ADC schematic for 10.7Mhz IF

Hi,
 I'm triyng to assemble an SDR using my old ICOM receiver as RF front
end.
From it I've captured an IF signal at 10.7Mhz (1.3V).
I need to sampling this signal and then use the samples as input for a
Spartan 3 FPGA.
The difficult part of the project for me is the ADC circuit. I've seen
the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
chip is good (in my case) or maybe there are better ones ?
Do I need to use an OPAMP in front of ADC chip or a transformer ?
Can you suggest me some schematic for the ADC part ?

Thanks
Simone

On Tue, 25 Mar 2008 20:13:54 -0700, simone.rotondo wrote:

> Hi,
>  I'm triyng to assemble an SDR using my old ICOM receiver as RF front
> end.
> From it I've captured an IF signal at 10.7Mhz (1.3V). I need to sampling
> this signal and then use the samples as input for a Spartan 3 FPGA.
> The difficult part of the project for me is the ADC circuit. I've seen
> the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
> chip is good (in my case) or maybe there are better ones ? Do I need to
> use an OPAMP in front of ADC chip or a transformer ? Can you suggest me
> some schematic for the ADC part ?
> 
> Thanks
> Simone

There can be quite an art to keeping a signal clean going into a high-
speed ADC.  There's some charge coupling that happens with the sampling.

Go digging for app notes, and if you find any really good ones that 
aren't Maxim then reward that vendor by buying their parts!

Analog Devices has (IIRC) a 60MSPS, 14-bit ADC.  I know it works very 
well for video, but I couldn't tell you how it works in a radio.

Sample as fast as you can -- the ADC front-end is much broader band than 
the sampling rate on those things, so the noise energy/sample is pretty 
constant.  The faster you sample the lower the noise power/Hz gets, which 
gives you a wider dynamic range in the end.

-- 
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com

Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html

simone.rotondo@gmail.com wrote:
> Hi,
>  I'm triyng to assemble an SDR using my old ICOM receiver as RF front
> end.
> From it I've captured an IF signal at 10.7Mhz (1.3V).
> I need to sampling this signal and then use the samples as input for a
> Spartan 3 FPGA.
> The difficult part of the project for me is the ADC circuit. I've seen
> the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
> chip is good (in my case) or maybe there are better ones ?
> Do I need to use an OPAMP in front of ADC chip or a transformer ?
> Can you suggest me some schematic for the ADC part ?

I'm sure it's no coincidence that your IF has the same frequency as an 
FM receiver for the 88 to 108 MHz range. What bandwidth do you need? You 
woulf like to drive the ADC to about half its range for normal signals, 
so you might want to boost that 1.3# somewhat. Is that RMS, peak to 
peak, or what? If you need gain, a transistor and an FM IF transformer 
might be a good approach.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Jerry Avins wrote:
> simone.rotondo@gmail.com wrote:
>> Hi,
>>  I'm triyng to assemble an SDR using my old ICOM receiver as RF front
>> end.
>> From it I've captured an IF signal at 10.7Mhz (1.3V).
>> I need to sampling this signal and then use the samples as input for a
>> Spartan 3 FPGA.
>> The difficult part of the project for me is the ADC circuit. I've seen
>> the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
>> chip is good (in my case) or maybe there are better ones ?
>> Do I need to use an OPAMP in front of ADC chip or a transformer ?
>> Can you suggest me some schematic for the ADC part ?
> 
> I'm sure it's no coincidence that your IF has the same frequency as an 
> FM receiver for the 88 to 108 MHz range. What bandwidth do you need? You 
> woulf like to drive the ADC to about half its range for normal signals, 
> so you might want to boost that 1.3# somewhat. Is that RMS, peak to 
> peak, or what? If you need gain, a transistor and an FM IF transformer 
> might be a good approach.
> 

... and resist the temptation to split the ground plane. No matter what 
datasheets, professors and so on say. OTOH I wouldn't have as much work 
if nobdoy would split ground planes ;-)

-- 
Regards, Joerg

http://www.analogconsultants.com/

Joerg wrote:
(snip)

>> simone.rotondo@gmail.com wrote:
(snip)

>>> The difficult part of the project for me is the ADC circuit. I've seen
>>> the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
>>> chip is good (in my case) or maybe there are better ones ?
(snip)

> ... and resist the temptation to split the ground plane. No matter what 
> datasheets, professors and so on say. OTOH I wouldn't have as much work 
> if nobdoy would split ground planes ;-)

Since the OP didn't ask about split ground planes ...

Some ICs have separate analog and digital ground pins.  By split
ground plane do you mean a split between the two?

I understand the problem with split planes when signals have to
cross the split, but with proper split analog/digital parts of
the board that shouldn't happen.  Then again, real circuits never
come out the way they are supposed to come out.

Isn't it better for low noise to keep analog and digital signals
and power/ground lines separate?

-- glen

glen herrmannsfeldt wrote:
> Joerg wrote:
> (snip)
> 
>>> simone.rotondo@gmail.com wrote:
> (snip)
> 
>>>> The difficult part of the project for me is the ADC circuit. I've seen
>>>> the datasheet of MAX1421 (12bit 40MSPS) but I'd like to know if this
>>>> chip is good (in my case) or maybe there are better ones ?
> (snip)
> 
>> ... and resist the temptation to split the ground plane. No matter 
>> what datasheets, professors and so on say. OTOH I wouldn't have as 
>> much work if nobdoy would split ground planes ;-)
> 
> Since the OP didn't ask about split ground planes ...
> 

He didn't but IME this is the almost perfect scenario where people head 
off into that direction. I am usually the guy who has to come in as a 
consultant after it didn't work and clean things up.


