Roboclaw motor controller investigation

Hi @Pit: yes, motomonster is a good an proven choice combined with Arduino

but I really want to have a try with a controller board with PID inside to compare!
As JRK a bit expensive for me (200$ for 2 DOF), i'm looking at alternatives...

Another "Closed loop position control" board: Motion Mind 3

• 6-24VDC motors,
• 9A continuous,
• 25A peak
• 75$ per motor

http://www.solutions-cubed.com/content/Downloads/MotionMind3Docs/MOTM3DS_6.pdf

12V x 12A for JRK -> 144W
24V x 9A for MM3 -> 216W
It may be interesting to investigate
0,35 $/W
0,68 $/W

Hi

Sabertooth 2x32 (usb) + Kangaroo x2 two motors @ 32 amp cont 64 amp peak $150.00

24x32 = 768, = 0.20 $/W

regards Terry

Yes!
I was curious to compare the solutions on this $/W indicator basis.
I'm not sure it's perfectly relevant but it's another information

I'll update the list of PID controller (peak values, cost per watt) with these information.
yours
Mat

Remember with motion simulators, especially DIY, we run the motors at 2.5 to 3 times the rated current as the motors run at stall current.
so JRK is a 12v 40 amp continuous controller - 480watts, although Ive had mine at peaks of 530 watts.
12v @480 watts = 0.20 cents per watt. (jrk cost me $99 US, so AU dollar rate is higher cost)
The interest here is to find an alternative controller for sure. The JRK is the best around hands down, but to find a suitable equivalent at a cheaper cost is for sure the go.
And as @eaorobbie says, even though there are codes here written by others for the arduino, I for one, am code hopeless, even if someone writes a code for me. Hence the JRK is controller of choice for me.
Simple setup for simple person.
@RacingMat , maybe buy some of these new controllers yourself and explore how they work and share the info on site.
Not everyone is as clever as you and you need to remember this to make it easy for everyone.
People here are all for a "group" share, but nothing seems to eventuate from it. I know with the DIY actuator discussion, unless I am wrong, only myself actually went and bought the parts for a DIY actuator to start building one. Now people are interested, why? Because I spent the money. I didnt ask for a group share, but I am sharing the info in return.
That goes for anyone on site, if you think its good, buy it and test it and SHARE IT.

The continuous output current is approximately 12 A perhaps 15A (30 A max) but not 40A continuous.

@Pit , myself and @eaorobbie have had the JRK at 40 amps continuous. We have Both tested the JRK very hard. I dont post things like that unless I know for sure.

We can set the max current in the utility program at 40 amps, then run a HARSH profile and then watch the utility program spike at 40 amps or more.

@bsft : I'm only at the stage of gathering informations in this "investigation" topic
and next I'll synthetise the results and share what I'll buy and try!

PS: sorry google didn't succeed in translating "People here are all for a "group" share, but nothing seems to eventuate from it."

Your right but with good cooling and decent gauge wiring the Jrk can push way past its limits. The Mostor Moto wont push that hard (they shutdown) and I haven't seen an Ard code with Full PID control like the Jrks makes a massive difference to the motion, as in high speed stop starts that make you leave the seat or let go of ya steering wheel. LOL bracing your self for an impact in game on a motion sim, think I got the profile and setup pretty close.

Plus all you guys are comparing is $ for power, lol .
How about $ for quality (resolution) of movement, and end speed of movement, accurateness.
(mmm This requires an industrial quality PID controller. mmmm where do I get one of those.)

Yes, benchmarking would be a nice (and challenging) feature
I was wondering if Simtools could play a "standard" serie of data with shifts that would be noticeable (or not!) on the sim.

By using this smart tool from @RufusDufus, we could look if the signal replicated physically.
http://www.xsimulator.net/community/threads/simple-digital-scope-for-motion-sim-testing.5485/

The calibration signal would send shifts that would have reducing period of time: does the sim has enough reactivity to follow the signal? We could visualize this. And counting the number of first visible periods could be factual.



In a complementary manner, the signal could increase range of movement and see at what point the sim hasn't enough reactivity to reach the ends of movements:

@value1 @yobuddy : Do you think it would be possible to make Simtools (or a plugin) play a preset of data to benchmark our sims?

@RacingMat , what do you mean "benchmark"? in actual motion in the unit? Or actual processing of data via the controller?

