DRV8251: Minimum Coil Inductance for DRV8251 Current Regulation, and Minimum VREF Current Regulation Level?

Hello Nathan,

I will investigate the VREF current regulation issue and get back to you soon. I do have a few question to help my investigation:

1. Is the abnormal VREF behavior at various inductances repeatable in multiple DRV ICs? I want to understand if the issue is repeatable and observed in multiple devices and boards.
2. Are you using your own board design or an evaluation board from TI?

Nathan Cook said:

3) Why is there a minimum current of about 200mA when using the VREF current regulation?  Is this normal?  Again, I don't think the data sheet mentions this. Section 7.3 "Recommended Operation Conditions" lists the VREF voltage range as 0-5V.  It makes no mention that the device just ignores all voltages under 200mA.

Main reason is the lower the VREF, the more that offset voltage and other noise at the comparator affect current regulation. In the snippet below from the datasheet provides more information. Even when ISEN is being used, VREF should be 0.3V-5V. You can always resize the SENSE resistor to set the desired current limit.

Nathan Cook said:

2) What is the minimum coil inductance for VREF current regulation?  The data sheet does not mention minimum inductance (unless I missed it).

There is no minimum coil inductance. However, there can be abnormal behavior when using very large or small inductances. For example: current in larger inductance motors will change much slower so the PWM frequency should be low in order to allow the current to change before switching H-bridge state.

Regards,

Pablo Armet

Thank you for the quick reply.

I am using a custom circuit board for this testing.  I only have one built up now, so I can't check another chip for the inductance range.  

Thank you for pointing out the datasheet quote about VREF minimum voltage.  

I need to make a note here, that the highlighted datasheet text you quoted reads "The VREF voltage MUST still be 0.3V to 5V".  That directly contradicts the "Recommended Operating Conditions", which says the VREF Minimum = 0V.  You guys should really update the recommended operating conditions to list 0.3V as the minimum to more accurately represent how this chip must be used and how it will act below 0.3V.  It's inaccurate to recommend 0V as the minimum, then later have a note in a block of text that says the minimum is actually 0.3V (because the chip just doesn't respond to anything under 0.3V and instead outputs a constant minimum current).

Looking at the DRV8251A data sheet, I'm not seeing that note about VREF minimum of 0.3V.

https://www.ti.com/lit/ds/symlink/drv8251a.pdf

New Question:

Does the DRV8251A have that same restriction on minimum VREF voltage for current regulation?  Can I use the DRV8251A with VREF current regulation to control current all the way down to 0V VREF?

I ran some other tests.  I should have noted above, I am running VM at 12V.  That is the same for all testing here.

Test1: Increase the sense resistor to 2 Ohms.  

This caused the 3.6mH coil to also not regulate current with the VREF signal.  It acted like the 215uH coil (which is to say it didn't work and had a constant current no matter what voltage was on VREF).  

When I put the resistor back to 1Ohm, the 3.6mH coil again behaved as noted in the last post.

Test2: I changed the sense resistor to 0.5 Ohm.

In this configuration, with the 3.6mH coil, the DRV8251 device has the following behavior:

1) The current output does not respond to a VREF under 1V.  1V (not 0.3V) appears to be the minimum VREF voltage in this configuration.

2) While the VREF is under 1V, the current is constant at about 200mA.  Once the VREF exceeds 1V, the current starts to track VREF.

Here is a plot.  scales are correct. Note 200mA minimum current, and current starts tracking after VREF > 1V.

Yellow is VREF voltage (measured at the VREF pin). 

Blue is coil current (current probe).

If I'm doing something wrong here, I would love to have you tell me what it is.  (Too low inductance?  Wacky sense resistor value?  Something else?)

Thanks.

More tests:

Test3: Sense resistor back to 1 Ohm (This is the same condition as the initial post)

The current only starts to track the VREF after the VREF voltage exceeds 2V.  Before VREF=2V, the current is constant at that same 200mA level.

Hi Nathan,

Thank you for the information. 

Pablo Armet said:

I will investigate the VREF current regulation issue and get back to you soon. I do have a few question to help my investigation:

The key limiting factors will be the driver's over current protection (OCP) and thermal protection mechanisms.  OCP will establish the minimum series resistance of the load, and obviously thermal protection will come into play over time.  High inductance will smooth out current peaks, and actually help with OCP.

