AMC1302: Application in H Bridge

Part Number: AMC1302

Hello.

I'm trying to simulate the TINA model of the AMC1302 and I had some issues and weird responses of it on a H bridge but when I erased these sources the reading became completely great. What these sources reproduces and the impacts of it in real applications? 

The design of the reference model of the AMC1302 have some sources and need to be used to simulate this IC, right?

Hi Vinicius!

Welcome to our e2e forum!  Since your two posts are related, I joind them together to keep the conversation in one location.  Can you supply the actual TSC file you are using?  It would be interesting to see what else you have in the circuit.

Hello.

I was trying to simulate an ideal half bridge with amp ops and the real full bridge with MOS. The PWM generator I simulate likely generic with the amp ops.

The half bridge works better but it's ideal operation cannot be considered that much when applied. Other question I have stated is that when the sinewave is at lower frequency the wave is distorced. I would use it between 10Hz and 10kHz of signal, but lower than 500Hz the AMC1302's output becomes very distorced. 

Regards,

Vinicius

Hello Tom,

Thanks for the reply. I connected the grounds as the reference design says. I was simulating without them and so the results were wrong. So let me see if I understood:

• The GND1 must be connect on one side of the shunt resistor and also needs to be connected on the ground of the whole circuit? I'm saying that because on the datasheet the GND1 is only on the shunt but on the simulation reference design it is connected on both. Won’t it be dangerous connecting one side of the H bridge on the ground of other source?
• I tried with a voltage source between the ground and GND2 but it was still noisy. I have the right signal when I cut off the isolator and read IN but I can't recover the signal even changing values and the structures. The IN read keeps different and so the output.
• I tried the voltage source between the ground and GND2 but it is too much sensitive. Any change on frequency of the signal inserted on the PWM the stability is lost.

The focus is keep the integrity of the signal and have the right reading of it but I can’t none of them. I put the ground on GND1 and GND2 on PWM(2). If you could verify what is the matter, I appreciate.

Regards,

Vinicius

AMC1302.tsc8081.PWM(2).TSC

Hi Vinicius,

The way you had GND1 originally is right, when you connect it to TINA ground you will violate the common mode input.  Take a look at the attached changes and let me know if it makes sense to you.Vinicius.TSC

Hello, Tom

Thanks a lot for the changes. It really works now.

I just didn't get the meaning of the source I marked just as the picture. It is as the differential of the source to VDD1? It is just for the simulation and I don't need to do this on real application?

On real application I should so wire it like this to use like my simulation?

Regards,

Vinicius

Hi Vinicius,

You could set the V1 source marked in the square to 0V or 1000V - if you know what your end isolation needs are, you can adjust that as necessary to whatever voltage level you need between GND1 and GND2.  You could even remove the source (for simulation purposes) altogether. TINA does not have a good way that I am aware of to run simulations based on two different grounds so you have to artificially provide that connection to the simulation.

In your end (actual) circuit, the VDD1 source is a floating power supply connected to the 'load' of your H-BRIDGE (the way you originally drew it).  GND1 and GND2 would be completely isolated from each other.

Hello Tom,

I got it about the isolation in the simulation.

But I didn't understand your point about the circuit. The way that it's drawn will work as the same as the simulation? How should be the final real circuit to work like the simulation on the H bridge?

Regards,

Vinicius

Hi Vinicius,

The way its drawn in the datasheet is how you implement the AMC1302 in the end.  You need a 5V power source for VDD1 that is referenced to GND1.  That is a floating supply tied to the center of the H bridge.  VDD2 and GND 2 would be referenced to the rest of your control circuitry.

Hello, Tom,

Your answer sounded like my circuit is wrong about the floating supply. Indeed it is like the datasheet.

So actually I should keep VDD1 and GND1 separated and not let power supply floating?

Regards,

Vinicius

Hi Vinicius,

I think we are confusing each other!

The 5V source to VDD1 must be referenced to GND1.  The low side of your h-bridge is not going to be at the same voltage potential as the mipoint of the bridge, but this is the reference point for GND1 as shown in the ADS1302 datasheet.  The VDD1 source is in fact a floating supply in this case.  The normal application is to tie GND1 of the AMC1302 to the 'INN' pin as you have done in the TINA circuit.

Hello, Tom.

So I actually need to connect the GND1 to get it floating and besides that, needs to be referenced on the 5V?

The 5V would be supplied by a voltage regulator so I need to reference it on the voltage regulator and not on the ground of the circuit?

Regards,

Vinicius

Correct! 

