Looking to build an OSW setup with Simucube and Kollmorgen AKM motor, could use some help.
2 2

54 posts in this topic

Hello,

If you're impatient just skip down to the questions bit. :-)

I've been lurking this community for a long time, but now I have a good reason to post.  My old Fanatec GT3 RS V2 is dying and I've decided to go all out for an OSW (Open Sim Wheel) setup.  I've spent weeks researching like crazy and reading literally hundreds of forum pages here and at many other forums.  I've even been using Google Translate to read French, German, and Italian forums which seem to have a lot of good information as well, even though it is tedious.  I'm a big computer geek, but I'm not an electrical engineer so I'm learning as I go when it comes to power supplies and motor specs.  I'll present a summary of the things I've learned to help others as I learn myself.

Here are some examples of forum resources I've found on OSW primarily contributed by the famous "Beano":

http://www.isrtv.com/forums/topic/19813-leo-bodnar-simsteering-v2/
http://isiforums.net/f/showthread.php/24986-Project-BruteForce2-IONI-based-DIY-DD-FFB-Wheel
https://forum.virtualracing.org/showthread.php/95803-Project-BruteForce2-IONI-based-DIY-DD-FFB-Wheel-(Extensive-worklog)
http://forums.overclockers.com.au/showthread.php?t=1139613
http://www.racingfr.com/forum/index.php?showtopic=51856

That last link is of particular interest as it's a fairly recent summary from Beano on the current state of things and it makes it sound like I can use an AKM motor with Simucube.  After all the research it looks like I've gone through the entire history of OSW and much of the information is no longer up to date.  Today I'm looking at a Simucube based build hopefully using a Kollmorgen AKM motor.  I want the highest quality OSW wheel I can reasonably build and ideally I want 1:1 torque.  I've checked out a number of vendors who sell OSW kits and found lots of Mige stuff but no Kollmorgen kits aside from what Dennis Reimer sold using an Argon controller (he closed up shop so that isn't even an option now).  I'm located in Canada and the cost of parts with shipping and currency exchange is high so I'm trying to be as smart as possible about how I source everything.  I'm going to see what my options are for motors locally before buying everything else foreign.

When it comes to picking a motor I know the Mige is a great option, however if I'm doing so much work to get this researched and built I might as well upgrade a notch and not have to upgrade again.  The Lenze might be a good choice and I might actually go with that, but can I do better?  So I watched Barry at "Sim Racing Garage" and his reviews of the high end Bodnar systems and here's what I learned:

Sim Racing Garage:
https://www.youtube.com/channel/UCT_50ZiRFWSm7oziV1OZloQ

Bodnar SimSteering Specs:

Bodnar SimSteering V1 Motor = Kollmorgen AKM52G-ANCNEJ00 - Encoder based (EJ encoder)
Bodnar SimSteering V2 Motor = Kollmorgen AKM52G-ANCNC-00 - New resolver (C resolver – AKA "Smart feedback device")

Both V1 and V2 use the XP DNR480 Power Supply for AKM52 based motor.

Barry also tested an OSW kit with a Kollmorgan AKM53G-ANCNR-00 motor which was popular (also resolver based ("R-" in the part number), powered via Argon).

Apparently this new Bodnar V2 system is a step ahead of the V1 system using the new "smart" Resolver based feedback according to Barry.

How to read a Kollmorgen part number:

AKM53G-ANCNC-00 translates to:

AKM series
Frame size 5
Rotor stack length 3
Winding Type G
International Standard Mount (A)
Smooth Shaft (N)
IP65 connectors (C)
No brake (N)
Smart feedback device Resolver (SFD) (C-)
Standard motor without shaft seal (00)

The "C-" in the part number is the resolver.  Or, if the motor has the "EJ" code instead that means "10000 PPR digital encoder".

Where to find Kollmorgen motor details:

http://www.kollmorgen.com/en-us/products/motors/servo/akm-series/akm-series-ac-synchronous-motors/ac-synchronous-servo-motors/

I pulled the catalog and motor selection guide from here and dug in.  The numbers are daunting, and this is where I'm starting to get lost.

Power Supplies:

The "Mean Well SDR-480-48" seems to be a popular choice that is packaged with kits, however I want a power supply that has no risk of straining.  Beano presents some other options here:

https://forum.virtualracing.org/showthread.php/95803-Project-BruteForce2-IONI-based-DIY-DD-FFB-Wheel-(Extensive-worklog)?p=2137757&viewfull=1#post2137757

With that, I wonder about using something like this:

http://www.antekinc.com/ps-8n48-800w-48v-power-supply

However, it does appear that Bodnar is powering AKM motors with only 480 watts.