Final result: actual motion via pots reading (as described by Rufusdufus)

But this could let us to fine tune our sims through the whole processing chain

- windows com port speed,
- Simtools : delay of packet rate (ms) is it efficient to reduce it ? until when?
- code improvment (Arduino),
- voltage of power supply (12V -> 24)
- using a battery in addition of PSU
- using pot or hall sensor?
- adding springs?
- moving the center of gravity
- changing the rod lentgh (shorter or longer)?
- with or without additionnal loads (screens...)​

things like this

Ok then, I suppose that may help considering I have done just about ALL of that already. Exceptions being arduino coding and only using a voltage range of 12-15 volts.

But every sim is so different , what I love here, we don't just copy.

My answers -

1.Flat out
2. Flat out 10ms but no noticeable diff to motion at 20ms
3. Jrk doesn't need it. (hence I recommend these to newbies)
4. 16v is a must to get the ultimate performance from the Jrk, currently running at 14v constant.
5. A must when using the MonsterMoto or Jrk or you need a protected psu , if not psu will burn out eventually due to back current
6. Pots (wire wound) , hall effects are no good at high speed and show noise on a Jrk ( well fail after 40hrs of use to me that's a week of simming)
7. This is the most important part of the initial design , get this wrong and the sim will have poor reactions.
8. All depends on how much throw you use , not a simple thing to test with.
9. All comes back to No7.

Sorry my 2 cents.

@yobuddy has created a video plugin ("Video Ride Creator" and "Video Ride Player"), where the motion cues are hardcore for a video. There is an editor available to create the motion cues. I guess this is ± what you're looking for.
You can program like 0.2 sec of pitch rotation, then 3 sec of yaw, etc. or whatever you like.

@value1 : thanks!
Exactly: that sounds good!

Yes, this would work!

@RacingMat, Just note however depending on "what part" of the system you want to compare you may need other approaches. This will compare everything including Simtools scaling and individual sim settings so if you are trying to compare with someone elses rig they may have different scaling etc that drives the sim harder or softer with the same input signal. If all your testing is with one rig this would work great.

If you have the time, give the Digital scope a go, use it just running your sim with a game. Not having the opportunity to sit in any other sim to compare what it "could feel like", I found the software was a good way to see if my sim was at least tracking the output of simtools... I discovered even though it seemed OK to me there was initially alot of tracking error. Now I have fixed that I can definately notice the improvement.

But then again, if I didn't have the software, I would have just continued to think it was OK and been happy anyway! Maybe its better not to know?

What? No good? I have used hall effect units for over three years and hundreds of hours of use without any issues whatsoever! And what does high speed mean? Even 180 rpm is a relatively slow speed for them (If we are talking about the BI Technologies units they are rated at 500 rpm when using the cheapo bronze bearings). I foresee only three issues that could happen with them; the shafts bearing goes out, the magnet comes detached from the shaft, or more remotely, the electronics goes out from fatigue.

If a bearing goes out in 40 hours then it is because it is being overly stressed. As I stated in several different posts already, it is imperative that the pots housing is able to float with any out-of-round movements or other unwanted forces applied to it or it will fail much sooner than it should. If the magnet came loose it is probably just a defectively made unit. If this is the case, are we talking about only one unit giving you problems so far, or several? Same question if the problem was with the electronics? To my knowledge, no one else has stated any problems using them until now. But, considering the price of the units mentioned above ($15 to $20 usd), it is not out of the question. If the problem is the bearing, and you can’t correct the mounting to eliminate the fatigue problem, you may want to consider upgrading from standard bushings to ball bearing ones.

Everyone, give hall effect units a try! They are amazingly accurate and noiseless (unlike standard pots wire wound or not)! My two cents too!

ps While they are “noiseless”, due to the nature of digital electronics (square wave vs linear), they will sometimes appear to shift back and forth by 1/10 of a degree when they are on the edge between tenths of a degree.

pps This thread was split from another because it was taking a different path and now it appears it needs to be split again for the same reasons!

? what bearing , units I have had a stainless shaft in brass bush. Poor but at 21 buck were cheap, my wire wounds are 36 each and I have a sim still running them 5 years later and is used about 6 hrs daily, 7 days a week. As a training rig.
And not noiseless if your using a quality driver, lol they never reach 5v's too. Sitting still and zooming in on the output graph shows them fluxutating about 2% of total value. Wire wounds = zero.

I agree with how you said to mount them, lol used a lathe they are spot on. Actually had to spray a little off non conductive lube on them but got upset stuffing with them and went back to what I have been using for years.

Try hall effects but be warned not as simple as stated.

This still in very early beta, not supposed to be here in public yet, mmm.