Looking for advice building this quadrupole driver

[ikrase]

So, as you may have heard, I'm trying to make a DIY mass spec. I'm about 80% decided on using a Quadrupole Mass Filter.


For those who don't know, a quadrupole mass filter involves a set of metal electrodes that get driven with MF AC.

The voltage is what's important, not the power, and there are also DC offsets.

The two electrode pairs get driven at V_fa + V_dc + V_rf *sin(2pi*f*t) and V_fa - V_dc - V_rf*sin(2pi*f*t) relative to ground. These electrodes are rods in the range of 20 cm long and 3 mm apart, and are connected via coax cable, so you can count on at least a hundred pF capacitance. V_fa is going to be roughly 100 volts and both V_dc and V_rf need to linearly vary from near-zero to a couple of hundred volts. The frequency is somewhere in the 500 kHz to 3 MHz range, probably 2 MHz should be the target. Usually frequency is fixed (though benefits from being adjustable) while the voltages sweep.


My intent was to build something based on this paper: https://web.stanford.edu/group/Zarelab/publinks/798.pdf

However, I need to figure out how to do this. I'm experienced with hacking equipment in general, but not great when it comes to analog electronics of any complexity.

The first issue is that the MAX038 that generates the sine wave has been discontinued like, apparently, every other MAXIM chip that was useful for any purpose. My thought is to replace it with one of the Analog Devices DDS chips such as the AD9833, which would also fix one of the big flaws of this power supply by making it crystal controlled. I might need an op-amp to level-shift the output into the double-ended range needed for the next stage.

The second issue is the use of an Apex PA19 power amp. This is discontinued, and the Apex IC which I think is its successor costs over 200 USD. I'd be interested in advice as to whether there is a better option for this part. Normal op-amp plus large discrete transistor drive stage?

Likewise, am I correct about just "shrinking" the transformer to provide suitably lower inductance for a higher frequency system?

I notice they never said the value of the capacitor dividers used for the RF feedback detector. Intuitively it seems like these would probably be somewhere in the 100 pF and 100 nF range. A bit of a tight squeeze between loading the RF output too much and being so small that wiring exceeds capacitor capacitance.

I would also be glad of info on 1. component types and 2. whether its a good idea to wire-wrap this.

[dmills]

Looks to me like a transformer wound on a 43 mix ferrite ring driven by a pair of IRF510 or so from a push pull supply, build it dead bug style on a bit of copper clad (Much better then wire wrap for anything involving RF power).

I would wind the transform with bifiliar windings for both primary and secondary, but that is just me.
If you want an integrated amplifier, the ADSL line drivers might get it done, typically good for a few watts at those frequencies if you can keep them cool.

There is a ham amp design using the IRF510 on a 24V rail kicking around that would probably be a good starting point for something discrete transistor based.

Initially I would just use a 4MHz oscillator and 74HC74 to produce the quadrature outputs @ 2MHz, why make it harder then it needs to be?

Regards, Dan.

[David Hess]

Those requirements are very similar to the requirements for the horizontal CRT amplifier of an analog oscilloscope.  They usually had a bandwidth no higher than a 1 to 3 megahertz and drive up to about 100 volts into 100s of picofarads.

I would probably go with either a bootstrapped low voltage operational amplifier where it drives the output directly or a cascode design with the power supply pins driving a high voltage output stage through cascode transistors like that shown below.  The capacitive load will require attention to frequency compensation but is otherwise not a problem.

The example below is comparable to the complexity of a horizontal CRT amplifier but maybe I can suggest something better with more information.  Does the output need to be differential?  What are the requirements for precision?  If your budget is high, then there may be suitable integrated or hybrid parts.

Point to point wiring could certainly be used or Manhattan style construction.  Wire wrapping is not compatible with the round leads used for discrete components but might otherwise be possible with care to layout.

[kony]

CRT scope deflection plates drivers design is solid advice, when I had to design very similar drive stage, I derived it from what is published in ~ 1960s Tektonix technical manual about oscilloscopes which you can find freely on web. It's basically just discrete class B amp with cascoded output. If your frequency sweep range is limited, consider resonant output transformer. Apex modules are way less fuss, but expensive.

[ikrase]

Thank you everybody for your suggestions! I skimmed some literature from TeK about their deflection amp drivers today. Looks like a potential possibility -- and also notable for the fact that they do arbitrary waveforms (i.e. bandwidth rather higher than the sine wave used in this application).


Here's how somebody else did this, also with a transformer: http://www.rapp-instruments.de/Radioaktivitaet/div/quadropol/quadro.htm


@kony: What kind of very similar drive stage did you design -- what app? I looked up some Tektronix CRT drive schematics. Of course, those all require that you have high voltage DC supplies and have a harder time going above maybe 400 volts.

By a "resononant output transformer" you mean one that is self-resonant, not one like in the linked paper where there's a capacitor to make it resonant, right?

I note that the authors of the paper I linked described the "scope deflection" drive strategy as an expensive one -- is it the higher voltages (600-800+) that make it that way, or something else?