SimuCube / IONI:

https://granitedevices.com/wiki/About_SimuCUBE
https://granitedevices.com/wiki/SimuCUBE_hardware_requirements

Kollmorgen motors are not listed in the Simucube motor table here.

Questions:

  1. I'm pretty sure I want more than an AKM52 for torque.  Is an AKM53 enough, or do I dare consider an AKM54?  As the motor gets bigger you get more torque, sure, but you also lose out on other aspects such as speed and responsiveness, I think.
  2. I don't know much about the motor winding type differences.  It seems like most motors are using the "G" type winding (Like AKM53G), are there any other options I should consider?  I hear the "G" winding is optimized for more torque.  Beano has a setup that uses the "K" winding.
  3. What kind of power do these Kollmorgen motors take?  From the spec documents I gathered they are "Brushless servo motors, AC".  Can the Simucube / IONI power them?  I think so...but I'm not positive.
  4. On page 61 of the Kollmorgen catalog they list the AKM specification numbers in a big table, with different columns under different voltage values.  How does that voltage stuff work here?  Suppose I have an 800 Watt PSU supplying 48 V to the Simucube with IONI Pro HC, and the IONI supplies a max amperage of 25 A to the motor...um...something something voltage?  I know Watt = Amp X Volt, but I'm getting lost here.  How do I read that fancy table?
  5. Simucube does not support resolvers, it only supports encoders so I would be forced to use an Argon for a Resolver motor.  Should I stick with the encoder, or consider abandoning Simucube for Argon?  Simucube seems like the better choice, I just wonder if I'd be missing out on something with the encoder.
  6. I understand it may be possible to get an encoder with even higher resolution than 10000 PPR?  "You will get better positioning and thus feel finer FFB forces.....after 32k PPR though, no advantages......between 16k and 32k the best for is racing use I believe.... " - Beano
  7. Are there any other good forums or resources I haven't yet found that could help me out?

I suppose that's enough for a first post.  I haven't even begun to figure out wheel details yet, just getting this motor and Simucube stuff sorted out is my first step.

Thanks. :-)

 

JonnyK

Share this post


Link to post
Share on other sites

1: It is possible that the AKM52 could make more torque than the AKM53, you'd have to spec everything out to see that, too much work. The AKM54 is overkill from what I gather.

2: The "G" is the "optimal" winding as it offers a better torque per amp, than the "K" which is frequently found as well. 

3: These are AC Brushless motors, essentially you power the drive with AC power (Argon) or DC power with a PSU (IONI/Simucube). The drive typically uses an IGBT which 'chops' the DC power, simulating AC power with frequency modulation. This is done at various frequencies to regulate the speed. It is also done at the maximum voltage of the DC bus, creating the torque. If you look at a torque/speed map of the motor you select, you'll see that the torque output is hgih and fairly flat up to a certain speed and then it drops significantly. 

In the case of an OSW setup you have to look at the Nm/Amp, there's a formula that goes with this based on the motor selected. 

4: Kinda goes with the last tidbit in #3, there's some more math that goes on with amps/RMSx something. Sorry I have class shortly and am trying to help. A 800W PSU could push 16amps if the efficiency is high enough. I think the AKM motors are suggested to use 1000+.

5: That is true, even then the Argon drops the steps by a power of 4, not really worth it. I converted an AKM53K Resolver to a 10000ppr encoder.  I'm going to be putting a Simucube together in the spring.

6: The Bondar V2 use Kollmorgens SFD, which is capable of some obscene PPR, check the spec book for the exact number. At slow speeds the detail can probably show thru, but at speed I don't believe you'll be able to tell the difference. 

7: Forums: iRacing's if you have a subscription (3 or 4 HUGE threads for the Argon/IONI/Simucube builds. There's two more, don't recall the names off hand, one in French and German. X-Simulator might have some info. 

If you have more questions, I'll try to clarify later. If someone with more experience feels I'm wrong, please correct me. I''m still learning as well. :)

Share this post


Link to post
Share on other sites

Thanks for the responses guys, very useful.  Yes, I've heard the iRacing forums have a lot of good stuff in them, however I'm not a member so I can't access that.  A closed members only forum full of awesome "Open Sim Wheel" info?  Boo.  :?  However I am still stuck on the electrical details.  I'm fully aware that the more I talk about it, the stupider I'll probably sound, however let's get to it. :-P 

Attached is the full AKM motor specs, I've highlighted both the 52G and 53G:

AKM.png

And now here are the IONI electrical specs:

https://granitedevices.com/wiki/IONI_specifications

So the 52G has a torque constant of "1.70 Nm/A" while the 53G has "2.39 Nm/A".  Okay, so the 53 makes more torque for the same input power.  But now how many amps will I be sending to it from the IONI to calculate the torque I'll get?  The IONI Pro HC can output a max of 25 Amps.  So then...25 * 2.39 Nm = 59.75 Nm.  But that's not right.

