Rook's Pawn III
Rook's Pawn III
by bauermech
Mon Jun 04, 2007 2:33 am
16 HSR-5990TG's, 8 HSR-8498HB's, an MR-C3024 micro controller, and the rest 100% custom parts.
One mean mother
by bauermech
Mon Jun 04, 2007 7:10 pm
Thanks!
They do... they actually do suppport catch-n-play, unlike the 5995's. They're quieter as well. The 5995's screamed like wild banshees. These are not that much more audible over the 8498's.
Here's a few specs...
Weight: 2.26kg (5.0lbs)
Height (fully erect): 48cm (19”)
Height (full squat): 30cm (12”)
Still using the MR-C3024. Programming away... won't stop until RoboGames starts.
I'll post some video if I manage to find the time.
Do the 5990's allow catch and play?
They do... they actually do suppport catch-n-play, unlike the 5995's. They're quieter as well. The 5995's screamed like wild banshees. These are not that much more audible over the 8498's.
Here's a few specs...
Weight: 2.26kg (5.0lbs)
Height (fully erect): 48cm (19”)
Height (full squat): 30cm (12”)
What controller are you using have you progammed it yet or is that in the pipeline.
Still using the MR-C3024. Programming away... won't stop until RoboGames starts.
I'll post some video if I manage to find the time.
by i-Bot
Tue Jun 05, 2007 8:57 am
The 5990 uses the latest version of HMI software. It does all that the 8498 does. Also fixes a couple of things which did not work on the 8498 (motor stop) via serial control. The HW and SW is totally different from the 5995 ( the 5990 cannot be programmed by the HFP-10 servo programmer). The gains and offsets of the 5990 can be reprogammed via the HMI serial interface
The biggest difference is that the servo will switch off if overloaded for longer than 10 sec. The Hitec servos do not measure current directly. The current they report over serial HMI is the proportion of PWM applied to the motor. This in turn is proportional to the distance of actual position from desired position when not moving. If the servos switch off they have to be repowered to reset. This feature should be good to save servos, and we think it will also be in the new 5498 if that too has v1.1 HMI software.
The biggest difference is that the servo will switch off if overloaded for longer than 10 sec. The Hitec servos do not measure current directly. The current they report over serial HMI is the proportion of PWM applied to the motor. This in turn is proportional to the distance of actual position from desired position when not moving. If the servos switch off they have to be repowered to reset. This feature should be good to save servos, and we think it will also be in the new 5498 if that too has v1.1 HMI software.
by limor
Tue Jun 05, 2007 11:31 am
This sounds like important piece of information but i have a hard time following it. Can you kindly expand on this. what do you mean by "proportion of PWM applied to the motor"? why is it proportional to distance? how does externally applied torque, which creates more current consumption, influence the position..i-Bot wrote:The Hitec servos do not measure current directly. The current they report over serial HMI is the proportion of PWM applied to the motor. This in turn is proportional to the distance of actual position from desired position when not moving. If the servos switch off they have to be repowered to reset.
by JonHylands
Tue Jun 05, 2007 1:39 pm
I think the second sentence is more telling:
"This in turn is proportional to the distance of actual position from desired position when not moving."
If the sensed position is different from the desired position, and the servo isn't moving, then it must be applying torque, and the torque is going to be proportional to the amount of power being applied by the motor.
- Jon
"This in turn is proportional to the distance of actual position from desired position when not moving."
If the sensed position is different from the desired position, and the servo isn't moving, then it must be applying torque, and the torque is going to be proportional to the amount of power being applied by the motor.
- Jon
by i-Bot
Tue Jun 05, 2007 3:45 pm
The on to off ratio of the motor current from 0 to 100% on, is represented in the HMI serial interface as a value from 0 to 255. When you read the servo voltage and current, this value is what is returned.
The servo works internally in microsec from 550 to 2450 for the position which is the servo pulse width, and corresponds to the 10 to 190 value which is sent and read by the HMI pulse modes.
The 8498 default settings have a dead band of +/- 5 microsec, then the motor pwm current rises linearly over the next 100 microsec to 100% in each direction. These settings all appear the same on the 8498 , though they vary on the 5990, and are programmable on both. Note this is the static proportional behaviour, there is also a derivative term, but that does not matter when servo is not moving which is the case we normally want to consider for load sensing.
