DMM Dyn5

Lol you always mention the "lag" and that they wouldn't respond to you. IIRC you were asking them to send out free servo setups to like 3 or 4 different people. I don't know how that conversation went but who else can you get to do that?

Stephan from DMM has always responded to me when I asked about this or anything else. I can't say I was fully satisfied with the answers but nobody has avoided the question. I will say that the "lag issue" was fairly misrepresented going by what people said in the one thread about it vs the test that was posted at the end of this thread: forum.linuxcnc.org/27-driver-boards/4122...alog?start=30#268186

That halscope screenshot (on page 4) shows the analog hal command and the external encoder having about a 1-2ms delay. Nobody knows how much delay there is between the hal command and the actual analog output change because it wasn't checked with a scope. The DYN4 emulated encoder output shows about an extra 3ms delay which is all you can really blame on the drive since all that was posted was the halscope. So there's possibly a 3ms delay in the encoder output probably due to the CPU in the drive that scales the output. That's something that could possibly be fixed with drive firmware, and may even be fixed already.

Like I said, the guy does respond. He responded to me asking about that right away, several times. It took them several months to test it but he sent me a response randomly a couple weeks ago:

Hi Justin, sorry for delay on this, we finally had some time to go through this in detail and here is what we found:  We spent some time looking through the firmware and did some testing.  The encoder output response is ~1ms in the testing we performed as you can see below:
 

 Not quite sure why your controller setup is delayed so much.  Maybe it's the parameters? In our test, the parameters is as below: MAX_ACCELERATION = 127MAX_SPEED = 24OnPosRange = 127Torque Filter = 48GEAR_NUM = 4096LINE_NUM = 2500 The other parameters like the deadzone etc. has no real effect on the encoder output response.  The only parameter that can introduce delay into the encoder output is the MAX_ACCELERATION parameter, so make sure this parameter is set and saved into 127 for fastest response of the encoder output.

Now, I'm personally not quite satisfied with that response as the "controller" lag wasn't the point, the lag difference between the Haimer encoder and the DYN4 encoder output was the point. The scope picture he sent however does show only 1ms lag between the emulated encoder output and the actual analog input rise but for some reason they sent a 1.3v pulse. If that test is good then the encoder output lag itself is actually less than 1ms because the drive had to respond by moving the motor before the encoder had anything to output.

Point being, I have a DYN4 drive/motor that's been sitting on my bench for months. It's literally just a test set for something else I do. I could easily test this but why bother when for 1) people will keep saying the same thing about "lag" no matter what I post and 2) what good is it without some other servo drive setup to compare. Like if this drive has an undeniable 2ms lag on the emulated encoder output, how does that compare to a Yaskawa or anything else? Can anyone else do a good test with a scope on some other servo?.....we can put this to bed....TOmmy?

I'm sure I can get the story on the DYN5 if anyone wants to know about it. The manual has been on he website for like a year but it's not on the storefront. If I had to guess it's probably available but limited.

tommylight wrote:

LOL
This reply should have come from them, so thank you for taking the time to defend them as obviously they are not interested in doing that.
The lack of their response, and their insisting that it is always LinuxCNC's fault, while insisting it works properly in analog mode to potential customers, is what triggered me.
 

Their "lack of response" to who? You're not a DMM customer, and you weren't inquiring about purchasing a drive. I'm pretty sure DMM is a very small operation, I've only ever spoken to 1 person either on the phone or through emails. And I'm not defending them, my experience with them is different then yours, probably because I'm a customer, I've bought things. While everybody in that thread was throwing around "5ms", I saw 3ms. Since it was done with halscope it could be closer to 2ms depending on how the latencies stack up. Notice how nobody ever acknowleged that when I pointed it out?

Literally any company, from Texas Instruments that sent development boards and chips for free for the last 20 or more years to any Chinese company nowadays.

OK, you just referenced TI which is probably still worth at least a few billion US dollars and "any Chinese Company" which is making whatever it makes for peanuts. I get PCBs made in China for less than a pack of gum, China is not the real world it's like an all-you-can-eat buffet of everything you can't eat. You were asking DMM to send you drive/motor/cable sets that they sell for ~$600 to several people and given the size of DMM as a business you're somehow comparing that to dev boards from TI or chips from random Chinese companies? You gotta be a bit more fair on that one.

