Most instructions for measuring RC is in Ohms with no power/voltages added. That is with a dead circuit. Why/How can the measurements of RC here be with a powered circuit and in Volts?
RC in this context refers to time constant. Think Resistor and Capacitor connected in series, where the capacitor is charged and discharged thru the resistor. The actual on-chip implementation and usage is a wee bit more complicated, but the idea is the same.
See also Ed's Stepper Motor Analysis write-up. It helps shed some light the "why" behind different stepper driver REF settings for different motors.
I kinda have a cloudy on the why for REF and PFD. The RC is for reference and assist the other two thus I kinda get it as Resistor and Capacitor. I've looked at the schematics and see where they fit and what values they carry. This is why I've seen that the RC should be measured in Ohms w/o power. Anyway that is what the manuals and lessons instruct.
Oh - thanks for the link to Ed's blog article. I had not read this recent posting. I was bugging him in Mar/Apr about this when I was having issues with my steppers ringing, grinding and misbehaving. I was working through for a correct set of PFD, REF, RC1, & RC2 values but, did not understand enough nor did I have the test equipment to run/record events and conditions. I'd also read: Nophead's <a href="http://hydraraptor.blogspot.com/2009/12/motoring-on-with-a3977.html">"Motoring ON…"</a> article.
Bare in mind that you're setting a variable resistor that's in-circuit, not picking a resistor for insertion into the circuit. To read the resistance, a multimeter has to apply a voltage to measure current or apply a current to measure voltage, either approach of which has polarity. When it's in this particular circuit you can't directly measure the resistance because of all the other linear and non-linear element connections.
For folks in the field (aka you and I and the commercial PCB assembler) it's faster and more reproducible to provide tables of measured test-point voltages than a complex set of instructions for desoldering/measuring/replacing individual components.
We are not measuring individual components but 'in situ' circuit pairs. There is no need to dismantle circuits to test the RC(x) pair. The designers (Zach, Nophead or ???) placed pads onto the circuit board per the chip manufacturer's pre-design chip manual suggestions. (I think.) It sort of reads/looks that way. But, I am really not in my element being an ME instead of an EE. I do understand that a circuit is not always tested under power and in the case of checking for resistance between circuit sections and 'in situ' RC pairs this is such a situation.
From my reading the A3977 chip manual, the MBI Controller v3.4 circuit design, the design of such common stepper control circuits, and stepper motor settings plus how this particular controller's test points are in place for the R & C combination it should be tested as an RC component pair in Ohms without power. With this case - again - the test is for an RC pair measured in Ohm's, not Volts or Amps, with a multimeter on an un-powered circuit.
On reviewing the MBI circuit schematic - The ground of the board is on a bus, rail or common. So, using the circuit board ground or ground wire and the test pad at the RC(x) you would test Ohms.
To test: Unplug the PS and USB. With the multimeter set to read Ohms touch the ground (-) probe from the meter to the ground spot on the circuit board or to the ground (black) wire in the power connector of the board. The positive (+) probe is next touched to the test pad for the RC(x) combination on the circuit board. Read the Ohms for the RC(x) combination. The circuit pair are isolated by the test across the pair. Next, a math calculation is worked to determine values for other settings or as I've been told you need to individually tune the circuit via oscilloscope. But, I do not now believe the later requirement for this to be true. The circuit is standardized, the chips are known with spec. sheets, the engineering tolerances are consistent and the systems relatively similar. This should be a test once and set all or a calculate once and set all with maybe some tweaking.
That is how the Manual for the controller chip and the designers of other controller's describe the process. It is odd that the MB community does not follow common electric circuit practices nor proscribe to them. Maybe this odd view is why when behind the power curve things change long after the urgent discussions of the REF, PFD & RC(x) settings. [I'd had struggles for weeks in late March - early April about this with MBI and others.] This counter view of how things work is very detrimental to learning what is actual when there is not the attention to the process and science which directs the workings of components and process. ——
The circuit pair are isolated by the test across the pair.
Except, unfortunately, that's not the case. There's no isolation whatsoever: everything remains connected and you must analyze the "circuit" that way.
