Psuedo PIN Diodes

[Graeme Jury]

Hello Group,

I have been looking into the possibility of relay switching the filters using PIN diodes and have attached a summary of some tests that I have done to measure the reverse recovery time of diodes. I have presented the apparatus used which is very simple and can be quickly lashed up. Other factors like reverse bias capacitance, necessary forward bias, cost and IMD need to be addressed yet but indicators are good so far and of course we already know of some successful filter implementations and to some extent I am reinventing the wheel but I feel it is worth expending some time and effort the get the best diode possible hence the testing. The size advantage and probably the cost can't be ignored.

Claudio IN3OTD currently DK1CG has made some preliminary IMD tests at 1.8 MHz and I expect when he is a little less busy and gets them finalized will publish to this list. Also a set of reverse capacitance tests need to be made and again published to this list where a chart of diodes can be made up. A selection needs to be considered as receive diodes have different requirements from transmit diodes as does 5 watt Tx from 100 watt Tx.

73, Graeme ZL2APV

[Steve Haynal]

Hi Graeme,

Thanks for the document! It is very instructive.

73,

Steve

KF7O

[Graeme Jury]

Hi Steve and group,

Have just run trr test on a BAR64-04 PIN diode and blew my theory about the slope of the space charge time as the PIN diode is very slopey (is that a word?) and as can be seen is much shorter than the rectifier diodes. It is to spec which is 1.55 uSec but I will have it somewhat extended as I am using the same bias current as the rectifier diodes which is higher than the specified current for the data sheet tests and thus extends the carrier lifetime.

I guess the only way to predict the IMD is to actually measure it which means finding a selection of suitable diodes from other requirements e.g.capacitance, voltage and current carrying capacity and package and getting a sample of each and measuring the IMD. It would be good to have a chart of the best diodes for the job available to the whole ham community.

73, Graeme

[in3otd]

Hello,
I have measured several diodes to understand if some of them could be used as a switch for a QRP PA; the 1N4007 seems to work well (as more or less expected), even if its reverse isolation is not that good. For the SMD types,  the "M7" diode, sometimes advertised as "SMD 4007" seems to be quite similar.
The detailed results are on my website, summary at http://www.qsl.net/in3otd//electronics/PIN_diodes/PIN_diodes.html ; every diode has then its page with the usual graphs - without much text due to lack of time, hi. Hope you can find some useful information there.

73 de Claudio, IN3OTD / DK1CG

[Graeme Jury]

Hi Claudio,

Thanks for the brilliant work on these diodes. Quite a few enlightening points came out from your measurements but the big thing is that they show that in TX service cheap rectifier diodes will do the job just fine up to at least 10 watts. The S1FLM diodes look the best from the minimum capacitance point of view and all the figures seem to at least match the 1N4007's as you get above 1.8 Mhz with 50 mA being sufficient bias. In the case of 1.8 MHz it would be trivial to increase the forward bias by switching in another parallel bias resistor and keep the figures good for that band too.

For switching filters it needs to be borne in mind that 2 diodes need to be biased on and at 50 mA each, will consume 100 mA and of course if using 100 mA bias 200 mA will be consumed. On receive the TX filters can be switched off unless variable bandwidth is being generated by using the TX LP filter in conjunction with an RX HP filter.

In terms of the RX filters a different criteria apply and we need minimum coupling across the switched off filters due to the much higher dynamic range of the receiver so minimum capacitance is required. Probably any diode will do if we apply the criteria that forward bias is >= the current through the filter. We are very unlikely to encounter signals > +30 so biasing for this gives currents in the 10 - 20 mA range and maybe a look at 1N4148's in this service would be worth it. If not genuine PIN diodes at this power level are pretty cheap compared to relays aand if Trr is a bit short we can make up for this with a bit more than usual forward bias.

The M7 diode does not match the predicted capacitance from the data sheets being much lower (which is good) but I wonder if different batches and manufacturers would produce the higher capacitance on the data sheet? Maybe these diodes will have a large spread.

I read an opinion which might have been from Wes Hayward? that it is likely that 1N4007's have a light intrinsic layer between the PN junction to raise the voltage of the diode which contributes to its PIN like performance and I guess it is possible the other diodes that test well have this same manufacturing technique applied. It certainly seems that the diodes rated for 1 KV+ are better than the low voltage diodes with the same Trr. I wondered why the diodes with a high Trr did not go better than PIN diodes with a lesser Trr and I think it is due to the intrinsic layer carrying majority carriers vs the minority carriers on the PN junction under reverse bias conditions.

I think that thanks to Claudio we now have a sound base on which to make a decision on whether to use diodes or relays for the V2 support board and it would be good to make this decision sooner rather than later or Steve will have an SDR and there will be no PA or filters to match.

