1MHz circuit diodes, which ones?

[raff5184]

Hi all,

I am building two circuits, something quite simple, two diode bridges rectifiers, something like this: https://upload.wikimedia.org/wikipedia/commons/thumb/f/f5/Diode_bridge_alt_2.svg/2000px-Diode_bridge_alt_2.svg.png

The input frequenceis for the two circuits are 700kHz and 1MHz for the other one. What kind of diodes do I need? And which parameter do I have to look at?
I have some SM540 MIC but I don't know in which frequency ranges they work, or up to which frequency

I am working with small power and voltage values (mW, 1V)

Thank you

[danadak]

Basically look at Trr, reverse recovery time, and capacitance.

High speed circuits, much faster than yours, typically will use small
signal Shottky or Galium Arsenide doped diodes.


Regards, Dana.

[rstofer]

You didn't say anything about the load current and that's an important consideration.  For small signal levels, I don't see why a common 1N914 can't be used.  The Reverse Recovery Time is just 4 nS.  The 1N4148 can probably do the job as well.

[raff5184]

You didn't say anything about the load current and that's an important consideration.  For small signal levels, I don't see why a common 1N914 can't be used.  The Reverse Recovery Time is just 4 nS.  The 1N4148 can probably do the job as well.

honestly I just don't know what to use. I have these RS540 diodes but apparently I am not able to retify the signal at higher frequencies

[alsetalokin4017]

You didn't say anything about the load current and that's an important consideration.  For small signal levels, I don't see why a common 1N914 can't be used.  The Reverse Recovery Time is just 4 nS.  The 1N4148 can probably do the job as well.

honestly I just don't know what to use. I have these RS540 diodes but apparently I am not able to retify the signal at higher frequencies

I can't find a diode datasheet for the part numbers you've cited (SM540MIC and RS540). Are you sure you've got the part number right?

Please show the actual schematics you are working with. Yes, the picture you linked above is a FWB, we've got that. It's not enough information to be able to answer your question though.

I agree with the above posters. For small load currents 1n914/1n4148 should work. For larger loads the right-rated Schottky diodes would probably work.

[raff5184]

sorry my bad, I ment: SR 540 MIC

http://www.alldatasheet.com/datasheet-pdf/pdf/191379/WTE/SR540.html

I'll post the schematic asap

[amspire]

1N4148 (1N914) diodes are the kind of diode you should have - you can get a bag of 100 for hardly any money on eBay or Aliexpress.

But at 1V, you are better with Schottky diodes as they have a lower drop and have less stored charge every time they switch. 1N4148 switch on about 0.6-0.7V and so they are not great in a 1V bridge.

If you have any scrap motherboards, you can often salvage 3 pin surface mount chips commonly of the BAT54 family of Schottky diodes. These are 30V diodes and at 10mA, they have a 0.3V forward voltage drop. At 1mA, the drop is about 0.13V. You can of course buy them, but I like free. There are versions with different single/dual diode arrangements and all can be used in a bridged rectifier.

http://www.nxp.com/documents/data_sheet/BAT54_SER.pdf

You can identify them from other 3 pin devices by their low voltage drop. If you have a multimeter on the Diode range, transistor junctions will give a 0.6-0.7V drop, Schottky diodes will give a 0.1-0.3V drop. Mosfets like the 2N7002 (also common on motherboards) will not seem to have a junction at all.

[T3sl4co1l]

Impossible to recommend without:
- Voltage
- Current
- Circuit impedance
- other stats (switching or RF? Power range?)

We could go to great lengths to explain all the many properties that vary between types of diodes, but that probably wouldn't be very useful, and would end up saturating you with otherwise-useless information.

Tim

[rch]

Just to add to the questions to the OP,  what is the purpose of the rectifier?  Power transfer, accurate voltage measurement, demodulation?   And what is the impedance of the following circuit?   It's hard to prioritise characteristics if we don't know what it is for.

[bson]

honestly I just don't know what to use. I have these RS540 diodes but apparently I am not able to retify the signal at higher frequencies

Diodes cost pennies.  Just buy a strip of 25 or 50 of each of a few different ones and try them out.  1MHz is utterly mundane and doesn't require anything special; it's not "higher frequencies".  1N4148 will work just fine at 10x or even 25x this.  You can even evaluate a 1MHz small-signal circuit on a breadboard, no problem at all.

By the way, even 1N400x style rectifier diodes (x = current rating in amp, so 1N4005 is a 5A rectifier diode) are rated for junction capacitance and recovery at 1MHz.

