Also the dv/dt during deadtime don't change as function of switching freq.
Also the dv/dt during deadtime don't change as function of switching freq.
dV/dt and dead-time are two separated concepts, if you design for lower switching losses on the output stage, dv/dt will increase. The only way to control dv/dt at the motor terminals is by using Snubbers (or output filters). Of course, when you make shorter mosfet switching times, dead-time will increase if you don't change anything else to adjust it. The rate of current decay during the dead-time depends on the Power supply voltage and motor inductance (fast decay) in this case (UHU board), due to the PWM signal being applied to the higher mosfet and opposite's H leg lower mosfets at the same time.
Also, low inductance motors will have higher losses, and in some cases will be unstable when employing blanking time on the current sense circuit.
I've just received the Embeddedtronics PC board for the UHU controller and I'm waiting for the processor to arrive. Meanwhile, I'm collecting the bits and pieces needed to build it.
On the motor side I have a question about the encoder for the servo motor. I'm planning on connecting the UHU to a lathe Z axis driven by my Electronic Lead Screw. It's limited to a 20kHz stepping rate.
So here's the issue. With a 200 line encoder, the UHU controller appears to use that in quadrature so that's 800 ppr. The implication if I understand the configuration then is that 20000/800 = 25 rev/sec or only 1500 RPM. That's pretty slow for a servo but more than adequate for a stepper drive.
If the servo has a top speed of say 6000 RPM then I expect I'll want to either have a 50 line encoder or have the motor move 4 steps for every one step input. But I'm having trouble interpreting the parameter Step Multiplier (M) which appears to suggest that each step is really one encoder line or 4 quadrature steps.
A 200 line encoder in quadrature mode I think is the same as the document's 4-fold interpreted encoder line. So does a one step input create a 4 quadrature step (one encoder line) and if M is 2 then one step does 3 encoder lines or 12 quadrature lines?
If that's the case then I can still stay with my max 20Khz stepping rate but have the servo move 4 times as far with M=3 and use a 4:1 pulley to bring the lead screw RPM down to 1500 RPM max.
UHU's decoder is X4, and as you stated, will produce 800 counts per revolution with a 200 CPR encoder.
On the UHU's user's Manual it says:
"Step multiplier (M)
The provided integer value is added supplementary with every step pulse. I.e. if 'M' is set to '2', the
motor moves exactly three 4-fold interpreted encoder lines with every step pulse.
This makes it possible to drive setups with high encoder resolutions to high speeds even if the
commanding PC is limited in Terms of stepping rate. As it also reduces overhead operations, it is
always desirable to increase this parameter if the resolution of the encoders is not needed.
Attention: Setting 'M' to a high value leads to higher tracking errors with every step. The tracking error
shutoff 'E' has to be increased accordingly.
Useful values are 0 to 20."
Yes. and as an added bonus, you will be using all the motor's available torque at the maximum lead speed (1500 rpm).If that's the case then I can still stay with my max 20Khz stepping rate but have the servo move 4 times as far with M=3 and use a 4:1 pulley to bring the lead screw RPM down to 1500 RPM max.
At the moment I'm not considering assembling these boards. They're not really labour friendly in that it takes a while to assemble them.
Not having a Gecko to play with I don't know how much better it would be over a Gecko and they're on sale right now.
I did post a question asking whether tuning via software (UHU) or via trim pots (Gecko) is better. No one has answered that question.
I wouldn't buy another UHU-Servo Embeddedtronics board. Although it's a compact layout, the TO-220 mounting holes are way to small with undersized pads for the kind of current that this system could deliver. Blow a transistor and it's quite likely the board could be damaged replacing it.
I bought the larger IRFP260N devices and they won't even fit in the holes much less have the right spacing for the holes. I had IRF640s in stock so that's what I ended up using. That's fine for this motor and experiment because 18A is max stall current for the motor anyway so I just have to limit it to 15A or so to protect the transistors.
Each time I build up a through hole board I remember how much I dislike through hole boards. Surface mount is so much easier.
I've attached a photo of the completed board and motor. Now I need to build up a good solid 24V power supply and get the proper pulleys so I can hook it to my lathe and try it out with the Electronic Lead Screw.
We only designed the board layout for TO-220 sized FET's. The size of the pads and traces should be more than adequate for 20amps. IRFP260N are much bigger devices.
I put a new disclaimer on our website that reads...
"*UHU board has only been tested with IRF540N FET's at 60volt/20amps. Using any other type of fet driver and/or higher voltage or current is the sole responsibility of the end user. Using high voltage and currents is dangerous. Building and using the UHU servo board is at your own risk. Embeddedtronics cannot be held liable or responsible or will accept any type of liability in any event, in case of injury or even death by building and/or using or misuse of this device or any other high-voltage device posted on this web site. By accessing, reading, and/or printing the articles presented here you agree to be solely responsible as stated in the above disclaimer and exempt Embeddedtronics from any criminal and/or liability suit. Safety is a primary concern when working with high voltage circuits."
We care for your safety so officially we will just limit our version of the UHU board to 60volt 20amps. However we cannot control what type of fet, voltage and current that people use once they buy our board. Please be safe when using high voltages!
I'm pretty sure IRF540N is the same fet that gecko uses. They conservatively rate their drive at 80volts 20amps. I don't have a power supply bigger than 60volts so that is the rating I will use for our version of the UHU board.
If you do a Rev. D layout here are a few of suggestions.
1. Make the board a bit larger so that the mounting holes don't need insulating washers to avoid shorting the traces nearby. Make the holes work for 4-40 screws and nuts with room around the nut. Also allows the board to be held in an assembly frame so the foam panel can hold the parts down while soldering. See point 4.
2. Separate the 12VDC power in connector a bit more from the Data and Encoder input connector. I had to bend the pins on the connector to make it fit.
3. Spread the TO-220 package pins further apart so that wider anular rings can be used with larger holes. Both of those will contribute to serviceability when a FET is toasted and the magic smoke comes out. In fact, design it for the IRF260 foot print. Large holes, large pads.
4. I prefer to lay resistors down rather than the vertical method used here. But I think that's a personal preference and your method certainly makes the board smaller.
5. Set up the board to also have the ability to use an encoder with differential pair drivers for noise imunity.
Other than that. Nicely done!!!! Appears to work quite well.
I'm glad the servo board works for you. Thanks, for the comments. We will look into making the changes in our next rev of the board. We haven't sold that many so it may be awhile before we need to order more boards.
We welcome any comments on our version of the UHU servo board, both negative and positive. It will only help us design a better one the next time. It's hard to get things right the first time or even the 3rd rev. The first 2 had silly errors in the layout that we totally missed.
Been there done that. I'm fixing all the bugs on the second rev of the ELS board this weekend. http://groups.yahoo.com/group/E-LeadScrew/
There's been a few silly mistakes. And now changes to accomodate a larger range of buttons, digital pot for setting stepper motor current. It's really easy to get carried away with 'features'.
BTW, I was very happy to see that the pinout of the 3 pin RS232 connector on your UHU-Servo board matched the RS232 pinout on the pins of my ELS design. Saves keeping yet another DB-9 to 3 pin adaptor.