Tuned LCR circuit for humidifier

[unclejed613]

an LCR calculator i have with LTSpice shows the following calculations:
given:
1.67Mhz
60uh

results:
151.375668pf Cx
629.575166 XC
629.575166 XL


for 100khz you want XC and XL in the same neighborhood as the first driver circuit.
given:
100Khz
1mh

results:
2.53303nf Cx
628.318561 XC
628.318561 XL

Cr in the above circuit is not part of a seies resonance wit Lr, but is a load on the crystal. increase it to 27nf to maintain the capacitance ratio with the crystal.

 

[antknee]

Hi Eric,

I have the same board. I'm fairly sure I measured C6 at 96uF (on the board) and guessed the notation was .001 Exp5 uF. Which seems to add up. I assumed .047 was .047 Exp0 uF which is 47nF and I measured it at 60nF.

My LCR meter is a little temperamental tho, it gives 'sensible' readings in the uF and nF range so i'm never sure which it is. I'm sure you're readings are right!

I tried to fire up the atomiser last night and promptly fried my power adapter. I bought that especially and it took a week to arrive, so fairly annoyed! I've been careful with the board but not the adapter. What do you use to provide AC/DC?

I've also been wondering how to get the water level up to 2cm as is the minimum. The provided black water reservoir is 1cm deep.

 

[antknee]

Hi unclejed,

That went a bit above my head. I will have to come back to that!

Thanks again.

 

[unclejed613]

if you want to reconfigure the board for 100khz, you want to keep the impedances (actually reactances) of the inductor and the cap from the collector to ground about the same at the new frequency. going to 1mH does that for the coil, and going to 27nf does that for the cap. that maintains about 630 ohms reactance for the coil, and about 60 ohms reactance for the cap. this maintains the voltage and current ratios required for the oscillator to function reliably.

 

[antknee]

Hi unclejed,

I see now. That is very clever! I'm going to come back to this and study it.

I'm in a library looking up wireless internet acces because my telephone company are now going to take a month to put broadband on the landline.

Hi eric,

I took the components off the board last night. The markings are correct as you noted above. I've been using a 16V AC supply to get something off the board (as my 36V supply is toast). I'm only getting 100HZ at the atomiser tho. Perhaps it needs the full 36V.

Thanks and regards,

Antknee

 

[erictuve]

Hi Antknee,

I am using a variable 60 VDC power supply connected after the full wave bridge on the PCB. I too have been running it below the 36V since I'm on the bench and don't have the piezo underwater. At about 20V the circuit does oscillate as expected and piezo buzzes. The board has a full wave bridge and filter capacitor so you can run with 36VAC transformed down from line power. A transformer from a discarded TV or other old equipment should work for you in a pinch.

The piezo assembly will need to be connected to a water bath. What I have done in the past is cut a 39mm hole in thick acrylic and then glue together a tank. The four mounting holes around the piezo are for attaching the piezo to the tank so that the orange silicone seal around the lip seals the water [Important!].

I removed parts last night to see what was important for resonance.
The circuit ran with:
C4 shorted - This is DC blocker for base voltage. Low impedance at resonance.
L3 shorted
D5 open - Diode between collector and base. Collector is always higher DC level than base so diode is reverse biased in normal operation (off). Likely it is there for protection.

The circuit did not run with:
C6 open
C5 open

Measured the capacitance of the piezo: 2.25nF = 0.002uF
Diode D5 is an FR10 and 15pF from datasheet.
Removed L1 and measured capacitance of rest of circuit with power leads shorted = 4.16nF

Eric

 

[antknee]

Hi eric,

I have a box containing about a dozen power adapters. I keep them/collect them for times like this. The best I can do is 16V both AC and DC so I'll just have to buy another. I had to laugh when I saw your [Important!] remark because I can be fairly clumsy, hence short circuiting my new 36V supply, I had visions of the atomiser dripping water and me getting a shock as i tried to turn the power off!

I'm going to have a look to see if I have any transformers. I might be able to find one.

Ultimately I have to rapid prototype the bath/container for the transducer/atomiser I'm building. I have RP'ed a holder and gasket for that purpose, i'm awaiting the pzt so can't build my own atomiser yet, the atomiser on this board won't fit my exising RP parts...must check that for sure. So yes i'm into the realm of building a bath for this board out of bits and pieces. I'm probably going to end up using plastic milk cartons and siliconing it together. I really just want to see it working. Doesn't have to look great...and won't!