[T3sl4co1l]

Capacitance depends on diffusion of charge carriers, so has a lossy diffusion characteristic at quite high frequencies (100s MHz?), regardless of doping.  (Purpose made varactors are low voltage devices, so have stronger doping and thinner junctions, and are good to much higher frequencies. They're probably also constructed differently, having lower series resistance.)   At modest frequencies, and low enough amplitudes, it truly does look like a pure capacitance.

Typical recovery for a 1N400x series is several microseconds, so one wouldn't be able to be tested with a pulse repetition rate much over 200kHz.

Tim

[raff5184]

thank you for all your replies.
So basically this is the circuit, very simple:
http://www.radio-electronics.com/info/circuits/diode-rectifier/diode-rectifier-full-wave-bridge-01.gif

- the source is a 700kHz or 1Mhz continous wave, with less than 2V peak-to-peak (can be  between 300mV and 2V) coming from a waveform generator or some ultrasonic transducer.
- the load is  1 MOhm

These are the characteristics of my diodes: http://www.alldatasheet.com/datasheet-pdf/pdf/191379/WTE/SR540.html

The question is, should it work? I  am asking to understand if I'm doing something wrong or if it depends on the diodes.


I'll try with other types like 1N4148 (1N914)...

[raff5184]

what is the purpose of the rectifier?  Power transfer, accurate voltage measurement, demodulation?   And what is the impedance of the following circuit?   It's hard to prioritise characteristics if we don't know what it is for.

It is for power transfer, I don't need accuracy

[T3sl4co1l]

Then a smaller diode like 1N4148 or BAT85 should be fine.

Tim

[bson]

1. 1N400x are 1A diodes, x denotes voltage, where the highest voltage variant, 1N4007, can withstand 1kV.
2. No matter how high frequency they are tested at, try to pump it with 1MHz at even half the rated current and 10% the rated voltage, I guarantee it fries.

You're absolutely right, what was I thinking...  Anyway, I'm sure they'd still be fine at smaller voltages, like 9VAC at 1MHz, maybe even 1A.  (Or, hmm, tesla seems to think recovery is too slow... clearly 1MHz power rectification is a really odd application, even low voltage.)  For small signal use they're a bit capacitive though.

Edit: I just tested a 1N4004 and it absolutely SUCKS for small-signal use.  Meh.  It's not usable past a few kHz, at 5kHz it leaks a ton of AC through due to overshoot.  The 1N4148 is perfectly fine at 1MHz.

[amspire]

thank you for all your replies.
So basically this is the circuit, very simple:
http://www.radio-electronics.com/info/circuits/diode-rectifier/diode-rectifier-full-wave-bridge-01.gif

- the source is a 700kHz or 1Mhz continous wave, with less than 2V peak-to-peak (can be  between 300mV and 2V) coming from a waveform generator or some ultrasonic transducer.
- the load is  1 MOhm

These are the characteristics of my diodes: http://www.alldatasheet.com/datasheet-pdf/pdf/191379/WTE/SR540.html

The question is, should it work? I  am asking to understand if I'm doing something wrong or if it depends on the diodes.


I'll try with other types like 1N4148 (1N914)...

The SR540 have too much capacitance. The 1N4148 will be OK, but they will not handle the low voltages well. Remember, there are two diode drops in the diode bridge, and even with a 1M load, there is still about 0.5V across each 1N4148.  Ideally, you want to look for Schottky signal diodes with as low a capacitance as possible.

[bson]

For 2V, Shottkys might be a good choice though due to the lower forward voltages.

[rch]

Did you really mean that the load is 1 megohm?   This circuit is not going to work at 1 MHz with such a large value load, as even the capacitance of a small signal diode is going to be a fraction of this impedance, so the diodes will look more like capacitors then diodes, and conduct more or less equally in both directions.   If you want power, you are only going to get a microWatt at this load.  If you put a capacitor in parallel with the load resistance (lowering the load impedance at 1MHz) you will probably get the peak to peak DC voltage across the load.   If you want to see the full wave rectified output on an oscilloscope as in your circuit diagram, try a load of 1kOhm and signal diodes.

[alsetalokin4017]

Just for fun....

Scoping a commercial FWB (D3SBA60, nominal rating 600V 4A) using a low voltage input and differential output measurement (A-B) at about 60 Hz and about 600 kHz.

Comments? 

[alsetalokin4017]

Same as above but with 1n4148 diodes and 1 megohm load.

[alsetalokin4017]

Now with 1n4148 diodes and 1000 ohm load, with reduced input voltage.

