need help for FPGA connections to high voltage circuit

[Spidyelectronics]

Hi everyone,
I am doing a project in which I want to take feedback of high voltages (in the range of 200V AC and DC also) and provide this signals to FPGA. I can convert these high voltage feed backs to lower voltages in the range of 2V to 5V. But I am confuse for one thing that is can I connect these 2V to 5V signals to FPGA?
In this case, the GND for high voltages will be the common for FPGA also. So is their any way, by which I can protect to FPGA.

 

[harii74]

Hello
Convert 230V into 5V by simple potential divider and zener diode. Connect opto coupler between 5v and FPGA kit.

Added after 3 minutes:

Hello
Convert 230V into 5V by simple potential divider and zener diode. Connect opto coupler between 5v and FPGA kit.

 

[karesz]

Hi Spidy,
I think its an important question; how much propagation time do you need?
Optocoppler are today not the best isolating copplers, and they have limited lifetime too!
You can select ADUM12xx i.e. from ADI, but they have in some situations by power up/down falsh impulses_these can make relevant problems for you (i.e. by PWM are both FETs at same time on! :-()
Top actuell are capacitive copplers i.e. from TI, they are not only less expensive, works as multi channal types up to 150Mbps, but simpler to handle and dont have lifetime problems as both befor discussed coppler types...
K.

Added after 1 hours 7 minutes:

Hi Spidy,

tnx!
I have overseen that you have DC tocoppel too; but you need so & so some "primary-side" electronic, with a little pwr supply...
I think, the simplest way is for dc to apply first an U/F converter (only 1 IC) and than some isolating coppler.
Some one can say: its possible direct to transmit with an analogue coppler too_ yes, but you need minimum selected (some more expensive) optocoppler & "a feed back" to stabilize the amplitude changes vs altering and temp effects...
To end it can not be simpler & better as an U/F converter + coppler.
K.

 

[Spidyelectronics]

Thanks karesz and harii74,
I also think that, directly using opto coupler is not a good idea, because optocoupler wont work for DC voltages. Also the amplitude is important for me and if we give variable amplitude analog signal, then at the output of isolator we get the constant amplitude signal.
I think the use of U/F converter + coupler will be good option. But I want to ask you one thing, that will it able to work for both analog signal and DC signals? the amplitude is important here and the second thing is I havent used any U/F converter practically.

Thanks,
Spidyelectronics

 

[FvM]

optocoupler wont work for DC voltages

No idea, where you get this? Standard optocouplers won't actually work for analog quantities, either AC or DC. FPGA don't provide ADC capabilities, so you have to add external converters in any case.

The V/f conversion idea is one of many ways to digitize analog quantities at the primary side and transmit them digitally across the isolation barrier. Optical isolated analog transmission would be another option, but the accuracy is limited. For AC signals, transformers are always an option. To choose a suitable converter and coupler technology, the intended amplitude and time resolution should be known, however.

 

[karesz]

Hi,
Of course (in a range) its equal for your U/F converter if its input is a DC or a changing signal, i.e. your 50Hz, but be care pls, on the receiver side you will need a(or two) F/U or F counter & its not needed for the DC signal if you have an ADC eventually on chip of uP, but_OK_ you have an FPGA, it will not have an ADC .....
Shorter said; I thank with U/F conversion only over your DC parts, but its possible (its depending of especially used IC or circuit) to use for 50Hz too.
K....

 

[hakeen]

Interesting discussion. Just my 2 cents.

Optocouplers will work in DC. They are level triggered devices, not edge. On the contrary non-optical technologies nwould have diffculty to pass DC information. However, in IC design you can get rid of this with some tricks, I guess.

I think you are looking for analog input analog output kind of isolator. In this case you should analog optocouplers. You can use parts like HCNR200 or isolation amplifier like ACPL-C790. I would prefer former one, as it has also feedback photodiode. You can adjust LED current over time. However package is very big to achieve higher working voltages up to 1414Vpeak. It is very probably too bulky for your app. To be able to make a further comment, I would need more data about your requirements. What is max speed of feedback, do you need reinforced or basic insulation etc.

Currently there are several kind of technologies for isolation (or for your case more correctly insulation ). I strongly believe in nature there is always ying and yang . In case of EEE and more explicitly isolators, every technology has its own advantages and disadvantages. Not mentioning any disadvantages would be a marketing strategy, not technical reality.

Cheers

Added 5 min later: I also don't know why you have to use additional components and adding complexity to the design with using V/f conversion.

 

[karesz]

Hey,
we arnt writing that optocopuler dont work for DC, It was writen: they arnt enough linear & hase problem over longtime...
So is a sinple U/F conversion better for linearity and hasnt problem safter some years.

Tese system will be than more cheap too_ it was all ...
regards!
K.

 

[hakeen]

Hi Karesz,

Dont worry. I am just sharing my thoughts and dont put your knowledge into question. Sometimes it might be difficult to see different perspectives of different options.

I believe there linear enough optocouplers working in pure analog domain. This is why they are often used for current/voltage measurement. HCNR201 is another kind of good example for linearity with additional error feedback.

Being cheap can be changing under different conditions. If it is not for hobby project and really for production, I would always recommend asking to authorized distributors for pricing. I was very surprised to see how the pricing was changing then. Moreover I also saw many ICs which had chepaer sell price, but in mid and long term it was costing more than initial value.

Again these are my thoughts and experiences after several projects.

Cheers

 

[karesz]

Hi Karesz,
Dont worry...
... I would always recommend asking to authorized distributors for pricing. I was very surprised to see how the pricing was changing then. Moreover I also saw many ICs which had chepaer sell price, but in mid and long term it was costing more than initial value...

Hi,
OK_Im (relative) flexible (or believe it minimum)!
Yes, the price are an extra subject..
Sometimes ago I had a project for an well knowed global player wher had so 300-500`000 pices only from my equipment to produce pro annum.
Their "global buyer net" has prices, their are for us unrealistic cheap & not everytime to understand, as you are writng it too.
For normal situations I do (over longer time) check by https://octopart.com the possible prices_their are most enough realistic & lited by their!
K.

 

[FvM]

I fear, it's somewhat "bloodless" to discuss analog signal isolation methods without a least a rough idea of intended accuracy and bandwidth. The requirements may be as different as 10 to 24 bit and a few Hz to 100 Khz or even MHz. Obviously, this matters a lot when identifying your favourite isolation technique.