Pretty sure it is a Chinese made one as the form factor and materials of the case are very very similar to that of the encoders that Mige currently uses… I know MIge right now does NOT offer a SinCOS encoder but that of course doesn’t mean that the company supplying them encoders does not… I have actually been searching AliExpress to see if something pops up and there is one there but they give so little information I a m not sure if it is the same or not… If one can be found that is Chinese based I’m sure it will become the next standard. Yes it might not be as quality as the Heidenhain but it will be significantly less expensive as in reality a NEW 2048 line Heidenhain would cost about $350 - $400.00 so about double what that one costs from them.
Yes I think that might be the case. Anyway it’s nice to see some cheap alternatives as well.
We were happy with our 40K Chineese encoders until now.
Ofcourse I have to say a high quality encoder truly changes that overall quality of the servo.
It feels like a different make now.
I would say this one will be satisfactory as well… and will make it not worth the price of the Super High end encoders if it works well. + no real adaptor kit will be needed if the Encoder shaft mounting is the correct size…
A part of me wants to get my hands on one of these just to find out if there is any real difference in feel between the highend encoders and this…
That would be very good to know.
I would love to test one too.
Hi been told that Mige will be doing the encoders as a special order so you can order your Mige with the 2048 Line SinCos Encoder Upgrade pre fitted.
Think you cam order them from Today 
I believe it’s same Yuheng encoder as before http://en.yu-heng.cn - they do sin/cos
What I don’t understand is that why we don’t have abs encoders support still - Tamagawa T-format specifically?
Absolute encoder completely removes question of center point, doesn’t have noise errors as sin/cos, resolution 8M per turn + 66K turns counter, costs 80USD in China (standard 40K costs 30 USD fyi).
Tamagawa TS5700N8401 https://item.taobao.com/item.htm?id=43620072496. It makes 20K updates/second of digital position with speed of up to 6000rpm.
Or is it a plan that everybody should buy 2500cpr, 5000cpr, 10000cpr, sin/cos and finally absolute?
The price of a motor plus that encoder gets you pretty close to buying a kollmorgen no?
Who would benefit from such a scheme as to sell 2,500, then 5,000, then 10,000, and lastly SinCOS?
It is just tech progression and availability/affordability that drives the market, and thus the suppliers, like MiGe, to make them available on their servos.
Wrt the ABS encoder support, Tero can weigh in if that will be a possibility real soon. Please note that you can only safely use ~2M /21bit encoder so as to ensure we don’t run out of bandwidth on communication link between SimuCUBE and IONI.
You can already use BiSS encoder to get to those levels, or SinCOS, but I am sure you are aware of that.
Cheers,
Beano
I’m certainly joking about anyone intentionally dragging us through the generations of encoders. Ioni pro can work with 15M for, so not sure why it’s 21 bit. This Tamagawa is 23 bit. What I said, though is that cheap high resolution full digital absolute encoders are available already and interface to them is very simple. And from specs they are superior to quads and sin/cos by far. So if anyone from Granite can chime in about potential of such support - we all might benefit skipping one more intermediate encoder, fighting analog signal noise issues and so on. I know some biss are supported, but they are expensive. Realistically sin/cos won’t have stable 2M precision - that’s a hope to get clean repeatable 8 bit ADC conversion which would seriously depend on environment and luck.
Anyway, I wanted to get everyone’s opinion on a matter. I would like to see Simucube getting more and more traction, eventually returning to OSS. And this needs more affordable and reliable high performance support. And encoder from Tamagawa looks like one. Meanwhile I don’t mind buying now sin/cos, but it’s me and I happen to be in China now, so it’s easier.
Btw, I’m sure that abs encoder would not have any sand crosstalk noise with TBW above 680hz on a Mige. Like zero of it. Nil.
Price of encoder 80 USD. Link is on taobao page in CNY…
Hey Mash,
Ioni can support higher resolution, in fact, I am using 22-bit /4M CPR on my setup, but…there is a challenge to stream all the data back/forth between SimuCUBE and IONI and still ensure we do 100% hit-rate with the FFB data…
Once we go past the 21-bit /2M CPR mark, a few guys have experienced ‘hiccups’ in the FFB feel, almost like micro-stutter, and Mika and Tero has clamped the officially supported resolution on simuCUBE side to 2M CPR…
The 2M limit comed from this:
Ioni gets updated at 2500 Hz. At every update, the amount of change of the wheel position compared to the previous update must fit into signed 16bit integer. The angular speed of the wheel required to surpass this can be calculated. The update rate should be enough as that speed is actually quite crazy high, but still people do see some loss-of-center events at >2M CPR when using high forces on bumby track -
DW12 at Sebring, for example.
We think there are some hickups still in the firmware that may cause the 2500 Hz rate to momentarily dip, causing this. But as these crazy high resolution encoders don’t have any effect to the feedback feel, we decided to put a limitation somewhere. This is, in part, to save us from spending development time early on in something that has 0 effect to feel and that would “benefit” only those that bought the most expensive encoder.
We have had exactly 1 interested party contact us to get involved with open source development. That development is hopefully starting soon. What makes you think releasing the code as totally open source would change the number of interested people? What is your view on how we could release the firmware?
Definately interesting.
Ha, thanx for setting me straight, Mika, so it is to do with positional update and fitting new data to integer value.
So perhaps some small timing issue there.
But anyway, no dramas, we know this limitation and can work around it, as you said, no benefit going past 2M CPR anyway.
Cheers,
Beano
Eh, is having 32bit integer going to solve those limits? Is it really necessary to have 16 bit? In terms of cheap good abs encoders - I see 17 bit which is not a lot and 23 bit.
Mika, I believe that everybody afraid that eventually you won’t have enough time to keep pushing it like it was at the beginning of Simucube time. Having it OSS would just allow to a) increase speed of feature development b) calm people that support won’t become stall. You are free to use some GPL license to make sure competition won’t release 30$ Simucube with embedded Ioni, but you understand - those who really want to make a clone they can decompile firmware or just make exact hardware clone. What stops that is a community that is dedicated to the specific brand and OSS certainly helps to that. Easy dev-env, good documentation, set of unit tests and people would be submitting their patches with new features eventually significantly overcoming internal resources.
Btw, I believe that main advantage of abs encoder support here would be no TBW limit and probably smoother feel.
This limit comes from the SimpleMotionV2 API, and changing that just for this SimuCUBE usage case does not bring enough benefits to invest our time to do that, as the feel would not improve. Also, changing the API could have wide-ranging effects for our other customers.
- Having it OSS would just allow to a) increase speed of feature development
If there is someone who wanted to to implement support for device X or feature Y using STM32 HAL API, we are listening. So far, even while advertising this opportunity, we haven’t gotten any contacts for the device side of things.
- b) calm people that support won’t become stall.
DirectInput development will become stalled as that API does not get any changes any more, and hasn’t gotten since many years.
- You are free to use some GPL license to make sure competition won’t release 30$ Simucube with embedded Ioni, but you understand - those who really want to make a clone they can decompile firmware or just make exact hardware clone.
We don’t want to make it easy for the chinese. Their current copy uses different motor control technology, but we don’t want to give the DirectInput device side either.