Okay, what about voltage?  Let's see..."48V supply, the maximum three-phase motor output voltage is 32.2 V RMS phase-to-phase".  Okay then...so I can only output 32.2 volts to the motor?  But that doesn't even register on the table (which goes up to 480V).  Umm...

Okay, let's look at input power.  The Simucube takes a maximum of 48 VDC power in, but then the IONI Pro HC can take 55 VDC in.  So...that's 7 VDC I can't access?

While looking for PSUs that output more power but are still under that 48 VDC I find this:  http://www.antekinc.com/ps-15n44-1500w-44v-power-supply/

1500 Watts sounds like way more than I need, but the 44 V seems to fit, maybe?  Is it a problem if it's 44v instead of 48?  But wait, that PSU has the outputs split to output 17A each, and I'd have to wire them in series for more, which would put me at way too high voltage.  Right?

And now for another stupid question - if the power outlets in my house supply 15 amps, and I've got a crazy 1500 watt PSU plugged into them that provides 44 amps, that's probably not a good idea, is it?  Or is it just 44 amps at a max output, but the input power will trickle in at a lower rate?

 

That's probably enough for tonight.  Devil Inc, you might have to give me the "electronics for dummies" lecture here.

Thanks,

 

JonnyK

Share this post


Link to post
Share on other sites

there is a guy website -> boxthislap or "Beano" on iRacing, he works closely with Granite and has build every single OSw combination there is. (at least to my knowledge) he is using a simucube + lenze and his 2nd rig is a LB SS2. you could google him too if you need some help. just wanted to let you know. good luck! 

Share this post


Link to post
Share on other sites
11 minutes ago, rocafella1978 said:

there is a guy website -> boxthislap or "Beano" on iRacing, he works closely with Granite and has build every single OSw combination there is. (at least to my knowledge) he is using a simucube + lenze and his 2nd rig is a LB SS2. you could google him too if you need some help. just wanted to let you know. good luck! 

Yes rocafella, I'm familiar with Beano.  I've learned a lot from reading his enormous forum posts!  Recently he seems to be playing with an AKM53K servo: http://www.racingfr.com/forum/index.php?showtopic=51856

Share this post


Link to post
Share on other sites

Jonny,

I have one of the Batch 1 Ollie DIY kits with the small MiGe.

At the time, the only info available was on I-racing, so to further my research I took out a subscription to join the forums.

In hind sight, I would do the same again, it was absolutely worth it and small money compared to the cost of the components.

Should be some special deals this time of the year.

Good luck and please share your build with the forum.

Cheers!

J.J.

Share this post


Link to post
Share on other sites

@JonnyK

Take a look at this. This applies to all drives, though it mentions the IONI. So the AKM53G is 4.77 RMS X 2.39 = 11.4003NM

So to drive the servo to that rating in IONI/Simucube: 4.77 RMS X 1.4142 = 6.745A POS

If you wanted the full 29.7NM out of the motor: (29.7NM / 2.39) X 1.4142 = 17.574 POS in Granity

You would need a Simucube paired with the IONI PRO (Max 18A Output). It could draw 25A at 48vDC if needed with the IONI Pro HC installed. 

An RSP-1000-48 ([email protected] PSU) draws [email protected] MAX and is able to output [email protected] according to the datasheet.

That should get you full torque from the servo. I have an AKM53K, so I'd need a bit more power. I'd need a setup that could power 35A to get the full 30.3NM.

The motor is rated in RMS current (Root Mean Square)

The drive output is rated in Peak of Sine current (POS).

If memory serves, the small Mige is rated at 9.1A RMS peak and 2.20NM/A torque constant. 9.1 X 2.2 = 20.02NM torque.

To get the full servo peak torque in IONI, you need to set the POS current level in IONI at 9.1A RMS X 1.4142 = 12.87A POS.

Thus IONI MMC of 12.87A POS setting will drive the servo at 9.1A RMS, which is ~=20NM.

MMC = 12.87
MCC = 12.77

Now, the above is true of you set MMC and MCC in Granite as per the above, MMoS to 100% and the IRacing slider to 20NM.

