Delta Pulse

[porkluvr]

73us is low... no need to use the diode for this case. yes you should redo the sim for 500us, this is maximum PULSE WIDTH for DP.
let be inductance 400 uH and resistanse 680 Ohm.

thanks

Thanks for the insight, Tepco.

I previously was of the mind that because the DP seems more like a relic hunter than a gold machine, an added diode in series with the coil would be superfluous. I no longer hold that opinion because I do plainly see what looks like a faster flyback with the diode in place. But there is more to it than what appears on the surface...

Here is a point I would make. Adding the diode by itself may not be magic, and perhaps only adding the diode is not going to make much difference, but if you can increase the value of the damping resistor and still finish with a slightly faster end to the flyback period, that is something worth investigating.

KT315, I was actually inquiring about your choice for coil winding resistance, not the damper. But I have to admit my question was odd. Still, damping R should be adjusted according to coil parameters, and not the other way around, right? I wanted to have a concrete number for coil resistance to start out with. Damping R would be a variable after selecting coil wire.

Here's my "design" process. I am not intending to belabor any particular issue but only trying to make it easier for somebody to follow my (often derailed or just plain wrong) train of thought:

Using MiscEl's rectangular coil calculator: basing an imaginary coil on a 0.47m x 0.47m form, MiscEl tells me that 14t will get me about 400uH. With 0.4mm cross section wire (~26AWG) the resistance will be ~3.5Ω.

With PW set to 500us (12V supply and 3.5Ω coil resistance) a 500us pulse-width may be too long. LTspice showed a flat-top on the current waveform which I believe represents wasted power. (So, it's "back to the drawing board".) There are no pictures of this shown here.

To eliminate wasted power I could either reduce PW by about 70us~100us, or else I could use a larger gauge wire. Since "virtual" wire is very inexpensive, I will increase wire size.

Next I'll try 0.5mm. MiscEl tells me that the same 14t of wire should exhibit 2.3Ω resistance.

After making the necessary change in the simulation file, it looks like 0.5mm wire (which is only slightly smaller than 24AWG) might work very well for a 400uH coil, with PW set to 500us. As I am comparing one circuit with a diode and one without, I am having to split the difference when figuring out where I want to be... it is not a huge difference in peak coil current when using a diode as compared to not using it, but the difference is measurable.

Now to run the sim and take some pictures. Note that the 40us mark in the pictures represents about 14.5us from end of tx pulse.

After comparing things, it seems apparent to me that using a diode can definitely allow you to use a higher value damping resistor and still sample earlier than without. In retrospect, a series diode will affect the coil current final value and perhaps adding a small resistance in the test circuit on the left side would have equaled things out, making the comparison more valid, but I have no time for that particular exercise.

I want to understand this because my design in progress uses two transmit channels, alternately. A diode in series with each coil is required for the two channels to have some degree of isolation. I hope to see it come together before the end of summer.

Sorry about the long post.

[calebsone]

So I know this is off current topic of the thread but by tune on the 680ohm resistor and the 560ohm do you mean exact to that number or find something close for the coil resistance and meter type. Or is it the number or nothing I'm all most done just needed this last bit.

[Tepco]

Possible diode avalanche mode?

This may be important. Following this discussion, (all sides apparently have something to say or claim) I get one interesting idea do explain discrepancy.



I never tried to add diode in Delta, especially not specified type, but did tried on many other designs, getting results exactly opposite of what is claimed here. Namely, on long pulse detector, something like 100-120uS or more, there is practically no difference, during MOSFET breakdown period diode is forward biased, coil voltage will never exceed this level, so there is no 1500V peak present (but KT well may be right in it's claim?). Some small advantage can be expected in short pulse, high repetition rate detector, having less energy stored per pulse and smaller flyback, below breakdown level. Then diode will isolate MOSFET from circuit faster but improvement is minute. Exactly opposite is claimed here, and one particular diode type recommended, giving me some idea:


Under very long pulse condition, large amount of energy released, diode may enter so called “avalanche mode”, at least this particular type. Normally, computer simulation will fail to predict this. Under this condition, circuit behavior can change slightly, producing that famous 1500V peak and significantly influencing detector behavior. Someone with suitable hardware should provide scope pictures, eventually experiment with other diode types under various conditions. Some modern fast diodes are specifically designed to avoid avalanche, ordinary slow types can reach this mode easily, recommended diode is probably somewhere between. It is less known that ordinary 1N5408 rectifier, otherwise too slow to rectify even at audio frequencies can produce very high power nS pulses in avalanche mode. Experiment with different types can be interesting, at this point ignore capacitance and switching speed, try some ordinary ones too.


