BUF634A: overvoltage protection of the output

Paweł Tomulik

Part Number: BUF634A
Other Parts Discussed in Thread: BUF634

Hi,

section 8.3.3 of the BUF634A datasheet describes ESD protection feature and provides the following schematics. It's said that "current limiting resistors must be added at the inputs if common-mode voltages higher than the supply voltages are possible". RIN is a current limiting resistor foor VIN in the schematic.

Image 8.1 from the datasheet

It's also said, that protection diodes can typically support 10mA continuous currents. Does it also apply to the output Vo? How can I protect the oputput from overvoltage in an application where low-output-impedance is required (i.e. current limiting resistor can't be used)?

over 2 years ago

0 kai klaas69 over 2 years ago

Guru 140200 points

Hi Pawel,

how can an overvoltage enter the output of BUF634A? And what overvoltage will be expected? Please describe in detail.

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

The output is exposed to user, DC coupling, so a mistake is expected. Let's assume the voltage 10V above supply rail (for example, +25V for a +/-15V supply rails).

0 kai klaas69 over 2 years ago in reply to Paweł Tomulik

Guru 140200 points

Yes, but how shall +25V be applied? From which source and why?

Kai

0 kai klaas69 over 2 years ago in reply to kai klaas69

Guru 140200 points

Hi Pawel,

you could add AC coupling by the help of a cap. But I guess you don't want that.

Another remedy is to use a fast acting melting fuse. You could also use a polyfuse but trip time of such a polyfuse is huge.

In any case some series resistance would be helpful, to make an isolaton resistance. Togther with the internal protection diodes this could act as a simple diode clamp protecting the output until the fuse breaks.

You could even add external protection diodes. But diode capacitance could be critical.

Kai

0 kai klaas69 over 2 years ago in reply to kai klaas69

Guru 140200 points

Best you show us a schematic and tell a bit about your application

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

kai klaas69 said:
Yes, but how shall +25V be applied? From which source and why?

I think about situation as below

Externally applied voltage

The device with opamp will be given to students, they'll have access to other stuff such as power supplies, so anything can happen.

0 kai klaas69 over 2 years ago in reply to Paweł Tomulik

Guru 140200 points

Don't give it to medicine students. They will kill everything

I have studied physics and have supervised a beginner's practical training for medicine students. So I know what I'm talking about...

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

kai klaas69 said:
You could even add external protection diodes. But diode capacitance could be critical.

Yes, this was my initial approach, ext. diodes + fuse. But... aside of capacitance, there is another possible issue - the forward voltage of the internal ESD diodes which is unknown and if it by any chanse it's lower than that for external diodes, the external diodes will simply be baypassed.

0 kai klaas69 over 2 years ago in reply to Paweł Tomulik

Guru 140200 points

Hi Pawel,

an additional series resistor can help. But to give a proper recommendation I need your schematic. That's why I asked for your schematic

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

kai klaas69 said:
an additional series resistor can help. But to give a proper recommendation I need your schematic. That's why I asked for your schematic

Well, it's still pretty incomplete, but the amplifier part is like the following

The above is intended to be used as a buffer for a function generator. In default application it shall drive RC/RLC circuits of relatively low impedances (e.g. RLC circuit with R component ~150Ω, typical function generator with 50Ω output impedance would have significant amplitude drop). Test frequencies for these RLC circuit are between 10Hz and 30kHz. I'd also like to have a possibility to use the buffer at higher frequencies to drive other circuits.

0 kai klaas69 over 2 years ago in reply to Paweł Tomulik

Guru 140200 points

Hi Pawel,

unfortunately, your circuit isn't stable:

pawel_buf634.TSC

Kai

+1 kai klaas69 over 2 years ago in reply to kai klaas69

Guru 140200 points

More this way?

pawel_buf634_1.TSC

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

I've just realized, that for BUF634A model, the output VO is offset by about +200mV from input VIN. Is it normal? According to datasheet, typical input offset is 36mV.

buf634_test_circuit.tsc

0 Paweł Tomulik over 2 years ago in reply to Paweł Tomulik

Prodigy 20 points

Seems like "Rev A." models generate +200mV offset, while the "Rev B." sets it to 36mV.

www.ti.com/.../BUF634A

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

kai klaas69 said:
Hi Pawel,

unfortunately, your circuit isn't stable:

Thanks! Now it becomes evident, that I'm trying to solve issues in wrong order. Looks like I need to resolve first issue first - to read a book. ;)

+1 SimaJalaleddine over 2 years ago in reply to Paweł Tomulik

TI__Mastermind 22360 points

Hello Pawel,

Thanks for bringing this to our attention. I let the model engineer know about this issue, and he is looking into fixing it to make it more evident that Rev B. is the correct version of the BUF634A.

There are internal ESD protection diodes to the power supply for all pins on the device. If further protection is needed for overvoltage protection at the output of the amplifiers, I agree with Kai if you can you can add capacitor at the output to block DC signal. If not, you could do something similar to this:

Schottky diodes turn on earlier than the silicon ESD diodes and prevent them from turning on and conducting too high of current. The ESD diodes are rated for a maximum dc current of 10 mA. Either way, would need to be simulated to determine stability especially with the added output capacitance. You might need isolation resistors which will depend on the capacitance at the output.

Thank you,

Sima

+1 kai klaas69 over 2 years ago in reply to Paweł Tomulik

Guru 140200 points

Hi Pawel,

if you insist in using the B340 which is a very robust Schottky diode but with a very high junction capacitance unfortunately, then I would give this scheme a try:

pawel_buf634_3.TSC

For enhanced protection R4 is introduced. It additionally limits the current through the internal ESD protection diodes of the output of BUF634.

R4 and the junction capacitance of B340 form a phase lag in the feedback loop which is compensated by C1.

To limit the voltage drop across B340 in overvoltage condition R6 is added. R6 is nothing else than the fuse, of course, concretely spoken the "nominal cold resistance" of fuse.

Take care with the fuse: Fast acting fuses can show an unbelievably high cold resistance and voltage drop at rated current. You would be surprised. A semi time lag fuse might be the better choice then.

I think a 100...160mA melting fuse, fast acting or semi time lag should be appropriate here.

As Sima already mentioned, add TVS across the supply voltage pins of BUF634!!! The TVS must be able to absolutely safely limit the overvoltage to less than +/-18V, but must not relevantly conduct already at +/-15V.

In any case the performance of protection circuit should be thoroughly tested by experiment. Fuses can be tricky and are always good for a surprise...

Kai

0 Paweł Tomulik over 2 years ago in reply to kai klaas69

Prodigy 20 points

Thank you. This clarifies a lot. In the meantime I also found TI precision lab series which explains several EOS (Electrical OverStress) protection techniques. One crucial thing, I learn, is that ESD protection is not EOS protection .

Amplifiers

Amplifiers forum

BUF634A: overvoltage protection of the output