LM5106: Stepper motor weird current output

Hi, Nima

Thanks for your interest in TI here. I've contacted the appropriate product group. You should hear from them soon.

Regards,
Yunsheng

Nima,

I am an applications engineer for the LM5106 and will be working with you on this item. To begin with, I would like to get a better understanding of the application:

- Where in the circuit is the shunt located?
- Are the plots above actual collected data or a tool simulation?
- What do the waveforms for the LM5106 outputs look like? Do you suspect the LM5106 is the culprit?
- A partial schematic showing the part of the circuit with the driver, FETs, shunt, and motor would be quite useful.
- What value of RDT is being used to set the dead time? Have you confirmed that this dead time setting meets application needs through the RPM range?

Hopefully we can resolve your issues and help your team meet your deadline.

Regards,

- Daniel

Hi

- Where in the circuit is the shunt located?
we use INA240 inline with motor to collect the current information . the good thing is that INA240 is bidirectional current sense IC. therefore I don't have low side current issue stuff.

- Are the plots above actual collected data or a tool simulation?
Actually we created a software which work as oscilloscope . every time we trigger the software it will provide us the last 2000 samples from the reference and feedback and output of current loop. If you need any other imortant inforamtion I could still add to software. so please let me know if you need any other readings .

- What do the waveforms for the LM5106 outputs look like? Do you suspect the LM5106 is the culprit?
LM5106 is perfect, no issue. my PWM after LM5106 also clear . definetly the issue is not from LM5106. Most probably our issue comes from some mistake in the way we control the motor on the algoritm side and etc.


- A partial schematic showing the part of the circuit with the driver, FETs, shunt, and motor would be quite useful.
I wonder if you could provide me your email address or something. since this is a company project. not very convenient to disclose the schematics in public.

- What value of RDT is being used to set the dead time? Have you confirmed that this dead time setting meets application needs through the RPM range?
at this moment I dont reameber the value but I m sure is not dead time issue. the LM5106 is working perfectly so far at 20khz. at higher frequency LM5106 under 4 amp will burn and I have many issues with LM5106 at any speed higher than 25KHZ but that one I will most probably will create another page for it.

Nima,

Thank you for providing further details. I have sent you an e2e private message/friend request so you can securely share the schematic with me.

One possibility to consider is an issue with the PID controller sample rate vs the system bandwidth if you are using digital control. Could it be that the PID loop is not running fast enough to keep up with the plant? What is the controller being used?

I will take a look at the schematics once I receive them and let you know if I see anything that could help. Please understand that I may not be able to debug issues that are not related to the TI device (s).

Regards,

- Daniel

Hi Nima,

Thanks for considering TI part for any and all of your needs.

Would it be possible for you to share the schematic as much as possible?

Regards,

Ritesh

Hi Daniel

Actually we run PID at the same speed rate as PWM frequency which is 25KHZ. And we have tried higher speed PID and PWMs . Up to 100khz which is quite high. Still the result is same and no we achieve no improvement.

The PID get saturated in high speeds and it forces the PWM to 100% duty cycle. the response is very fast which is correct but the current output cannot even get close to to request reference, not reaching to refrense is not an issue but the shape of current get disorted , and this shape distortion is an issue . In my humble opinion, this is not PID response time issue. I have to mention we have competitor which we test bench our product with at the same speed they fully reach to 100% pwm exactly like us. but when we measure their current output with our very expensive accurate current probe their output is really meaningful oppose of us ,our current output at full speed and under 100% PWM duty cycle is really messed up.

Our original PWM was centered aligned. but yesterday we tried a PWM signal which is exactly look like slow decay PWMs inspired by TI reference designs. As a result the current output become more stable, but after some time hardware get unstable and started to burn . we tested the new pwm few times, one time INA240 burns another time LM5106 burns, another Time one of the mosfet burns. still testing the Slow decay PWM signal once I get better result I will update you again regarding it.

But still my competitor runs a centered aligned 25KHZ pwm and they set PWM duty to 100% and they have very meaningfull current output and we don't .I really dont know where the problem comes from.

Still thanks for you prompt reply and I hope you could guide us through this mess :D .

Nima,

I received the schematic via private message, and I have reviewed it with Ritesh. We did not identify any issues related to the LM5106 or the power MOSFET stages. The schematic I received did not show any bulk capacitance for the power supply of the H-bridges, but I imagine they are shown somewhere else in the full schematic document. Insufficient bypassing could be one reason for not being able to draw enough current when running at higher RPM

Our suspicion is that you may be experiencing a control issue. Some items to consider:

- Check scaling of any feedback values and ensure nothing is clamped

- In PID loop check if the constants and multipliers are as they need to be

- Generally switching frequency and loop bandwidth are not the same - have you tried speeding up the PID with respect to the PWM?

- Does the current measured with your custom software instrument match what you measure with the accurate current probe?

Since the issue does not seem to be related to the TI device, we offer the ideas above in good faith, but we may not be able to further debug the system issues.

Best of luck, and please do let us know if you find out what the problem is, as it will benefit the entire e2e community.

Regards,

- Daniel

Hi Daniel

for the question you asked for control side I think I m quite ok. but regarding bulk cap I m not sure. in the second message which I send you today as you can see we use only one capacitor at the 24 volt input which is 330uf/50V. I frighten as you said it might not be enough. But PWM output is clean. might the problem cause by this capacitor?