, then go with the next larger servo. So the data we provided is good relative data, but it in no way duplicated what the manufacturers data is. WHAT SERVO TORQUE DO YOU NEED? The actual torque a servo puts out is worthless information to most modelers. It gives you a relative power between servos, but tells you little about whether it will work in your application or not. Do you know that you need 300 oz in max torque on a 35% aileron when it is deflected 38 degrees at an airspeed of 78mph? No, of course not. You can only go by trial and error. What has worked for others? If that's not enough for you, then which servo should you use? If you try a HS-5645, and you are pushing it to its limit, then you can move up to a HS-5955, and if that still is at or near its limit, then go to a JR-8711 and if that is not enough start doubling up the servos!

TBM conducted its own servo extension testing using the equipment shown above. We found that adding a single servo extension (22 gauge with universal connectors) between the Rx and the servo will cost you 1.5% - 5% of your power to that one servo. Adding another extension will cost you another 1.5% - 5% and so on. In one of the many tests run, I put ten 6"extensions in series and measured a total drop of 15% of the maximum torque that a single JR-8711 could put out. In another test, I put ten 48" extensions in series (40' of extensions) and found the drop in power to be 55%. Therefore losses due to servo extensions is a combination of plug losses and wire resistance losses.

POWER LOSSES IN SERVO EXTENSIONS

WHAT'S THE BOTTOM LINE OF THE ABOVE TESTING?:

SERVO POWER: You will see a significant drop in power of servos under constant use HOWEVER you will only notice this drop in the performance of the rudder. The rudder is the only surface which is pushed to its limit for extended periods of time. No matter what, the first time you use your rudder it will have more power then than later on in the flight. If your first maneuver is a knife edge loop, the second knife edge loop will not have the same power available to the rudder. Therefore, overpowering of the rudder is necessary so that you don't need full servo power to do a knife edge loop (or knife edge pass or whatever you use the rudder for). You must not stall the servo because stalling the servo causes it to heat up and lose power which changes the inputs required. Of course for those who don't use the rudder much, then the drop off in power is of little consequence to you, and you won't notice any difference.

SERVO EXTENSIONS: With proper servo sizing, you will not experience any difficulties using extensions. * Never use extensions between the battery and the Rx. If you need more length, splice in additional 18 gauge wire. * Use the shortest length extensions possible. * Use one long extension instead of 2 or more shorter ones.

The servo power required for your plane's control surfaces varies with the following:
The Obvious:
* airplane size
* control surface size
* mechanical advantage of linkage
* control surface throws
* airplane speed

The not so Obvious, but equally important:
* current available to the servo under load which can vary with battery, regulator, switch, servo extensions, plugs and more.
* the reduction of torque due to the servo heating up or wearing out
* the weight of the plane. The more weight, the faster it must fly, and the more servo power is required.

* For 85cc and 100cc planes, if the aileron is fully sheeted, you can use one servo. If the ailerons have lightening holes, you must use two servos.
* For 210cc planes, if the aileron is fully sheeted, you can use two servos. If the ailerons have lightening holes, you must use three servos.

*For 150cc planes, there are many options, and the weight of the planes available vary considerable, as do the sizes of the rudders. If you use a push-pull set up in the back of the plane using pushrods, you will gain power, and have a more effective rudder versus using a pull-pull set up. The problem with servos in the tail is weight. All planes are tail-heavy, so adding weight to the tail by putting servos in the tail must be accounted for. With the Extra 260 and Yak cowls not having nose rings molded into the cowls, and thus the cowls being open in front, you can move the engine out without it looking bad, and more easily counteract the weight of the servos in the tail. Most manufacturers do not offer servo mounts in the tail, so you need to make these yourself. A Seiko servo has tons of power and is less expensive, though it weighs 10 oz plus an additional 5 oz for a separate 3-cell battery.

SERVO SUGGESTIONS FOR MAX PERFORMANCE

Where I am recommending the HS-7955 servo above, I do so because of the following:
* The HS-7955 servo is not much more money than the HS-7985, and with its titanium gears and higher torque, it is well worth the money even if the HS-7985 will do the job such as in 50cc ailerons and elevators.
* The HS-7955 and the JR-8611 use the same motor and similar gear trains, so the actual power is similar. The drawback to the JR-8611 is that its gears wear quickly and you will have gear slop issues.
* You can upgrade to the JR-8711 for even more power, but you will not notice the change in most applications because the HS-7955 is already an overkill in the above situations.
** For the 150cc planes, the HS-7955 servos work well, but in many instances you will notice a performance increase going to the JR-8711 servos on the ailerons and elevators. Weight does play a factor. If your plane is 40 lbs or more, use the JR-8711 servos. If you can swing the extra money for the JR servos, and can put up with potential gear slop, then do so.