ADS8688: Asking if ADS8688 is enough in terms of speed to be used in FOC

Josel Go

Part Number: ADS8688

Tool/software:

Hi team,

I am using ADS8688 in my design. I am designing PMSM FOC but am unable to tune my FOC.

I need to ask if ADS8688 is enough in terms of speed to be used in FOC. I am implementing FOC PMSM
It requires sensing inline current along with shaft angle to compute d and q currents
I am currently using ads8688 at around 8Mhz and it is taking roughly 22us for 3 channels. Then I measure my shaft angle, So which add a phase delay between the shaft angle and phase currents.

Please advise

Regards,

Josel

over 1 year ago

0 Tom Hendrick over 1 year ago

TI__Guru**** 181981 points

Hi Josel,

can we get more detail? Perhaps a schematic or block diagram of the proposed setup? The ADS8688 has a fixed 15kHz low pass filter built into the device, which may be causing the tuning issue.

0 Josel Go over 1 year ago in reply to Tom Hendrick

TI__Expert 6355 points

Hi Tom,

Yes I'm aware of that. But the max electrical cycle will of 250Hz at 3500rpm.

So is the 15kHz filter an issue? Plus its taking me 20us to measure three channels for ads8688 at ~8Mhz. Can i get any faster using auto mode?

Can i somehow use ads8688 to sample analogue data and would read the shaft angle at the same time, so there won't be any delay between angle and current readings?

Regards,

Josel

0 Tom Hendrick over 1 year ago in reply to Josel Go

TI__Guru**** 181981 points

If you have fast edges on the input signal, yes, the 15kHz LPF could be an issue. The ADS8688 does not have simultaneous sampling capabilities, so there is no way to sample the signals at the same time. The MAX SCLK on the ADS8688 is 17MHz so running at that speed would put you at sub 10uS to sample three inputs.

0 Josel Go over 1 year ago in reply to Tom Hendrick

TI__Expert 6355 points

Hi Tom,

I think the problem is that I sampled the electrical angle and then measured phase currents. It gives me a delay of ~20us between the angle and currents. Do you think this might be an issue?

Regards,

Josel

0 Tom Hendrick over 1 year ago in reply to Josel Go

TI__Guru**** 181981 points

Josel,

It could be, usually there are (at least) two phase currents that are simultaneously sampled. Here are a couple of app notes for your reference.

https://www.ti.com/lit/ug/tiducy7/tiducy7.pdf

https://www.ti.com/lit/an/sprabz0a/sprabz0a.pdf

0 Josel Go over 1 year ago in reply to Tom Hendrick

TI__Expert 6355 points

Hi Tom,

I have already read them and had a look at them again.
The first paper doesn't give a detailed guide/explanation of the current inline sense. And I am using inline current sense. But it seems very clear that both currents and angles should be sampled simultaneously.

Regards,

Josel

0 Tom Hendrick over 1 year ago in reply to Josel Go

TI__Guru**** 181981 points

Simultaneously sampled would be the best option in my opinion.

0 Josel Go over 1 year ago in reply to Tom Hendrick

TI__Expert 6355 points

Hi Tom,

So should I be shifted to a built in adc? or you can name a TI adc that has a sampling option?

0 Tom Hendrick over 1 year ago in reply to Josel Go

TI__Guru**** 181981 points

Here are some 4 and 8 channel simultaneous sampling options for you to take a look at.

Precision ADCs-parametrics-20240821070742.xlsx

Data converters

Data converters forum

ADS8688: Asking if ADS8688 is enough in terms of speed to be used in FOC