I presume that to provide the (sometimes rather large) DC offset voltages for this type of amplifier, I would just add them all up using op-amps before the input stage (after generating the sine wave)


@David Hess: 100 volts is probably not enough. Also, i'm a bit worried -- my research suggests that CRT plates are usually no more than a few tens of pF, while my system will probably be over 200.

My budget is not high -- but I am willing to spend at least 150 dollars of BOM cost on this thing, more if there is a high chance of success or the parts are usefully re-usable.

The output does indeed need to be differential -- but it's not just differential -- its a differential sine wave, plus a differential DC offset, plus a common-mode offset. All three of these voltages need to be pretty precise (1%? I don't even know) although what matters is that they are accurately known (feedback?) than that the gain is perfectly controlled. And the sine and offset need to linearly scale from zero to max.

Doesn't one normally put round-pinned parts on sockets or stub pins? IDK, I can definitely see how wire-wrap is likely not suitable for the HF AC portions.


@dmills: Can you provide me a resource that advises on designing such a transformer? And would it be resonant (or otherwise tuned) with a capacitor, self-resonant, or just non-resonant direct driving? Why a ferrite transformer rather than the paper's aircore one?

Also, I am afraid I'm not equal to understanding what you are proposing I do with the 74HC74? Would the flipflops be used to create a 2-bit sine wave approximation (sorta like the "modified sine" from cheap AC power inverters)? Or is it to create sine waves in quadrature? (that is not used in this application).





- If I do end up using a transformer, does anybody have any comments on the power measurement / feedback system used by the paper I linked?

[kony]

Sorry, did that for my previous employer - I cannot disclose exact useacase and design details for obvious reasons, but suffice to say it was analytical chemistry instrumentation, but not mass spectroscopy.

Additional capacitor (even adjustable) is pefrectly fine to tune the secondary, just be sure not to produce too high Q of the tuned circuit, otherwise your useable output frequency sweep range will be too narrow. Sense winding for amplitude feedback works, but the coupling is not perfect, however as starting point it should suffice for you. The galvanic insulation via trasformer gives you advantage of easy DC offset control on the output and also simplifies the drive stage, as it no longer needs to operate on the full high voltage swing, but rather into low impedance output.
Having resonant output load also lends itself nicely for class-C operation amplifier, as the harmonics of the primary switching are partially filtered - whenever this is acceptable or not for quadrupole I don't know. Playing it safe with linear mode operation would be my safest bet.

Your budget rules out any of the readymade Apex hybrid modules.

[David Hess]

By a "resononant output transformer" you mean one that is self-resonant, not one like in the linked paper where there's a capacitor to make it resonant, right?

The transformer is usually part of an LC circuit with the inductance and capacitance determining the resonate frequency.  The resonate Royer converters used to drive cold cathode lamps for LCD displays are an example.  Some oscilloscopes also used these to generate their high CRT voltages with the Tektronix 7603 being a good example of a push-pull design.  Most oscilloscopes that had these used a single-ended design which requires a transformer with a higher DC saturation current.  Secondary side feedback closely controlled the output amplitude although this is not the only way.

I presume that to provide the (sometimes rather large) DC offset voltages for this type of amplifier, I would just add them all up using op-amps before the input stage (after generating the sine wave)

That works but it adds a requirement that the output amplifier have good DC stability which is not always the case.  Enclosing the output amplifier within the feedback loop of an operational amplifier completely solves this.

@kony: What kind of very similar drive stage did you design -- what app? I looked up some Tektronix CRT drive schematics. Of course, those all require that you have high voltage DC supplies and have a harder time going above maybe 400 volts.

I note that the authors of the paper I linked described the "scope deflection" drive strategy as an expensive one -- is it the higher voltages (600-800+) that make it that way, or something else?

There is nothing to preclude extending the operating voltage of the typical horizontal CRT amplifier up to 300 or even 600 volts.  The limit is the breakdown voltage of the output transistors which is chiefly limited by the PNP bipolars or p-channel MOSFETs if used which are not available above about 600 volts.  If you need a higher voltage than this, then either the PNPs and p-channel transistors need to be cascoded to divide the output voltage or an all NPNs or n-channel MOSFETs need to be used.

Of course a high voltage DC supply is also required.

Older oscilloscope CRTs had lower deflection sensitivity so required higher deflection drive voltages.  A major factor in higher deflection bandwidth was increasing deflection sensitivity so that lower voltage but higher bandwidth devices could be used in the CRT amplifier.

To me the deflection amplifier route seems easier but I think that is just a matter of familiarity.

@David Hess: 100 volts is probably not enough. Also, i'm a bit worried -- my research suggests that CRT plates are usually no more than a few tens of pF, while my system will probably be over 200.

The usual methods for frequency compensating a feedback amplifier which is required to drive a capacitive load apply.  Solid deflection plates were commonly used up to 100 MHz so 200 picofarads should be relatively easy to handle at 1 MHz.  Testing of the frequency compensation and performance can be done at low voltages.

