BUF634A: Parallel configuration

Hi Anant-san,

>Even the small offset voltage at the output can cause a very large circulating current between the outputs of the two buffer. 
Let me confirm about below because I may have caused misunderstanding.

* What I would like you to confirm is below.

- Parallel connection two BUF634A is acceptable or not.
  (This means short input and output connection.)
- I can apply same configuration to LMH6321 as well or not. 

I did NOT mean "parallel connection between BUF634A and LMH6321".
In this situmation, is your answer unchanged ?

Best Regards,

Hi Anant-san,

Thank you for your reply.
>it's better to put isolation resistor at the output
Purpose which customer will try to connect BUF634A(or LMH6321) in parallel is to reduce output impedance.
(Customer will drive MOSFET, load capacitance of MOSFET is large, so they will try to improve frequency response by reducing output impedance of opamp.)
According to datasheet, case of BUF634A, output impedance is defined as "5ohm".
In my understanding, in general,  "tens of ohm" will be used as isolation resistor. However, if this resistance value is used, I think customer can not realize their purpose. What do you think about this ?

Best Regards,

Ryuuichi-san,

Can you please provide the following information

1. What the max current they want to drive?

2. Bandwidth required?

3. What's the load capacitor (mosfet gate capacitor)?

Based on these i will try to suggest the best possible solution.

Regards

Anant

Hi Anant-san,

Thank you for your reply.
Here is information I heard from customer.

2. Bandwidth required?
=> Tens of "kHz". (Less than 100kHz.)

3. What's the load capacitor (mosfet gate capacitor)?
=> Yes. And load cap is approx 0.5uF.

1. What the max current they want to drive?
=> As I described, drive target is gate of mosfet. And they said DC load current is almost 0. (drive as AC current.)

Best Regards,

Ryuuichi,

Let me look into this, will get back to you by tomorrow.

Regards

Anant  

Hi Anant-san,

>Let me look into this, will get back to you by tomorrow.
Do you have any update ?

In my opinion, if user can accept dissipation due to offset voltage difference, I think user can adopt this parallel configuration.
In fact, it seems that LT1010 is also have same concern and I can see following description in datasheet of LT1010.

What do you think ?

Best Regards,

Ryuuichi San,

Yes , if dissipation due to current caused by offset error is ok then we can use this method. We can put a series isolation resistance of around 7 ohm. And the output impedance of BUF634A itself is around 3ohm so the total impedance per device will be around 10ohm. When we parallel two such device, the effective bandwidth when driving 0.5uF cap will be around 63Khz. If that is sufficient then we can definitely use it.

Other than this, we have some current feedback amplifier  (like THS3470, THS3491 and THS3120) which are high voltage and high output current amplifier and can be used as buffer but please note to drive 0.5uF of capacitive load we need isolation resistance here as well.

Regards

Anant

Hi,

Thank you for your reply.
Understood.

Customer have following question.
When they were using different OPAMP(not BUF634A and LMH6321), they observed following pulse current due to parastic inductance and capacitance of MOSFET.

* 400mA with 10us pulse

For BUF634A, according to datasheet, you described short circuit current(up to 550mA) as shown below. So, I think this current may accept but could you please give your comment about this ?


Best Regards,

Ryuuichi San,

If 400mA of current is only required for short pulse then yes, we can just use one BUF634A to drive the mosfet gate. In that way we dont need the isloation resistor so the bandwidth will also be high.

Regards
Anant

Hello Anant-san,

Thank you for your reply.
I have following additional two questions.

Q1.
>If 400mA of current is only required for short pulse then yes, 
According to your reply, I understood that 10us pulse can be accepted, however is there any criteria to assess acceptable pulse or not ?

Q2.
One another concern about this is instantaneous power consumption due to short current pulse.
In above customer's case, customer input approx 7V to drive MOS. The case which customer observed short pulse current is when they perform "turn off" behavior to MOS. And they observed this condition approx gate voltage was 5V.
They were using +/-15V as power supply to different OPAMP(not BUF634A and LMH6321). So they observed approx 8W(400mA * (5V - (-15V)) instantaneous power consumption in this case.

At first, when they use BUF634A in this same configuration, we are not sure whether they will observe same power consumption.
However, if they observe same/similar instantaneous power consumption, since all package option of BUF634A have fairly high thermal resistance, so we are concerned that thermal shutdown will be observed. 
On the other hands, from following datasheet of BUF634A, it seems we can estimate this instantaneous power consumption as RMS. 
---
When determining whether or not the device satisfies the maximum power dissipation requirement, make sure to consider not only quiescent power dissipation, but dynamic power dissipation. Often times, this dissipation is difficult to quantify because the signal pattern is inconsistent, but an estimate of the RMS power dissipation provides visibility into a possible problem.
---

I would like to get your advice about how we should consider this instantaneous power comsuption.
Could you please give your comment ?

Best Regards,

Ryuuichi San,

It all depends on the pulse frequency and duty cycle. If the average power in not large, then the instantaneous power of 8W should be ok.

Also, I would recommend to use the package with thermal pad so as to have better heat dissipation to the PCB

Regards

Anant