Under Driving the servo:

With all else equal, but moving the IRacing slider to 30NM:
Real servo torque then becomes 20/30x 20NM = 13.33NM

Moving the IRacing slider to 40NM:
Real servo output then becomes 20/40x 20NM = 10NM

Overdriving the servo:
For example, moving the slider to 15NM
20/15x 20= 26.67NM.

Under driving the servo, no other changes need in the setup. Overdriving the servo, you can see that more current is needed in the IONI setup. How much more?

Well, 26.67NM with a torque constant of 2.2NM/A = 12.12A RMS.

12.12A RMS X 1.4142 = 17.15A POS ( this is what would be set in Granity MMC, perhaps 17.00 for MCC)

Above just showing the calculations in how it all works, I won't advise overdriving the servos though wink.gif if you want to play with overdriving the small Mige to 25NM, you will need (25/2.2)X 1.4142 =16.1A POS setting in Granity, and then move the slider to 16NM in IRacing.

Again, not a good plan, as you will induce clipping, much better to get the correct servo that suits your physical abilities.....

Above all for the small Mige, the big Mige have a torque constant of 1.57, the Lenze = 1.4 and the AKM53K = 1.24, you need to recalculate granite settings.

Share this post


Link to post
Share on other sites

@Devil Inc, thanks for the update.  I've read through the details you posted and that fills in some details.

Meanwhile, I contacted a local Canadian supplier and was able to get a quote for the motor.  The motor actually came in a little less than I had expected (~2000 CAD) but the power and data cables that go with it then add a lot to the price (~$375 a cable).  I'll try to get a few quotes to compare.

I hadn't spent any time researching the cables, but it looks like now I need to do that too.  I'm thinking that if I have to spend so much for the cable, maybe I should get the proper connectors to go with it so I can wire up the connectors in a box and then cleanly plug in the cable to the box (like the Bodnar setup) rather than having loose wires coming out of the cable end and wiring into the SImuCube.  I understand there are different cable specs based on how much power is moving through them, so I'll need to understand those requirements too.

I'll keep you guys updated.

 

JonnyK

Share this post


Link to post
Share on other sites
4 hours ago, JonnyK said:

@Devil Inc, thanks for the update.  I've read through the details you posted and that fills in some details.

Meanwhile, I contacted a local Canadian supplier and was able to get a quote for the motor.  The motor actually came in a little less than I had expected (~2000 CAD) but the power and data cables that go with it then add a lot to the price (~$375 a cable).  I'll try to get a few quotes to compare.

I hadn't spent any time researching the cables, but it looks like now I need to do that too.  I'm thinking that if I have to spend so much for the cable, maybe I should get the proper connectors to go with it so I can wire up the connectors in a box and then cleanly plug in the cable to the box (like the Bodnar setup) rather than having loose wires coming out of the cable end and wiring into the SImuCube.  I understand there are different cable specs based on how much power is moving through them, so I'll need to understand those requirements too.

I'll keep you guys updated.

 

JonnyK

If you need the pin out for a kollmorgen servo cables, I can help you out. Also, if you want to roll your own cables you can save a ton of cash. The ones that come with the kollmorgen servos are overkill (this is an understatement to say the least... but I digress). 

Another thing to be mindful of is that the bigger servos are more suceptable to "cogging" because of the greater rotation intertia of the motor. Also, keep in mind that anything over 2048 ppr is barely detectable for dd servo wheels. Also, keep in mind that you can only run the osw ffb with an argon based drive... everything else is mmos based with an spi interface like the discovery board or the simucube's all in one spi controller. 

Share this post


Link to post
Share on other sites
7 hours ago, Anthony A. said:

Which motor did you get the quote for?

I got a quotes for both the  AKM52G-ANCNEJ00 and  AKM53G-ANCNEJ00.  The price difference was only a few hundred to move up to the 53.  The company that quoted me asks that I don't discuss the quote, which is why I'm not naming exact pricing.

 

JonnyK

Share this post


Link to post
Share on other sites
18 minutes ago, JonnyK said:

I got a quotes for both the  AKM52G-ANCNEJ00 and  AKM53G-ANCNEJ00.  The price difference was only a few hundred to move up to the 53.  The company that quoted me asks that I don't discuss the quote, which is why I'm not naming exact pricing.

 

JonnyK

Not to completely hyjack this thread and drag it down a rabbit hole... (wait for it...) buuuuut, that seems to be a common occorance with industrial equipment (and servo based direct drive kits. You may find that shopping around will save you a few bucks (on the cables especially... I've seen them for $150, and even that's high for what they actually consist of (shielded 15 wire twisted pair copper wire with a drain and proprietary plug on one side) especially if you are going to have to solder the other end of it (hopefully correctly). Then again, it's not like you can make these things yourself in the garage, unless you have a foundry in the shed in your back yard. 

