You missed the point of suggesting valves, which was a response to the incorrect notion of "it's gonna take a lot of stages", which it will not if done with a high gain valve. I am not suggesting that there are no other ways, just pointing out that valves can achieve high gain with relative simplicity.
As this is a DIY site the implication is that the apparatus will be home made....
Yes it is quite possible to do the job with solid state technology but valves would be easier and lend themselves to this kind of application far better than solid state.
Yes it is quite possible to do the job with solid state technology but valves would be easier and lend themselves to this kind of application far better than solid state.
@Jez - I've seriously considered tubes, perhaps I should start a post in the other (tube amplifier) section?
@wakibaki - I've looked into using an OTS RF amp, and have a 50w for that purpose. It drives peak-to-peak about 25v, and will put out full power into 50 ohms.
The problem is that at 50 ohms the drop across the load is not 200v, it's 25v. Finding a toroid transformer rated at 50ohm, 10w+, and 20Mhz is not easy, and even if there was one the secondary inductance is so high it'd be hard to match output impedance without a ridiculous tank anyway.
BTW - I have an amateur radio license for 15 meter so I can legally transmit on 20Mhz (abet w/callsign, which would be a pain).
@wakibaki - I've looked into using an OTS RF amp, and have a 50w for that purpose. It drives peak-to-peak about 25v, and will put out full power into 50 ohms.
The problem is that at 50 ohms the drop across the load is not 200v, it's 25v. Finding a toroid transformer rated at 50ohm, 10w+, and 20Mhz is not easy, and even if there was one the secondary inductance is so high it'd be hard to match output impedance without a ridiculous tank anyway.
BTW - I have an amateur radio license for 15 meter so I can legally transmit on 20Mhz (abet w/callsign, which would be a pain).
Right, if you haven't got the raw power you should easy be able to get the voltage you want with an antenna tuner, it's just an impedance transformer. You can use a primitive one with an exposed coil and a wandering croc clip, or it's a simple match to make with a few (3?) components, I just haven't thought about Smith charts for a while. I'll have a look in the handbook.
Stopping it radiating? Well, hopefully you won't be giving it an antenna.
w
Forget the toroid, it'll be an air-core autotransformer.
Stopping it radiating? Well, hopefully you won't be giving it an antenna.
w
Forget the toroid, it'll be an air-core autotransformer.
Last edited:
L match with complex load cap absorbtion
If you take a 250R dummy load, parallel it with your 30pF device + an additional 30pF cap that's 250R||60pF which transforms to ~55R (close enough to 50) in series with 77pF, which is ~55 -j103. An 800nH inductor in series (with a reactance of 101 ohms at 20MHz) will resonate out the notional series cap, leaving all the power (50W) appearing across the 250R dummy load, for ~111v RMS.
w
If you take a 250R dummy load, parallel it with your 30pF device + an additional 30pF cap that's 250R||60pF which transforms to ~55R (close enough to 50) in series with 77pF, which is ~55 -j103. An 800nH inductor in series (with a reactance of 101 ohms at 20MHz) will resonate out the notional series cap, leaving all the power (50W) appearing across the 250R dummy load, for ~111v RMS.
An externally hosted image should be here but it was not working when we last tested it.
w
Double checked the math and ran some simulations, dead on. 0.6 A in, 200v Pk-Pk out...
So the resistor needs to be rated at least 50w, if not 100w, so something like 5 series 684-MP930-50 each dissipating 10w into a heatsink would work?
I guess 50w isn't that bad (compared to some of the class-A amps on here), but it just seems so inefficient 🙁 cest la vie
I'd like 100W for the load if you go this route.
Dummy loads are normally co-axial, which results in the minimum radiation. For example you could take 10*2500R carbon resistors (for 250R), twist them all together at one end and solder that end to the centre pin of a BNC plug. Then you would separate them out like the petals of a flower and bend all the other unsoldered ends back down to where they could be soldered to the BNC outer, in a 'squirrel cage' kind of arrangement. The more symmetrical and compact the arrangement, the better the fields cancel, the less the parasitics and the less radiation.
