Ultra fast triggering?

Thread Starter

Leo Silver

Joined Apr 27, 2016
46
Hi everyone, I need to create a triggering ttl pulse out of an event (led light) with a ps (if possible) up to maybe 10ns delay of the event. Does anyone have an idea of how might this be possible to implement in a circuit? Any material or guidance would be much appreciated!
 

RichardO

Joined May 4, 2013
2,270
I think the hard part will be to detect the light quickly. How bright is the light? How fast does the light turn on?

By "ps" do you mean picosecond? Even 10 ns of delay is going to be challenging. My first guess is that you will need an avalanche diode to detect the pulse of light. If that does not work then we are talking about using a photomultplier tube. :eek:

A TTL output is going to limit how fast you can respond unless the signal in embedded and processed in a gate array.
 

crutschow

Joined Mar 14, 2008
34,281
As RichardO noted, getting a fast enough detector is the hardest part.
Linear Technology and ohers do make analog comparators with a few ns response time, so that part should be doable.
 

RichardO

Joined May 4, 2013
2,270
As RichardO noted, getting a fast enough detector is the hardest part.
Linear Technology and ohers do make analog comparators with a few ns response time, so that part should be doable.
Unless the TS really did mean picosecond. :eek:

I have been looking at sub-nanosecond PECL voltage comparators. Analog Devices and Maxim make them. Look at ADCMP561, ADCMP580 and MAX9601.
Under 500ps delay if there is hefty enough input overdrive. :D
 

Janis59

Joined Aug 21, 2017
1,834
The 74AS series provides the ca 1 ns delay, best scenario. For example simple logics 74AS02D at Avnet/Farnell.
Or triggers SN74LVC1G80 (2,5 ns) and SN74AUP1674 (5ns) / Farnell.
Fast comparators are much, but damn expensive. On 7 USD edge are balansing just AD9687 (2,5 ns) and ADCMP604 (1,6 ns) / Farnell.
Cheaper (~3 USD) are TLV3501 (4,5 ns) and LT1719 (4,5) and Max9011 (5ns) and LT6752 (3,5ns) /Farnell.
There are even 170 picosec comparators, however cost over 20 USD per piece.
 

Janis59

Joined Aug 21, 2017
1,834
RE:""but if I want to use it, I can't buy it in Taiwan.""

How to read this?? Firstly, the 95% of all our Planet electronic components are produced of course at Taiwan. Secondly, Octopart.com gives a links to practically all sellers on the World. I cant believe if they says NO to parcel to Taiwan if they says Yes to parcels to India, Zambia, Finland, Russia and Australia operating from USA. Its not a strategic missile detail having (probably) some regional restrictions.
 

ScottWang

Joined Aug 23, 2012
7,397
@Janis59,
What I mean was that I can't buy it from the local stores and that is my way to get the parts, but the local stores almost no such special parts.

If I want to pay some more money then yes I can get the parts from the online shop of International companies , because the quantity is too less, so I didn't use that way to buy, maybe I will do it if I need more parts or I have to use it with some special design in the future.
 

Janis59

Joined Aug 21, 2017
1,834
Aaaah! Okay!
At my land we have only six small component shops what rarely has something useful at the place, so anything must be ordered anyway. I had no even idea at head that shop may keep the needed details on the shelve - it then must cost than the gold consequently.

However, let try the Octopart.com, they are not selling, they only give a tracking links to each part pathway out of the plant gates. So, there You see which client has bought how much, for how much price, and full contact data to him. So probably You may find a rather near seller in Your region what was completely unknown for You beforehand.

And about the quantities - not all firms practice the lowest amount limits, and not all components are bounded with those limits. Try the smaller firms like Farnell, Lemona, Elfa, Ebay and many more companies which has no limits at all, or limit only total sum of order (ca 20 USD or so).
 

DickCappels

Joined Aug 21, 2008
10,153

Janis59

Joined Aug 21, 2017
1,834
@Janis59,
You were suggested a component "SN74AUP1674 (5ns) / Farnell", do you have any info about it?
Just get into Farnell.com, choose Your region, press into search field the name of component You are looking for, allow window to show the parameters, and when choose the tradeoff between price, quantity and goodness, then press the button "show datasheet". The same order in thye Octopart.com. Just push the name, say comparator, then push - show rise time (or something else) and pick up the elements for which push the button - show datasheet. Yes, there are in both some rare components without of datasheets, indeed rare. Then copy and paste tyhe element name into mamma-mia-GOOGLE-boogle and add the word `datasheet`.
Your case is SN74AUP1674 + datasheet pdf in Google searc. But if You are in the Farnell, then happens readdressing to https://www.avnet.com/shop/SearchDi...-LV-FARNELL-ZERO-CROSS-SELL&avnSearchType=all
what gives a datasheet relocation at https://4donline.ihs.com/images/VipMasterIC/IC/TXII/TXII-S-A0003379244/TXII-S-A0003379339-1.pdf
Then see that minimum quantity for all them is 250+ what is not well, so try up the Octopart luck. It answers that it have no info about such, but there is replacement namely SN74LVC2G74
https://octopart.com/search?q=sn74lvc2g74&start=0
in 26 different places on Planet. One of it is
https://octopart.com/search?q=sn74lvc2g74&start=0 and what is strangest that price for one piece is just half with 21 cent (at Rochester electronics)
Farnell in the same case shows the best deal for alone piece the 40.5 eurocents, so the octopart in this case works better.
 

Janis59

Joined Aug 21, 2017
1,834
1) RE: DickCappels, ""One is to use a photomultiplier tube""
You are right about nanosec pulses, however me everyday working with FEMs know well how painful is that techique, from buying the tube up to balancing the resistor matrix, adjusting the photon counting regime, acquiring the good power supply, make all turning bench and grinding jobs for enclosure etc etc. If I had not a full shelf of cheap but very old russian FEYs, our institute budget would crash about that 4-digit price sets. Never I can reccomend the FEM for any kitchen size projects.

2) ""The other is to use an optically triggered avalanche device""
Thanks for the link, nice idea, but who will drill the hole into transistor to push the lightquide in? Such was done about 1950 not nowadays, when there are enough data transfer IC for range betw 150-500 mbit/sec what corresponds to 1 GHz. For example, if You may read the EGAS conference materials, you may find in year ago on my surname the article on how to link nanosecond scale strobe (pulse) to be sync between two subsystems at hundred kilowatt nanosecond laser EMI-impacted region. There I used pair of HFBR1528 (or 1533) as radiating the strobe, and HFBR2523 (or 28) for receiving the nanosecond trigger pulse. Circuit was prost, stable, and gave a fronts below the 1 ns. The price at Elfa was few Eur for each. Note that probably the best thing is there that devices has standard connector for glass-fibre so "no drilling" police and consequently the mechanical stability may be is the best of outcomes.
 
Top