Share this post


Link to post
Share on other sites
20 hours ago, posimosh said:

If you need the pin out for a kollmorgen servo cables, I can help you out. Also, if you want to roll your own cables you can save a ton of cash. The ones that come with the kollmorgen servos are overkill (this is an understatement to say the least... but I digress). 

Another thing to be mindful of is that the bigger servos are more suceptable to "cogging" because of the greater rotation intertia of the motor. Also, keep in mind that anything over 2048 ppr is barely detectable for dd servo wheels. Also, keep in mind that you can only run the osw ffb with an argon based drive... everything else is mmos based with an spi interface like the discovery board or the simucube's all in one spi controller. 

@posimosh, I missed this response from you earlier.  Regarding your statements:

Cogging - So this is where you can feel little steps as the motor turns at low speed, correct?  At high speed it smooths out, but at low speed the magnetic attractions within the motor causes a jerkiness in the motion.  Right?  From what I've heard, the Kollmorgen motors are supposed to be much smoother than other options, so even if this were the case I'd assume it would be a lesser problem with these motors, wouldn't it?

Over 2048 ppr is barely detectable - Really?  What's the deal with people upgrading their 5000 ppr resolvers to 10,000 ppr parts then?  Is it all snake oil?  Whether this is truly the case or not, since I'm going crazy building this thing to begin with, I might as well get a high res resolver for it whether it makes a noticeable difference or not.  My arm muscles will surely be the weakest link of this whole system. :-)

Can only run OSW FFB with Argon - I haven't looked much into the software side of things yet, however from my reading it appears a lot of people have been saying things like "If I had known about Simucube before I did [thing] I would have done that instead!" or "Simucube is definitely the way to go now...".  Are there any notable things I'm losing out on by not being able to run OSW FFB?  Isn't it the case that OSW wheels are simply identified as generic wheel controllers by the system and all games can make use of them?  Is MMos a problem?  Looks like I'll need to download it from forum.virtualracing.org, which is a bit scary, the idea that I'm building this crazy expensive wheel that depends on software hosted in a forum post...

Edit - I just noticed a post here:  https://boxthislap.org/mmos-afterwards/  Apparently Simucube will have it's own open source firmware soon.

 

JonnyK

Edited by JonnyK

Share this post


Link to post
Share on other sites

I've been spending time reading and learning about power supplies and electronics in general, and things are finally coming together in my mind, at least partly.

Watts = Voltage * Amps

Voltage is pushed from a power supply to a device and doesn't change, where amps are pulled from the device and can vary all the time.  The total amount that can be pulled can't exceed the total wattage of the PSU, and since voltage is fixed, that lets you calculate the maximum amps that can be pulled.  So, my household electrical provides 120 volts on a 15 amp circuit, so that's 1800 watts total.  If I plug in a 1000 watt power supply that outputs 48 Volts DC it will use up to 1000 of those input watts and the output will then be maximum 1000 watts at 48 volts, so 20.8 amps.  But it could also be using 48 volts and 2 amps (96 watts) or any other amount of power drawn based on the device, that amperage draw varies.

So, now looking at the motor specs again - The AKM53G is rated using "RMS (Root Mean Square) Amps", but we need to convert those values to "POS (Peak Of Sign) Amps" by multiplying RMS values by 1.414.  https://granitedevices.com/wiki/Peak_value_of_sine

Standstill current = 4.77 RMS Amps -> 6.75 POS Amps
Peak current = 14.3 RMS Amps -> 20.22 POS Amps

At this point, something doesn't add up.  The peak current is 20.2 amps, but from the calculation example from a few posts back we should actually have a limit of 17.57 amps, since the motor max torque is 29.7 which is then divided by the torque constant (2.39) and multiplied by 1.414.  So for maximum torque we need 17.57 amps, but the motor has a peak draw of 20.2 amps?  Is that correct?

Assuming we use 20.2 amps, the maximum power draw from the PSU would be 48 V * 20.2 amps = 969 watts.  Therefore, I need a 1000 watt PSU.