These are nice looking non-inductive types you have found. The challenge is to achieve something that physically resembles what I have described using them. You could also look at SMT types, you can get quite high dissipations, I have used them with success in the past. You can get up to the dissipation you want by using series-parallel to multiply up the total number of resistors.
Obviously you can get away with less power if you make a different impedance transformation, I just went with a dumbass brute-force quick fix that gave me easy numbers and meant I could pretty much ignore the Q, which is ultimately what will limit how little power you can get down to.
You're not actually dissipating much (anything ideally) in the capacitance. The smart thing really is to include it in a tank circuit, where the energy just shuffles back and forth from cap to inductor, and the dissipation is in the parasitic resistances. Despite what I said about ferrites, air-core will be feasable. You need to get some devices and do some trial builds.
Couple of ideas... series fed Armstrong oscillators...
...or you could drive them with a 5V 20MHz computer clock module. The transistor will have to withstand 200V+ though. I think either of these will run in class C (no standing current) with sufficient drive.
You can build a circuit with a 30pF cap in it, get it going and then sub in your device.
w
Dummy loads are normally co-axial, which results in the minimum radiation. For example you could take 10*2500R carbon resistors (for 250R), twist them all together at one end and solder that end to the centre pin of a BNC plug. Then you would separate them out like the petals of a flower and bend all the other unsoldered ends back down to where they could be soldered to the BNC outer, in a 'squirrel cage' kind of arrangement. The more symmetrical and compact the arrangement, the better the fields cancel, the less the parasitics and the less radiation.
These are nice looking non-inductive types you have found. The challenge is to achieve something that physically resembles what I have described using them. You could also look at SMT types, you can get quite high dissipations, I have used them with success in the past. You can get up to the dissipation you want by using series-parallel to multiply up the total number of resistors.
Obviously you can get away with less power if you make a different impedance transformation, I just went with a dumbass brute-force quick fix that gave me easy numbers and meant I could pretty much ignore the Q, which is ultimately what will limit how little power you can get down to.
You're not actually dissipating much (anything ideally) in the capacitance. The smart thing really is to include it in a tank circuit, where the energy just shuffles back and forth from cap to inductor, and the dissipation is in the parasitic resistances. Despite what I said about ferrites, air-core will be feasable. You need to get some devices and do some trial builds.
Couple of ideas... series fed Armstrong oscillators...
An externally hosted image should be here but it was not working when we last tested it.
...or you could drive them with a 5V 20MHz computer clock module. The transistor will have to withstand 200V+ though. I think either of these will run in class C (no standing current) with sufficient drive.
You can build a circuit with a 30pF cap in it, get it going and then sub in your device.
w
Can't run an oscillator - at least not an astable one - the 20mhz comes from a splitter that runs both the laser diode (somewhere else) and this crystal so they are in-sync (drift, etc... phase isn't that important).
The triode idea is very interesting - I just have no experience whatsoever with tubes. Seems like a good place to start though, can't think of a simpler active circuit; vs, vin, vout, ground.
In the beginning I had hoped to do something like this:
with the cap between the 50 ohm and ground, but that doesn't seem to slew near fast enough (from SPICE anyway).
The triode idea is very interesting - I just have no experience whatsoever with tubes. Seems like a good place to start though, can't think of a simpler active circuit; vs, vin, vout, ground.
In the beginning I had hoped to do something like this:
with the cap between the 50 ohm and ground, but that doesn't seem to slew near fast enough (from SPICE anyway).
Sounds like a beefed-up oscilloscope deflection driver.
If you want to make this non-resonant, your outputs will have to be cascodes or the input capacitances of the semiconductors will be much higher than those of the load, and will also see full voltage swing or heavy Miller effect. Also, some form of neutralization will be needed due to series inductances of the component pins, as SMD will be a problem to use and cool simply, at least in the output stage.