I had it in mind that C5 C6, L1 and L3 were important for resonance and impedance matching based on earlier posts. I'm a little surprised L3 can be shorted. I've done numerous LC calculations with series and parallel components to try and get the numbers to add up and felt sure it was more important!

You certainly know your humidifiers. Why are you looking into this circuit in particular?

Regards,

Antknee.

 

[antknee]

Hi unclejed,

The penny has dropped! Thanks.

I bought a wireless broadband stick. So i've had a proper chance to read everything through.

L3 and C6 form the resonance circuit and determine the working frequency. (These are also noted as Lr and Cr)

Currently: L3 and C6 = 4uH and 1.5nF = resonance at 1.6MHz

I need to change to: L3 and C6 = 115uH and 27nF = resonance at 100KHz

In order to match reactances L1 should also change to 1mH

So essentially I'm scaling these 3 parts up by 16 (approximately)

The 25Watts i'll have to start off with should mean I can err with the reactances and not have to worry. I'll probably only need 15-20W.

Cheers! More testing tomorrow.

Antknee.

 

[erictuve]

Hi Antknee,

I’ve done a lot of work on this circuit. For research of how it functions there are many good sources. A similar circuit is labeled: “Figure 6. Fundamental Mode Pierce Oscillator”
in a paper titled “Oscillator Design Techniques…” at https://www.electro-tech-online.com/custompdfs/2010/02/saosc.pdf

Read crystal oscillator circuit design at:
https://www.electro-tech-online.com/custompdfs/2010/02/2001SEP06_AMD_ANPDF.pdf

These US patents describe the operation and design of piezo drive circuits very similar to the one under examination.

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

Here’s one that describes a 100Khz circuit!

**broken link removed**

Useful calculator sites:
Series / Parallel Capacitor Calculator
**broken link removed**
**broken link removed**

The capacitance to calculate the resonant frequency is the series combination of these parts:
C6 – 1.5nF
C5 – 22nF
and the piezo (2.25nF) in series with the 0.1uF and both of these in parallel with the diode (15pF) = 2.215nF (effectively the piezo element itself).

The series combination of these = 859.4pF. [C6, C5 and piezo combo]
The inductor L1 = 60uH
The oscillator starts up at the LC tank frequency of 700 kHz. The piezo resonates at 1.67Mhz. It is modeled as a capacitor in parallel with a series combination of capacitor, resistor and inductor. It is positioned between the collector and the base of Q1 providing feedback. Once the circuit begins oscillating the piezo takes over as the crystal oscillator element. This is observed on the scope as a much larger amplitude wave at a 1.67Mhz frequency and the piezo element buzzes (the low frequency envelope is audible). The piezo operates as an inductive element and the circuit provides the capacitance.

Attached are some screen captures of the Q1 collector voltage at start up and after oscillating. Unfortunately, the device ran too long dry and the piezo element died. Now the circuit only resonates at the RC frequency (a learning opportunity…).

The component values are also chosen for their reactance values at the operating frequency of 1.67 MHz so that the transistor biasing is correct for stable oscillator operation. These are:
L1 – 60uH – 630 ohms
L3 – 3.4uH – 35.7 ohms
C6 – 1.5nF – 63.5 ohms
C5 – 22nF – 4.3 ohms

The patent writers of the listed Japanese companies seem to propagate an error in calling this a Colpitts oscillator circuit. The Colpitts design has one side of the resonator connected to ground. This circuit is a Pierce oscillator – (see Oscillator Design Techniques…).

A poem:

Ran the humidifier circuit dry
Turned the voltage up too high
Scope trace showed the how and why
I watched a piezo crystal die!

…and now I need another 1,666,666 Hz piezo resonator. Any good sources?

Eric

 

[antknee]

Hi eric,

Ah yes, you are an expert in this field. The LC frequency and reactance make sense to me now. I had the circuit working last night and it resonates at 1.72MHz for me, my C5 is actually 47nF, it is marked and measures at that cap, perhaps that explains the difference. I was in an electronics store yesterday and happened across a fogger this one:-

**broken link removed**

The fogger I bought certainly produced a lot of atomisation, more than I was expecting, it also looked quite cool, impressive even. So I won't now need to make my own bath for this board, I know what they are supposed to do.