These app notes might be helpful:

https://www.ti.com/lit/an/slva505/slva505.pdf

https://www.ti.com/lit/an/slva504/slva504.pdf

Nathan Cook said:

I need to make a note here, that the highlighted datasheet text you quoted reads "The VREF voltage MUST still be 0.3V to 5V".  That directly contradicts the "Recommended Operating Conditions", which says the VREF Minimum = 0V

Apologies for the confusion here. The abs max and operating max specifications provide the safe values that the device can operate to before the device becomes damaged or operates abnormally.

Nathan Cook said:

1) The current output does not respond to a VREF under 1V.  1V (not 0.3V) appears to be the minimum VREF voltage in this configuration.

To reiterate, the 0.3V minimum is not a strict minimum limit. We recommend minimum of 0.3V WHEN current regulation is not being used. As shown in the diagram below, the offset voltage of the comparator can cause unwanted regulation when ISEN is tied to GND (disabled current regulation). However, VREF can be < 0.3V when RSENSE > 0Ω but it will depend on various motor parameters such as inductance. 

Nathan Cook said:

If I'm doing something wrong here, I would love to have you tell me what it is.  (Too low inductance?  Wacky sense resistor value?  Something else?)

Is it possible to share the driver schematic? I can check if schematic design is not the root cause. I am going to perform a few experiments on my end to confirm.

Regards,

Pablo Armet

One more question: What is the resistance of your motor?

Interesting.  

I am most certainly using current regulation, originally with the 1 Ohm sense resistor.  In this application, I need a very low peak current (under 1A) so I should not be running into any thermal issues.  I can confirm that from the tests above with peak currents in the 0.65A range, nothing is getting hot to the touch.

Apart from some signal conditioning on the incoming VREF signal (as you can see from the plots I'm doing an absolute value and scaling with some opamps), the DRV8251 circuit is right out of the data sheet.  Note: the scope plots are measuring the VREF signal after the signal conditioning, right at the VREF pin on the chip.  I have the resistor to ground on pin1 (as noted in the datasheet figure 9-17) in case I wanted to try the internal sensing parts at some point, but it's presently a 0 Ohm resistor for the DRV8251.  The IN1, IN2, OUT1, OUT2 lines go straight to a terminal block right next to the chip.  As I noted above, IN1 is tied to 3.3V and IN2 is grounded.  The OUT1 and OUT2 lines are connected to my load coil.  

Load parameters (measured):

Coil Inductance = 3.6 mH

Coil Resistance (dc) = 2.3 Ohm

Questions:

1) So you are saying in this configuration (current regulation with a sense resistor) that there should be no VREF minimum voltage?  So I should, in theory, be able to regulate close to 0A using the VREF input?  

2) You mention comparator offsets and such.  Can you quantify this or is is random and unpredictable part to part?  It appears from my last tests that there is a linear relationship between the sense resistor value and the lowest VREF voltage that the chip will respond to.  

3) In both of my test cases, below that voltage level that VREF starts responding and actively regulating the current, the constant current level is about 200mA.  Is this in fact the lowest that the VREF current regulation can go?  Should I ever expect to get a current lower than this when using VREF current regulation with a sense resistor with the DRV8251?

Hello Nathan,

Nathan Cook said:

I have the resistor to ground on pin1 (as noted in the datasheet figure 9-17) in case I wanted to try the internal sensing parts at some point, but it's presently a 0 Ohm resistor for the DRV8251

Current sensing is done using the ISEN resistor. The voltage at ISEN pin is a voltage proportional to the motor current since it is connected internally to the source of the LOW side FETs. The 0ohm between pin1 and common GND is not required for current sensing. A CSA is needed to amplify the ISEN voltage so it is within the dynamic range of your ADC. Please refer to this FAQ for more information on configuring current regulation and sensing. Note that the FAQ focus on integrated current sensing, which we call IPROPI, that allows one to sense the current without the need of a big sense power resistor or external amplifier.

Hey Pablo,

I know you must be very busy responding to a lot of e2e messages so it's understandable that you didn't fully read my messages or look at my schematic snippit before responding.  Let me be more clear so there is no further confusion.  

If you look at the DRV8251 data sheet figures 9-16 and 9-17 as I noted, you would see this graphic:

What you can see from this graphic (from your data sheet) is a dual schematic configuration, which given different component loading options, can be used with either the DRV8870, DRV8231, or DRV8251 with a sense resistor...... OR.... The DRV8231A or DRV8251A with the IPROPI output.  