The meter applies a (relatively) low voltage to drive a current through the device-under-test, which is the parallel combination of the resistor and the A3977 input. It has no way of knowing you're interested only in the trimpot.
The trimpots are about 47 k and the series resistors are 12 k, so if the chip presents a relatively low resistance, you'll get a bogus result: the 47 k trimpot shunted by 12 k. If it's a relatively high resistance, you'll get something closer to the actual resistance. The point is, you do not know what's going on inside the (unpowered) chip.
Better to simply multiply the trimpot's resistance by the fraction you've turned it. That'll get you close enough: CCW = 47 k, CW = 0, midway = 23.5 k.
Normally, you figure the resistance value from the requirements of the motor & driver, then select a suitable resistor. Adjusting it after the fact works the same way: figure out what resistance you need, subtract the 12 k series resistor from the calculated value, then set the trimpot accordingly.
Or, watching the motor current (not the voltage) on a 'scope, you tune for best picture…
What Ed said. :)
Since we're talking cross-disciplines (I'm an analog IC design EE, not an ME) trying to measure and set the trim-point resistance directly in-circuit when it's off is a bit like… trying to tune a car's idling speed with the engine off by measuring how much of the idle screw is poking out of its mount point. I don't (given how little I know about car engines); I watch the tachometer RPM and adjust the idle screw until it reaches the target.
Measuring the voltages at those test nodes (==tachometer RPM) when the stepper is powered (==engine is running) but not doing anything (==in park) is how you & I make adjustments to these stepper boards in the field.
Or, watching the motor current (not the voltage) on a 'scope, you tune for best picture…
This is in fact how we came up with these potentiometer values. Hooking up a current meter, looking at the waveform in an oscilloscope, and tuning for best picture. Here are some of the screen captures if anyone is interested: http://www.flickr.com/photos/nickstarno/sets/72157626766309234/
looking at the waveform in an oscilloscope, and tuning for best picture
If you're running at 1/8 microstepping, then one of those 30 ms cycles represents four full motor steps. You can certainly set the overall current that way, although it's really not necessary: simply use the desired motor current to set the REF voltage level and you're done.
However, the RCx and PFD adjustments adjust the transitions between successive microsteps and there's simply not enough resolution in those pix for a meaningful measurement of what's happening there. In order to see those transitions, you'd trigger on the chip's Home output, sweep-delay to the step in question, blow up one transition to full screen, and then tune for best picture. Iterate over several different steps to find the optimax settings.
I suspect you haven't done that. [grin]
FWIW, old Tek 7400 mainframe scopes had dual triggering plugins that made short work of such setups. Back then, I could use those scopes in the lab, but they cost more than my house…
In an actual design, you would choose the external capacitor to set the comparator blanking time, decide the proper fixed off time based on the actual motor L/R, then choose the resistor to suit. You'd make some measurements to verify that the current is behaving as expected, then you'd be done.
Which is why I suspect those trimpot settings, while well-intentioned, are largely irrelevant: nobody has done the engineering behind the motor-driver system to justify any particular setting.
The RC(x) and PFD values that were initially published had been determined by looking at the voltage waveform across the sense resistors on the stepper driver while attempting to eliminate noise and keep the wave clean. I'm not going to pretend to be an EE, and at the time this post by Nophead had been a huge influence: http://hydraraptor.blogspot.com/2009/07/lessons-from-a3977.html
After purchasing a current meter and adjusting the REF settings to what they are currently, I noticed that the RC(x) values did not seem to have a significant effect, so they have not changed from what was previously working.
If you think any of the RC(x) values are problematic and have a scientific method of adjusting them for better results, I would be happy to work with you in doing so. Feel free to email me moc.tobrekam|kcin#moc.tobrekam|kcin, and we can go from there.
I noticed that the RC(x) values did not seem to have a significant effect,
We're in violent agreement!
Let 'em stand until someone grinds through the motor L/R, the driver board R & C, and the datasheet equations…
The RC value sets the off time.
If it is too high the switching frequency drops into the audio range and becomes annoying. Lower still it beats with the step frequency and starts to perturb the stepping.
If too low it can prevent the lowest microstep target current being achieved, so distorts the sine wave drive and produces positional inaccuracies.
