|
Re: Pitfalls of measuring components with the NanoVNA
#measurement
Mike,
how do you characterize ferrites and powder iron (rod and toroid) for useableAs Jim pointed out, you wind a few turns on the candidate core, and measure its complex impedance over the frequency range you are interested in, and a little around. But that's only the first step. In that measurement you can see how much inductance and how much resistance you get at any specific frequency. These will scale in proportion to the square of the number of turns. When you use many turns, you will also get strong effects from stray capacitance between all those turns, and you can see the capacitance too if you measure such an inductor. You can then decide in which frequency range you want to use each core, depending on application. For example, if you want to build resonant circuits, lowpass filters and the like, you want to use a core only in the (low) frequency range where it provides nearly pure inductance. You can use the software to get the Q values at different frequencies. You will typically want a Q of 30 at the very least, to get low filter losses, and much higher than that in power applications. Certainly above 100, while the best you can achieve might be something like 300 to 400, using a good core in its best frequency range. On the other hand, if you want to build a transformer, you want a core that provides a high enough impedance, and you can trade quite a bit between this impedance being provided mainly by inductance, resistance, or by a mix of them. For a broadband transformer typically you would look for a material that at the low frequency end provides most impedance by inductance, and use a number of turns that provides an impedance of at least about 3 times the load resistance. Some people prefer 5 times, perfectionists prefer 10 times. It depends on how much inductive reactance your circuit can handle. You also need to watch that the resistance associated to that inductance is low enough to cause only an acceptable amount of loss. At the high end of your frequency range, this same core will typically provide a high, although almost purely resistive impedance. So it has very little effect on the circuit, but still you need to calculate that the loss will be low enough. But you always have to be aware that the NanoVNA measures with very small signals. You can apply those results directly to small-signal circuits, but when you are dealing with large signals you need to do additional measurements, using a high drive level that the NanoVNA cannot provide. This is required because core materials typically are non-linear with drive level. The kind and amount of nonlinearity varies with the material, and is often poorly documented by the manufacturer. While several manufacturers do give loss graphs for their power materials, these graphs often don't extend into the frequency ranges we need. A "PowerVNA" could be used to measure thereactance and resistance at various drive levels, to draw curves for the power behavior of a core. Lacking such an instrument, what I do is applying various drive levels from a ham transmitter, and directly measuring the rate at which the core heats up. Then I calculate the power loss from this, and use it to draw loss curves. The result is that some materials are great at low drive levels but less good at high drive, while others tend to be mediocre in terms of loss at low drive but don't worsen much at higher drive. And all this is highly frequency-dependent. In applications where the transformer or inductor has to carry DC in addition to your signal, it's essential to also evaluate how much flux density will be caused by this DC, and how much that will affect the core's behavior towards the signal. You need to make sure that the core can handle the DC while staying well away from saturation. That requires using a long enough core (in terms of path length), of a material that has low enough permeability. It helps if the material has a high saturation flux density, which you can get from the datasheet, or you can measure it with a power inductor tester. When you need a core for EMI suppression, you want it to provide the highest possible impedance over the frequency range you need to suppress, and it's a bonus if this impedance is mainly resistive. Attached are impedance measurements up to 250MHz for two large toroids with 2 turns on each: A T-200-2 powdered iron toroid, and an FT-240-61 ferrite one. You can see that the powdered iron toroid provides almost all of the impedance in the form of reactance, up to a frequency of at least 200MHz. It's high Q material. But the impedance is rather low, at 20MHz it provides only 27? or so. The bending-up of the reactance curve at VHF probably means that we are nearing the parallel resonant frequency with the stray capacitance, while the resistance bending up is due to increased core losses towards VHF. The graph for the ferrite toroid shows such a high Q zone only up to about 15MHz, but it provides a lot more impedance, around 130? at 20MHz . Above that frequency the core starts providing a quickly increasing amount of resistance. At 250MHz essentially all of the impedance comes from resistance. What does this mean? Well, the #2 powdered iron material is great for making high Q inductors in the HF range, and even into VHF, at somewhat lower Q. But it's a poor transformer material and would be lousy for EMI suppression over the entire range. The #61 ferrite instead can be used to make inductors in the range up to 15MHz or so, and is a good transformer material from a few MHz to the end of the measured range and beyond. In the VHF range it's also a decent EMI suppression material, and towards UHF it's excellent for EMI suppression. That makes the question arise, how can the same material be good for EMI suppression (which requires energy absorption, that is high loss) and for transformers (which require low loss), in the same frequency range? Simple: It's a matter of number of turns versus circuit impedance. For transformer use, you wind enough turns to get an impedance that's very much higher than the circuit impedance, so the resulting loss is low despite the core providing almost a pure resistance. For EMI absorption instead you use so few turns that the choking impedance is similar to the circuit impedance. That will make the core "suck up and destroy" unwanted RF signals. It will turn them into heat. Enough writing for now. For loss measurements at higher drive levels, see And if you need a power inductor tester, you can copy mine: So you can complement the NanoVNA with those tools, to more completely characterize any mystery cores you might have laying around. Manfred |
|
Re: Nano VNA: An Antenna Stethoscope ( pdf file latest edition de k3eu )
It was posted and recorded by AJ3DI back a few messages
This recording (about an hour long) missed the first few minutes, which I thought were essential to compare the Nano VNA to an Antenna Stethoscope and a Smith Chart to an EKG of your heart. But it is what it is 73 Barry k3eui K3euibarry at gmail.com |
|
Re: Bricked NanoVNA-H
Hi Richard,
So now that you have performed the operations I described in the message /g/nanovna-users/message/20945, you have not forgotten to remove the BOOT0-VDD jumper and it will log in white the next time you turn it on. What you wrote next, I don't understand: In the message /g/nanovna-users/message/20957, he writes that he wanted to join VNA-QT and could not. Why would you want to use a program made for another device for this NanoVNA-H ??? Each device starting with NanoVNA uses different firmware and the PC programs created for it do not match. Here you may need a terminal program to check if your device is accessible via the COM port, even though you log in with a white screen. Such a terminal program is Tera Term or PuTTY or if you know another one. If you don¡¯t know, choose from the two above. If you have a problem with the display, you just don't know that the foil connection between the display and the pcb has ever dropped and broken, so a white display is possible. So telemedicine is difficult because there is always someone to tell His story to help, disrupting the successful error delineation. So there we are, there is a chance that there is hardware damage if you still see a white screen and there is no information on the screen. 73, Gyula HA3HZ -- *** If you are not part of the solution, then you are the problem. ( ) *** |
Gyula Molnar
Barry K3EUI
Gregg Messenger VE6WGM