With all that said, I still don't understand the AKM53 spec chart where it has voltages down the left side (see post #4).  It lists 75VDC, 115V, 230V, 400V, and 480V.  What does it mean?  It sounds like I can power this motor directly from a power supply with 480 volts, and the more voltage I give it the more speed I get (but not torque).  But how much voltage does the IONI send to the motor?  It appears to be limited to 32.2 V RMS?  See the "Motor output" section here: https://granitedevices.com/wiki/IONI_specifications

 

JonnyK

Share this post


Link to post
Share on other sites
2 hours ago, JonnyK said:

@posimosh, I missed this response from you earlier.  Regarding your statements:

Cogging - So this is where you can feel little steps as the motor turns at low speed, correct?  At high speed it smooths out, but at low speed the magnetic attractions within the motor causes a jerkiness in the motion.  Right?  From what I've heard, the Kollmorgen motors are supposed to be much smoother than other options, so even if this were the case I'd assume it would be a lesser problem with these motors, wouldn't it?

Over 2048 ppr is barely detectable - Really?  What's the deal with people upgrading their 5000 ppr resolvers to 10,000 ppr parts then?  Is it all snake oil?  Whether this is truly the case or not, since I'm going crazy building this thing to begin with, I might as well get a high res resolver for it whether it makes a noticeable difference or not.  My arm muscles will surely be the weakest link of this whole system. :-)

Can only run OSW FFB with Argon - I haven't looked much into the software side of things yet, however from my reading it appears a lot of people have been saying things like "If I had known about Simucube before I did [thing] I would have done that instead!" or "Simucube is definitely the way to go now...".  Are there any notable things I'm losing out on by not being able to run OSW FFB?  Isn't it the case that OSW wheels are simply identified as generic wheel controllers by the system and all games can make use of them?  Is MMos a problem?  Looks like I'll need to download it from forum.virtualracing.org, which is a bit scary, the idea that I'm building this crazy expensive wheel that depends on software hosted in a forum post...

Edit - I just noticed a post here:  https://boxthislap.org/mmos-afterwards/  Apparently Simucube will have it's own open source firmware soon.

 

JonnyK

Regarding cogging, it seems like you can tune any notchyness out of even the akm54g if you mess with the filtering long enough. And I believe your description is of it is basically correct. And it gets more noticeable the more ffb effect you are feeling. So like the few times I've noticed it with my akm52g and small mige is when I'm trying to save a car that is destined to spin or fly in to a wall and then many pieces anyway. But I run relatively low ffb, and no filtering of any kind. I can't really speak to it anymore than what I've read (and it sounds like you have read as well)

so... as far as encoders and resolution are concerned, you are going to have to trust me on this... I've used a 5k mige, a 10k mige, a v2 "baby" bodnar, and a 5k kollmorgen osw/dd... you can't tell the difference between any of them in practal terms. That is, the simulator has to estimate the input between the steps anyway, because the input is digitial, and not analog. To the simulator, it's just not that important if you change your wheel angle 1/100th of 1 degree or 1/10th of one degree, or really even 1 degree. It is only at 10 degrees that it will start making a difference while catching a slide, hitting a turn in point, or whatever. 

As for the firmware stuff and simucube, if you are building a brand new, direct drive wheel from scratch, and don't have unlimited money to spend, buy a simucube. It's simple-ish, purpose built for direct drive wheels, and inexpensive relative to any other option. It's a streamlined solution for direct drive wheels, and a great product. It doesn't, however, make anything better in practice except for the easier assembly and wiring. That's it. Once it's built, who cares... 

As for the distinction between an osw, and a direct drive wheel, is semantical really. Bernhard Berger designed a ffb API using only a servo and argon from the ground up, and made the code open to anyone who wanted to use and build on it as long as they followed the open source license. Everything else uses other methods some of which borrow code from other projects to make it more user friendly. The problem is, they are breaking the liscene agreement of some very large companies (that make... windows). Long term, this could/would be problematic for anyone selling kits using this method/hack explicitly.... which makes the open source solution that granite devices may or may not come up with more realistic for any commercial entity wishing to sell kits ready to play. 

***edit... sorta trailed off at the end... getting tired... 

Edited by posimosh

Share this post


Link to post
Share on other sites

I've read on multiple occasions a 10k encoder even over a 5k one provides smoother feedback over curbs and a more progressive feel when under steering for example. The comments from folks going from a 2.5k encoder to 10k is overwhelmingly positive in this regard. 

Can you confirm this from your experience?  

