AASD15 Alternative - Thanos

Hello there reader! Hope you doing well

My brother and I are slowly shopping for our 6DoF rig. One of still missing things is Servo+Driver.

We are using @Thanos board and found interesting drivers, see P100S-40 (by Sinrobot).
Their price is half of AASD, so we dug deeper. It seems that all pins/features required by ThanosBoard are also provided by this particular driver.

Does anyone have experience with other Servo Drivers instead of AASDs?

Thanks!!

I have the same question, seems like that might be a secret. I'm getting ready to start my build and plan on buying 1 servo drive and motor kit to try out and see what happens!!

Are we talking about theese?

I also bought 6 of them recently. Mine are from a diffferent company called "PFDE". The motors are smaller and have a 17 bit magnetic encoder instead instead of the optical incremental encoder of the AASD types.

The disadvantage is that the drivers have a 44pin high density connector instead of the standard SUBD25. So you need to make a special cable to connect them to the Thanos. Shouldn't be too hard though. Cables with the 44pin HD connector on one side and an open end on the other were already included. So all you have to do is to solder a D25 to the other side.

Thank You for the info! Which models do you purchase?

I've bought theese drives. But I don't want to advertise too much before I have tested them thoroughly.

Please let us know if they have that audible screech

The PCBs for the adaptors for the SUBD44HD connectors have just arrived. I hope I can assemble them soon and make the first tests with the P100S servos.

Yes, they make some audible noise in the >10kHz range. I don't care much as I'm old and don't hear much above 10kHz anyway.

But I've tested the drives and they work smooth and well. There are some differences to the AASD drives that need attention. I already mentioned the command signal connector is DB44HD instead of DB25. So anyone who owns a Thanos or M4S controller has to make custom cables. I've made some little adaptor PCBs to fit the RJ45 connectors of my controller.

Some hints for the parameter settings:

• PA53 needs to set to 0, otherwise the Servo-ON input is disabled (forced to ON) which is not mentioned in the manual. (Chinese mentality: a safety risk is better than customers complaining about drives not working)
• PA11 selects the resolution (steps/revolution). The default of 10,000 means you need 500kHz for 3000RPM which is close to the maximum the input can handle. If your controller can output only 100 or 200kHz this has to be lowered to 2000 or 4000.
• I've set P3-20 to 7 (arrival torque) so that the DO1 output can be used as homing signal for a hard stop.
• and P3-21 to 9 (run) so DO2 outputs "ready for command input".
• Set PA89 and PA90 to the torque threshold for the hard stop. Something around 20 to 50% should be adequate.

A funny thing is that the encoders are of magnetic type similar to what we use as "hall effect potentiometers" for joysticks or as feedback for rotary actuators built with geared DC motors. If you put a strong magnet near the rear end of the servo motor it moves by several degrees because of the interfering magnetic field. There is also a little bit of noise you can feel when applying torque to the motor. Higher quality servos with 17bit optical encoders (from Sanyo Denki for example) don't have that noise but run totally silent at low speed. But for the amazingly low price we can't complain.

If anybody needs help to connect the P100S servo drives to a Thanos controller please contact me. The step+dir signals should be no problem. But I'm not sure about the digital inputs. I think the Thanos and M4S controller use 5V to drive theese. I've used an optocoupler and 24V.

I've just finished the control cabinet.

Test run look here

Another thing we found out: Connecting all 6 drives in parallel to a single 220V 16A socket immediately pops the circuit breaker when switched on. One to three drives max. are OK. There already is a soft start circuit in the drive, at least I can hear a relay clicking 1-2s after applying power. But the inrush current is still too high.

We solved that by adding an external 100 ohms resistor in series to all L1 inputs and a timing relay with a delay of 5s which shorts the resistor. And as already mentioned, it's important to set PA-53 to 0 so that the motors are not instantly enabled.

Or you could use a industrial current limiter...

Thanks, so "Meanwell ICL-16R" is the keyword. I think that's also a timing relay with a resistor. But it has thermal protection and costs less than the Finder timing relay.

Oh man this is gold!! Thank you for the info

So that link takes me to a 750W no brakes motor/driver set for ~ $150 US with shipping. I had always resisted AC servos because they were so expensive but I guess not any more.

So did you buy six more of these for Sven’s rig? Does that mean you are ready to recommend?

I checked Amazon and this appears to be the same item for a few bucks less for Prime members:

https://a.co/d/9cNvlbn

They are even less expensive here in Europe (~120€ with shipping).

Theese are different from mine. But they seem to have a D25 connector with pinout compatible to Thanos. That's even better and you don't have to make a special cable for the D44HD connector of the P100S drives.

I hate making cables. I have been working on a DC ball screw for a yaw axes and have not produced an electrically reliable result as yet.

Working with 220V would be different for me. I am used to 24VDC and that’s not an issue. 120V here in the US will bite you but generally not hurt you. This is from a person who has gotten bit many times.

Years ago we were working on a project for Europe which was 220V. One of the techs got bit and they hauled him away in an ambulance. I’m just saying to anyone thinking about this needs to respect 220V.

Sabertooth is my plan B I prefer a custom solution.

I’m about ready to order one of these for some single axis bench testing. I’d like to validate my assumptions first:

1) You set the desired velocity via -10 +10v analog input.

2) The AC servo will provide whatever power profile is required to run at the set speed.

3) You monitor the AB quadrature output (the 200khz signal). You don’t actually pay any attention to Z.

4) You use 2 or 3 processors because the AB interrupt processing is heavy. What processor/os do you use?

Most of us use those servo drives together with a motion controller that outputs step and direction signals like the Thanos or M4S controller. But it's also possible to run it in analogue speed command mode (+/-10V). However, for the P100S driver I haven't found any way to get emulated encoder signals (AB quadrature) out of it. So I fear you have to use some other feedback like hall sensor potis or an external incremental encoder.

Ok that’s a revelation. I was looking at manuals from more expensive servos since the knock offs have none. The one I was looking at had ABZ encoder outputs. I have been using DC servos since 1982. Yes sim old. They say to learn new things as you get older is good for your brain. Thank you for the revelation.

New assumptions:
1) Step direction is confusing (stepper motors). Pulse direction yields useful search results.

2) The Thanos is controlling accel/decel by controlling the frequency.

3) Thanos assumes that the position X steps is Y degrees.

I will consult my favorite AI and start over. Grok is down. Sucks a bit for me but people are starting to count on it to be up.