LMH0318: Problem with CDR

Sergey Avdeev

Part Number: LMH0318
Other Parts Discussed in Thread: LMH0397

Hello!

I use LMH0318 in my development. The HD-SDI signal with an impedance of 100 Ohms is input to IN0 of this chip. Wire length from the camera to LMH0318 20 cm.

Code for initialization:

writeRegister(0xFF, 0x04, 0x07); //Select Channel Registers

writeRegister(0x16, 0x25, 0xFF); //Enable Full Temperature Range

writeRegister(0x3E, 0x00, 0x80); //Initialize CDR State Machine Control

writeRegister(0x55, 0x02, 0x80);

writeRegister(0x6A, 0x00, 0xFF);

writeRegister(0x03, 0x80, 0xFF);

writeRegister(0x0A, 0x0C, 0x0C); //Reset CDR

writeRegister(0x0A, 0x00, 0x0C); //Release CDR reset

writeRegister(0x31, 0x03, 0x03); //Select IN1*/

writeRegister(0x0A, 0x0C, 0x0C); //Reset CDR

writeRegister(0x0A, 0x00, 0xC); //Release CDR reset

writeRegister(0xFF, 0x04, 0x07); //Select Channel Registers

writeRegister(0x09, 0x20, 0x20);//Enable over-ride

writeRegister(0x1C, 0x80, 0xE0);//OUT0 Retimed Data

writeRegister(0x1C, 0x00, 0x0C);

Read_HEO_VEO();

readRegister(0x02);

With these settings, I get an output signal:

HEO-52dec (81% open)

VEO-152dec (475mV)

Reg 0x02 - 11: CDR locked.

In Pin16 “LOCK”- High.

The device is operating normally.

Problems begin when I try to connect a 30-meter wire from the camera to LMH0318. The program code does not change. The following parameters are changed:

Reg 0x02 - 00: CDR not locked.

In Pin16 “LOCK”- Low.

The video signal is not received at the OUT0. But the device recognizes that the input signal has arrived.

LOS_INT_N –Low.

I tried different settings Boost, but not help.

What could be the problem? Thank you!

over 5 years ago

0 Nasser Mohammadi over 5 years ago

TI__Guru* 93670 points

Hi Sergey,

I believe you had mentioned when using 20cm differential cable on IN0 - from the output of the HD-SDI camera - this worked fine. However, you are having problem when using 30m cable. Suggestion is to increase reg 0x03 value. At 750MHz - nyquist for 1.5Gbps - device has about 10-12dB of gain with maximum settings. This means as long as your cable has about 12dB of loss at 750MHz this should work. It is possible that your 30m cable is too lossy - loss is more than 12dB at 750MHz - and device cannot handle this loss. If this is the case, you may consider using LMH0397 on both output of the camera and on the receiving side. This would provide much more flexibility and would enable higher data rates such as 3G SDI as well.

Regards,, Nasser

0 Sergey Avdeev over 5 years ago in reply to Nasser Mohammadi

Prodigy 60 points

I am using a coaxial cable with an impedance of 75 ohms. At the input to the IN0, I convert the signal into a differential pair of 100 Ohms using a RF transformer.
I tried different combinations of register 0x03, but this did not give positive results.
It seems to me that the LMH0318 is blocking the signal when I connect a 30 meter wire.

If you exclude the LMH0318 from the circuit (camera-- wire 30 m --video capture), then the video signal reaches the receiver, the image appears.
For what reason can a chip block a signal at the input?

0 Nasser Mohammadi over 5 years ago in reply to Sergey Avdeev

TI__Guru* 93670 points

Hi Sergey,

Thanks for clarification. Given 20cm works and we are having problem with 30m cable, i believe the same root cause or issue is still valid. That is we do not have enough gain at the nyquist frequency and thus LMH0318 does not have enough eye opening to lock to the incoming signal and enable the PLL to lock. Also, transformer may play a role here as well since it could be reducing the signal bandwidth. With short cable like 20cm LMH0318 can handle this but with longer cable or insertion loss it is out of its capability.

Do we know the loss of your 30m cable at about 750MHz?

Regards,, nasser

0 Sergey Avdeev over 5 years ago in reply to Nasser Mohammadi

Prodigy 60 points

Hi Nasser,

To test the device, I used a cable with an attenuation of 85dB per 100 meters at a frequency of 800 MHz. I changed the cable to a cable with an attenuation coefficient of 34 dB per 100 meters at a frequency of 800 MHz, then the CDR started to work. When configuring the register 0x03 (EQ_Boost) 0xFF received values HEO-47dec and VEO-120dec. This result completely satisfies me.

Thanks for the help.

Regards, Sergey

0 Nasser Mohammadi over 5 years ago in reply to Sergey Avdeev

TI__Guru* 93670 points

Hi Sergey,

Thanks for the update.

We were expecting 10 to 12dB at 30m and your new cable seems to be in line with this estimate.

Regards,, Nasser

Interface

Interface forum

LMH0318: Problem with CDR