If the servo is to apply any torque, current must be applied to the motor. Torque is proportional to current when the motor is not moving and no back emf. The current is pretty much proportional to the on to off ratio in the lower current part, guess there are some inductive effects in the motor
If the servo is 30 microseconds off position ( 3 degrees) , it is loaded to about 30%.
Of course with serial HMI you have accuracy 10 times greater than over pulse HMI, and more accurate communication. However measuring the position is still a good way to measure the torque. Also reading the position via serial HMI does not turn the servo off, like the pulse HMI position read.
The servo works internally in microsec from 550 to 2450 for the position which is the servo pulse width, and corresponds to the 10 to 190 value which is sent and read by the HMI pulse modes.
The 8498 default settings have a dead band of +/- 5 microsec, then the motor pwm current rises linearly over the next 100 microsec to 100% in each direction. These settings all appear the same on the 8498 , though they vary on the 5990, and are programmable on both. Note this is the static proportional behaviour, there is also a derivative term, but that does not matter when servo is not moving which is the case we normally want to consider for load sensing.
If the servo is to apply any torque, current must be applied to the motor. Torque is proportional to current when the motor is not moving and no back emf. The current is pretty much proportional to the on to off ratio in the lower current part, guess there are some inductive effects in the motor
If the servo is 30 microseconds off position ( 3 degrees) , it is loaded to about 30%.
Of course with serial HMI you have accuracy 10 times greater than over pulse HMI, and more accurate communication. However measuring the position is still a good way to measure the torque. Also reading the position via serial HMI does not turn the servo off, like the pulse HMI position read.
by AI-Bot
Wed Jun 13, 2007 5:01 pm
HI
This is probably has nothing to do with anything.
But I notice that Rook Pawn 3 uses Hitec 5990 servo
I have a quick questions.
Hitec 5590 servo and Hitec HS-5995TG - DIGITAL
I know that the speed is 0.12 for 60degree
and their torque is very good.
But how is the noise on them... are they noisy like any other servo.
They are very expensive, just curious before I buy them.
p.s. Nice looking robot.
This is probably has nothing to do with anything.
But I notice that Rook Pawn 3 uses Hitec 5990 servo
I have a quick questions.
Hitec 5590 servo and Hitec HS-5995TG - DIGITAL
I know that the speed is 0.12 for 60degree
and their torque is very good.
But how is the noise on them... are they noisy like any other servo.
They are very expensive, just curious before I buy them.
p.s. Nice looking robot.
by i-Bot
Wed Jun 13, 2007 5:59 pm
The HSR-5990TG has HMI, the HSR-5995TG does not, so avoid it. The 5990 seems more quiet than the 5995, though both are significantly more noisy than most servos. The 5990 also has motor overload protection and better heatsink.
The 5990 has a metallic ring to its sound, like large coins in a bucket, which is rather appropriate !
The 5990 has a metallic ring to its sound, like large coins in a bucket, which is rather appropriate !
by StuartL
Mon Jun 18, 2007 5:34 am
Having now seen this robot in person this weekend at the RoboGames I can confirm it's a monster. It's a superb piece of engineering and Matt's work on the PC side of software is easily as good as the hardware and software on the robot.
I can't comment on the 8498s but the robot as a whole whistles at an uncomfortable frequency when it's stood/straining. When Matt was doing extended testing my ears felt like they were being drilled from the inside with the tiniest drill conceivable. When in motion the servo noise drowns out the whistle but the servos are still surprisingly quiet considering the load they're under.
You lie! You carried on programming all the way to the end of RoboGames!
It was a pleasure to meet you and a pleasure to see Rook's Pawn in person. I only wish my stuff had arrived in time to give you something else to beat up!
bauermech wrote:They're quieter as well. The 5995's screamed like wild banshees. These are not that much more audible over the 8498's.
I can't comment on the 8498s but the robot as a whole whistles at an uncomfortable frequency when it's stood/straining. When Matt was doing extended testing my ears felt like they were being drilled from the inside with the tiniest drill conceivable. When in motion the servo noise drowns out the whistle but the servos are still surprisingly quiet considering the load they're under.
Still using the MR-C3024. Programming away... won't stop until RoboGames starts.
You lie! You carried on programming all the way to the end of RoboGames!
It was a pleasure to meet you and a pleasure to see Rook's Pawn in person. I only wish my stuff had arrived in time to give you something else to beat up!