Share this post


Link to post
Share on other sites

So, I'm beginning to have some serious doubts about Simucube and Kollmorgen together.  I've been reading a lot about motors and going through Simucube / IONI specs.  The Kollmorgen AKM53G seems to be rated between 230V and 640V.  Below 230V specs are not even listed.  When it comes to motor voltage, if you over-volt or under-volt the motor you get worse performance, and you also damage the motor.  IONI was not built to power high voltage motors:

Unsuitable/non-optimum motors

  • Motor with very high voltage, such as 200 VAC AC servo motor - it will work if it has incremental encoder feedback, but maximum rotation speed is limited due to drive voltage. We recommend Argon for motors above 100 volts rating.

http://granitedevices.com/wiki/IONI_%26_IONICUBE_user_guide/Motor_compatibility

 

Does this make any sense?  So then I started examining the specs on the Lenze MCS12H15L, which is actually a recommended Simucube motor.  It's rated voltage is 158 V.  That's much lower, but still higher than an IONI can output.  I think I'm going to ask Granite Devices for some input, because it's just not adding up.

 

JonnyK

Share this post


Link to post
Share on other sites
On 12/5/2016 at 6:44 AM, sebj said:

I've read on multiple occasions a 10k encoder even over a 5k one provides smoother feedback over curbs and a more progressive feel when under steering for example. The comments from folks going from a 2.5k encoder to 10k is overwhelmingly positive in this regard. 

Can you confirm this from your experience?  

That was not my experience. But I didn't have the placebo effect that I suspect at least some people are experiencing pertaining to the resolution of the quadrature encoder in the servo. That said, it's not like getting a higher resolution encoder will hurt anything (actually, this may not be the case as it sorta depends on the physics engine and the way it interprets steering input...) and, generally, more is better... it just may not be noticeable if you don't know it is there... but I digress.

Stuff like magnetic pole count, efficiency and speed of direction change, initial force/torque and intertia, and signal refresh is more noticeable to me... but again, what feels best to one won't always feel best for all. 

Mm keep in mind that with a quaderature encoder, the listed points per revolution number should be multiplied by 4 to get the real points of input per revolution. Next time you are at your wheel, try and move it either direction and count how many different points you can hit while moving it ever so slightly in the same direction, over and over until you get to a quarter of the way around... 

Share this post


Link to post
Share on other sites

For today, a note on power supplies.  I spent a bunch of time researching PSU options to try and get around 1000 watts.  There really aren't a lot of options available for 48 volts PSUs.  I wasn't able to find a linear PSU that could produce 1000 watts without substantial cost as linear models are much more expensive than switching units.  As for switching power supplies there were only a few good options in the end without high costs as well.  I even looked at bench test power supplies (the ones where you can configure the voltage with a dial, typically used in electronics labs) but those were much more expensive.  I did find some crazy used power supplies for sale on Ebay, the kind of things that were made to charge industrial machinery batteries but I was too afraid to seriously go for anything like that.

Switching power supplies VS linear:  https://www.circuitspecialists.com/blog/power-supplies-switch-mode-vs-linear/

Ultimately, it came down to these two switching options:

Mean Well RSP-1000-48
http://ca.mouser.com/Search/ProductDetail.aspx?qs=Xb8IjHhkxj5mmeT6EC3r2w%3d%3d

Mean Well SE-1000-48
http://ca.mouser.com/ProductDetail/Mean-Well/SE-1000-48/?qs=%2b6mEGs9UJHzUPduULJoqcg%3D%3D

After doing some research to understand the difference I discovered the main thing is that the RSP model has "Active PFC".  Without a long story, this basically means it's much more efficient and therefore pulls much less power from the wall to produce the same 1000 watts.  In fact, the SE model will typically pull 17.5 amps from the mains while the RSP model pulls 12 amps to produce the same power.  Regardless of your belief in energy efficiency, the fact that the RSP won't cause my home's breakers to blow makes it the clear winner here, even though it's more money.

https://www.rcgroups.com/forums/showthread.php?1541412-What-s-the-difference-between-meanwell-SE-and-RSP-PSU

 

JonnyK

Share this post


Link to post
Share on other sites

I messaged Granite Devices and asked about motor voltages, referring to the Kollmorgen AKM53G and the Lenze MCS12H15L.  This is the response:

___
Basic operation of an electric motor is, that the operating voltage limits the rotational speed, and amperage limits the torque.

To simucube applications, the recommended motors are specified to higher voltage, but it is not a limitation. The motors are selected so, that their can produce enough torque with the range of speed they are intended to (low in this case).

If the motor is specified to 230 V, and is driven with 32.2 V, and if the nominal speed at the specified voltage is 3000 rpm, then from this the simucube can drive the motor at 32.2/230 * 3000 rpm = 420 rpm.

