Caught out by the Coss - Effect of the output capacitance of a mosfet TX in PI.

[dbanner]

Ferric, your new found high voltage probe is curious to me, is there something special involved in such a probe. I find your observations very intrigueing. It took you long enough!!!

[Ferric Toes]

Ferric, your new found high voltage probe is curious to me, is there something special involved in such a probe. I find your observations very intrigueing. It took you long enough!!!

High voltage probes are a standard accessory for oscilloscopes. x100, 100Megohm impedance, and 1-2.5kV for max applied voltage. I bought one for 12.00 GBP on eBay. Today I found I had one already by Hewlett Packard. Both give identical results.

It took me so long from when? 1982 when I started using mosfets, or 2019 at the beginning of November? I think this investigation has moved quite rapidly since my time on it has been little more that 1 hour per day. I have been looking for a while at high voltage mosfets that don't run into avalanche in the PI circuits used and also have a low Rds resistance. Low Rds resistance results in less losses in the device and much lower operating temperature. Important for what I am working on at the moment. Some of the later offerings I tried did not work due to serious ringing that required a low value of damping restance. Not good. That is where it started and it is still under investigation. Just tried a new device called a Superfet II. 650V, Rds ON is 0.133 milliohms typical at 12A. Works great with the diode and additional drain source capacitor.

Are they expensive? Fairchild FCP150N65F - 1.095 GBP. Vishay IRF740 - 2.00 GBP. Both are one off price.

Eric.

[dbanner]

I meant getting back to what you do best.

[Ferric Toes]

I meant getting back to what you do best.

Aw, Thanks.

Someone left a 0 off the data sheet graph showing capacitance-v-Vds on page 3. The last number on the X axis should be 500V, not 50. Anything over 100V maintained on the drain during the off period is good, but the addition of the diode assures this. Adding a capacitor helps to avoid any significance leakage between pulses.

Eric.

[dbanner]

I was wondering, are there any other suitable diodes besides the MUR460 which can be used here?

Is it just a case of a diode with a fast Trr, or low capacitance (<10pF for MUR460) at high reverse voltage? Does a lower junction capacitance imply faster recovery?

I would like to search for alternative diodes which might be suitable, but unsure of what electrical characteristics (besides the V and I) to look for.

Nice to have some options, otherwise MUR460 will become too famous!

[green]

I was wondering, are there any other suitable diodes besides the MUR460 which can be used here?

Is it just a case of a diode with a fast Trr, or low capacitance (<10pF for MUR460) at high reverse voltage? Does a lower junction capacitance imply faster recovery?

I would like to search for alternative diodes which might be suitable, but unsure of what electrical characteristics (besides the V and I) to look for.

Nice to have some options, otherwise MUR460 will become too famous!

KingJL suggested some diodes in the early replies. A test I posted awhile back with some diodes. Thinking X10 probe across coil. IN4937 has a slightly higher SRF but looks like larger drop in voltage first cycle decay. Someone had posted using MUR460 diode and it looked good so I quite experimenting.

[dbanner]

Here is an interesting set of parameters on this one. Although 200mA is low. I guess I'm going to dig through a pile of circuits boards from some switching power supplies and other stuff to see what gems may be salvaged for some experiments.
I also put the MUR440/460 datasheet.

[KingJL]

I was wondering, are there any other suitable diodes besides the MUR460 which can be used here?

Is it just a case of a diode with a fast Trr, or low capacitance (<10pF for MUR460) at high reverse voltage? Does a lower junction capacitance imply faster recovery?

I would like to search for alternative diodes which might be suitable, but unsure of what electrical characteristics (besides the V and I) to look for.

Nice to have some options, otherwise MUR460 will become too famous!

There are actually quite a few... the only issue is how much voltage do you want. For through hole designs the MUR460 has the highest voltage (600 V), which is why it is so popular. The first diode that I experimented with that exibited this charactaristic was a BYV28-200 Sinterglass diode... but the limitation here (if you veiw it as a limitation) is the "200" V. For SMD apploications the MURS360 (600 V) and the MURS320 (200 V) will work.
I have not had success with extremely low ( ~1p ) capacitance diodes. It appears that there has to be enough capacitance to hold a charge large enough that it will not be dissipated off too easily. All of the successfully used diodes (by me) are high speed or ultra fast recovery diodes and seem to have around 10pf capacitance when reversed biased.

[6666]

HER208

[Ferric Toes]

I thought I would try to determine how much capacitance the coil actually sees when looking back to the drive circuit with the diode and extra drain/source capacitor. I inductively pulsed the 300uH coil plus 1.6m of low capacitance coax. The resulting figure between the peaks of the sine wave generated was 1.47uS. With the coil and cable connected to the Tx circuit with diode and additional 2200pF, the period increased to 1.6uS. With the 2200pF disconnected the period changed only slightly to 1.59uS. The next thing was to take the disconnected coil and cable and see how much capacitance needs to be added to the inductively pulsed coil to bring the 1.47uS period up to 1.6uS. The result was 30pF. Now the question is, where does 30pF come from as it is three times more than expected from the MUR460 with 450V of reverse bias? The input circuitry for the preamp was disconnected so it was not that. One thing I have yet to try is to have two MUR460's in series, which if it is the diodes, should give me 15pF. My x100 scope probe only contributes 5pF.

Eric.

[Ferric Toes]

HER208

Looks good with 6pF at 100V.

[6666]

Looks good with 6pF at 100V.

And the legs are normal thickness.

[dbanner]

I thought I would try to determine how much capacitance the coil actually sees when looking back to the drive circuit with the diode and extra drain/source capacitor. I inductively pulsed the 300uH coil plus 1.6m of low capacitance coax. The resulting figure between the peaks of the sine wave generated was 1.47uS. With the coil and cable connected to the Tx circuit with diode and additional 2200pF, the period increased to 1.6uS. With the 2200pF disconnected the period changed only slightly to 1.59uS. The next thing was to take the disconnected coil and cable and see how much capacitance needs to be added to the inductively pulsed coil to bring the 1.47uS period up to 1.6uS. The result was 30pF. Now the question is, where does 30pF come from as it is three times more than expected from the MUR460 with 450V of reverse bias? The input circuitry for the preamp was disconnected so it was not that. One thing I have yet to try is to have two MUR460's in series, which if it is the diodes, should give me 15pF. My x100 scope probe only contributes 5pF.

Eric.

One thing I remember from watching tutorial video on diodes is that no diode seem to be exactly to published spec, even diodes from same batch appear to have slight variation of parameters. Hence the need for, in certain applications , to match pairs of diodes, or to try different ones of the same.
Anyways, maybe your MUR460 is a little "puffed up", excuse the pun.

[Ferric Toes]

And the legs are normal thickness.

I tried a HER208 diode today and it worked well. It is more convenient with the smaller body and thinner leads too. The odd thing is that that the ringing frequency was identical to that of the MUR460 although the data indicates that the capacitance is lower. There must be another factor at play somewhere. Again I disconnected the input resistor to the Rx preamp and did notice a drop in the period to 1.55uS. The input resistor is 1K and then I have two BAV21 protection diodes; one to the +5V rail and the other to -5V rail. This way they are reverse biased when not conducting on the Tx transients. The capacitance of these should not be significant. The input resistor and protection diodes do form part of the damping circuit, whether on the inverting or non-inverting input of the preamp, so the best result is with a high voltage probe across the coil terminals with the Rx resistor disconnected. That is where I measured the 1.55uS compared to the coil and cable when being inductively pulsed at 1.47uS.

Eric.

[6666]

Glad it worked ok, I have been using them for a while.