When the board is running the test point has a complex voltage waveform on it, so I can't see how you get a meaningful reading with a voltmeter.
You need to set the resistance with the power off. Best option is to measure the resistance with the meter both ways round and take the higher reading.
Wall - O - Words….
Finally —Thank You Chris — My questions were leading to what nophead is beginning to point to regarding settings which DO have purpose and ARE important. This is what I've been learning and asking for someone to put forth the values to get the relationships of the PFD, REF, RC1 & RC2 to have the bot hum - not groan.
The resistance (measured Ohms) is for the RC(x) pair configured in parallel and isolated via the two (2) test points on the circuit board. Refer to the schematic for clarity. The A3977 chip is not being tested or within the test points. We are trying to set the circuit/chip to be able to effectively use Slow, Fast or Mixed decay to best benefit our specific applications. REF sets the bar - PFD compares REF for cut-off values to assign the type of decay and RC(x) adjusts (?) to compare to the sense level? I've not understood that far yet. But, The values are all a working collective interrelated and to be as a song and dance. They are and can be calculated. They can all be chosen by desired use and purpose. They all can be adjusted and tuned for various motors and applications. It is a good board.
As Ed stated, "Let 'em stand until someone grinds through the motor L/R, the driver board R & C, and the datasheet equations…" I disagree with the, "Let them stand…" But, then Ed did not say just to leave them alone but, that someone (spelled MBI) has work to be done - Note again to MBI - whoever you are? Collective or now real company with assigned responsibilities and accountabilities?
In Mar & Apr '11 - I'd been asking then and again now to find who, how and why, within MBI, regarding the correct controller board values. I'd brought up the topic within the google group and to individuals. But, much information received was not accurate, a lot was loosely related, some smoke and mirrors, other irrelevant regarding electronics and some was related. Sadly to a newby's understanding that equates to - noise.
The trim pot is part of the R in RC and is used to adjust the pair in parallel. The A3977 is not in the tests but, is the recipient of the values from the settings made to the exterior feed circuits (PFD, VREF, RC1 & RC2). The RC(x) pair can be tested on an unpowered circuit with a multimeter set to Ohms for resistance. As Chris pointed out the RC(x) circuit under power is in flux and a meter attached will read irrelevant values. The circuit must be measured - off - and in - OHMS. —— I think then Ohm's law V=IR will work for relating to other portions ???
The ultimate responsibility was and is on the manufacturer (MBI) of the application and circuit for the closer/correct values and they should have or should be cranking the numbers and giving us better values for the components supplied. nicksterno does state that they started with values and those have been shown to be wrong. He has shown some tests but, the values were not correct and there should be more diligent work to give better values. I thank him for the information because it is enlightening to the evolution and to some testing of our machines. It is very evident that no one clearly understands the relationship of - REF, PFD, RC1 & RC2 for the operation of the A3977 chip. But - we are getting closer. When the relationships of the values are understood and expertly used this violin will not screech but sing. Maybe even beat a Dimension?
It is known that we all have different environments (Hot, Humid, Dry, Cold, always printing, seldom printing, printing complicated configurations, printing simple figures, etc…) and the settings may not be exact - but - the settings currently published are not (I don't believe from the lack of understanding of what was being asked in Mar and now.) a proper relation for the circuit, motors or system. I am glad that the answers to my questions are getting much closer to an understanding that the PFD, REF, RC1 & RC2 are important to the A3977, are interrelated and have a teamwork effort for a finely tuned bot. The motors are not to always whine, grind, groan or be loud. Just as an F1 race car does not chug, clunk and bang.
I also understand this is a hacking club with the dues to pay both in money, time, pain and work to put forth but, some of us do not have all the skills yet and the time to make the bot work right to print appreciably well. Yes, The T-O-M is very good. I really like my machine and what the MBI crew have done and are doing. Since we are paying ($$$) then I'd expect a little more engineering.
Hey ALL —- Thank you. Good discussion. I've been aggravated with the insistence of not being able to get the correct dynamics but - we all are trying. I have enjoyed the exercise and am happy that we all are getting along, can discuss a real topic and learn about fun things. …. :-)