OPA2544: Voice Coil Driver using transconductance amp

Hey Thomas,

Great work on this circuit! This looks very much like what I would expect to see for this type of circuit. 

thomas mcgilvary said:

1.   Should I achieve the required gain with U1, or by the sense resistor or by the R3/R2 ratio.    Currently using R3/R2 ratio.

R3/R2 will be applying initial signal gain, and the U1 path will be applying positive feedback to set the current drive for your coil drive. 

It looks like your design is operating well in the current configuration, so maybe we can stick with the current R3/R2 ratio for now. 

thomas mcgilvary said:

2.  Does one of these make stability easier.    I have not analysed stability in transconductance amplifiers before, but followed a similar post on the forum to do this.

This type of configuration can make the stability more complicated as we now need to be concerned with the positive feedback causing ringing through the circuit. If the U1 control signal has too much phase shift in the frequency response, we will find that the circuit will destabilize. The good news is in that we can always slow the system down with components like C1. slower response does not always help, but in many cases it does. 

thomas mcgilvary said:

3.  The snubber circuit seems to be important in the stability analsyis.    I can change these value once metal is cut, but what value should I use for this simulation?

Snubbers appear simple in nature, but in reality it is not uncommon to have some trial and error in finding the ideal value. These values you have selected are in the range for what i would expect to see. Fortunately, SPICE will get us an approximate range of values, it just may require some real world testing to get the rest of the way there. 

thomas mcgilvary said:

I have not used bypass caps in simuilation tools before.   Are they really required?

Bypass caps in SPICE simulations are almost always added to encourage designers to not forget them when creating the PCB schematic. It is possible however to allow bypass capacitors to operate in SPICE, but it requires modeling power "traces" to have real impedance and to make the supply sources non-ideal. You could pull C10 and C7 off with no impact to the simulations. Of course, it is very important in the board design to include these capacitors. We often recommend including more than one bypass capacitor to ensure low AC impedance for return currents. 

thomas mcgilvary said:

5.   I can procure 0.01% resistors for matching.    Is this precision required,   Is 0.1% good enough?

I would shoot for .1% if possible, .01 is likely overkill. The important matching is R4 to R7 and R3 to R2. Ideally these should be similar ratios for a stable response. 

thomas mcgilvary said:

Below is ac analysis to examine stability.     I notice noise at low freq in the phase.    THere is also a bump in the phase around 4khz.    Would appreciate comments and improvement in the circuit.

I see the 4k bump, but I this looks to be well corrected with the snubber network (no concern here). 

For fun, I might add an equivalent cap of C1 value to GND at the IN+ node. This will better match the AC response of the feedback paths. 1pF Will not make any real world difference, but in the event we change C1, it will be nice to have the ability to math the positive feedback side. 

When I ran the simulation, I see that +-2.5V in is not giving me 2A out. I assume this is the nature of your Rsense question. 

It looks like increasing your ratio of R2 to R3 and R7 to R4 solves this problem well. 

Quick questions, what supply rails are you using here for your application? I was not able to get the +-2A response without increasing the supply rails due to the impedance of the load.  

Thanks,

Jacob

Hey Thomas,

Perfect, please let me know how the testing goes!

Best,

Jacob