For the output stage, you will probably need MOSFETs. There is a number of BJTs that will do the frequency, designed for CRT outputs, but you might have trouble with the current and SOA. There are also hybrid and monolithic CRT driver chips that may do what you need, National Semi had a whole line of them that would do in excess of 50MHz. Each has 3 outputs that could be operated in parallel and this way should be able to drive your 30pF load. Just checked their site: have a look at LM2422, using a differential configuration...
If you want to make this non-resonant, your outputs will have to be cascodes or the input capacitances of the semiconductors will be much higher than those of the load, and will also see full voltage swing or heavy Miller effect. Also, some form of neutralization will be needed due to series inductances of the component pins, as SMD will be a problem to use and cool simply, at least in the output stage.
For the output stage, you will probably need MOSFETs. There is a number of BJTs that will do the frequency, designed for CRT outputs, but you might have trouble with the current and SOA. There are also hybrid and monolithic CRT driver chips that may do what you need, National Semi had a whole line of them that would do in excess of 50MHz. Each has 3 outputs that could be operated in parallel and this way should be able to drive your 30pF load. Just checked their site: have a look at LM2422, using a differential configuration...
Last edited:
@lilmzn - That CRT chip is exactly what I need - except it's 130vp-p, 70v short. Looking around (mostly National, am I missing something?) it seems the monolithic CRT drivers all cap out around 200v in. Am I missing something, or is there no easy way to get two -130v channels to 260v without a CT transformer?
I've tried half a dozen cascodes (BJT, JFET, MOSFET) and can't get PSPICE to show any useful gain. It's got to be a biasing issue in the BJT model, or a lack of gate capacitance parameters for the MOSFETs, or something. At 200+V it's not something I feel comfortable just building to see what would happen. Monolithic is nice since they take care of all the hard stuff for you...
@wakibaki - Off the top of your head, is there a decent way to get decent voltage gain out of a triode without it set up with feedback?
I've tried half a dozen cascodes (BJT, JFET, MOSFET) and can't get PSPICE to show any useful gain. It's got to be a biasing issue in the BJT model, or a lack of gate capacitance parameters for the MOSFETs, or something. At 200+V it's not something I feel comfortable just building to see what would happen. Monolithic is nice since they take care of all the hard stuff for you...
@wakibaki - Off the top of your head, is there a decent way to get decent voltage gain out of a triode without it set up with feedback?
What the hell kind of AO crystal are you driving that needs 50W up it?
Even the Q switches on large yags normally make do with half that.
The trick to this is to correctly design the matching network between the amplifier collector/drain load impedance and the rock. Normally the best way is to go to 50 ohms resistive as an intermediate stage as it means that off the shelf coax and connectors can be used between the amp and the final matching network mounted right at the crystal.
Dye and Granburg "Radio Frequency Transistors" has good coverage on the issues when designing RF power stages.
At HF the toroids (if one is even required, often a few turns in air is all that is needed for an inductor), are tiny powdered iron or ferrite affairs.
You should be easily able to do this with an IRF510 or two off a 28V rail, there are good broadband 50 ohm amp designs for this combination in the ham radio literature.
Unless you need linear AM (you probably dont unless this is an AOTF of some form), I would run the amp in class C (much less heat) the matching networks will take out most of the harmonics.
The design of the step up matching network has already been discussed, just rework for 50 ohm input and whatever the actual parameters of your AO are (the 30pf will have a series loss element that takes the place of that resistor).
Regards, Dan.
Even the Q switches on large yags normally make do with half that.
The trick to this is to correctly design the matching network between the amplifier collector/drain load impedance and the rock. Normally the best way is to go to 50 ohms resistive as an intermediate stage as it means that off the shelf coax and connectors can be used between the amp and the final matching network mounted right at the crystal.
Dye and Granburg "Radio Frequency Transistors" has good coverage on the issues when designing RF power stages.
At HF the toroids (if one is even required, often a few turns in air is all that is needed for an inductor), are tiny powdered iron or ferrite affairs.
