Forums

Hello,

Most applications are going to supply 3.7V to all the inputs.  In theory the having a 3.7V+250mV on IVIN/IREF and 3.7V-250mV on QVIN/QREF would probably work, but you may end up with some reduced linearity specs as its not at the recommended common mode of 3.7V.

If you can let me know some of your setup conditions, maybe I can try to reproduce your setup in lab.

voltage on power down pin(pin7):??

Supply voltage: ??

LO frequency: ??

LO amplitude: ??

Baseband frequency(single tone): ??

phase difference on IREF and QREF for the baseband signal: ?? (I'm assuming these are 90 degrees out of phase, but wanted to double check)

is the thermal pad soldered to ground?

Thanks

Chris

Hello Chris,

thanks for your reply. Here some additional info about our problems using the TRF3701:

- voltage on power down pin(pin7): 5.1 V
- Supply voltage: 5.1 V (both VCC pins 6 and 10 are decoupled to ground with 1pF and 100 nF ceramic capacitors very close to the chip)
- LO frequency: 867 MHz
- LO amplitude: ca. -3 dBm
- Baseband frequency(single tone): we use a 2-ASK (simple on-off-keying) with a datarate of 80 kbps
- Baseband voltages: DC offset = 3.7 V, Vpp = ca. 350 mV (using a phase of 45 degrees so the amplitudes for I and Q are the same)
- Phase difference IREF/QREF: Yes, 90 degrees. To make it complete: IVIN/IREF 180 degrees as well as QVIN/QREF, both are differential
- is the thermal pad soldered to ground? Yes!
- It's on a four layer board with 50 ohms microstrip used for LO and RFOUT with a ground plane below the top layer.

I've added a screenshot to make clear what's going wrong on our setup:
- The blue trace shows one channel of the baseband signal at the input of the TRF3701. A (filtered) 2-ASK-signal with a datarate of 80 kbps centered at 3.7 V
- The pink channel shows the corresponding RF output of the TRF3701. You see no difference whether it's a modulated '0' or a '1'.
If I understood the datasheet correctly, I should have a full amplitude for '1' and zero amplitude for '0' with this setup.

Now when I change the DC offset on the baseband input, I'd at least expect the amplitudes of the modulated signal for a '1' to go up and for a '0' to go down (or vice versa). But both still stay the same (checked with DC-offset levels from ca. 3.2 to 4.2 V).

Now this somehow looks like a ground problem to me, but the thermal pad is soldered to ground, as well as the other ground pins are connected to the ground plane through vias.

Thanks,
Thomas

In the meantime I replaced the TRF3701 with another one, but the problem remains the same - so I must be doing something wrong...

Thomas,

I did not get lab setup today to evaluate this more.  However, after reading your VCM comments with the offsets not increasing the amplitude the 1st thing I want to check tomorrow is the following

My understanding of your setup:

IREF VCM = QREF VCM = 3.95V  & IVIN VCM = QVIN VCM = 3.45V

What if QVIN/QREF were flipped?

IREF VCM = QVIN VCM = 3.95V  & IVIN VCM = QREF VCM = 3.45V

If for some reason either QVIN/QREF or IVIN/IREF were flipped in the documentation somewhere this would result in the offset detla would cancel each other out.

Is this something you can check easily also on your board?

Thanks

Chris

Chris,

sorry, I didn't want to confuse you with my description - no, VCM is the same for each of the four baseband inputs on our board.

So IREF VCM = QVIN VCM = IVIN VCM = QREF VCM = 3.7 V

Watching the modulator output signal as on the screenshot I posted in my previous posting, I first thought this might be wrong - thus my initial question about different VCM's for REF & VIN inputs.

Thanks,

Thomas 

I've now set it up with single ended baseband signals and have chosen a phase of 0°, so that I = I_max and Q = 0, so: 

IREF = QREF = QVIN = VCM (no baseband signal, just DC) 
IVIN = baseband signal + VCM

I've again varied VCM in the range from 3.2 to 4.2 V. But the output signal again looks as posted yesterday - the same (equal) amplitudes for '0's and '1's regardless of the chosen VCM.

Thomas 

Thomas,

I think what I'll do is friend you through the e2e and that way you can send me your phone number via an email, and than we can talk real time tomorrow.  What time zone are you in?

My understanding of your desired output is that you want some level of signal out when your word =1.  Conversely you want no signal or a much smaller level when your word =0.  From what I can tell from your plot is something like

for 1:  I/QVIN=3.85 (guessing I/QREF = 3.55V)  --> amplitude = 300mV

for 0: I/QVIN = 3.55V (assuming i/QREF = 3.85V) --> amplitude = -300mV

IF this is what is happening than I understand why you have the same amplitude for 1's and 0's..  You may have a slight phase change...

I think I understand why you wanted to modify the VCM to something like

for 1:  I/QVIN=3.85 (guessing I/QREF = 3.55V)  --> amplitude = 300mV

for 0: I/QVIN = 3.7V (assuming i/QREF = 3.7V) --> amplitude = 0mV

Chris

Chris,

thanks again four your reply. I'm in germany, so UTC + 1 - I suppose that might be a problem to talk on the phone. I'm also not at work today (I'll be back at work on monday).

Chris Pearson

My understanding of your desired output is that you want some level of signal out when your word =1.  Conversely you want no signal or a much smaller level when your word =0.

Yes, that's correct.
I'd like a '1' in my baseband signal converted to maximum signal amplitude in the modulator output and a '0' converted to (ideally) 0 V output. 

Chris Pearson

From what I can tell from your plot is something like

for 1:  I/QVIN=3.85 (guessing I/QREF = 3.55V)  --> amplitude = 300mV
for 0: I/QVIN = 3.55V (assuming i/QREF = 3.85V) --> amplitude = -300mV

IF this is what is happening than I understand why you have the same amplitude for 1's and 0's..  You may have a slight phase change...

Yes, that's exactly what's happening and why I thought I'm doing something wrong with my offset voltage.

Which phase change do you mean, a phase change between the differential signals (180) and I/Q (90)? Each of the baseband signals goes through an op amp circuit to apply the 3.7 V offset voltage and to do some filtering, so there could be a small phase / amplitude error - but I'd say this error stays < 5 degrees / few mV.
 

Or is it maybe correct for the TRF3701 that I get those output signals? I'm starting to think that I have a misunderstanding on how the part actually works.

Thomas