[evo4wrx]

I purchased 4 x VSDE drivers a few months ago.

I Have just recently finished wiring up a custom built 4 axis lathe and have begun to iron out the issues.

I have 3 x 180 Volt PMDC motors they are 4 pole and are working perfect.
I have 1 x 200 Volt (7amp peak) BLDC control techniques motor with 24 VDC brake and its having a few issues.

Ive gone through and set it all up......
It begins to turn and makes about 10 full rotations then it just feezes and doesnt go any further. Ive tried a few other VSDE in its place to confirm it wasnt the driver and they all do the same thing.

I was wondering how far down the road your firmware is for the BLDC motor or does it have some bugs.

If I was to take a guess from limited knowledge it appears as though the phases are swapping to begin the rotation but the timing goes out and causes the motor to stall. I have a 1000 CPR encoder on the top which has been tested and works fine. I can see the counts going up and down correctly in yoru diagnostics section. of GDTOOL.

[Xerxes]

Hi,

That kind of "freezing" may happen with synchronous motors (BLDC or AC) by two reasons:
1. Encoder count is not set exactly correctly in drive parameters
2. There is some noise in encoder signals so drive loses some encoder counts and eventually motor goes out of sync

The most probable case is #1. Are you sure your encoder is not for example 1024 CPR instead of 1000?

[evo4wrx]

According to the shipping sheet from US digital it is a

E5-1000-250-I-S-D-D-B

Ive checked the website to make 100% sure....

1000CPR with 1/4 inch bore. Single ended output.

I suppose noise is an option being a single ended output.

[Xerxes]

Ok,

Please test if motor can you rotate motor in opposite direction when it has frozen. You may need to help it by hand to make first revolution (this may be needed because torque can go to zero if motor is out of sync). Or can you rotate motor within +/- 5 revolutions infinitely?

If it's very repeatable, then it's bit strange to be noise. You may also check encoder line count from encoder disc directly (in case of usdigital error).

[evo4wrx]

From mach3 i can turn it back and forth without the use of my hand. I can only do this for a few spins in each direction until I get a complete lock up.

Sometimes the card goes into a blue light with red green flash sometimes it just stays blue but I lose control over it.

I had a filter put onto the encoder side and scoped it again the voltages seem to be very clean.

Upon checking out the encoder settings and running the motor drive test the faults coming up are;

2615 ish encoder counts detected and 4 poles.
4 poles is 2/3rds of the poles being a 6 pole motor and 2615 is about 2/3rds of the PPR count.


I placed this motor back on my large mill shortly after and it cranked first time again without issue.

I did take a backup of the config file using gdtool which I can send if you could be of any assistance. I also have remote access into my machine and camera setup on it.

[Xerxes]

Here is what you can do to verify number of encoder counts:

Turn down HV supply and turn motor exactly 5 turns by hand (place/draw some mark on shaft to get precise revolutions) and see how many counts the GDtool counter changes. With 1000 PPR encoder value should change almost exactly 20000 counts. If it's closer to 20480 then you have 1024 PPR encoder disk.

[evo4wrx]

19975 I take it as a 1000PPR.

[evo4wrx]

When I issue rotation commands from mach3 I watch the encoder counts in GDTOOL it manages to make it to around 8500 counts in either direction before it starts to stutter.
As long as i stay between -8200 to +8200 I can spin the motor back and forth all day long.

Incedently the encoder counts before error is set to its max of 16344.

[Xerxes]

The GDtool is not at the moment designed to be used simultaneously with other control sources (it may work, though).

Try unplugging CMD cable from drive and using USB adapter thru GDtool only. In your case can use moveinc 4000 commands in event log to rotate motor one revolution at a time. That way you can see if it turns full revolutions. If not, then the encoder signal quality (maybe shielding) is the problem.

[evo4wrx]

I played about with that command.

Sometimes I get what looks to b3 a fw rotations (quite quick)
Other times I get a 1/4 rotation in any direction, other times I get a half rotation in any direction.