So what this means is if you use the schematic with a sense resistor in Rsense, and a 0 Ohm resistor on pin1 then you can use the DRV8251 in that configuration with an external sense resistor.  If you instead use the schematic as shown with a 0 ohm in place of the Rsense on pin7, and a 1.5k in the Ripropi resistor then you can use the DRV8251A which has the internal sense output.  NOTE: these two component configurations are for two different chips.  

In my case, I am using the DRV8251.  NOT the DRV8251A.  This means I need to include the sense resistor and tie the Pin1 to GND with the 0ohm.  My last post was very clear about the fact that I have the circuit configured this way because at some point I wanted to experiment with the other parts that use the internal sense and IPROPI output (in this case the DRV8251A, which I understand is a totally different part).

Ok.  Great.  Now that we got that out of the way.... 

As I noted above, and in my schematic configuration, I am using the configuration for the DRV8251 (which has a sense resistor in Rsense, and a 0 Ohm in what would be the Ripropi resistor).  Because I am not actually monitoring the current in my application, I have omitted the sense amp part of that circuit.  That is why I don't have the sense amp part of the circuit in my schematic.  I am not monitoring the current from the DRV8251. 

As I mentioned in multiple messages above, I am holding IN1 high at 3.3V and IN2 low at GND.  I am then using an analog voltage on the VREF pin to regulate the current output from the DRV8251.  As you can see from this configuration, you don't need to amplify the sense resistor voltage and feed it into any microcontroller ADC for the DRV8251 to work.  This configuration as described and who's output is shown in the attached plots (IN1=high, IN2=low, VREF=current command) regulates current, but ONLY above 200mA and when VREF exceeds some minimum voltage which appears to be related to what value you use for the Rsense resistor.  All of my questions are related to this configuration, which I have consistently described in all of my messages.

I hope that better clarifies the situation.

So given all the above information, my previous questions still stand. I will repost them here along with my load info you requested:

Load parameters (measured):

Coil Inductance = 3.6 mH

Coil Resistance (dc) = 2.3 Ohm

Questions:

1) So you are saying in this configuration (current regulation with a sense resistor) that there should be no VREF minimum voltage?  So I should, in theory, be able to regulate close to 0A using the VREF input?  

2) You mention comparator offsets and such.  Can you quantify this or is it random and unpredictable part to part?  It appears from my last tests that there is a linear relationship between the sense resistor value and the lowest VREF voltage that the chip will respond to.  

3) In both of my test cases, below that voltage level that VREF starts responding and actively regulating the current, the constant current level is about 200mA.  Is this in fact the lowest that the VREF current regulation can go?  Should I ever expect to get a current lower than this when using VREF current regulation with a sense resistor with the DRV8251?

One more point of clarification:

Above you mentioned the difference between the Absolute Maximum and Recommended operating conditions.  I am very aware of the differences between those two.  I specifically asked above about the recommended operating conditions related to VREF, and made a point that I knew they were the recommended operating conditions by calling them the recommended operating conditions. 

Let me in fact post them both here as screen shots from the data sheet:

Absolute Maximum Ratings:

Recommended Operating Conditions:

Again, note that the recommended operating conditions list the Vref voltage as 0-5V.  

The recommended operating conditions does give a spec for the Iout, which is at best 0-4.1A.

It looks like the recommended operating conditions section is missing another important spec, which is the Iout when using the VREF input to regulate current.  It looks like that spec should be Min=200mA.  If this isn't the case, it would be great if you could answer my questions from the last message and let me know how to get less than 200mA when using VREF in "current regulation mode" to regulate current. 

Thanks!

Hello Nathan,

Apologies for the misunderstanding.

Nathan Cook said:

It looks like the recommended operating conditions section is missing another important spec, which is the Iout when using the VREF input to regulate current.  It looks like that spec should be Min=200mA

Let me clarify. The reason for not including the "functional" minimum VREF spec is because it varies depending on your operation conditions (VM, current, frequency, etc.) and also on your motor parameters. 

