I've posted the same problem about the HMC392LC4. We were able to remove the part and bypass that section. However, the same problem is now occurring with our HMC753LP4E LNA. This issue needs to be resolved before we spin a new PC board. Here is the issue:
For each case (and there are three cases so far: 2 with the HMC392LC4 and 1 with the HMC753LP4E), the amplifiers biased up perfectly. They were drawing the typical current shown on the datasheet. The problem is that the gain response, even though very similar in shape, was lowered by a significant amount. This has happened in all three cases so far. The HMC753 is actually part of a three stage amplifier section where the other two RF amplifiers are HMC311SC70's which are leaded parts. The order of the gain section goes: LNA - RFA - RFA . I cannot test the amplifiers individually, but because two other leadless amplifiers have been shown to cause a similar problem, I am leaning toward the LNA being the problematic amplifier in the gain section. I will attempt to isolate the LNA to test it, and update this post when I do. I've also attached the 2D footprint, the 3D footprint and the 3D footprint flipped and showing the simple body of the LNA. Hoping to hear from you soon. Thank you.
Could you share your apps ciruit and PCB layout with vias on pdf?
What are the power up and down bias sequences you follow?
What is the max RF input power level?
The power up sequence is vgg1= -0.7V , vdd = 5V, vgg2 = 1.5V... Power down is reversed.
I've been testing with a VNA at -40 dBm source power.
Debugging the issue with the complete chain is challenging. I suggest to remove HMC311`s cause they are fairly easy to de solder, and connect thru lines instead of them and test the LNA by itself. Soldering a coax cable to C76, after removing RFA4 is another option. Then if the LNA works properly, same process can be repeated by adding the other amps on the signal chain.
Hi Kagan, I was able to resolve the issue with the LNA. It was in fact due to a small break in the RF line hidden by the SMA connector. However, we are still unsure as to why the 2 HMC392LC4's are problematic.