You should be easily able to do this with an IRF510 or two off a 28V rail, there are good broadband 50 ohm amp designs for this combination in the ham radio literature.
Unless you need linear AM (you probably dont unless this is an AOTF of some form), I would run the amp in class C (much less heat) the matching networks will take out most of the harmonics.
The design of the step up matching network has already been discussed, just rework for 50 ohm input and whatever the actual parameters of your AO are (the 30pf will have a series loss element that takes the place of that resistor).
Regards, Dan.
Dan -
It's a EOM, not AOTF - the key difference in driving is that the effect in the EOM is caused by the electric field across the crystal, where as an AOTF physically oscillates and produces standing acoustical waves. (Functionally, an EOM is designed for a monochromatic source (laser) where as the AOTF is designed to separate or join light of different wavelengths, from what I understand).
The reliance on the E-field (rather than just RF input power) is what produces the need for such high voltages - according to the spec on this device, and using the wavelength laser they are using, a full 200V potential is required for a 180 degree phase shift (past which is indistinguishable from a negative phase shift).
There's no need to pass 50w through the device - in fact it's only rated for 1w - and in the circuit above the 50w is dissipated in the resistor and not the device. The only reason 50w came into the discussion is they have an extra 50w RF broadband amp lying about. I'm sure there are ways to do it with less power, but wakibaki's network will at least get them started...
The hard part is the 200V - the power is totally irrelevant, so long as it's sufficient to charge and discharge the 20-30 (about 26) pF crystal to 200V at 20MHz. It must be low distortion, since any non-sinusoidal input will cause additional (unwanted) sidebands in the output. It seems that this is non-trivial, since it implies the output stage must be high impedance with a 8000 Volt-per-microsecond slew rate. A run-of-the mill RF amp just won't cut it.
I have looked at using a (toroidal) transformer to get a N1:N2=V1:V2 boost, but there are not many 20:1 or even 10:1 transformers with a self-resonance frequency well above the working frequency (200MHz +) commercially available.
Any suggestions?
Jacob
It's a EOM, not AOTF - the key difference in driving is that the effect in the EOM is caused by the electric field across the crystal, where as an AOTF physically oscillates and produces standing acoustical waves. (Functionally, an EOM is designed for a monochromatic source (laser) where as the AOTF is designed to separate or join light of different wavelengths, from what I understand).
The reliance on the E-field (rather than just RF input power) is what produces the need for such high voltages - according to the spec on this device, and using the wavelength laser they are using, a full 200V potential is required for a 180 degree phase shift (past which is indistinguishable from a negative phase shift).
There's no need to pass 50w through the device - in fact it's only rated for 1w - and in the circuit above the 50w is dissipated in the resistor and not the device. The only reason 50w came into the discussion is they have an extra 50w RF broadband amp lying about. I'm sure there are ways to do it with less power, but wakibaki's network will at least get them started...
The hard part is the 200V - the power is totally irrelevant, so long as it's sufficient to charge and discharge the 20-30 (about 26) pF crystal to 200V at 20MHz. It must be low distortion, since any non-sinusoidal input will cause additional (unwanted) sidebands in the output. It seems that this is non-trivial, since it implies the output stage must be high impedance with a 8000 Volt-per-microsecond slew rate. A run-of-the mill RF amp just won't cut it.
I have looked at using a (toroidal) transformer to get a N1:N2=V1:V2 boost, but there are not many 20:1 or even 10:1 transformers with a self-resonance frequency well above the working frequency (200MHz +) commercially available.
Any suggestions?
Jacob
Here's a thread with a load of HV BJTs:-
http://www.diyaudio.com/forums/solid-state/150021-high-voltage-bjts-pre-driver-stages-high-power-amplifier-overview.html
Here's one with 300V, 7 watts & 150MHz:-
http://www.fairchildsemi.com/ds/2S/2SC3503.pdf
We're really concerned with power gain, not voltage gain. RF design always is.
If the design frequency changes you will have to change the circuit unlesss you have a totally brute-force solution @ 50 ohms. This is not a tragedy, just a case of retuning a cap.
