current feedback circuit

Hi! I am working on a unipolar stepper drive based on a ATMega16. I want the driver to have current limiting and microstepping. For that I need to read the current through the coils. The most common circuit used for that is the one in this figure( attachment ) . When I use the oscilloscope to measure from point A to ground I get what you see in the figure. The current through the coil during the ON period of the MOSFET is as expected, but during the OFF period it drops to zero (which is logical, right? If the FET is off the scope is connected to ground through a 1 ohm resistor). The problem is I want to read the decay current of the coil during the OFF period as well. How do I do this? What circuit do I use?
Thank you.

There is a thread "Mardus-Kreutz" on this site that covers your design objectives.
It will give the insight you need since Kreutz designed the whole thing in the long thread, including the firmware. You appear to be doing a similar project.

See here ... http://www.cnczone.com/forums/steppe...rnative-5.html

Edit: This the one I was looking for where a lot of details are provided. http://www.cnczone.com/forums/steppe...epping-42.html

Drive the coil with a H-bridge circuit, and then you can control where the decay current goes. It is possible to use a common resistor in the bottom of the H-bridge.
You need to ensure that the parasitic diodes in the mosfets don't mislead you.
H bridge - Wikipedia, the free encyclopedia

I haven't worked with unipolar for a long time. Bipolar is better and easier if you use a chip such as the tb6600 or thb6064.

In a unipolar drive, the normal practice is to share one resistor between the two half windings. When the transistor switches off, the resistor current goes negative because it freewheels through the other coil.

There is no real need to measure the decaying current. Typical drives simply turn the transistor off for a fixed duration.

I answered the question of how to measure the decaying current.
It is certainly easiest to use a chip that is designed for current mode use and there are too many to consider counting.

My motors are 6 wire so if I were to use H-bridges I would drive them bipolar, I was hoping to go unipolar and use just 4 FETs per motor, if that would be possible. I will read your links tomorrow (its 5 AM now), and I will post another circuit that shows the current decay as well but there is a small problem whit it. Thank you for all the info!

You can still use 4 FETs per motor, if you want to.

See here http://www.cnczone.com/forums/273089-post243.html

Here is the circuit, it shows the current during the ON and OFF period.The problem is that the amplitude is not correct and it doesn't change if I change the supply voltage but it does change if I change the PWM voltage. Is this because I am using N-Chanel MOSFET and I should be using P-Chanel? In the end (after some more reading)I think I will go bipolar, the extra torque is so temping, but still I am curious if this circuit is any good.
Thank you.

My idea was to use the uC to generate the PWM for the FETs and use the ADC of the uC to read the current, no comparators no reference voltage 1 uC, 4 FETs and some passive components (complicate the code and simplify the circuit).I am interested in the decay current because I don’t just want to simply turn the transistor off for a fixed duration, as H500 suggested. As far as I can tell the current sensing part of the Mardus-Kreutz circuit is the same as the first circuit I posted and will not show the decay.

Your circuit looks like a source follower. This means your coil voltage can't be higher than your drive voltage.

If you want good results, guaranteed to work, with not too much research and low cost try this, under $15, and all the research has been done.
DRV8825PWP - TEXAS INSTRUMENTS - STEPPER MOTOR DRIVER, 2.5A, 28HTSSOP | element14 Australia
data:

It sounds to me that you are attempting to use a linear mode rather than a chopper mode for current control. In that case you want a controlled constant current driver, something like the Linistepper idea.

Like post #8 here...
http://www.cnczone.com/forums/steppe...per_usage.html

The Linistepper doesn't seem to go much beyond 2A coil current possibly due to the heating of the drivers and the need for big heatsinks.

The circuit won't work because the mcu will need to drive the gate several volts higher than vcc to turn it on. A pchan fet would have a better chance, you will likely need additional parts to protect the fet.

Why do you wish to do it this way? The extra power resistors, 4 fast power diodes and gate protection circuitry will cost far more than a comparator.

The chip I mention in post #11 has it's own hi-side drive via internal charge pump, so has 4 N chan mosfets in each output.
How hard is it to do it the easy way, for $15? Have I missed something?