Farnell are my preferred supplier of electronic kit. They don't sell many atomisers but I did find this one at 1.65MHz:

PROWAVE|M165D25|ATOMISER, ULTRASONIC, LIQUID, | Farnell United Kingdom

I bought half a dozen atomisers from Steminc a while ago, collectively they weren't too expensive but individually they would be.

STEMINC - STEINER & MARTINS, INC. - PIEZO ELECTRIC CERAMIC TRANSDUCERS

Alternatively if you private message me a forwarding address I will send you one.

Regards,

Antknee.

 

[erictuve]

Hi Antknee,

Thanks for the great links and very kind offer! I just now ordered a few replacement piezo elements and another nebulizer from Steiner & Martins. They sell the exact unit I am working on: SMUTK1660RS111 I bought it from them on ebay 4~5 years ago.

Sounds like your circuit uses a 1.7Mhz piezo element.

Here is a patent that describes the **broken link removed**. These are popular today. Here is an interesting application:
Fountain Project

Eric

 

[antknee]

Hi Eric,

Steiner and Martins are a good company for piezo parts. I got my board from them too. At the same time I also got the 2.5MHz atomiser and board (SMUT2500RS112). I mapped the schematic for that circuit at the same time as the one we've been looking at and it's a lot simpler, so I'm not sure why I started with the more difficult one! I will adapt both circuits and see which I prefer.

I can't actually make the final circuit I want because until i've built the atomiser I don't know what capacitance and reactance it is going to have. What I will do though is to get the adapted circuits working with steminc's SMISTW3118K16 atomiser, it resonates at 130KHz, picture below. The capacitance is known at 2800pF and so is the impedance at ≤50Ω. I'll have to worry about the impedance later. I know there is a difference between reactance and impedance but I'm not sure what it is yet.

https://www.steminc.com/piezo/images/SMISTW3118K16.jpg

I'm going to be writing up the theory of piezo drive circuits to make sure I don't forget or lose the information.

The fountain project was quite cool. Water features are a draw for people, I'm sure that will go on sale at some point.

Do you have an ultimate goal for your piezo work?

Regards,

Antknee.

 

[erictuve]

Antknee,

Reactance is the opposition to current flow caused by capacitors and inductors.
Impedance is the overall opposition to current flow. This includes reactance from capacitors and inductors as well as resistances. That's all. So for a capacitor or inductor, the two terms are interchangeable. A piezo has capacitance, inductance as well as resistance and so a piezo has an impedance (not just reactance).

Your 130Khz circuit: the capacitance will be 2.8nF from the spec. At 130Khz (resonance) the piezo impedance is less than or equal to 50 ohms. So you can calculate the other parts. The difference of this design from the one I have been looking at is that the piezo is less than or equal to 2 ohms at resonance (1.67Mhz). What is the impedance at resonance of your 1.7Mhz piezo? The 130Khz element is also lower power and lower voltage.

Good idea to document your efforts.
Why am I looking at this circuit?
I have a humidification application but don't want to use a $30. [US] nebulizer. It can be built for less.

Eric

 

[antknee]

Hi eric,

I did write up my notes last night. A lot of it is fairly mundane, for example, map schematic, take prictures to zoom in on values and remember diodes have capacitance, its important I do this because I'm not experienced with electronics and I could come back to this in a few months and have forgotten some of it. I referenced the notes with appendix, mostly the documents you linked to (thanks), I am up to appendix 15 now. Today I need to finish writing those, I got up to tank circuits and their purpose, I need to check know how to calculate this for a given circuit. Then I need to understand impedance a little better, R15 fits into the equation somewhere. Finally I need to decide which capacitors and inductors to buy, with regard to getting the circuit working at 133KHz. I'll also do the same for SMUT2500RS112.

I haven't thought about the power requirements in depth, it is interesting you brought that up. I assumed I'd be able to just put fewer volts into either of these circuits.

I just tried to measure the impedance of the 1.7MHz atomiser but I couldn't get a reading, I don't have an impedance analyser and the multimeter and LCR meter I have couldn't tell me anything. It is listed at ≤2 ohms on the datasheet. I keep meaning to make a post asking for the best and cheapest oscilloscope and impedance analyser but haven't got around to it yet.

I look forward to seeing your humidifier application in the shops, I assume you'll be taking out a patent? I'm sure it will look very impressive too!