The greatest voltage swing can be obtained by having an inductor in the output circuit. The voltage swing in this case can be ~2 * Vcc. (B+)
Normally, other than in broadband circuits, the best gain regardless of device (triode, tetrode, BJT, FET) is obtained with a resonant tank. This is because the tank has a low DC resistance, allowing the (collector, anode) current to be determined by other factors, and a high impedance (ideally infinite) at AC resonance. The output impedance and power of the stage are inferred from the voltage swing and the RMS supply current.
If you run with enough standing current and a clean input you can expect a reasonable sine across the tank, but the harmonics will increase as you move into class C.
Normally, feedback reduces gain unless you mean positive feedback, as in the case of the super-regen, which is a special case.
Bear in mind that if you do create an arrangement that produces a substantial fraction of the RF power of which a device is capable, then if it is not matched into and dissipated in a resistive load there is every likelihood that it will destroy the device.
w
http://www.diyaudio.com/forums/solid-state/150021-high-voltage-bjts-pre-driver-stages-high-power-amplifier-overview.html
Here's one with 300V, 7 watts & 150MHz:-
http://www.fairchildsemi.com/ds/2S/2SC3503.pdf
We're really concerned with power gain, not voltage gain. RF design always is.
If the design frequency changes you will have to change the circuit unlesss you have a totally brute-force solution @ 50 ohms. This is not a tragedy, just a case of retuning a cap.
The greatest voltage swing can be obtained by having an inductor in the output circuit. The voltage swing in this case can be ~2 * Vcc. (B+)
Normally, other than in broadband circuits, the best gain regardless of device (triode, tetrode, BJT, FET) is obtained with a resonant tank. This is because the tank has a low DC resistance, allowing the (collector, anode) current to be determined by other factors, and a high impedance (ideally infinite) at AC resonance. The output impedance and power of the stage are inferred from the voltage swing and the RMS supply current.
If you run with enough standing current and a clean input you can expect a reasonable sine across the tank, but the harmonics will increase as you move into class C.
Normally, feedback reduces gain unless you mean positive feedback, as in the case of the super-regen, which is a special case.
Bear in mind that if you do create an arrangement that produces a substantial fraction of the RF power of which a device is capable, then if it is not matched into and dissipated in a resistive load there is every likelihood that it will destroy the device.
w
Ah, Pockels or Kerr?
So you are trying to phase modulate a laser beam at 200Mhz? Going to be rather noisy unless you have a very good single longitudinal mode rig of way long coherence length (200Mhz is bloody close to the carrier, so the resulting sidebands will have a lot of optical phase noise present as well).
You don't use a broadband transformer for this (Well, you could but why would you?), instead use a tuned matching network designed to be loaded by say 27pf in series with a few ohms. Very little power will be required assuming the cell has a reasonable loss tangent.
As to making it a low distortion sine wave, that is trivial, just use a tuned output stage, job done (A class C amp with a narrow band output stage should have 3rd harmonic at about -30dbC or so, add a trivial lowpass or use a tuned match to the load and you can easily be at -70dbC or so (-90 is not that hard).
I can do that with a IRF510 easily, and if you can design a more efficient matching network then even something like a '4427 or '3866 will get you half a watt or so.
I am betting that as long as that 50W rf amp module is good for 20Mhz, it will do just fine, you will need a matching network and might need a filter, but the RF brick itself sounds entirely acceptable.
I get the feeling that you are used to thinking about low frequency designs where things like broadband transformers are the norm and harmonics are within the passband, RF is different, there you work in a space where narrow band impedance transformations are cheap, think a cap or two and a coil (and most node impedances are complex), resonance is your friend.
One trick for prototyping, do it dead bug style over a solid copper ground plane, an inch of wire in a low Z location (like say an amplifier devices source or emitter) can completely screw you up due to its inductance.
Silver mica and ATC style ceramics rule here, COG dielectrics are acceptable in non critical locations, but if you have not done serious RF before you have a lot to learn.
