Looking for Space-grade TIA

So the MAX3658 is an OC-X type digital transimission type device intended for serial data reception - is that what you really want, or a more analog type transimpedance device? 

Hello Michael,

I need to drive high speed photodiode, up to 1Gbps bitrate.

Best regards

Dmitry

Hey Dmitry,

We don't have a dedicated TIA in our space portfolio, you don't see very many of these in commercial either. You can configure a standard op amp, or in this case a fully differential amplifier as a transimpedance amplifier though. This blog post shows the basic connection: https://e2e.ti.com/blogs_/b/analogwire/archive/2013/10/25/how-to-build-a-tia-circuit-using-fully-differential-amplifiers

The best space part to use for this is going to be the LMH5401-SP.

Best,

Zak

Hello Zak,

Nice to see you.

Our meeting in Germany last January: as I remember you had some plans on TIA. Let me know of you need NDA.

LMH5401-SP looks overkill for my application.

Best regards

Dmitry Tcholovsky

Also, I did check out that 2013 e2e by Xavier - the THS4520 circuit looks wrong - that 1uF on the non-signal side is going to mess this - king of looks like it works, 

Makes more sense that it is matched to the diode side, this is what the curve looks like in that e2e so I suspect the circuit was changed to this to run that plot, 

And the LMH5401 might not be overkill 

1. Source capacitance?

2. desired gain?

3. can you map your 1Gbit to a bandwidth requirement

Again is really sound like you are doing a optical digital comm requirement - the OCx amplifiers are transimpedance + comparator to diff out devices - digital output in essence. 

Hello Michael,

Probably you are right about application :)

Capacitance: 1-3pF

Gain: my photodiode provides up to 5microA and I need ~250mV output signal.

With Maxim 580MHz Bandwidth I have about 650mbps.That is OK now.  Looking for HiRel solution.

Best regards

Dmitry

Ok, that is a 50kohm gain assuming say 4pF max with FDA parasitic, the noise gain zero is at about 800kHz. To get 600Mhz closed loop Butterworth that will require a GBP of 4.5e14Mhz. 

If I am doing this right, you might need two stages where something like the OPA855 front end at 1kohm transimpedance gain - but then that next gain of 50V/V with say 1GHz is not going to be one stage either. 

So again, maybe this is really a much lower analog signal out of the transimpedance stage and then a comparator to take it to that differential 250mV digital output. 

I'm open for different options. I'm looking for HiRel TIA and based on that chips around will be chosen.

Best regards

Dmitry

If you are constrained to space grade, just apps solutions won't get there - probably should be talking to the HiRel group in Dallas. 

1. Do they have a fast single in to diff out comparator

2. Are there plans to make the OPA855 space grade?

But here is a starting point with the LMH5401 - using the commercial GBP, I can only get this much gain to provide around 600MHz GBP - I had to guess on the differential input C, 1pF, not specified and this sim suggests it is a little higher than that, this is only showing about 678ohm gain, so your 0 to 5uA will only be 3.3mV output. 

here is what a +/-2.5uA at 100Mhz looks like 

going to be tough noise wise, SNR to 1GHz is only 14dB

total output noise, which I think is RMS, is pretty high also -eventually would need like a 1GHz sharp filter to even get this good, if so, this would be 1.8mVpp noise 6X crest factor. 

here is that TINA file, 

LMH5401 transimpedance.TSC

Thanks, Zak.

Best regards

Dmitry

Again, the MAX3658 puts out a differential digital signal - that 2nd stage internally is more like a comparator. If you wanted to follow the LMH5401-S with a space grade comparator, look at the AD8561S