Nathan Cook said:

1) So you are saying in this configuration (current regulation with a sense resistor) that there should be no VREF minimum voltage?  So I should, in theory, be able to regulate close to 0A using the VREF input?  

correct, VREF only lowers the comparator limit which in turns sets the current regulation limit (ITRIP) based on equation (1). Setting VREF to zero regardless of the RSENSE value will cause the current to drop near 0A. Although because of the way the current regulation scheme in this driver, the outputs are re-enabled after TOFF expires. But then immediately disables the outputs if actual motor current is higher than ITRIP. Therefore, when VREF=0V, the output voltage and current will not be equal to 0 but instead will have periodic transients when the outputs are briefly re-enabled. (Side note: always set VREF>0.3V when RSENSE=0ohm to avoid the voltage offset of the internal comparator from tripping the current regulation)

Nathan Cook said:

2) You mention comparator offsets and such.  Can you quantify this or is it random and unpredictable part to part?  It appears from my last tests that there is a linear relationship between the sense resistor value and the lowest VREF voltage that the chip will respond to.  

I explained why the input offset voltage of the comparator in my reply from April 13. Essentially input voltage offset can cause the current regulation to trip when RSENSE=0ohms and VREF <0.3V. This is a limitation of the comparator. It was calculated 0.3V to be the minimum value needed to be confident that no regulation occurs by mistake. Let me be clarify that this setup and conditions are only applicable if current regulation is disabled or ISEN voltage and VREF are similar in voltage value. During normal operation using current regulation, there will be some hysteresis to prevent inadvertently regulation turn ON when there is noise in ISEN voltage or when ISEN or VREF values are close in value.

Regarding the second part of this questions, I have reached out to our design team to run a couple simulations and tests to understand the relationship between RSENSE and minimum VREF and also why the minimum current is 200mA. I will let you know as soon as I have some data to share. expect a reply within 24-48 hours. Our design team is based in Bangalore and will take sometime to receive data due to time zone difference. 

Nathan Cook said:

3) In both of my test cases, below that voltage level that VREF starts responding and actively regulating the current, the constant current level is about 200mA.  Is this in fact the lowest that the VREF current regulation can go?  Should I ever expect to get a current lower than this when using VREF current regulation with a sense resistor with the DRV8251?

The short answer is NO. Please see my second answer in this reply. As I said in my answer above, our design team can run IC level simulations and look at the internal circuitry to explain this phenomena. 

Again, Apologies for the misunderstanding. 

Regards,

Pablo armet

I still think we aren't on the same page here.  I'll try to ask one simple question with a clearly defined test condition.  

Test Condition:

1) 1 Ohm Sense Resistor on ISENSE pin.

2) Load:  Coil Inductance = 3.6 mH, Coil Resistance (dc) = 2.3 Ohm

3) IN1 fixed high at 3.3V and IN2 fixed low at GND

4) Using VREF voltage input to set current regulation output

5) (Sorry I forgot this)  VM=12V

Test results from this configuration:

This is the same plot from above with the given configuration here.  Yellow = VREF voltage, Blue = Current.

Note that the current never goes below about 200mA.

One Simple Question:

In this configuration, why does the current never go below 200mA?

Hi Nathan,

I have talked with our design team and they will let me know if this 200mA is expected. I will let you know within 2 business days.

Thank you in advance for your patience.

Regards,

Pablo Armet

Hi Nathan,

We are still investigating. But we have an initial theory. The current regulation in this device works such that the outputs are disabled when motor current reaches the ITRIP threshold. The outputs remained disabled until TOFF expires and then the outputs are re-enabled but will quickly disabled the outputs again if the motor current crosses ITRIP. This small ON time and TOFF time cycle will repeat until the current regulation condition goes away. 

So this small ON pulse will cause the motor to drive current for a short amount of time which will cause small ripples on the motor current. Depending on your motor's resistance and inductance, it may look as though the current is not exactly 0-A. 

Regards,

Pablo Armet

Hey Pablo,

That is an interesting idea, but I am not seeing any ripple.  And it's certainly not "small ripples" resulting in "current is not exactly 0A".  The lowest current is 200mA!  That is not "small current ripple".  

Please reference the plots from my previous posts.... Note that immediately after the VREF voltage passes the threshold that current regulation start working, the regulated current tracks the VREF signal very closely.  If your "ripple" theory was correct, I would expect to see 200mA of ripple on the current output all the time.  That is not the case.  

Instead what I see is a minimum current of 200mA, and then when VREF exceeds the threshold where it starts tracking, it tracks very well until the VREF dips below the threshold and then.... 200mA again.  

Nathan,

The ripple will depend on your motor inductance. if you have other motors with different inductances, you can try and see if this 200mA is constant regardless of motor inductance.

Regards,

Pablo Armet