CD74HC4051-EP: estimating power consumption

OK makes sense thanks a lot for clarifying this.

Also, if I had multiple of these multiplexers in a digital circuit connected to a microcontroller, and each multiplexer is enabled sequentially, for the total power consumption in the circuit would I still need to account for both of them at the same time (or just the one that is enabled)?

Thanks again!

Hi,

Thanks a lot for your reply, this was really helpful!

I'm still a bit unsure about the dynamic power consumption calculation, as the data sheet doesn't provide information on the input frequency (fin) or output frequency (fout).

Are these related to fmax (180MHz) given in the data sheet, and if so how? There don't seem to be any other frequency values given which is why I was thinking this one had to somehow be use, but this may be wrong...

I am also assuming that the fmax given is the maximum sampling frequency of the multiplexer, so then maybe using the Nyquist sampling theorem the input frequency to the multiplexer (fin) would have to be fmax/2?

Again, I am not sure if this is a valid assumption, or how the input and output frequencies for the multiplexer would be calculated, as all of the other key parameters for the dynamic power consumption are given in the data sheet apart from these.

For context, the multiplexer is connected to a matrix of force sensitive resistors, although I've read that the sampling frequency of the sensors and the input frequency of the multiplexer are independent so don't think that needs to be used.

Would be extremely grateful for any help on this, and thank you again!

Sama

Hi Sama,

Sorry for the confusion here! Let me back track a bit. Your original estimation of current consumption was pretty close. This is looking at the static power consumption which I believe is what you're interested in for your application. We want to look at Icc though, which is spec'd 160uA or 320 uA based on supply conditions. So you're looking at 5*160uA or 5*320uA depending on your application. Of course, this leads to a very small value because the current through the switch generally does not flow through V_CC or GND, so it is not really consumed.

Calculating power dissipation is a different calculation so let me know if that is needed too!

Regards,

Alex

Hi Alex,

Many thanks for your reply!

Ok, so I used the 'continuous current through VCC or GND' of 50mA for the current, but this is different to Icc that you've given above so just wanted to confirm that I should use those lower values for static power consumption?

I am also struggling with the overall power consumption (i.e. static + dynamic) so need to find the dynamic part.

The part I am not sure about from the equations in the link you attached showing how to calculate power consumption, is the input and output frequencies (fin and fout).

So for context, I actually have multiplexers connected to the rows of my sensor matrix, and shift registers connected to each column that sample the data. I was thinking that the input frequency(fin) for both the shift registers and multiplexers would be same as the sampling frequency I decide for my sensor matrix - would this be a valid assumption?

Also would the output frequency for both these components also be equal to the input frequency (same as sampling frequency of sensors)?

I was thinking this as the sensors are directly connected to the shift registers and multiplexers the shift register frequency would set the sampling frequency of the sensors and also the multiplexers.

My circuit would look something similar to this if that helps at all (multiplexers connecting rows, shift registers connecting columns):

The shift registers I am using are: https://www.ti.com/lit/ds/symlink/sn54hc595-sp.pdf?ts=1703224270900&ref_url=https%253A%252F%252Fwww.google.com%252F

Thanks again for your help, would be great if you had any advice on this to be able to find the dynamic power consumption!

Sama

Hi Sama,

Correct that you would use the smaller Icc values. We're interested in supply current in this case. In the previously attached FAQ, you'll see a calculator for obtaining power consumption near the bottom of the page. The sample Icc provided is in the uA range, so this checks out. 

In fact, the previous FAQ is related to CMOS logic devices. Our switches are passive devices so you won't need to worry about factors like input/output frequencies here. Sorry for the confusion. 

The power consumption for our switches can actually be found using the following formula as found in this E2E

which simplified to 

So, we can use 5V * 160uA + 1 * 50mA^2 * 240 = 0.6758W

Regards,

Alex

Hi Alex,

Thanks again for your reply!

Ah ok, so are the multiplexers and shift registers aren't CMOS components? I was confused by this, as both the multiplexer (https://www.ti.com/lit/ds/symlink/cd74hc4051-ep.pdf?ts=1703216081375) and shift register (https://www.ti.com/lit/ds/symlink/sn54hc595-sp.pdf?ts=1703224270900&ref_url=https%253A%252F%252Fwww.google.com%252F)  data sheets do mention they use CMOS technology, but it also makes sense that they are passive devices...

