The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Other Measurements >> On Wafer Measurement
https://designers-guide.org/forum/YaBB.pl?num=1183451527

Message started by aaron_do on Jul 3^rd, 2007, 1:32am

Title: On Wafer Measurement
Post by aaron_do on Jul 3^rd, 2007, 1:32am

Hi all,

I need to do some on wafer measurement for low power IF circuits such as IF amplifiers. The problem is the 50 ohm spectrum analyzers are loading down the circuit. So i'm thinking i can either design an on chip buffer or get an off chip buffer. For on-chip, the buffer would require much more power than my circuit which is not good. For off-chip, i simply haven't been able to find anything to buffer the signal with. Does anybody how i can get around this problem, what equipment to look for etc?

thanks,
Aaron

Title: Re: On Wafer Measurement
Post by mg777 on Jul 3^rd, 2007, 7:54am

The simplest solution is to put a 0604 470 Ω in series with your output, so you can still use your 50 Ohm GS or GSG probe. Given that you are at IF (maybe < 300 MHz) your tech should be able to put together a suitable SMA resistor barrel in no time. The barrel should likely be placed at the probe end. You'll get a 20 dB attenuation, hopefully that won't hurt your noise floor. If you're concerned about a match at the probe end throw in a 50Ω shunt after the 470 ohmer.

If you want to get a proper high-impedance probe, check with GGB: www.picoprobe.com. They used to do custom work, I suppose they still do - check with them.

M.G.Rajan

Title: Re: On Wafer Measurement
Post by didac on Jul 4^th, 2007, 12:14am

Hi,
If it's just a test chip and you have free pins you can put the buffer in the die with an independent power pin(I suppose that your concern about not putting the buffer it's due to the power consumption that will "mask" the IF circuits power consumption), if you don't have free pins for the power supply of the buffer one solution could be to perform a differential measurement of the power consumption, first you turn off all the biasing voltages of the IF circuits and measure the current at the power pin(this current will be all consumed by the buffer) then you start sweeping the bias voltages until the increased current equals to your expected consumption, one problem on this method is that you must "compensate" the losses introduced by the buffer by hand so it's a good method if you can test the buffer indepently to characterize it.
PS:I confirm that picoprobe continue to do custom work(in my opinion with good quality),but also expect a few weeks of delay while they manufacture and send the items so plan with anticipation(depending on which part of the world are you living now).
PS2:for the equipment question which measurements are you planning to do?, I think the basics are a Signal analyzer,a calibrated noise source, an arbitrary waveform generator,precision voltage sources,precision multimeters, baluns if you ghave a differential topology...

Title: Re: On Wafer Measurement
Post by aaron_do on Jul 4^th, 2007, 8:57pm

thanks for the replies.

For the equipment question, picoprobe.com answered it perfectly. I'm wondering though, how would such a probe affect noise measurements? I actually did use a separate supply pin for my buffer, but i'm worried about always having to use the buffer. In my latest design I went with an ultrahigh power buffer to get good performance...I used a simple source follower with about 30 mA or so to drive the 50 ohm load :D the test circuit less than 1 mA so i'm a bit worried...

BTW rather than going out and buying a new probe, would it be possible to connect an external buffer? I have a microwave probe (up to 50 GHz) so i guess the probe tip has very low capacitance. So after the probe tip is a 50 ohm cable. Could I simply attach a high impedance buffer at that point? I guess the 50 ohm cable could be modeled by an ideal conductor. If so, anybody know where I could get one?

thanks,
Aaron

Title: Re: On Wafer Measurement
Post by didac on Jul 5^th, 2007, 11:14pm

Hi,
I think that some instrumentation vendors sell modules to provide a high impedance to the circuit and 50 ohms to the instrument(at least Tektronix and Agilent sells this kind of things,Rhode I'm not sure and other vendors no idea) but they are for my taste quite expensive,maybe it could be possible if your frequency is below 100MHz to mount in a small PCB a good OP_AMP like the LM6181 as a voltage follower, it will provide a high input impedance and you can put a 50 ohms load at the output. Making this PCB with SMD components and SMA conectors on both ends will allow to fix it directly to the probe output and the output of the PCB connected to the coax that goes to the instrument.
Hope it helps,