Hello,
We are trying to use MMZ09332BT1 which is a 130-1000 MHz HBT RF amplifier in the frequency range of 136-174 MHz. There is an application circuit and relating S parameter graphs for the operation in this frequency band at the datasheet. But the question is whether we can achieve the output power of at least 31 dBm (preferably 32 dBm) throughout the band. According to the extrapolation of the figure of Gain vs Pout which is drawn up to 30dBm, we are going to experience a gain drop of nearly 3 dB (G = 28dB) at Pout = 31 dBm. But the gain is not important for us. I would like to know the saturation power for this operational band. Thank you..
Hi Andrei,
As an addition to my last message, I am now trying to design the layout of the sample boards for AFT05MS003N and AFIC901N based on the figures given in their respective datasheets. The problem is that the dimensions are not given in the datasheet? Do you have idea where we can find it? Should we again make a service request for that and wait for them to answer?
I am not sure what you mean.
Package dimentions for AFT05MS003N can be fond in the AFT05MS003N Datasheet, Page 16
Recommended layout can be found on the Page 15
http://cache.nxp.com/files/rf_if/doc/data_sheet/AFT05MS003N.pdf
Package dimentions for AFIC901N can be fond in the AFIC901N Datasheet, Page 16
Recommended layout can be found on the Page 15
http://cache.nxp.com/files/rf_if/doc/data_sheet/AFIC901N.pdf
Hi Andrei,
Will you please send us the Gerber files of the designed application boards at 136-174 MHz for the following three products, as given in their respective datasheets:
MMZ09332B
AFIC901N
AFT05MS003N
Thank you..
We can not share design files on this public community
Please create a Servuce request for this question.
The P1dB of this device in the 700 MHz band is typically 32.8 dBm with a saturated, P3dB point of 33.2 dBm. These numbers will only be better in the 174 MHz range.
Based on Figure 41 of the datasheet, with the given matching circuit, the p1dB is about 29 dBm. It seems that in order to obtain p1dB value of more than 30dBm, the matching conditions should be changed.
Did you adjust the transmission lines to accommodate the difference in the board materials dielectric constant? Rogers RO4350 has a er of 3.66 and most FR4 has an er of around 4.7. This change would require different trace widths and different matching component placements and values.
In order to provide 50 ohm lines, I adjusted the input and output CPW transmission lines based on the er and the thickness (1.6mm) of the substrate. I chose the line width and the gap as 1.7mm and 0.3mm, respectively. But about the placement of the matching components I just tried to make them as similar as possible to the given board in the datasheet. I thought the placement of the components and the short transmission line pieces in between may not be that much effective at our frequency of interest. Here is the photograph of our board:
Oscillations are typically due to poor back side ground. Did you follow the figure 42 recommended pad layout from the data sheet with similar via grounding underneath the active device?
Yes I followed the recommended pad layout from the datasheet. We could actually solved the oscillation problem. The input power of the network analyzer that we were performing the measurements was set to 0dBm and the number of the sweep points were about 400. When we decreased the number of sweep points to under 50 and decreased the input power level, we could obtain the stability. Our problem right now is not achieving 1W output power.
It seems that I should perform some kind of load-pull analysis in order to find out the proper load value and then design the output matching network accordingly. What is your opinion about this?
Yes it seems that load pulling is necessary. How can you send the working sample? Should we order from somewhere in the website? How much will it cost?
About heatsink, we actually did not use any heatsink before, but upon your advice, we attached an Aluminum heatsink using thermal tape under the via farm at the back side of the board just to see the effect. We don't have access to the copper plate as you recommended. Unfortunately, we could not see any improvement in the output power of the device.
By the way, I created a service request on Monday and another one on Wednesday for sending the Gerber files of the working samples. It was said the answer may come between 24 and 48 hours but there is no reply from Monday up until now:(. What else can we do about it, in your opinion. I think if the Gerbers are not available, the best choice would be samples themselves to be purchased.
We typically do not charge for our small signal evaluation boards. Have you worked with any local NXP or distribution sales contacts? I can enter in a request on your behalf but I would have to send it directly to an NXP employee.
Yes, some form of load pull needs to be performed to verify that you have a good complex conjugate load match to the device.
What sort of PAE are you measuring? It should be in the 45% range at P1dB with 5V operation. If the load match is poor, typically the PAE will be poor too.
I am assuming you are using the 5Vdd. The part is only good for 30 dBm at 3.6Vdd
Thank you Andrei.
Today we measured the output power on the test board which I talked about in my previous post. We obtained the Gain vs. Pout graphics as follows:
As can be seen, this is much different than the graph given in the datasheet. I did not expect this much of difference only because of the PCB material. Certainly there must be something wrong in our design.
Hello,
Based on the test circuit provided in the datasheet for the 136-174 MHz operation, I manufactured the test board much like what is proposed in the datasheet, but on FR4 substrate rather than Rogers RO4350B. I am experiencing an oscillation with a variable bias current and the gain is below 20 dB. Can you also comment on this please? What can be the reason of the oscillation?
I would like to let you know we are working on your request. I will provide answers to your questions as soon as possible.
This thread was updated by a live chat. The responsible team was informed to follow up as soon as possible.
Dear Amin,
We are very sorry for a delay in responding
Thank you for your patience
Have a nice day
Pavla
NXP Support team
Dear Pavla,
Thank you very much. I will be waiting for their answer.