[SOLVED] ADL5380 (Demodulator) Conversion Loss

[@YY]

Hi,

I am using Analog Devices's ADL5380 Demodulator for my down-convert mixer. In the datasheet, it promised a conversion gain of 5.8dB gain at 5.8GHz. However, I tried for many times but I got a conversion loss of 9dB. Is there anyone has the same problem? Can I have some idea for this?

 

[BigBoss]

Nobody can know without the details that you have implemented.If the datasheet has written 5.8dB@5.8GHz, you did probably an error.

 

[albbg]

What are the loads on I and Q outputs ? Did you use differential or single ended configuration ? What is the formula you used to calculate the gain ?

 

[@YY]

Hi BigBoss, yes, that's why i trying to troubleshoot the circuit. Is there any idea can be shared? I have out of clues actually.

Hi albbg, yes, I did have a balun to convert the differential to single ended.

There is some problem I not sure was, I have my 1st design many months back. The board is well functioning but after 1 months the conversion gain drop to -9dB loss. Then I try to reconstruct the new circuit board but it still give the same result. So I have some touch up for the circuit board where I external grounded the GND patch of the IC (demodulator). Then the board is functioning again. So I thought I have solved the problem, until last week I powered up it turns back to -9dB loss again. Then i tried to construct the new board with the external GND. Somehow the board still give the -9dB result.

I should admit I have really lack of the RF knowledge. But I have tried to debug everything i can. Is there anything important I should have concerned during the design but I left out?

 

[vfone]

Because you still get the mixing process, but with gain about 15dB lower than should be, for me looks like one of the RF or LO inputs lost the balanced configuration.

The gain of the mixer is easy to calculate.
The difference between RF signal input in dBm, and IF signal output also in dBm, is the gain (or loss) in dB.

 

[albbg]

From what you said I understood the board was working properly, then after one moth suddenly the gain dropped from +5.8 dB down to -9 dB. Reworking the board you obtained the correct gain, but after a while you have had -9 dB again: correct ?
It seems to me you have some soldering problem, possibly related to the ground. Be sure the exposed PAD on the underside of the package is well soldered to a low impedance ground plane.

 

[rf1008]

5.8dB is voltage conversion gain not power conversion gain, you may calculate the power conversion gain from it.In general,demodulator has negative power conversion gain.

 

[albbg]

5.8dB is voltage conversion gain not power conversion gain, you may calculate the power conversion gain from it.In general,demodulator has negative power conversion gain.

This is why I asked configuration, impedences and used formula, but from what said by @YY it seems at the beginning the gain was in line with the expected one; after a while it dropped down to -9 dB.

 

[@YY]

Hi,

I have tried to construct a new board but it still gave the loss instead of gain. I have checked the power gain shall be around 15dB for 5.8dB voltage gain. Hence, i have a big trouble where I have -9dB loss, which is 24dB difference from the expected results. I have made the grounded the exposed PAD underside of the package.

I used 3.3pF cap for the DC block at the RF side and 270pF cap at the IF side.

I able to see the correct down-converted signal but only the power loss problem. I tried to increase the power input at RF port by 1dBm each step, the down-converted signal shows increment by 1dBm accordingly. Beside, I tried to decrease the power input at RF port by 1dBm each step,and again the down-converted signal shows decrement by 1dBm accordingly too. I think the linearity of the down-convert mixer is there right? I think the circuit is functioning well, just the i couldnt get a conversion gain.

Personally I thought the circuit is simple and only plug and play. I think I was wrong. Too bad i lack of experience and debugging skill. Any more suggestion? I would try my best to work it out.

 

[albbg]

Could you please specify the the impedance you have at I and Q ports ? What about the IF frequency (I and Q) ? Are you sure the balun can pass the IF frequency ?

 

[vfone]

ADL5380 use differential inputs for RF and LO ports, and perhaps you did something wrong with the balance of the differential ports. Be sure you are using the balun transformers as shown in fig.80 and fig.81 of the datasheet.

https://www.analog.com/media/en/technical-documentation/data-sheets/ADL5380.pdf

If still cannot get the gain try to use the single ended option for the RF and LO ports mentioned in fig.84. In this situation have to match the ports to 25 ohms.

 

[@YY]

Hi albbg,

I am expecting a 40MHz (single tone) I and Q signal. I have checked the impedance is closed to 50ohm.

Hi vfone,

Yes, I have the correct balun, but i am using a 3.3pF instead of 100pF because I am dealing with frequency at 5GHz.
I have tried to make it single ended but it still give the same result.

Besides, I also had tried to change the balun at the baseband but it doesnt work too.

Is there any internal amp in this mixer? Could it be the amp not function?

 

[vfone]

There are three VCC's on the chip, to supply and bias the internal buffers of IF, LO, and baseband. I assume on your board all are connected to +5V.

According to the ADI system evaluation datasheet (check Table 1), ADL5380 has a power loss of -4.6dB (for RFin dBm - IFout dBm)

https://www.analog.com/media/en/reference-design-documentation/reference-designs/CN0374.pdf

 

[mtwieg]

You should just post your schematic at this point.

