[SOLVED] Very fast power Schottky diode

[FvM]

I did Google search for power Schottky diode and looked through many results.

I would be satified with one or two original datasheet citations...

I thought that the body diode in a MOSFET will always conduct when the MOSFET is reverse-biased, so I do not see why I would need to switch it on.

Good point. It actually rectifies without turning it on. But the latter makes the synchronous efficiency and suppresses reverse recovery effects of the body diode.

P.S.: I forgot to mention another point. You should take a look at the IXYS GaAs-Schottky diodes. They offer junction capacitances that are really suited for MHz rectifier applications. Unfortunately, they have considerable forward voltages, so they are mainly good for higher voltage operation.

 

[htg]

I tried Google search again, and I cannot find trr values, except minority carrier lifetime for MBD301:
https://www.onsemi.com/pub_link/Collateral/MBD301-D.PDF
which is 15 ps at very low current, 100 ps at 80 mA.

I looked at IXYS GaAs Schottky diodes. DGSK28-025CS Datasheet: Power GaAs Schottky Diodes-Low Leakage Current - Datasheets, Search, Parameters
I found 18ns trr specification in addition to a statement "fast recovery <= 500 ns".

Regarding MOSFET rectifiers: should I not worry about 0.5V reverse voltage specification, because it will be diode's Vf?

 

[jiripolivka]

Dear htg:

I think you are staying on the phantasy side. Today there is no way to "rectify" VHF in the "kilowatt range". I wanted to help with what is now used at HF- synchronous FET rectifiers. But as you insist there is a "diode" in a FET, I quit.

 

[htg]

Dear htg:

I think you are staying on the fantasy side. Today there is no way to "rectify" VHF in the "kilowatt range". I wanted to help with what is now used at HF- synchronous FET rectifiers. But as you insist there is a "diode" in a FET, I quit.

If there is no better way, I can use thousands of small-signal fast diodes.
The synchronous MOSFET rectifier seems enticing, but if you insist there is no body diode in a MOSFET, you contradict reputable manufacturers. I haven't experimentally verified it, but they all say there is a body diode in a MOSFET
- for examplelook at: http://www.datasheetcatalog.org/datasheet/irf/jantx2n6766.pdf
- they show diodes in the symbol of MOSFET and they specify "peak diode recovery" in V/ns.

 

[FvM]

You should better refer to current IXYS data than to second handsheets from the internet.

http://www.ixysrf.com/pdf/diodes/gs150t_25104.pdf

Congratulations for showing a schottky datasheet with a trr specification. Interestingsly, IXYS prefers in other publications the view, that schottky diodes don't have a trr, see e.g. the above linked datasheet. The problem is however, that the junction capacitance also causes a reverse current peak during switch-off. If you only look at the current, it appears like bipolar trr. If you look at both voltage and current, it's different, particularly involving less switching losses.

 

[htg]

I haven't experimentally investigated the issue of trr. But my problem is rectification at VHF. Are synchronous MOSFETs promising at 200 MHz?

 

[jiripolivka]

Not to my knowledge but the development of switching DC/DC converters looks promising.
Up to now I have not seen any real rectenna with Schottkys or other diodes capable to meet your phantasy wishes. You can certainly use MANY individual diodes but again, such rectenna would barely be able to efficiently rectify a high power density of a planar wave front: moreover, rectenna sections which would be driven to saturation ( and this is what you should expect) will reflect some RF power back and disperse it, so I do not see any efficient solution.
Like in those DC/DC converters, designers dislike the knee voltage in P/N and Schottky junctions, so now they started using synchronous MOSFET rectifiers; such devices when open have ~ 10 mOhms channel resistance, and in reverse (and closed) they stand 60-100 V and possibly more. I see this way for RF power detection as one future solution.
But- as I wrote- I do not see ANY really efficient rectification solution for VHF and higher- the best use of such power is to heat water )

 

[htg]

Can you name just one RF power MOSFET that can stand 60V reverse voltage?
The ones I checked can stand only 0.5V.

 

[jiripolivka]

You can try the same as I can on Google:
the first I found is " www./freescale.com"; they have a wide RF LDMOS selection up to 3 GHz; one for 470-860 MHz. MRF 373ALR1 has 75W, 32 V peak, 60% in linear mode.

Try and see for yourself: I fail to understand your "0.5V limit"

 

[htg]

I cannot get to see the datasheet, but clearly the 32V is the FORWARD voltage.
Look at the datasheet of MRF6V2300 its DS voltage is -0.5 V to +110V.

 

[FvM]

You may want to review the previous explanations about reverse and forward bias in synchronous rectifier operation. Not to trust too much in reading, I try to visualize the basic operation of synchronous rectifier with a (simplified) circuit. For optimal performance, active gate driver circuits are preferred, however.

 

[htg]

So is it rectification by the body diodes of the MOSFETs?

 

[FvM]

Yes, if the MOSFETs are not switched on by a gate voltage, the body diodes still carries the current. But in synchronous operation, the body diode is shorted by the active FET and the diode doesn't conduct at all. So it doesn't contribute to reverse recovery.

 

[htg]

So the MOSFETs conduct in reverse? Looks like I can rectify a lot of power at VHF this way. What surprises me about MOSFETs is that the GS voltage cannot be significantly reversed - for MRF6V2300 it is specified as -0.5V to +10V. But I guess it is not a problem.

 

[FvM]

So the MOSFETs conduct in reverse?

Yes that's a consequence of the conduction mechanism that involves an unipolar channel rather than a junction. The -0.5 V Vgs specification of RF LDMOS transistors seems to indicate an additional diode junction, possibly an overvoltage protection.

Generally, synchronous rectifier operation of regular MOSFET is widely utilized in power electronics, e.g buck converters or H-bridges. I don't know about rectifier applications of RF FETs. and I don't feel able to predict there behaviour. I think, you shouldn't imagine these devices as ideal switches.