LMH6629 open loop and phase

Hello,

I could not find the LMH6629 open loop gain / phase data over temperature to share with you!

I plan to take this data. What package (SOT-23 or WSON) is being considered (since the package type may affect things)?

I won't be able to measure a lot of units (just a couple of them and most likely from the same lot / run). I'm not sure how to satisfy the question about "the part to part variation with temperature". Also, this is not something measured in production on the ATE.

Please let me know about the package, supply voltage used, and what type of information would be needed to spec part-to-part variation?

Regards,

Hooman

Hi Shinnichi1,

Which LMH6629 package type (and what supply voltage) were you planning to use (SOT-23 or WSON-8) so that I can measure the open loop gain / phase over temperature for you with high priority / first? With WSON-8 package, the COMP pin setting is another variable that would be good to know if you have already decided?

I can do the other package and other supply voltages later.

Please let me know.

Regards,

Hooman

Dear Hooman

Thank you for your email.

I apologize my tardy reply.For my personal reason I was very busy because my father passed away last week.

Today I come back to the office.

I confirmed the customer, he is considering the following.

Could you please see as detailed below.

1. The package is SOT23.

2. deviation from lot to lot  (If it is difficult, could you please give your expertise.)

3.deviation of over temperature.

4.conclusion if the device is satisfy 40dB at 10MHz　under all lots and temperatures or not.

In addition, he is also considering LMH6624.

I'll post as another question.

Could you please advise, too. 

Best regards,

Shinichi

Hi Shinichi,

As I noted in the other post, I will measure the LMH6629 Bode plot over temperature by next week:

• http://e2e.ti.com/support/amplifiers/high_speed_amplifiers/f/10/t/367644.aspx

Regards,

Hooman

Hi Shinichi,

As promised, I've measured the LMH6629 open loop Gain / Phase response that you've been looking for.

I've got data at -40C, 25C, 85C, and 125C for both Gain and Phase in the SOT23-5 package:

At close to 1GHz, my setup may not have the best response as the Gain response seems to behave a little unexpectedly. But, this should provide you with the overall over temperature performance and to get some confidence that things are well-behaved over the entire operating temperature range.

Here is the actual data if you are interested:

1488.LMH6629 open loop gain _ phase response E2E Response 9_17_14.xlsx

Regards,

Hooman

Dear Hooman,

Thank you for your reply and measure.

It's a great data, I'll send the data to the customer.

Also could you advise some about my understanding

1. AOL is over 50dB at 10MHz at all temperature. 

2. How much should I estimate piece to piece and lot to lot. Could you tell me approximate deviation.

General information is OK.

Best regards,

Shinichi

Dear Hooman,

Thank you for your reply.

It's all right. if you find the information, could you please send me.

This is only my image, sample data is over 50dB at all temperature.

 I think It may be enough to achieve over 40dB, if any variations are existed. 

If you find the evidence, it's perfect. 

Or general information is also good, Lot to lot deviation is +/-3dB and piece to piece is +/-2dB for example.

Best regards,

Shinichi

Dear Hooman,

I'm sorry to send again.

I discussed to the customer.

LMH6629's AOL is good, but it was oscillated by customer's circuit.

LMH6629 is passed over.

So additional question is not needed.

I appreciate your dedicate support.

Instead, customer is considering LMH6624.

He wants to know minimum AOL. His  minimum AOL is change to 30dB. 

I will post another. 

Could you please advise me.

Best regards,

Shinichi

Hi Shinichi,

I would not have expected that you run into any instability / oscillation with the LMH6629 when operated at close to 30dB or 40dB closed loop gain! Generally, at lower closed loop gains, the amplifier would have larger bandwidth and be more prone to instability, if you are not careful. But, not at these higher gains. That's unexpected.

If you like, please supply more details about the way the circuit is configured (resistor values, load, input impedance, layout, supplies, decoupling, method of bread-boarding or circuit layout, etc.) so that I can possibly help you debug the oscillation issue.

Regards,

Hooman

Dear Hooman

Thank you for your reply.

At this time, customer has already changed circuit spec., minimum AOL is 30dB,  he is considering LMH6624 and LMH6611.   

Just in case, I ask him the circuit's detail information.