The datasheet is somewhat ambiguous on the measurement method used to get that 5.8dB number. It looks like they just terminate the baseband outputs with a differential 450 ohms, and probably measure the differential voltage appearing there as the output. So if you have a balun on the output which steps the impedance down to 50ohms and connect it to a SA, then the conversion gain will be lower than that.

 

[@YY]

Hi vfone,
Thanks for the evaluation board info. I am wondering the ADL5380 power output and Vpp is not compatible. Is the answer correct? Beside, I not quite understand about a sentence on page 3 regarding the conversion gain, which is as below,

"Table 1 assumes a 500 Ω load, 5.3573 dB conversion gain, and −4.643 dB power gain for the ADL5380."

Does it mean the ADL5380 output have to be 500 ohms load? Why is there a 5.3dB gain but the table shows a -4dB loss?

Hi mtwieg,

As per attached is my circuit. Please feel free to comment. Besides, i have tried both the 50 ohms and 450 ohms differential output, but both gave the same results.

Please advise.

 

[mtwieg]

Hi vfone,
Thanks for the evaluation board info. I am wondering the ADL5380 power output and Vpp is not compatible. Is the answer correct? Beside, I not quite understand about a sentence on page 3 regarding the conversion gain, which is as below,

"Table 1 assumes a 500 Ω load, 5.3573 dB conversion gain, and −4.643 dB power gain for the ADL5380."

Does it mean the ADL5380 output have to be 500 ohms load? Why is there a 5.3dB gain but the table shows a -4dB loss?

Basically yes, you should load the IF outputs with near 500 ohms (minimum 200), as the datasheet recommends.

5.3dB of gain refers to the voltage gain from the RF input to the IF outputs. But that is in the case where the IF load resistance if about 500ohms. The tenfold increase in port impedance means that there is a tenfold decrease in power gain (but not necessarily voltage gain). This is why the power gain is -4.6dB.

 

[@YY]

Hi mtweig,

I have tried to both the balun below,

T1-6T ratio 1:1
https://www.minicircuits.com/pdfs/T1-6T-KK81.pdf

and

ADt4-6T ratio 1:4
https://www.minicircuits.com/pdfs/ADT4-6T.pdf

However, both has the same results with -7~-9 dB power loss.

For the 5.3dB voltage gain, if converts the voltage gain to power gain, is it suppose to be around ~12 dB power gain?
Besides, in the ADL5380 evaluation board circuit note, the output of of the power level and voltage peak-to-peak in Table 1 (pg 3) is not equal. Did i calculate wrongly?

I have a huge trouble and curios because I did get the gain before but now my circuit shows the loss. Also, if I able to prof the problem is the limitation due to the IC, then at least I have the explanation.

Please advise.

 

[vfone]

In my opinion (and according to ADI evaluation datasheets) you cannot get a power gain with this chip. You could ask ADI in this matter.
Perhaps this is the reason that the system design is using an amplifier after the mixer, before the the A/D converter.

 

[mtwieg]

Hi mtweig,

I have tried to both the balun below,

T1-6T ratio 1:1
https://www.minicircuits.com/pdfs/T1-6T-KK81.pdf

and

ADt4-6T ratio 1:4
https://www.minicircuits.com/pdfs/ADT4-6T.pdf

However, both has the same results with -7~-9 dB power loss.

Okay, when using a 4:1 balun on the output you should be presenting the IF port with 200ohms. The datasheet says that the voltage conversion gain with 200ohms should be around 5.9dB, so the power conversion gain should be around -0.1dB. So something besides the IF transformer is wrong. What is your input return loss on the RF port? How much current is the thing drawing?

For the 5.3dB voltage gain, if converts the voltage gain to power gain, is it suppose to be around ~12 dB power gain?

No, power gain and voltage gain are different, and you can't relate the two without knowing the impedance at both ports.

Vgain=20log(Vout/Vin)
Pgain=10log(Pout/Pin)
Pout=Vout²/Rout
Pin=Vin²/Rin

So if Rin and Rout are equal, then Vgain and Pgain are also equal. Now lets say you have your downconverter with Rin=50 but on the output you have your 4:1 balun which is loaded with 50 ohms, which presents the IF port of the downconverter chip with 200ohms. At the interface between the chip and the balun, where you see 200ohms, you may get a Vgain of 6dB, but your Pgain will be 0dB. After the balun, where you see 50ohm, your Vgain and Pgain will both be 0dB.

 

[@YY]

Hi Vfone,
Thanks for the advise.

Hi mtweig,
Thanks for the information.

For a 1:1 balun, I would see a 50ohms input and 50ohms output. For a 1:4 balun, I would aspect a 50ohms input and 200ohms output?

I have developed both circuits with 1:1 balun and 1:4 balun, but i am curious that both the circuits shows the same result.

Besides, I have also checked that the RFin is actually 30ohms impedance. would this caused the conversion loss? Is this critical?