Please feel free to let us review your final schematic before you commit your PCB to the fabricators.

Hello, Tom

Thanks a lot for all your support.

I'm finishing my schematics and a doubt emerged. The floating power supply can be delivering power to others components or should I keep it exclusively to the AMC1302?

As I commented before, I will use the power supply regulator. And so it should supply only the node VDD1 and GND1?

Regards,

Vinicius

Hi Vinicius,

More often than not, you would only be powering the AMC1302, without seeing your schematic, I can't think of what else you might need to power from that supply.  You might also look into the AMC3302, this is a new part which does not require an external high side supply - that might make your design a little easier in the end.

Hi, Tom

I was wondering about this possibility because I have other ICs and the voltage regulator of 5V to supply the VDD1 would be used to supply them. But I also guess the floating of AMC1302 is going to vary the delivered power to those other components. Would this affirmation be right?

Regards,

Vinicius

It should not vary the power necessarily, but the other chips would all have to be referenced to the GND1 node.

Hello, Tom

So as I said before, to create a floating supply with the voltage regulator should I use like this?

Regards,

Vinicius

That's right!  You got it.

Hello, Tom

About the PCB layout of this scheme, if I'm going to reference de commom on the GN1 the die of the voltage regulator used also to dissipate heat should be wired with the same GND1 or reference it on the GND of the circuit?

And, if I would use two AMC1302, each one should have a regulator induvidually or I could tie them on the same reference of the regulator's commom?

Regards,

Vinicius

Hi Vinicius,

Take a look at figure 56 in the AMC1302 datasheet.  The circuit there taps off the high side gate driver power supply with a 5V zener to provide VDD1.  The power source for VDD1 must be referenced to GND1.  Next, take a look at Figure 52 - this is a 3 phase setup it should be similar to what you are doing (just one extra pair of transistors).  Each of those AMC1302 devices in Figure 52 would need to have their own power supply.

Hello, Tom

I got your point. So I need to individualize the VDD1 supply of each AMC1302. 

About the output, the real reading is given by the voltimeter between the OUTP and OUTN? And actually, the normalized gain reading is to match the output with the input and have both at the same level, right?

On figure 57, it has a label of floating power supply but both references the same ground and would it be the commom of the regulator, right? But then the ground referenced on the VDD2 is the same ground? I understood with the schematics analisys we did so far but this layout got me confused when everything is referenced on the ground plane... Or I need to use other way to float the power supply? Do you have any recommended way to do that?

Regards,

Vinicius

Bom dia Vinicius,

Yes - seperate isolated supplies for each AMC1302.  

Yes - the differential output is read between OUTp and OUTn.  The input is +/-50mV and the gain is fixed at 41V/V.  The outputs have a common mode voltage of 1.44V so relative to GND2, you would see VOUTp and VOUTn swing between .5V and 2.5V.

Figure 57 has two ground planes - one for the high voltge side (GND1) and one for the low voltage side (GND2).  If they were both using the same ground plane, there would essentially be no isolation.

Hello again, Tom

About the PCB layout of this scheme, if I'm going to reference de commom on the GN1 the die of the voltage regulator used also to dissipate heat should be wired with the same GND1 or reference it on the GND of the circuit?

I'm kind of confused about the use of the regulator. I made a drawn showing the use of the regulator voltage before and you said that it was correct. But my doubt is if it will work correctly if I don't reference it on the real ground as some datasheet metions about avoiding floating regulators. The datasheet recommends also the of a LDO but the schematics aren't clear.

The previous design:

The doubt generated this one:

Regards,

Vinicius 

Hi Vinicius,

I think we might be back to where we started.  In the second circuit, you are using a ground symbol tied into GND1 which is how you started off the TINA circuit way back when.  That ground symbol was also associated with the bottom half of your H-Bridge (refer again to Figure 52 in the AMC1302 datasheet - note that (-)Vbus is considered the high voltage "ground")  Usually there is an isolated gate drive supply that 'floats' relative to the high voltage ground and its return path is essentially the mid point of the H-Bridge - call this node 'ISOGND'.  In the first circuit above, the gate drive voltage, the LM7805 and the AMC1302 are connected to this ISOGND node.  Assuming your gate drive voltage is 12-18V (relative to ISOGND), that would be the input to your 7805 regulator.

Hello, Tom 

I evolved the schematic and now I think I'm thinking the same as you. Sorry about the misunderstandings.

Could you review this sketch?

Regards,

Vinicius

Nice!  Looks good Vinicius!