high-speed photodiodes? (~32GHz)

[morgnk]

I'm looking for high-speed SMD or through-hole photodiodes that are fast enough for ~32GHz bandwidth. The usual suppliers (Mouser,Digikey...) only seem to have 110ps rise time photodiodes which are only good for around 9GHz.
There are some suppliers of ready units but I'm just looking for the bare diodes. Are such diodes available for purchase or are they expensive components only sold to corporations?

[mixedgas]

Good Mornin, Krazer...
Long Time No See...


https://youtu.be/kerUbfOQTW0?t=37

Steve

[krazer]

You certainly are not going to find a through-hole 32ghz photodiode (why that is left as an exercise for the reader...), and I have never seen a packaged SMD one for sale above 10G. The good news is that suitable devices do exist and are used in the newer 40G/100G telecom products, the bad news is that these are usually waveguide integrated devices and only work at a narrow band around 1550nm so if you are operating at a different wavelength you are screwed. I bet if you ask nicely newport or thorlabs will sell you a bare chip for only slightly more than the cost of a packaged unit, but you will be on your own for the rf design/wirebonding and optical coupling.

[mixedgas]

I'm looking for high-speed SMD or through-hole photodiodes that are fast enough for ~32GHz bandwidth. The usual suppliers (Mouser,Digikey...) only seem to have 110ps rise time photodiodes which are only good for around 9GHz.
There are some suppliers of ready units but I'm just looking for the bare diodes. Are such diodes available for purchase or are they expensive components only sold to corporations?

I'm looking for high-speed SMD or through-hole photodiodes that are fast enough for ~32GHz bandwidth. The usual suppliers (Mouser,Digikey...) only seem to have 110ps rise time photodiodes which are only good for around 9GHz.
There are some suppliers of ready units but I'm just looking for the bare diodes. Are such diodes available for purchase or are they expensive components only sold to corporations?

Thorlabs has them down to 35 ps as a catalog item, and faster if you buy a "packaged" OEM part.

Consider the laboratory manufacturing process for 50 Ghz and low cost used to be :
~
1. Buy a high speed microwave diode in a plastic pill package.
2. Solder it to a board with properly designed and matched microstrip.
3. Position a nanosecond pulsed laser based drill over the junction, basically a laser micromachining station.
4. Drill pulse by pulse till sufficient light response is just obtained by just barely burning through plastic and partially into the gold metallization.
~
I think due to the chip sizes needed to keep junction size down and obtain good matching, it is a "Industry Only" part. HP/Agilent/Broadcom no longer makes the part number referenced in the "DIY" Review of Scientific Instruments paper.
The fiber optics industry gets them someplace, but that does not mean "Inexpensive"
~
Good Luck...
~


Steve

[krazer]

It's nice to be back, I have a 'real job' now so a lot more time for posting on internet forums than when I was in school

I always did wonder how well it would work to crack open something like a 40G modulator driver that is wirebonded internal construction and hit it with a short laser pulse. I imagine that the metal layers would make it hard to couple into the interesting parts of the amplifier, but when measuring short pulses one rarely is concerned about photodiode quantum efficiency... I never thought of trying to micromachine away the metal layer, I I might give it a try I ever come across a multistage amp with dead input stage but working output stages.

[VDX]

... removing or "drilling" plastic housings is better done with a "plasma-torch" than with a laser

Viktor

[kecked]

Autocorrelator. I get down to 1fs with it.

[BeliaEllet]

Hi...the OPA380 might be excessively delayed for this 30ns beat application. The large number rate is just 80V/us with 100ns settling time.

You need to track down a trans impedance amp about 10x quicker. The speed will be a component of you input impedance. On the off chance that you attempt to utilize too high an input resistor, the wanderer capacitance will restrict the speed of the speaker. TI has a few plan apparatuses that will help you pick a superior part. Regularly, a course of a couple of speakers, each with lower gain will assist you with getting the speed you need.

I would search for a 500MHz GBW, 500V/us slew rate and 10ns settling time enhancer.

pcb assembly in usa

[kecked]

Hi...the OPA380 might be excessively delayed for this 30ns beat application. The large number rate is just 80V/us with 100ns settling time.

You need to track down a trans impedance amp about 10x quicker. The speed will be a component of you input impedance. On the off chance that you attempt to utilize too high an input resistor, the wanderer capacitance will restrict the speed of the speaker. TI has a few plan apparatuses that will help you pick a superior part. Regularly, a course of a couple of speakers, each with lower gain will assist you with getting the speed you need.

I would search for a 500MHz GBW, 500V/us slew rate and 10ns settling time enhancer.

I ended up doing it with an autocorrelator so I observed the diffraction pattern and signal off a SHG crystal. Hooked to OSCSOPE. Worked well. Base pulse was 80MHZ with 20 hz chopper rate.

[femtoman]

[QUOTE = kecked; 356383] J'ai fini par le faire avec un autocorrélateur alors j'ai utilisé le diagramme de diffraction et le signal d'un cristal SHG. Accroché à OSCSOPE. Un bien fonctionné. L'impulsion de base était de 80 MHz avec une fréquence de découpage de 20 Hz. [/ QUOTE]


un petit lien http://www.swissrocketman.fr/mesure-...2c8%2c146.html