But the customer is secretive company.

I apologize in advance, if I can't obtain.

Best regards,

Shinichi

Hi guys!

I think this would be the best place to ask about this.

I've been trying to plot the AOL for the LMH6629 using the SPICE model in LTSpice and TINA-TI and in both cases I got something around 49 dB flat until 500 KHz or so. It was something unexpected, instead of something similar to the AOL plots that Hooman posted.

What am I doing wrong? The circuit is very simple, just a voltage source to the non-inv input, neg input to ground and an output port at the output pin of the OpAmp

Thanks

Hi Gefeba,

It seems that you are using a single-supply of 5V and having the inverting and non-inverting inputs to tied to GND, which is outside the input common mode range of the op-amp. Make the DC source voltage on the non-inverting inputs as  2.5V and the same on the inverting input instead of GND.

I have attached TINA-TI simulation file if you want to take a look. LMH6629_Openloopresponse.TSC

Best Regards,

Rohit

Thanks a lot for the plots!

But... does it mean that the LMH6629 does not have the same speed when it's powered with a single 5V supply than when two 2.5V and -2.5V supplies are used?

Can you please provide a schematic for clarity. What bias is the non-inverting input connected to when you have VSup+ = GND and Vsup- = -5V? I think you may be violating the common-mode spec of the amplifier here. On Page 6 of the datasheet the CMVR spec is -0.3V to 3.8V on 0V to 5V supplies. This means than in your setup, you should have Vin+ connected to -1.2V (1.2V from the positive supply). This is why your circuit starts working at Vsup+ = 1.2V....indicating that you have Vin+ connected to GND in your application.
-Samir

Just to speed up, I attach the schematic of my setup. The input will have a current source that will generate a negative voltage at the output (hence the need of -5V in V- = V2). Is this possible? Because the LTSpice model does not allow me to do it in this way.

Hello,
Looks like the error is what I mentioned earlier, namely, you are violating the input common-mode spec of the amplifier. In this configuration, you will need to keep pin 3 @ -1.2V to satisfy the input common mode requirements of the amplifier. Since pin 3 is high impedance pin, you can generate the -1.2V through a resistive divider between GND and -5V...you can use a 1.2KOhm and 3.8KOhm resistor to generate the -1.2V.
-Samir

Hi Samir!

Thanks a lot for the quick replies. Now it's a bit more clear to me what the requirements of the OpAmp are. But using your suggestion I have another issue in my simulation, more specifically in the noise calculation. I basically want to use it as a low noise photodiode amplifier. The amount of noise I get using directly -1.2V at Vin+ is way less than using a resistive divider to output -1.2V from the original -5V. I played around and the noise gets similar when the rsistors are very low, 1.2 Ohms and 3.8 Ohms instead of KOhms, but ofc the current and power consumption is prohibitive. My feeling is that it could be relate to the voltage noise produced by the resistors, which increases with R. Do you agree?

It seems that the design I had in mind with the LMH6629 had more issues than I thought...

Hi again!

I'm reading the datasheet and in the page 23 it says that Rb (at the non-inverting input) "should to be no less than 25 Ω for optimum LMH6629 performance. A shunt capacitor (not shown) can minimize the additional noise of Rb."

This confuses me because it seems to indicate an opposite behavior to what I saw in the simulations. They recommend a higher (no less than 24 Ohms) resistor for optimum (i guess, less noise) performance.

Hello,

I am not quite sure I understand why Rb needs to be > 25Ohms. Hooman is an expert on this part. He may be able to clarify when he gets back. Getting back to your earlier questions - you are correct, the noise is higher due to the resistor noise AND the amplifiers current noise * resistance. The solution to this is to have a shunt capacitance in parallel with the resistor to reduce the high frequency impedance seen at the node. Because of the Ib direction in the LMH6629 (current flows out), I would put the capacitor  on the 1.2KOhm resitor to GND. A capacitance of 0.01uF - 1uF should suffice.

-Samir

Hello,

I'm trying to jump into this discussion and see if I can offer any help, but I'm not entirely sure what the issues are?

If you still need some help, I recommend that you post your entire schematic along with specific notes of what the issues are, and my colleagues and I will try to lend some help, as best as we can.

Regards,

Hooman