Regards,

Antknee

 

[unclejed613]

ok, a quick primer on impedance vs reactance. reactance is the opposition to
ac current flow of a capacitor or inductor. there is also a phase shift associated with reactance. voltage across a capacitor lags the current, and voltage across an inductor leads the current. if you have a cap and inductor in series, the reactances add, but not directly because of the phase shifts (which also have their own addition or subtraction). this makes for a rather complex impedance. there is a special case where the phase shifts are 180 degrees apart and the total impedance is (theoretically) zero. the special case is resonance. however at resonance, the voltage that develops across each component depends on it's reactance at that frequency. so with a series resonant circuit, you might have just about zero volts across it, but assuming a 100 ohm reactance (reactances of the two components are equal at resonance) and a 1 amp source, this would give you 100V!! scross each component, even though the total voltage is near zero. so the total impedance is zero, but the reactances of the coil and the cap in this example is each 100 ohms.

with a piezo, the cap and inductor are actually not real quantities, but electrical analogs of the physical properties of the crystal. the total impedance can be measured from the outside, but the actual values of the "series resonant" circuit is an unknown quantity. the values that can be known for sure are the series resistance and the capacitor in parallel with the resonant circuit. the resonant frequency can easily be determined.


if you are wanting to build an impedance analyzer, look at my blog in the "ac ohmmeter" posts. unfortunately i haven't got a working model yet that can do over 100khz. the one i've built so far goes up to about 75khz. actually using a TL2072 should get it working at up to about 200-300khz

 

[antknee]

Hi Unclejed,

I pretty much get impedance now, thanks. I have done some calculations.

The original circuit at 1.67MHz has the impedances below from Eric, see attached schematic.



L1 – 60uH – 630 ohms
L3 – 3.4uH – 35.7 ohms
C6 – 1.5nF – 63.5 ohms
C5 – 22nF – 4.3 ohms

L3 and C6 determine the resonant frequency - 1.67MHz
C4 is DC blocking
L1 and C5, C6, D5 and the piezo determine the tank circuit - 700KHz

My calcs for 130KHz

L1 - 1mH -628 ohms
L3 - 60uH - 38 ohms
C6 - 27nF - 61 ohms
C5 - 350nF - 4.5 ohms
C4 - 1.6uF - 1 ohm

I've just realised i did these calculations without taking into account that the atomiser I want to use at 130KHz has an impedance of 50 ohms. As opposed to 2 ohms for the 1.67MHz one above.

I will have to have a think about how that will change impedances. I hope this shifts down the tank frequency because I did some rough calcs and they came pretty close to 130KHz. So many things to remember...

Your impedance analyser looks good. I do remember complex numbers from many years ago, so the concept of impedance isn't lost on me. Translating that into electronics is difficult. My brain is swimming just thinking about it!

Thanks for your input.

Regards,

Antknee.

 

[unclejed613]

it's a very simple to use device. if you put 1V rms in and read the result with an rms voltmeter. it's a direct reading AC impedance meter (well almost, as you need to multiply the result by the value of Rset). for higher frequencies that may be beyond the capability of the multimeter, use an oscope, and set to either 1Vpeak or 1Vp-p, and read the result using the same terms.

 

[antknee]

Hi unclejed,

I posted some thoughts on the impedance analyser on your blog. I think I'll have to buy an oscilloscope!

Thanks,

Antknee.

 

[antknee]

I just worked out the tank circuit frequency for the components i'd use to start up the 130KHz pierce resonator, it comes in at around 16KHz.

Working out how I can best match impedances with the new load of 50 ohms is proving harder. I understand I need the load and source to have the same impedance to best transfer power, but its tricky. It would be handy if I could increase R15 from 2 ohms to 50 ohms and it all worked fine, but that would be too simple. I've been looking at the theory of impedance matching and I come up against smith charts, it doesn't help that I don't understand in depth how the pierce oscillator works. So I don't know which components/impedances I'm supposed to be matching. I will get back to the theory.

Regards,

Antknee

 

[unclejed613]

i think you're trying to "over-think" this one... it doesn't need to be perfect to work. the difference in crystal impedance isn't going to be that much of a problem. just build it and try it. the worst that can happen is that the oscillator stalls. all you need to do in that event is increase the positive feedback. i think you'll find that it works just fine. all it has to do is "thump" the crystal a few times, and the crystal then takes over as the resonant feedback element. it's not that "well oiled" a machine. the power supply is just this side of brute force rectification (as evidenced by the 60hz (or 50hz) "chirping" the thing makes in operation). it really is not a delicately balanced oscillator. it's an electronic monkey wrench rather than a fine pair of tweezers (for lack of a better illustration).