So the formula above adds the static power consumption with the losses of the device itself (using resistance of switch), so should N not be the number of channels (8) or is it 1 as only one channel is active at a time?

I am also not sure how this would be applied for the shift register, as the data sheet does not give an 'Ron' resistance? - would you please be able to help with this component too?

Thank you again,

Sama

Hi Sama,

The passive nature of this device allows us to simplify many calculations, such as power consumption. As far as N is concerned in the calculation, we only have one active channel at a time so we just choose 1 for this value. 

Secondly, are you asking about power consumption for the shift register? If so, I'd need to move this thread to the appropriate group that handles this device. Let me know if I'm understanding correctly and I will do so!

Regards,

Alex

Hi Alex,

Many thanks for your reply!

Ok this makes sense for the multiplexer, is there a reference to where this equation for power consumption (e.g. a paper or website) comes from at all? (no worries if not!)

Yes please, well I wasn't sure if the same equation could be applied for the shift register due to not having the same parameter values in the data sheet... but essentially yes, I'm also stuck on the power consumption for the shift register!

Thank you :)

Hi Sama,

This formula was derived by a previous applications engineer from a simplified linear model of the mux (i.e. an RC model). 

It comes from the idea that the power consumption on the part is a combination of the products of current x voltage at every pin. 

I'll go ahead and loop in the shift register team now!

Regards,

Alex

Hi Sama,

Use the original FAQ to calculate the power consumption for the shift register: https://e2e.ti.com/support/logic-group/logic/f/logic-forum/875721/faq-how-do-i-calculate-power-consumption-or-current-consumption-for-my-cmos-logic-device

If you have any problems please let me know. It would also be helpful to give me the part number for the shift register.

Best,

Malcolm

Hi,

Many thanks for your reply.

I am using a 74HC595 shift register (https://www.ti.com/lit/ds/symlink/sn54hc595-sp.pdf?ts=1703224270900&ref_url=https%253A%252F%252Fwww.google.com%252F).

I think this is a CMOS component, but am struggling to work out the dynamic power consumption using the values given in the data sheet.

It would be great if you could please help with the power consumption calculation - thank you!

Hi Alex,

Sorry to double check this again, but for the multiplexer (https://www.ti.com/lit/ds/symlink/cd74hc4051-ep.pdf?ts=1703216081375), the data sheet shows this equation for the dynamic power calculation:

Can this then be ignored, i.e. still don't need to find the dynamic power consumption using input and output frequencies (which not sure about) in this way?

I wanted to ask this again just to be sure, otherwise will use the equation from before with only static power consumption and power consumption of the multiplexer itself while switching.

Thanks,

Sama

Hi Malcolm,

Thanks a lot this is really helpful!

I had a couple of questions:

First, for the number of inputs switching is that always the total number instead of say 1 if only one input is changing at a time for the shift register?

Also, you mention that the input frequency is something that can be set - if the shift register is directly connected to sensors that are sampled, would the lower bound of the input frequency to the shift register be equal to the sampling frequency of the sensors (assuming the input frequency of the sensors can't be lower than how quickly the sensors are sampled)? - wanted to ask if this is a valid assumption to make.

Thank you :)

Hi Sama,

1. It's how many inputs are changing per unit of time, usually defined as one clock cycle. So how many inputs change per clock cycle fed into the device.
2. It is hard to see your circuit picture above. But assuming you mean the inputs of the shift register are connected to the outputs of the sensor, then yes, I would use the sampling rate of the sensors here.

Best,

Malcolm

Hi,

Many thanks for your reply, this is really helpful.

I am also confused on the communication/data transfer for the shift register. If the shift register is connected to the microcontroller in this type of setup (see diagram below) what communication protocol is being used?

I am a bit confused as to how data is transferred to the microcontroller after the shift registers help to read specific sensor elements in the matrix - I have looked into SPI, I2C etc but not sure if any of those apply for the shift register?

Thank you.