Forums

Taizhong,

Figure 77 in the PGA112 datasheet shows two different approaches to scale a bipolar input signal to a useable voltage range for the PGA112, Vin0 is AC-coupled through the capacitor while Vin1 is DC-coupled. In the AC coupled application (Vin0) the capacitor and the resistor RA form a high-pass filter with a corner frequency at 1/ (2*pi*R*C) therefore the value of RA is determined by the necessary low-frequency response and the value of the input capacitor. As you noticed in the figure, there is no DC voltage drop across resistor RA, this is because the input impedance of the PGA112 is very high (10 Gigaohms typ) and so very little current flows through this resistor, producing negligible voltage drop. This DOES NOT mean that RA is 0 ohm, in fact making RA a small resistance will create a very high corner frequency as shown in the previous equation.

As for your second question, the amount of disturbance between the two channels depends on the output impedance of your voltage reference (should be very low) as well as the common-mode rejection ratio of the input amplifier (usually very high). Because a good DC voltage reference will have a very low output impedance I would not expect much disturbance between the channels however, as frequency increases inter-channel crosstalk will increase.

Dear John,

Thank you very much for your answer!

This issue have confused me for more than a month and now I understand how to choose Ra. For the disturbance, I'll follow your instructions and try again, I may add a jump between input1 Vin1 and Vref(PGA), if it disturb too much, I'll break it.

Next edition will be ready in 4-8wks, I'll update the result then.

By the way, I find your suggestion is very helpful so I tried to connect "submit documentation feedback" on the bottom of pga datasheet, but I failed to open the web. I hope your suggestion appeared in pga datasheet so the other reader can understand the AC-Couple easily, furthermore, I guess there are written mistake about Ra,Rb in Fig78. Another suggestion is name Ra in Ch0 Ry or something else, as it's not necessary be same with Ra in Ch1.

Thanks again!

Taizhong

Dear John,

Today I had my new edition of analog board ready, let's name it A3.

Here it's the result: PGA112 in A3 doesn't work properly. The gain can not be changed, it's always be gain==1;the channel can not be changed, it's always be Channel==Ch0;

Background:

I use C2000 TMS320F28027PT to drive PGA112, and the 28027 also drive pga112 in old edition, let's say it's A2 well. Both the gain and channel can be selected, with the same program, the 28027 can not drive PGA112 in A3 well.

The 28027 program is modified from TI Example(V126, spi-loopback).

Here it's today's test procedure:

1. Check A3, confirm there is no welding problem, right voltage everywhere
2. Using 28027 to drive A3, using 28027 to take signal from pga112 pin 5(Vout), I found the Vout==Ch0, the gain can not be changed, the channel can not be changed. Signal out of pga112 is 1.65V centered(half of 28027 ADC sample voltage range.)
3. Using 28027 to drive A2, using 28027 to take signal from pag112 pin 5(Vout), program unchanged, I found Vout is equal to Ch0*8 or Ch1*8 as the program designed, both 28027 and A2 works well except the signal center is far low than 1.65V(half of the 28027 ADC sample voltage).
4. Switch pga112 in A2 and A3. (pga112 in A3 is taken out of A3 and be placed in A2;pga112 in A2 is taken out of A2 and be placed in A3)
5. Using 28027 to drive A3, program unchanged. Gain==1 and Channel==Ch0, Vout(pga) is 1.65V centered.
6. Using 28027 to drive A2, program unchanged. Gain==1 and Channel==Ch0,Vout(pga) is about 1.65V centered.

In early of February, similar situation happened with A2, at that time, gain and channel are also can not be changged, then I took off pga112, channged with a new pga112, it works well.

Here is my question:

1. Is pga112 is so sensitive that it's easily be damaged in welding?
2. If the pga112 is demaged, why Vout(pga)=Ch0, per my understanding, if pga112 is demaged, mostly, Vout should be zero.

Please comments at your convience.

Thanks!

Tyler

Hello,

It would be very helpful to have a schematic of your circuit employing the PGA112 in order for me to determine what is happening. To clarify your message above, it sounds as though the PGA112 that is installed in board A3 is being damaged because swapping the part into board A2 does not restore functionality of the part. Is this correct?

Dear John,

Schematic

A3: PGA112 and Upper side

Lower side:

J8,J9 is connector to 28027.

A2:

Confrim:

About your clarify: You are right, PGA112 in A3 board could be demaged.

Working Frenquency:

My signal is 40KHz, so C13=0.01uF and R29+R30=532ohm, the pass frequency =29.9kHz

Today's Test:

1. Take off pga112 from A3
2. Take off pga112 from A2
3. Weld the pga112 from A3 in A2, after the 3 steps, the A2 got it's original pga112
4. Test A2, it works well. The Gain and Channel can be programmed, the voltage center is also far low than 1.65V as in February.

It looks like:

1. Board A3 has some problem, either the problem is with schematic or print board. Maybe the problem is withen the board as Schematic of A2 and A3 are same about PGA112.
2. Pga112 in board A3 could be in demaged as when it's placed in A2, A2 doesn't work well in gain and channel slection.
3. If the pga112 in board A3 is in demage, where is the wrong? If we bought the demaged ones, the fail rate is too high, as there was similiar situation in February, and also I surf in Chinese website, I found a post showing similar situation happened in Nov 2010.  If the PGA112 is right when I got it, there could be welding problem in A3 or there is board problem with A3.

Plan:

I ordered new PGA112 and will have another A3 board made.

Thank you for your help very much and look forward to your comments!

Tyler

Dear John,

It looks like you can not find my schematic. Actually I posted my schematic and I found it in the confirm email e2e writing to me. Anyway, the schematic is not so important now as the pga works.

It looks like the problem is with "shut down" mode. As PGA112 has no hardware shutdown and I have not set shutdown mode, so I did not think about it. One Chinese TI support engineer remind me my problem is similar to shut down mode. After check all hardware issue, I tried to write E100H to disable shut down mode, then write the SPI command data sucessfully, the pga will work well and badly in a few cycles, after that it works well.

This afternoon, I found the pga works well sometimes and doesn't work sometimes, then I found it's necessary to give pga enough time to receive spi data as spi is in low speed, the best way is have my 28027 null loops for some time after SPI data writing.

Your suggestion on Ra selection is right! The signal voltage is centered in 1.65V now leave alone the gain.

Thanks for your advice!

Tyler