[Xerxes]

So it is encoder signal problem. Have you used shielded cables the way VSD-E manual figure 12 suggests? Have you also connected also motor frame to drive FG?

There are rumors that some USdigital encoders are extremely sensitive to noise (possibly by defects). Several CNCzoners have had troubles with them (also with Geckodrives, UHUs etc). You may search about it. Some of them had to purchase another kind of encoder to solve the problem.

[evo4wrx]

Have you got an encoder that you recommend.

[Xerxes]

I think Renco makes quite nice encoders and haven't yet heard anybody having trouble with them. That is just one example, there are really lots of encoder manufacturers.

Check out also CUI AMT102 from digikey. It's robust and low cost, but it's not optical and may have lower dynamic performance (longer response time to sudden rotation). For most users it's still perfectly Ok.

[H.O]

Hi,
I had severe problems with the USDigital E7P encoders. I tested three brands/models (AMT102, Hengstler RI36 and Renco 35i) which all worked fine while the E7P's didn't. Same wiring, same power supplies same everything. I ended up using the R35i and haven't had a problem since.

/Henrik.

[evo4wrx]

Thanks for the heads up.
What configuration options did you use on the right hand side of the website?

Ive basically wasted a week with scopes tearing my hear out and a lot of wires...I cant believe they are putting this rubbish out if thats the case with everyone else too.

My motors run 130 mm bases or there abouts...so I was going to run with the 50 mm option and 1/4 inch through hole as thats what I have right now...m3 mounts.

Differential with index for home 5vdc.

I was thinking the RCH50i.

Im not sure about commutation, Output interface or Gating option.

[H.O]

Hi,
The gating option is just how the index pulse relates to the A and B channels of the encoder - it doesn't matter in this case.

Commutation channels is not needed with the VSD-E but I belive it does support it, it provides the signals normaly given by the hall-sensors on the motor. If you opt to use it I think it should match the number of poles you have on your motors. Tero will have to chime in on what the benefits of having commutation channels are as opposed to just using the A and B channels.

As for the output interface you definetely want the LD option (LineDriver).

Be prepared to pay quite a bit more for these though. I paid $73/pcs for the R35i at qty 5 anf the RCH50 is probably more judging by its "looks", that could of course be wrong.

/Henrik.


Sidenote regarding the USDigital crap: Fitting a single bypass capacitor across the powersupply pins right on the encoder seems to cure the problem for most people but in my case it didn't make any difference what so ever. I tried every trick in the book when it comes to shielding, grounding, filtering, linedrivers, twisted pairs, powersupplies, isolators, different motor voltages - you name it. It simply wouldn't work with the E7P. Yet, whenever I tried ANY of the other three encoders it ALWAYS worked.

All signals on the E7P was verified with the scope, nice crisp pulses, phase shift, rise time etc - nothing specific to complain about but it simply refused to work.

[Xerxes]

Your contribution is greatly appreciated, Henrik!

Hall signals become useful if axis is vertical (Z axis) with heavy enough load that attempts rotate motor by gravity. With Halls motor energizes instantly after power-up and without it takes few seconds. Sensorless mode also requires that motor can turn freely at least 1/2 turns in both directions from initial position.

However, retrofitting hall sensor to motor may be time consuming for first timer because it must be aligned correctly to rotor magnets. But anyway, motor polepair count should match sensor polepair count.

[evo4wrx]

The axis that im having the main issue with is the Z axis on a lathe I have built. it is a large lathe and the Z axis is weight offset by a pnematic ram. It also a 24vdc brake.

Thanks again for both of your help.

[evo4wrx]

On the subject of hall sensors it would seem some of those renco encoders provide that. Would it be worth getting one of those?

If not is ther eanother one that makes a decent one that would be worth trying.

[Xerxes]

BTW, have you tried your current motor & encoder when motor shaft is free from all load? There is couple more things to try before dumping the encoder: try testing motor as free running and try also increasing "phasing current" from GDtool (also make sure you selected AC motor, not BLDC).

Halls are nice to have as backup if sensorless operation has troubles. However, as said aligning halls will be probably tricky and may need lots of trial and error.