73, Graeme zl2apv

[Steve Haynal]

Hi Graeme and Claudio,

Thanks for the discussion and measurements on psuedo-pin diodes. The 1N4007 is available in surface mount package. The RS-HFIQ uses the MRA4007T3G. Also, in the RS-HFIQ, the RX path does not go through the TX filters switched by the 1N4007s, perhaps for the minimum capacitance requirement Graeme brought up. For RX and prePA TX switching through band pass filters, the RS-HFIQ uses the BAP64-06 which is a nice low cost PIN diode, 2 in a package. 

I don't like the high voltage biasing required by this type of switching. One can use a switching boost converter such as the MIC2619, but the 1.2 MHz switching frequency leads to a potential third harmonic problem in the middle of 80M. Many boost converters run with switching frequencies in the 100 kHz range, but then you may see hash across the spectrum. I haven't found an inexpensive boost converter with high fixed switching frequency or adjustable frequency that I like yet. Rectifying the RF output avoids this.

I do like the space savings diode switching offers. I also like the space savings from the surface mount filters the RS-HFIQ is using for TX, although Jim did say that he sees losses of 1 to 2 dB. I really don't see the HL2 as offering anything out of the ordinary in terms of RF switching, filtering and amplification. What I do see as unique about the HL2 is the relative low price point enabled by the AD9866. Because of this perspective, I am inclined to just use proven low cost options that are already out there for RF switching, filtering and amplification. I am open to taking most of the RS-HFIQ TX switching and filter as is. (It is open source.) The RS-HFIQ schematic is here. The inductors used in the filters are below.

One wrinkle is that the RS-HFIQ TX filters are designed for a 5W PA. One would have to check or substitute new inductors to enable it for 10W. We could also limit the HL2 power to 5W. A power limit <10W also means the sky13774 may work for TX switching, without complicated biasing requirements and at <$0.50 a piece.

 

L1, L18, L19, L20, L32, L34, L35, L36, L37, L38, L39, L40, L41	IND 15UH 400MA 560 MOHM	MURATA	82153C	MOUSER	580-82153C
L12	IND 390NH 210MA 2.2 OHM	VISHAY	IMC0805ERR39J01	MOUSER	70-IMC0805ERR39J01
L13, L21, L30	IND 1UH 370MA 1.8 OHM SMD	ABRACON	AISC-1008-1R0G-T	MOUSER	815-AISC10081R0GT
L14, L23	IND 680NH 400MA 1.47 OHM	ABRACON	AISC-1008-R68G-T	MOUSER	815-AISC1008R68GT
L15, L25	IND 270NH 190MA 2 OHM SMD	MURATA	LQW2BHNR27J03L	MOUSER	81-LQW2BHNR27J03L
L16, L27	IND 180NH 250MA 1 OHM SMD	MURATA	LQW2BHNR18J03L	MOUSER	1-LQW2BHNR18J03L
L17, L29	IND 120NH 320MA 400 MOHM	MURATA	LQW2BHNR12J03L	MOUSER	81-LQW2BHNR12J03L
L2, L7, L8	IND FIXED 1.8UH 1.7A 100 MOHM	WURTH	7440450018	MOUSER	710-7440450018
L22	IND 6.8UH 200MA 7.7 OHM	ABRACON	AISC-1008-6R8J-T	MOUSER	815-AISC-1008-6R8J
L24	IND 3.3UH 450MA 1.7 OHM	ABRACON	AISC-1008F-3R3J-T	MOUSER	815-AISC-1008F-3R3J
L26	IND 1.8UH 600MA 840 MOHM	ABRACON	AISC-1008F-1R8J-T	MOUSER	815-AISC-1008F-1R8J
L28	IND 1.2UH 650MA 800 MOHM	ABRACON	AISC-1008F-1R2G-T	MOUSER	815-AISC1008F1R2GT
L3	IND 560NH 460MA 1.1 OHM	ABRACON	AISC-1210-R56J-T	MOUSER	815-AISC-1210-R56J-T
L31	FERRITE BEAD 1K OHM 500MA 0805	TDK	MMZ2012R102A	MOUSER	810-MMZ2012R102A
L33	IND 100NH 300MA 900 MOHM	BOURNS	CE201210-R10J	MOUSER	652-CE201210-R10J
L4, L5, L10	IND 270NH 630MA 400 MOHM	ABRACON	AISC-1210-R27J-T	MOUSER	815-AISC-1210-R27J-T
L6, L11	IND 180NH 700MA 300 MOHM	ABRACON	AISC-1210-R18J-T	MOUSER	815-AISC-1210-R18J-T
L9	IND 470NH 490MA 800 MOHM	ABRACON	AISC-1210-R47J-T	MOUSER	815-AISC-1210-R47J-T

73,

Steve

KF7O