10MHZ clock distribution

• 10MHZ clock distribution
  Posted by mj21 on 07 Aug, 2013 22:03
• i am trying to get a 10MHZ sign wave  distributed to four outputs. i have  few op amps at my disposal which i would like to use, these include:   MC33282,AD841,OPA627,MC34080,LM118.
  now looking at the data sheet for these op amps i can see that almost all can handle frequency of at least 10MHZ, so i did a search on Google and used a non inverting op amp circuit as an buffer to get the signal out. but this does not work. connecting the scope to the out put of the op amp after 50 ohm resistor to ground and 50 ohm resistor to the out put give me a sign wave which does not look right but the counter on my scope says that it is 10MHZ. the wave is very small about 100mv Ptp and it does not look like the original signal.
  
  the 10MHZ sign wave is generated via a oven controlled crystal oscillator. it uses  12 volts but the sign wave measures at 4.8v ptp.
  
  can some one please help let me understand what i am doing wrong?
• #1 Reply
  Posted by c4757p on 07 Aug, 2013 22:13
• Misspelling "sine"?
  
  Look at the datasheets. I didn't check all of them, but the ones I did check all didn't want to drive 100 ohms. There are amplifiers that are specifically meant for driving 50 ohm systems - look for those.
  
  Of course, you're just taking a sine wave and shoving it to outputs, this is an easy task for a few discrete transistors even at 10 MHz.
• #2 Reply
  Posted by mj21 on 07 Aug, 2013 22:21
• lol. its late in England. tnx for for the info, do you suggest i use transistors to design a buffer? ideally i want to feed the sine waves to 50 ohm loads like scopes and counters and alike, would using discrete components not effect that?
• #3 Reply
  Posted by c4757p on 07 Aug, 2013 22:28
• No effect that I can think of. They might introduce a small amount of distortion, but I doubt it's a very low-distortion sine wave anyway. They certainly won't shift the frequency.
• #4 Reply
  Posted by mj21 on 07 Aug, 2013 22:35
• thank you. do you have any suggestion to what transistor or a example circuit i can check out?
  
  also if i use transistors to make the buffer will it be able to drive more than one device or do i need to make separate buffers for each separate device?
  
• #5 Reply
  Posted by c4757p on 07 Aug, 2013 22:45
• If you don't need to boost the amplitude (you'll lose half because it's terminated into 50 ohm at the input and output), it really can be as simple as this. I've used something like this many times around 10 MHz.
  
  Edit: don't know where I got the idea that it's terminated at the input, you never said that...
  
  You do have to be careful about layout, though. Keep things close to eliminate inductance and watch out for capacitance between base and collector. This will drive at least one 50 ohm load unmodified, probably two. I'd do two loads per amplifier.
  
  If you need voltage gain it becomes slightly more complicated, but still not bad.

  ---

  
• #6 Reply
  Posted by plesa on 07 Aug, 2013 22:49
• Did you check http://gerrysweeney.com/build-a-10mhz-rubidium-frequency-standard-and-signal-distribution-amp-for-my-lab/
• #7 Reply
  Posted by mj21 on 07 Aug, 2013 22:54
• thank you so much (c4757p) for your prompt response.  i will breadboard this and test it out, lol after drinking 1L of diet coke. its almost midnight in London and last time i worked this late i blow up a cap in my face because i powered it wrong way round lol.
  
  
  
• #8 Reply
  Posted by c4757p on 07 Aug, 2013 22:58
• Ew!! Diet Coke?
  
