EEVblog Electronics Community Forum
General => General Technical Chat => Topic started by: onemilimeter on June 28, 2011, 09:58:17 am
-
I wish to measure the phase voltages of an induction motor as shown in figure below. Each phase voltage measuring circuit (PVMC) contains an isolation amplifier (i.e. Avago ACPL-790A), which requires two dc power supplies (i.e. VDD1 and VDD2). The post isolation amplifier circuitry of each PVMC is the same to the one shown in Figure 21 below. PVMC-1, PVMC-2 and PVMC-3 can share the same VDD2. My question is can PVMC-1, PVMC-2 and PVMC-3 share the same isolated VDD1 power supply? Or, each must have its own isolated power supply for VDD2?
Cheers.
-
It looks to me like all three PVMC units have their Vin- connected to the same common terminal. In that case, a single floating supply should be fine with the negative output connected to the common terminal. It looks like that isolation amp has an input range of +/- 200 mV, even with a single supply. This is nice -- you don't need a floating split supply to keep the common mode voltage in range. You probably need voltage divider to keep the signal in range unless your phase voltage is 200 mV.
-
It looks to me like all three PVMC units have their Vin- connected to the same common terminal. In that case, a single floating supply should be fine with the negative output connected to the common terminal. It looks like that isolation amp has an input range of +/- 200 mV, even with a single supply. This is nice -- you don't need a floating split supply to keep the common mode voltage in range. You probably need voltage divider to keep the signal in range unless your phase voltage is 200 mV.
Thanks. True... it needs a voltage divider to keep the input signal within the allowable range.
-
What is the purpose of the circuit (C6 and R4)? As low pass filter?
Cheers.
-
Is there an AC source available?
If so how about a small linear supply?
-
Is there an AC source available?
If so how about a small linear supply?
Yes... there will be a AC source. I'm planning to derive several isolated power supplies using linear regulator from it. Is that what you meant in your last post? Cheers.
-
What is the purpose of the circuit (C6 and R4)? As low pass filter?
U3 is configured as a standard op-amp differencing amplifier. C6 and R4 need to be matched to R3 and C5, just as R1 and R2 are matched. The resistor is necessary for the circuit to work, but the capacitor does act as a low-pass filter. I suspect that the intent of the capacitor is that it overwhelms the parasitic capacitance of the op-amp and allows for improved common mode rejection.
Also, make sure that in this configuration, where the sense resistor is on the high side of the motor and driven in a push-pull fashion the circuit 'ground' bounces back and forth between -HV and +HV. That means that each phase would have to have their own independent floating supply. In this circuit, they are getting 'double duty' out of the floating supply by also using it to drive the high-side gate voltage
-
It looks to me like all three PVMC units have their Vin- connected to the same common terminal. In that case, a single floating supply should be fine with the negative output connected to the common terminal. It looks like that isolation amp has an input range of +/- 200 mV, even with a single supply. This is nice -- you don't need a floating split supply to keep the common mode voltage in range. You probably need voltage divider to keep the signal in range unless your phase voltage is 200 mV.
"IF" the input side of the isolation amp requires +/- 5V supplies, do you think the PVMC-1, PVMC-2, and PVMC-3 can still share the same "floating split supply"? I don't understand the statement "... to keep the common mode voltage in range" in your comment above. Kindly advise.
Cheers.
-
Why do you need the complex isolation in the first place, wouldn't it be easier to put the µC/ADC on the non-isolated side and isolate the digital signals (or debug interface) instead, it is really much easier (and considerably cheaper) than trying to isolate the fragile and easily spoiled analog signals? Isolation makes sense for user safety, not for circuit protection. Furthermore, isolated is not really isolated at AC, it still has shunt capacitance across the isolation gap, which sometimes causes a great pain and agony for the system designer. Thus it is best to not have fast switching transients across the isolation.
Even the phase currents can be measured against the non-isolated ground (by putting sense resistors between HV ground and bridge switch), when sampling of it is timed exactly when lower switch in the bridge conducts (An usual feature of motor controller oriented DSCs).
I believe many motor drives are done in this way. You can get isolated debugging interfaces, at least for TI's C2000 series of digital signal controllers.
Regards,
Janne
-
I agree with Janne, the industrial and prototype drives I've seen don't use such isolation. The usual way to get isolation is using Hall sensors (magnetoresistive are a relatively recent option, don't know about the costs).
As a reference for this kind of design, I'd recommend you to take a look at the schematics and layout of this Texas dev kit http://focus.ti.com/docs/toolsw/folders/print/tmdshvmtrpfckit.html (http://focus.ti.com/docs/toolsw/folders/print/tmdshvmtrpfckit.html). It isn't the state of the art (IMHO there are also some small signal conditioning errors), but it's a complete solution that works quite well and is quite cheap as for measurements and conditioning.
-
I would try the circuit first with some batteries.
-
Why do you need the complex isolation in the first place, wouldn't it be easier to put the µC/ADC on the non-isolated side and isolate the digital signals (or debug interface) instead, it is really much easier (and considerably cheaper) than trying to isolate the fragile and easily spoiled analog signals?
Thanks jahonen. I've recorded your suggestion in my log book.
-
I agree with Janne, the industrial and prototype drives I've seen don't use such isolation. The usual way to get isolation is using Hall sensors (magnetoresistive are a relatively recent option, don't know about the costs).
As a reference for this kind of design, I'd recommend you to take a look at the schematics and layout of this Texas dev kit http://focus.ti.com/docs/toolsw/folders/print/tmdshvmtrpfckit.html (http://focus.ti.com/docs/toolsw/folders/print/tmdshvmtrpfckit.html). It isn't the state of the art (IMHO there are also some small signal conditioning errors), but it's a complete solution that works quite well and is quite cheap as for measurements and conditioning.
Thanks scrat. Currently LEM LV-25-P is used in my system. The response time of the LV-25-P is approximately 40us, which is found to be too much for a control algorithm I tested. It seems that a response time of less than 10us can be achieved using the Avago ACPL-790A (datasheet states the maximum propagation delay is 3.3us) and that's the reason I go for it. However, I agree with Janne and you that the circuit I propose may not be economical viable from commercialization point of view.
-
"IF" the input side of the isolation amp requires +/- 5V supplies, do you think the PVMC-1, PVMC-2, and PVMC-3 can still share the same "floating split supply"? I don't understand the statement "... to keep the common mode voltage in range" in your comment above. Kindly advise.
Cheers.
Yes, all three branches still share the same 'common' terminal, so they can share the same floating supply, whether split or single ended.