If avalanche mode behavior can be verified , along with it's impact on detector performance, this may have impact to overall detector design, replicating similar effect in more controllable fashion.

[kt315]

73us is low... no need to use the diode for this case. yes you should redo the sim for 500us, this is maximum PULSE WIDTH for DP.
let be inductance 400 uH and resistanse 680 Ohm.

thanks

[porkluvr]

how long was PULSE WIDTH in your interpetation? I ask because there has not deep sense to use the diode at range before 150us, but has after, near to 500us.

about 73us. Hmm, so I should redo the sim using a much longer pulse (allowing the coil current to rise higher towards the "max" value - battery voltage divided by resistance in coil and related components)?

OK


edit: KT315, what would be your choice of inductance and resistance for a DP coil?

[kt315]

how long was PULSE WIDTH in your interpetation? I ask because there has not deep sense to use the diode at range before 150us, but has after, near to 500us.

[porkluvr]

Here is a virtual comparison of using a diode and not having one. I see some apparent difference between the two circuits' performance, but how or if that equates to improved target detection can only be determined in the field .

It looks to me like adding the diode will cause flyback to decay much faster in the first few microseconds. If adding a diode works out for builders being able to sample earlier than they could have otherwise, that is a good thing. And if you can use a higher damping resistance than before, that should also allow a stronger signal to be seen at the receiver input.

I don't think the diode would need to have a 1500V breakdown voltage (although for all I know, the STTH1210 might still be the best choice for some reason or another), but it would be a bad idea to use something with a lower breakdown voltage as compared to the zener voltage of whatever MOSFET is being used.

In case it is not obvious, what I did with the progression in the photo was start with the "big" picture, and then zoom in sequentially towards the last few microseconds of the flyback's settling. Note that the 176us point is about 14us after the end of the transmit pulse.

[mohandes]

Maybe something like this: http://www.youtube.com/watch?v=NhF2v35qU2c

[turboaleks]

Maybe something like this: http://www.youtube.com/watch?v=NhF2v35qU2c

[mohandes]

hi
KT say right or my old friend porkluvr say OK ?
whatever i add this diode in my new PCB that i design it.
KT what is the meaning of FFFFFFFFFFFFFFFFFFFFFUUUUUUUUUUUUUUUUUUUUUUUUUUUUU UUUU ?
regards

[kt315]

Wow, you're a trip, KT.

[kt315]

Wow, you're a trip, KT.

thank you guys for attention to DP matter, I so did wait scilicet your attention in this thread all this time you are absent and ignoring, but I do not need in additional PR. continue to answer albeit. you can learn me telling about the impedance of MOS-FER in different parts of timing of 'delta pulse' with the HV fast recovery diode and without it.

[Tepco]

If someone really want to play with high voltage pulses in this, or any other detector, there are two simple ways, I tried them successfully. First is to replace MOSFET with high voltage IGBT transistor, drive requirement is about the same, 1200v device will clamp at some 1.35-1.4KV. Additional benefit is reduced capacitance, depending on transistor used (no need to be 50Amp rated, some modern smaller device in TO220 package). Disadvantage is increased loss, due to 2-3V saturation voltage. Another method is using MOSFET (can be low voltage type) and high voltage bipolar transistor, like used in CRT horizontal scan stages, in “cascode” connection. Works perfectly, flyback can reach 1.5-1.8kV or even more, capacitance is slightly lower. Disadvantage is somewhat higher current consumption, large base current is needed, due to very low current gain of HV bipolar, usually less than 10. In both cases, some precaution is needed, damping and protection resistor, diodes etc, also very short high energy pulses will tend to deteriorate metal film resistors after some time. I don’t think all this is particularly useful in detector like delta, attempt to speed it up this way, but in some other designs may be interesting to experiment. Most obvious “improvement” in delta can be redesign to some other type input amplifier, to get rid of lF357 (too noisy!), not tried.

[porkluvr]

Wow, you're a trip, KT.

[mickstv]

My post was correct.

Please do not post in this thread again, if you plan to continue shouting i.e. "using bold uppercase and stupid emotions" that is just bl00dy rude and stupid.

Hi kt315, can you please refrain from copying my posts and using them in your threads.