(Note horizontal Math error at 500 ns/div.)

[Kleinstein]

For such a small signal a full wave rectifier bridge is likely not the right circuit, as there is quite some signal lost below 0.5-1.5 V depending on the diodes.

For higher frequencies and not to much performance need one might use the 1N4007 as a low cost PIN diode - so not as a normal diode but as a DC current controlled HF switch / attenuator.

To see how slow a 1N400x is, just put a +-2 V rectangular signal from a 50 Ohms  source to the diode and watch the voltage - this is an easy way to measure and show reverse recovery.

[raff5184]

hi again,

so I got some 1N4148 diodes, applied a 2V pp voltage at 700 kHz in input to the rectifier. Load is 1 kOhm
The result is pretty bad.
If I reduce the input voltage the "rectified signal" it is similar, just smaller 

[amspire]

hi again,

so I got some 1N4148 diodes, applied a 2V pp voltage at 700 kHz in input to the rectifier. Load is 1 kOhm
The result is pretty bad.
If I reduce the input voltage the "rectified signal" it is similar, just smaller 

The shape is exactly as expected. The reason I suggested Schottky diodes is with the much lower turn on voltage, you will get a much better rectified waveform. The 1N4148s are only turning on at the peaks of the input signal. I don't understand why it isn't a clean waveform - it looks like you are using an AM modulated 700KHz.

[raff5184]

I don't know either. The input signal is a very clean sinusoid.
So if I understood correctly there isn't much I can do to see a rectified signal with 1N4148s (uless I work with much higher voltages)?

Schottky BAT54 are only with 3pins? Is there any equivalent with 2 pins?

[T3sl4co1l]

The AC input and DC output cannot BOTH be common to the same ground reference.

You need isolation.

Tim

[alsetalokin4017]

It's deja vu all over again.... 

[alsetalokin4017]

I don't know either. The input signal is a very clean sinusoid.
So if I understood correctly there isn't much I can do to see a rectified signal with 1N4148s (uless I work with much higher voltages)?

Schottky BAT54 are only with 3pins? Is there any equivalent with 2 pins?

It is likely that your Function Generator's output BNC shield is grounded to the FG's chassis which in turn is grounded to the mains, and so back through the mains to your scope probe's ground reference lead. So when you connect one scope probe to the output of the bridge in the "usual" manner (probe tip to bridge + and probe ground clip to bridge -) you are in fact creating a groundloop between the AC input from the FG to the output of the bridge. So you are seeing a distorted signal and only one-half of the bridge output.  To get around this problem you either need to use an expensive isolated differential voltage probe on the bridge output, OR make a differential measurement of the output using two scope channels, no probe ground connections, and math to subtract, as I showed in the scopetraces I posted on the previous page.

If you want your bridge to work well at 2V input then Schottky diodes are probably necessary. 1n5817 for example.

If you are set on using surface mount BAT54 series diodes you can use one BAT54A and one BAT54C, or two BAT54S, to construct your FWB. These contain two diodes each, in common-anode (A) and common-cathode (C)  and series (S) connections. The plain BAT54 (no letter suffix) has three leads also but only contains one diode and the third lead is not connected internally. But they are in tiny SOT-23 packages!

[raff5184]

thanks alsetalokin4017 I'll try with a differential measurement as you showed in other posts.

I saw the schematics of the different BAT54's, now it makes more sense why there are two diodes.
I'll first try with 4 BAT54 (I already ordered them) and than with the other solutions you suggested.
Since the third pin is not connected, can I ignore it?

Sorry guys if these questions are basic but I have no practical experience and don't know where to find this information, which diodes are used, what they are called and so on. I know the theory but apparently it doesn't help. But thanks to you I'm getting a grasp on these topics 

[alsetalokin4017]

No offence intended, but with "no practical experience" I'm wondering how you are going to be working with SOT-23 package devices..... but yes, you can ignore the non-connected pin on the BAT54. Usually it will be soldered to a pad for mounting purposes but not connected electrically to anything.

[raff5184]

no problem alsetalokin, no offense.
I'll learn step by step, read guides that I can find on the Internet, trial and error... I have to start somewhere.

Btw if you know any good PRACTICAL guidebook feel free to suggest it.
Thanks

[alsetalokin4017]

If you are set on using the BAT54 diodes you may want to look at getting some adapters or breakout boards that will allow you to mount your SOT23 devices and use them like standard DIP or SIP packages on your breadboard. There are many manufacturers and vendors of these things. You can solder the SOT23 pins with a fine-tipped iron and very thin solder. It may take some practice...