1.3V is the value scope triggered at, rising edge, right side of screenshot. 

You are correct, I just glanced at that and thought it was P2P voltage...that scope kinda sux.
 

tommylight wrote:

blazini36 wrote:

Their "lack of response" to who? You're not a DMM customer, and you weren't inquiring about purchasing a drive.
I am not, but i did inquire for 12 sets of DYN2 and was pointed to new DYN4, so i waited a bit as they were not available yet, and waited and waited for nearly 4 months, so i moved on.

blazini36 wrote:

You were asking DMM to send you drive/motor/cable sets that they sell for ~$600 to several people and given the size of DMM as a business you're somehow comparing that to dev boards from TI or chips from random Chinese companies? You gotta be a bit more fair on that one.

Yeah, that is also their fault as their website gives the impression they are a big company, with the included "call for a quote" and a price list with old stuff but no mention of new stuff that i was actually after.
$600 is something i can deal with, let alone a company, that is peanuts compared to the damage it did to their reputation.
I asked them to send those to other people, not necessarily me due to where i live and the trouble it would cause. This means i took the liberty to burden other people to waste time and effort to help DMM, i am sure their time spent on testing and analyzing and logging and making sense of what is going on in those drives would have cost them much more than $600.
I do not like burdening other people, but it was to help DMM, so i could justify it to a certain point.
There is no justification for their response to all this.
P.S.
Why are you protecting them so vigorously?
 

Lol for like the 3rd time, I'm not "protecting" or "defending" DMM. I just don't like bullshit perpetuated by bullshit. You just don't seem to realize how unreasonable you're being about the whole thing. Saying things like "Yeah, that is also their fault as their website gives the impression they are a big company, with the included "call for a quote" and a price list with old stuff but no mention of new stuff that i was actually after.".....Just reeks of vitriol. You're usually a pretty reasonable guy on the forum, but man it's like you're drooling to get at DMM for some reason, and you really don't even seem to know why. I'm a 1 man business, I "call for quotes" I send POs, I send quotes....what are you even saying here?

Another user just pointed out  in that other thread some pretty valid statements about Mitsubishi and Yaskawa servos having an even larger "lag" on the encoder output than what the DYN4 does. You can't tell me Mitsubishi/Yaskawa servos are junk and I think if you expect a servo drive on the cheap end of the spectrum to perform better than those, you're not really being reasonable.

So what we're left with here is an inflation of the actual problem and DMM being assholes because they don't send out free equipment. Get Yaskawa to send you a free setup so you can clear their name and we can compare it with what you say here and maybe we'll get a baseline for the issue. Ive got a DYN4 on the test stand right now, I just have to setup something with Windows so I can reconfigure the drive but I'll post scope shots and all that....No LinuxCNC, LinuxCNC is completely irrelevant for trying to figure out this issue.

-Have you tested the Ethercat version of DMM DYN5 drivers?
-Does they are compatible with LinuxCNC?

 

grijalvap wrote:

-Have you tested the Ethercat version of DMM DYN5 drivers?

 

No

-Does they are compatible with LinuxCNC?

Well Everything is compatible with LinuxCNC to some extent.....

I called DMM and asked about DYN5, they're all available buy they won't be sold on the website like the DYN4's and DYN2's. They use the same motors so nothing to talk about there but the drives are quite a bit more expensive. Just for reference I asked about a 750w drive without giving a specific model, it's about $650 (forgot to ask if that was US or CAD). The DYN5 manual only shows the Ethernet/IP version of the drives but the EtherCAT is available, it'll be a separate manual it's just not ready yet.

So basically you have to call and order or inquire by email as these are mostly intended towards industrial customers but it's still sold direct. If you wanted an Allen Bradley or something you'd be dealing with a distributor so it's not quite that. I'm personally not upto snuff with the whole EtherCAT thing but I was thinking about learning it and using it for something else I might cobble together with LinuxCNC. I did mention this to them and they said they are scheduling to test it with LinuxCNC maybe the end of this month.