question about LMH0384

Can you provide a block diagram or schematic of this application?  It is unclear to me how you have these devices configured.

The LMH0384 is a cable equalizer (receiver), and the LMH0302 is a cable driver (transmitter), but it appears you might be using these devices in different roles.

The LMH0302 cable driver will drive an 800 mV_P-P single-ended signal through 75Ω coaxial cable to the LMH0384 cable equalizer.  The LMH0384 automatically adjusts its gain to reverse the effects of the cable loss and restore the original signal.  The LMH0384 has limitations in how much gain it can apply, and this equates to a maximum cable length of about 140m Belden 1694A cable at 2.97 Gbps.  What length & type of cable are you using?

Are you driving the signal from the LMH0384 to the LMH0302 instead?  The interface when driving a signal from the LMH0384 to the LMH0302 should be differential, typically via 100Ω differential traces on a PCB.

Gary Melchior

Hi,

Sorry,,,my bad.the diagram is in the annex.

model A is my system model.the cable which I used is 75Ω coaxial cable,and the length is about 3m. I want to know the LMH0302 will rebuilt the signals or not? I tried to adjust the parameters of FPGA's serdes,such as swim voltage and pre-emphasis,I'm not clear with these parameters. It seems like any pre-emphasis is a bad idea,and lower swim voltage is a good way.the normal swim voltage is about 1020mv in my FPGA. The receiver will running good for a while when I lower the swim voltage between 780mv to 920mv.I don't understand why this happens.

In the transmitter PCB,length of PCB trace from FPGA to LMH0302 is about 7cm. I think the diff-signals on transmitter PCB doesn't affect the single-ended at 75Ω coaxial cable if the LMH0302 receives the diff-signals correctly,Is that right?

I've tested the receiver before by using another resource to replace the transmitter,It's OK for 3G SDI.

8270.model of my system.docx

This is still not clear.  Based on your diagram, it looks like you are using an FPGA to generate a signal, and then sending this signal differentially to an LMH0302 cable driver to drive the 75Ω coaxial cable.  (So far so good.)  But then, this signal is being received at the other end of the coax by another LMH0302 cable driver (should this be an LMH0384 equalizer instead?) and then sent differentially to another FPGA.

The LMH0302 is a cable driver and not a receiver (or equalizer).  Its input should be 100Ω differential, and its output is typically 75Ω single ended for driving coaxial cable.  The LMH0302 does not equalize, retime, or restructure the signal; its purpose is simply to drive the proper amplitude and edge rate into the coaxial cable for proper recovery by an SDI equalizer (e.g. LMH0384) at the other end of the cable.

The differential input to the LMH0302 should be via coupled traces with 100Ω differential impedance, with the differential termination resistor placed close to the LMH0302 input pins.  7 cm is not long and does not require pre-emphasis or other signal conditioning.  But it is important to follow proper high speed PCB design guidelines for the LMH0302 differential input traces to ensure they are balanced and the impedance is well controlled (especially for 3G data rate operation).

Gary Melchior

Sorry,I've made a mistake.Yes,The LMH0302 which was placed on the receiver PCB of my diagram should be LMH0384.I think I've got the issues of my system after many tests.

there are some issues on the power supply.

1:I found a swim voltage about 300mv v-pp on the 3.3v vcc.this maybe cause LMH0302 to stop working when passes a 3g-sdi signals.

2:there are 2 drivers on my transmitter PCB,and these two drivers are powered by one LDO(AMS1117) directly,A FB must be placed on PCB to separate their power supply,But we don't know the parameter of the FB,and placed a inductance by mistake.Now I place a 0R resistance on there and they can work normally and last for a while,But don't stable.

As you said:"7 cm is not long and does not require pre-emphasis or other signal conditioning",I translate it as another way:these SDI signals will be drived by LMH0302 independently,and any other way to change source's(FPGA's serdes) parameter won't affect the signal on the coaxial cable if these diff-signals is received correctly by LMH0302,Yes?

Best regards

Tom

Yes, you are right; as long as the differential signal is cleanly received by the LMH0302, then any signal conditioning applied prior to the LMH0302 should not have a noticeable effect on the LMH0302 output.  With such a short trace length (7 cm), signal conditioning should not be necessary anyway.