  Might not be the most breadboard-friendly circuit if you aren't careful. The capacitance between two rows between the base and collector pins of the transistors may well be enough to keep it from working. Though you can probably do it if you are careful.
• #9 Reply
  Posted by mj21 on 07 Aug, 2013 22:59
• ye pelsa i have. want to do the same thing, got the enclosure and the rest of the stuff. but being me i want to build everything my self with what ever i have at hand.
  gerry done a fine job on the rubidium thing he did, but as he always say he likes to build with what ever he got lying around. lol
• #10 Reply
  Posted by mj21 on 07 Aug, 2013 23:03
• lol. i would have had redbull but that stuff is like superman blood, very addictive. i will build the circuit and test it at lower frequency if it worked OK i will do a prototype board and test it with higher frequencies, just out of interest what is max frequency this circuit can handle?
• #11 Reply
  Posted by c4757p on 07 Aug, 2013 23:09
• Can't say off the top of my head. I've used it to 20 MHz before it starts needing extra "high frequency" attention. PN2222A in theory has a current GBW product of 300 MHz. Depending on the current gain you could get a usable frequency up to 60-100 MHz, but it would not be easy to do in practice.
  
  Red Bull is worse!
• #12 Reply
  Posted by mj21 on 08 Aug, 2013 00:15
• perfect, i bread boarded the circuit and it works fine, i tried it up to 60MHZ and no problem. now its time to do the pcb. i decided to make a distribution box for up to 20 outputs so i have to put 20 of this amps on the pcb, power it from linear reg as switching will induce noise prob.
  
  tnx again for your help.
• #13 Reply
  Posted by c4757p on 08 Aug, 2013 00:18
• Make sure when you test it you use a small enough input coupling capacitor that you know the amplifier is actually driving the output and you're not just getting the input bleeding through to the output.
  
  Also - I don't know how much output power the signal source has, but if you're worried that loading it could cause distortion, use 21 amps: one to drive the other twenty, rather than hanging all 20 off the one oscillator.
• #14 Reply
  Posted by mj21 on 08 Aug, 2013 00:31
• yepers. sound good. tnx
• #15 Reply
  Posted by c4757p on 08 Aug, 2013 00:39
• And like I said, each can most likely drive two outputs, so test that. Maybe even four, but keep an eye on power dissipation.
• #16 Reply
  Posted by c4757p on 08 Aug, 2013 01:14
• $25 for four outputs is a lot when you need twenty..... It's only 10 MHz!! That's hardly RF, let alone microwave... A PN2222 used properly will hardly blink at that.
  
  The next person to call anything under 40 or 50 MHz "RF" gets a free one-way ride on the TARDIS to 1960. I have tickets!
• #17 Reply
  Posted by vk6zgo on 08 Aug, 2013 01:58
• You mean that bloody big 55kW MF Broadcast Transmitter I used to work on wasn't operating at RF?
  So all those burns weren't RF burns?
  
  I must admit,that I have been known to say that anything under UHF is "damn near DC!"
  
  The OP may be able to pick up a rack full of surplus analog video amps dirt cheap.
  Some of those make it out to 10MHz without any mods,but others may need a bit of fiddling.
• #18 Reply
  Posted by c4757p on 08 Aug, 2013 02:00
• Nope. Sunburn. Must have been working outside.
  
  250pHz*-75kHz = DC
  75kHz - 1MHz = low AC
  1MHz - 50MHz = high AC
  50MHz - 500MHz = low RF
  500MHz - 5GHz = high RF
  5GHz - 100GHz = microwave
  100GHz+ = I reject your reality and substitute my own. It does not have 100GHz.
  
  *Radio waves were discovered in 1887, the longest entire wave period that could have existed in manmade devices in this time has a frequency of about 251pHz.
  
   
• #19 Reply
  Posted by bingo600 on 08 Aug, 2013 16:54
• I'd get one of these
  This box was recommended as nice , and easy to mod to 50 ohm on the T-nuts list.
  
  http://cgi.ebay.de/ws/eBayISAPI.dll?ViewItem&item=161082475549
  
  Change the 75ohm resistors in the outputstage to 50 ohm , and done ....
  
  It has all the needed BNC's , and if i remember correcty. There's even room in the box to mount the OXCO inside.
  
  
  Edit: Ohh i see it's the same that is used by G Sweeney
  /Bingo