> Some ICs have separate analog and digital ground pins.  By split
> ground plane do you mean a split between the two?
> 

Yes, that's what I meant.


> I understand the problem with split planes when signals have to
> cross the split, but with proper split analog/digital parts of
> the board that shouldn't happen.  Then again, real circuits never
> come out the way they are supposed to come out.
> 

The last sentence is the reason why it usually doesn't work. Circuits 
don't live in a vacuum bubble. On one side there is a power supply and 
thus a power line, then maybe a USB line, some other comm ports, a 
keyboard that someone with a conductive human body touches, a monitor 
etc. On the other side of the split there is the RF input which 
generally leads to some other place. Plus a user interface, usually. 
Then there is the whole topic of safety ground, you can't just leave one 
side off because it creates noise. Else the UL guys may not hand you the 
coveted blessings.


> Isn't it better for low noise to keep analog and digital signals
> and power/ground lines separate?
> 

Generally not. Mainly because some external connection will form an 
undesired or totally unexpected loop. The you have an efficient antenna 
and the brunt of the signal will be present at, voila, the split. 
Causing lots of headaches. You can combat all that with a box full of 
toroids but that gets old really fast.

There are situations where a star ground or a split can work, for 
example in audio but usually only until some strong RF source shows up 
in the neighborhood. Where is usually doesn't work is in the RF world.

There are situations in my field of work where a split is mandatory for 
safety reasons. For example when a part of a circuit gets connected to 
stuff that resides close to or inside someone's heart during a medical 
procedure.

-- 
Regards, Joerg

http://www.analogconsultants.com/

Thanks for your advices.
I found this schematic on the net:

http://oscar.dcarr.org/ssrp/hardware/LTC1746/LTC1746sch.png

Do you think is it good for my purpose ?
As you noticed I'm a beginner in electronic engineering, becouse I'm a
programmer, but I think that if I pass this phase of the project I
come on my land (vhdl programming and usb driver) so I hope that with
little help this project is not too ambitious for me.

> I'm sure it's no coincidence that your IF has the same frequency as an
> FM receiver for the 88 to 108 MHz range. What bandwidth do you need?

Yes. It's an FM receiver. BW is 50kHz or 443kHz.

> Is that RMS, peak to peak, or what?

I've read that value on the schematic of my receiver, in the point
where I grab the IF signal (at the emitter of NPN Transistor, Q85 on
the schematic below):

http://www.wentztech.com/radio/PCR1000/files/page9_4.jpg

I think that with a 0.1uF capacitor I can drop DC signal.


Simone

simone.rotondo@gmail.com wrote:
> Thanks for your advices.
> I found this schematic on the net:
> 
> http://oscar.dcarr.org/ssrp/hardware/LTC1746/LTC1746sch.png
> 
> Do you think is it good for my purpose ?


Probably. I don't know this ADC but as I mentioned before I do not like 
the split grounds. If you do that at least make sure it's a clean cut 
and that you have 0805 or 0603 pads across the split every 1/2 inch or 
so, for zero-ohm resistors. Those can spare you a lot of grief.


> As you noticed I'm a beginner in electronic engineering, becouse I'm a
> programmer, but I think that if I pass this phase of the project I
> come on my land (vhdl programming and usb driver) so I hope that with
> little help this project is not too ambitious for me.
> 
>> I'm sure it's no coincidence that your IF has the same frequency as an
>> FM receiver for the 88 to 108 MHz range. What bandwidth do you need?
> 
> Yes. It's an FM receiver. BW is 50kHz or 443kHz.
> 
>> Is that RMS, peak to peak, or what?
> 
> I've read that value on the schematic of my receiver, in the point
> where I grab the IF signal (at the emitter of NPN Transistor, Q85 on
> the schematic below):
> 
> http://www.wentztech.com/radio/PCR1000/files/page9_4.jpg
> 
> I think that with a 0.1uF capacitor I can drop DC signal.
> 

I don't think you can tap off into a 50ohm load there. It'll snuff out 
part of your signal. One solution would be a source or emitter follower 
up front.

-- 
Regards, Joerg

http://www.analogconsultants.com/

>
> I don't think you can tap off into a 50ohm load there. It'll snuff out
> part of your signal. One solution would be a source or emitter follower
> up front.
>
I think the Q85 is a emitter follower yet.
So maybe I can tap off directly from its output without loose of
signal. is it possible ?
(schematic: http://www.wentztech.com/radio/PCR1000/files/page9_4.jpg )

Simone

simone.rotondo@gmail.com wrote:
>> I don't think you can tap off into a 50ohm load there. It'll snuff out
>> part of your signal. One solution would be a source or emitter follower
>> up front.
>>
> I think the Q85 is a emitter follower yet.
> So maybe I can tap off directly from its output without loose of
> signal. is it possible ?
> (schematic: http://www.wentztech.com/radio/PCR1000/files/page9_4.jpg )
> 

Yes. Just don't expect a lot of distortion free amplitude range. The 
picture is too fuzzy to read DC levels. If it's just a few ten mV of 
signal level you might be ok.

-- 
Regards, Joerg

http://www.analogconsultants.com/