Please refer to the following application notes for useful guidelines on supplying power to SDI devices (such as the LMH0302 and LMH0384):

• Power Consideration for SDI Products
• Power Design for SDI and other Noise Sensitive Devices

Gary Melchior

Hi,

Guiding by datasheet of LMH0302, LMH0302 or LMH0307 are actually dual output cable driver devices.Now I want to use it as coaxial cable driver that should be a single-ended cable,So I use the SDO(positive) port as the single ended signal to transmit.what should I do to the SDO/(negative) port?

Thanks

Tom

I've drew the diagram of LMH0302,it's very important for us and I must let that be sure.All about this is on the annex.Please check that for me,Thank you very much.

2783.diagram of LMH0302.docx

Tom

Hello,

Thanks for your help,I've get the answer.

Thank you again!

Tom

Hello Gary

    I am having problem with 8 to 12 video multiplexer design, I thought maybe you can help me I appreciate it.

    I am using LMH0302 as the cable driver (single end) and I am getting slower rise and fall time (about 350 pico seconds), is the 4.7 uF cap suppose to be a specific part? is there anything else I should check for? appreciate your help.

appreciate your help

Perry

pfatehi2@gmail.com

Hi Perry,

Please check SD/HD pin(pin 10) and make sure this pin is low. Could you please confirm?

Regards,,nasser 

hello Nasser

      Thank you for your help, I had checked the SD/HD pin and it is low, I have been soldering tektronics probe and following the signal from input to output, in fact I am in the middle of it right now.

We are using LMH0344 for our receiver and at the output of the receiver signal looks good (rise and fall time appear to ne normal)

signal from the receiver goes to multiplexer which is DS25CP104A via pair of traces and at the input of the multiplexer signal does not look good, the rise time levels at about 70% from the start point for about 20 pico second  and then it goes up which results at lower rise time ( measuring between 20% and 80% of the signal)

I really really appreciate the help, I can Email you the scope image and schematic in pdf format if you give me your Email, I am stock and need help.

Thank you again

Perry

Hi Perry,

What type of cable are you using between the BNC of the LMH0302 and the waveform analyzer? Please note SMPTE calls for 1m B1694A cable and color bar data pattern. Is this what you are doing?

Also, please attach your gerber or layout to this case so we can comment further. My email address is nasser.mohammadi@ti.com

Regards,,nasser

Hello Nasser

    I looked at the signal in more detail and it looks like problem also exists right after our receiver and it gets magnified after the mux and the cable driver so the problem seems to be at the receiver which is LMH0344. What do you think I should be re-checking at the receiver circuitry.

Thank you again for all your help

Perry

pfatehi2@gmail.com

Hello Nasser

   Thank you for the fast response, we are using belden 8218 75 Ohm cable, I was not doing test pattern, I just did and looks the same. I will get the gerber file on Monday and send it but it looks like it is the LMH 0344 and circuitry around it.

Thank you again

Perry

Perry,

Given the problem exist at the output of the LMH0344, please check to make sure you comply with LMH0344 data sheet schematic and etc.

To debug the receiver further, you can use a short B1694A cable to see if you still see this issue?

Additionally you can pull Bypass high on LMH0344 to see if you still see this problem.

Regards,,nasser

Hello Nasser

      I have checked the data sheet the schematic complies

      I will try the short cable

      I will try the bypass pin

      Let you know as soon as I have the result

Thank you again

Perry

Hello Nasser

   I ran some test here are the result

  1) scope images

            a) image 1 is (single shot)            ch1= input of LMH0344, pin 2&3     ch2= input cable(75 ohm)

            b) image 2     same as image 1 but continious

            c) image 3 is (single shot)            ch1= input of LMH0344, pin 11 & 10     ch2= input cable(75 ohm)

            d) image 4 i    same as image 3 but continuous

            e) image 5 is (single shot)            ch1= input of multiplexer U25 pin 1 &2 (same point as image 3 with added trace)     ch2= input cable(75 ohm)

            f) image 6 i    same as image 5 but continuous

            g) image 7 is our final output at J9

Nasser if possible please give me your phone number so I can call you, my number at EON is (818)781-2185

I will try to attach the images, schematic, board layout, layout stack to this if failed I will email it to you

again thank you for your help

Perry Fatehi