Also consider the requirements for testing of the design.  Do you have an oscilloscope with some x100 high voltage probes?

My budget is not high -- but I am willing to spend at least 150 dollars of BOM cost on this thing, more if there is a high chance of success or the parts are usefully re-usable.

The parts for a discrete design are not all that expensive but the design itself may be tricky.

The output does indeed need to be differential -- but it's not just differential -- its a differential sine wave, plus a differential DC offset, plus a common-mode offset. All three of these voltages need to be pretty precise (1%? I don't even know) although what matters is that they are accurately known (feedback?) than that the gain is perfectly controlled. And the sine and offset need to linearly scale from zero to max.

Horizontal deflection amplifiers were designed for high linearity and achieving this now is trivial by enclosing the amplifier within the feedback loop of an operational amplifier.

Doesn't one normally put round-pinned parts on sockets or stub pins? IDK, I can definitely see how wire-wrap is likely not suitable for the HF AC portions.

Your frequency requirements are not *that* high but I would use point-to-point wiring before wire-wrap construction.

- If I do end up using a transformer, does anybody have any comments on the power measurement / feedback system used by the paper I linked?

It uses an fixed frequency square wave adjustable oscillator to drive a resonate transformer with DC bias applied to the output windings in series.  Amplitude feedback is via another winding on the transformer.

[T3sl4co1l]

If frequency adjustment via variable inductor/capacitor is acceptable, this gets much easier.  200Vrms on 100pF at 3MHz is 75 VAR.  With a Q of 20, you would only need 3.7W drive power, under 10W total if a class AB amplifier is used (and that's not really necessary, if a slightly distorted waveform (filtered square wave) is acceptable).  Probably a Q over 100 would be typical.

To drive this, you would use an amplifier with variable output amplitude (either a class D amp with variable pulse width or variable supply voltage, or a class AB amp with variable signal level), and an adjustable LC matching network.  The network both transforms the impedance to a suitable range (the amplifier requires a load of such-and-such impedance), and causes the capacitance to resonate at the desired frequency.  A voltage or current sensor provides feedback to the amplifier, allowing it to oscillate (for the linear amplifier, a slow, nonlinear ballast is necessary, classically a light bulb, but an analog multiplier could also be used; for the switching amplifier, a comparator is all that's needed).  Alternately, a PLL is used, which oscillates only within the desired frequency range, and tracks output phase (this is preferred for a wide amplitude range).

The coupling transformer design should be made carefully, to maintain balanced operation.  Typically, a single ended primary winding would be placed, in a single layer, followed by a slitted cylinder foil shield, and then the secondary windings, bifilar, single layer, connected in series to obtain a CT winding.  This is easiest on a cylindrical bobbin (various core shapes: E, U, P, etc.), but a toroid can be used.

The transformer should probably be in parallel with the quadrupole, to preserve that balance.  This requires a pretty high impedance transformer (100pF at 3MHz is ~500 ohms, so Zsec >> 500 ohms is needed), which is somewhat inconvenient at the high frequency here.  On the upside, that's the magnetizing impedance; the characteristic impedance can be quite low, indeed should be, so that it is dominant capacitive, and acts in parallel with the electrodes. 

On the primary side, you attach whatever matching impedance is needed.  You will probably add additional capacitance in parallel to lower the frequency.  A series matching inductor (thus making it an L match network) has a step-up feature which is valuable, though not so much with a transformer in use here.  Alternately, a parallel inductor can be used to null out reactance, without providing matching.  Note that a series inductor is still required for the class D amplifier (as its harmonics will not see a friendly impedance otherwise!), and may be desirable for the linear amplifier case.

Whatever the case, the tuning inductor should be adjustable as well, so as to keep Zo = sqrt(L/C) more or less constant, while varying L*C to set frequency.  It won't actually be constant, because loss in the system will depend on frequency, but you can assist this by swamping system losses with an explicit resistor (which is probably a good idea, to set Q to a lower, more stable value).

Now, you might note the transformer has inductance, and that could be useful.  But it's nonlinear (depends on core), and lossy (indeed, it may be the primary source of loss in the circuit).  I would prefer using a transformer that's high enough impedance not to care (i.e., a transformer as such, as opposed to coupled inductors), and adding components around it to get the desired response.

You could just as well use an air cored (or well gapped, e.g., ferrite rod core) transformer, wound in much the same way but made larger to achieve the required inductance.  It wouldn't be easy to adjust, though.  Best way is probably an adjustable core, which might be ferrite to increase the inductance, or copper to reduce the inductance.  I worry that this won't achieve enough adjustable range, or that the winding will be inconveniently large for the frequency range.  (The adjustable range is about 2-4x for either method, or twice that if both are used alternately.)

Tim

[ikrase]

I wish to give my greatest thanks to everybody who has offered advice here.

However, I'm afraid that I am somewhat out of my depth. Still, I've now got a lot of leads to research.