Of the motors you mention, if you check the winding resistance, the AKM53G has 3.97 Ohm, and the Lenze has 1.41 Ohm. From the basic formula U = R*I, it comes, that the maximum theoretical current from a 48 V PSU is 32.2/3.97 = 8.11 A, and the maximum torque is 2.39 * 8.11 = 19.4 Nm.

For the Lenze the theoretical maximum current is 22.8 A, and with the torque coefficient 1.4 Nm/A the maximum torque comes to 32 Nm.

There is no way an under-voltage could cause damage to a motor. With lower voltage, the motor turns slower.

For the 1 kW PSU, it will certainly be enough, most likely even excess. Wheel application doesn't require much continuous power, and so far the SDR-480-48 Meanwell PSU that has 3 second 720 W peak power capability is used widely without any complaints.
___

Interesting.  Based on this new knowledge it sounds as if the Lenze would be a much better match for a Simucube than the Kollmorgen would.  I'm going to refocus my energy on a Lenze build now.

Lenze MCS12H15L Specifications:

Lenze MCS12H15L.png

 

JonnyK

Share this post


Link to post
Share on other sites

Another message from Granite regarding power supplies.  I had asked them about using a 1000w PSU since I wanted to be sure I would have maximum torque and it seemed like 720 Watts wouldn't be enough.  It would appear my understanding of power usage was quite incorrect:

___

It is true, that with 48 V, we can only draw 15 A from a 720 W PSU. However, in this case, the motor would be turning at the maximum speed possible with that voltage.

The thing here is, that we are transferring power. Because of this, we cannot just use the basic formulas. IONI can be compared to a switching DC-DC power supply to some extent.

IONI has the power conversion efficiency of 93-97%, so let's assume 95%.

A rough estimation:

For the Lenze, to achieve the 29 Nm peak torque, we will need 29Nm/1.4Nm/A = 20.7A current. With the winding resistance of 1.41 Ohm this requires 20.7A * 1.41 Ohm = 29.2 V. Thus the power comes to 29.2 V * 20.7 A = 604.6 W. Taking the IONI power conversation efficiency into account, this will need 670 W from the PSU.
___

 

Looks like I can get by with a much smaller PSU than I had expected.

Share this post


Link to post
Share on other sites

So it turns out the Simucube is unable to push a Kollmorgen AKM53G to full torque, but it is capable of pushing a Lenze MCS12H15L to a full 29 Nm, so that's the choice for sure.  The exact part number I need is:

MCS12H15L-C40B0-A19N-ST5S00N-R0SU 

This breaks down as:

MCS Series
12H – Motor Size
15 – Rated Speed
L – 230V spec (Low)
C40 – TTL Incremental encoder with communications signal, IK4096-5V-T (Renco R35i)
B0 – No brake
A – Standard flange form A / FF (B5) shaft without keyway
19 – Shaft diameter
N – Normal vibration severity standard bearing
ST – Separate round plugs for power + brake, encoder + thermal sensor
5 – IP54 protection level
S00 = natural ventilation, no blower fan
N – Without flywheel
R – KTY temperature sensor
0 – International nameplate
S – Black
U – UL / CSA approval

I'm not sure about the C40 encoder though.  4096 PPR?  I thought I read somewhere that there is a 16,000 PPR encoder available?

Share this post


Link to post
Share on other sites

Hello there, i am also building my own OSW with a Kollmorgen AKM motor with a 2048 PPR encoder setup and found a really good "Post" that clearly shows how to build it with electrical drawings, i will share the link here.

FYI: keep in mind that Leo Bodnar Sim Steering Systems all are 24VDC power supply! so in my project i will be using a 24VDC PS.

Here is the link:

http://www.racingfr.com/forum/lofiversion/index.php/t51856.html

I hope this helps!

Best

Milton

 

Share this post


Link to post
Share on other sites
14 hours ago, Milton said:

Hello there, i am also building my own OSW with a Kollmorgen AKM motor with a 2048 PPR encoder setup and found a really good "Post" that clearly shows how to build it with electrical drawings, i will share the link here.

FYI: keep in mind that Leo Bodnar Sim Steering Systems all are 24VDC power supply! so in my project i will be using a 24VDC PS.

Here is the link:

http://www.racingfr.com/forum/lofiversion/index.php/t51856.html

I hope this helps!

Best

Milton

 

Milton,

Are you planning on using Simucube?  In my conversations with the Granite people I learned that Simucube is not able to get maximum torque from the AKM motors.  They even updated their Wiki page in response to our discussion.  See the lower table here:

https://granitedevices.com/wiki/List_of_motors_for_SimuCUBE

 

JonnyK

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
2 2