Regards, Dan.
So you are trying to phase modulate a laser beam at 200Mhz? Going to be rather noisy unless you have a very good single longitudinal mode rig of way long coherence length (200Mhz is bloody close to the carrier, so the resulting sidebands will have a lot of optical phase noise present as well).
You don't use a broadband transformer for this (Well, you could but why would you?), instead use a tuned matching network designed to be loaded by say 27pf in series with a few ohms. Very little power will be required assuming the cell has a reasonable loss tangent.
As to making it a low distortion sine wave, that is trivial, just use a tuned output stage, job done (A class C amp with a narrow band output stage should have 3rd harmonic at about -30dbC or so, add a trivial lowpass or use a tuned match to the load and you can easily be at -70dbC or so (-90 is not that hard).
I can do that with a IRF510 easily, and if you can design a more efficient matching network then even something like a '4427 or '3866 will get you half a watt or so.
I am betting that as long as that 50W rf amp module is good for 20Mhz, it will do just fine, you will need a matching network and might need a filter, but the RF brick itself sounds entirely acceptable.
I get the feeling that you are used to thinking about low frequency designs where things like broadband transformers are the norm and harmonics are within the passband, RF is different, there you work in a space where narrow band impedance transformations are cheap, think a cap or two and a coil (and most node impedances are complex), resonance is your friend.
One trick for prototyping, do it dead bug style over a solid copper ground plane, an inch of wire in a low Z location (like say an amplifier devices source or emitter) can completely screw you up due to its inductance.
Silver mica and ATC style ceramics rule here, COG dielectrics are acceptable in non critical locations, but if you have not done serious RF before you have a lot to learn.
Regards, Dan.
Right - I understand this (blown a few caps up in my day) but the quantum physicists do not. That's why I'd prefer something with a 300k resistor between B+ and the device, so that even if they "turn it up to 11" there's a hard cap on the power output (like that monolithic CRT driver). Hence my tendency towards the "brute force" and away from the tanks: broad-band repeatability is significantly more important than efficiency in this case.
SWR bridge, rigged to reduce the amplifier power if the mismatch becomes excessive, no big deal.
Or just use oversized output devices, and under run them, plenty of CB radio parts that should be suitable.
Work out what B+ you need with 300K in series with 27pf to give you a 200V swing @ 20Mhz, I suspect it is impractical without some sort of matching network in there.
Regards, Dan.
Or just use oversized output devices, and under run them, plenty of CB radio parts that should be suitable.
Work out what B+ you need with 300K in series with 27pf to give you a 200V swing @ 20Mhz, I suspect it is impractical without some sort of matching network in there.
Regards, Dan.
No idea, I could ask
20Mhz, not 200Mhz. But yes, the idea is to phase modulate the beam to produce specific sidebands for their experiment.
RF amp is good for a minimum of 44dBm (25w), with a typical of 47dBm (50w) up to 150Mhz, with an output swing of 26V.
You could say that - I've got almost no experience in RF other than hooking up antennas and working the sets. All the stuff I've built has been below 100kHz, heck my signal generator at home doesn't go above 2M, and I've never had it up there... All of this is
compared to audio stuff. Thanks for bearing with me here...
Ahh yes, but it is phase modulation so you will actually produce a complex set of bessel sidebands very close to the laser output frequency (20Mhz modulation by 180 degrees on a carrier measured in namometers?).
The laser itself will have some phase noise which I fear (unless it is a very good single longitudinal mode job) will swamp your modulation.
Of course you could use a lock in or phase sync. detector to lift the result out of the noise but you do need to account for the optical noise in your experimental noise budget.
Something a little funky with that amp spec, 50W into 50R is 50V RMS (so more like 70 odd peak), so in no case can the thing swing a mere 26V (but it may have a power rail somewhere in the 24 - 28V region, this does not limit the output swing as the impedance at the collector of the output device will be MUCH lower then the impedance at the final output.
Assuming 44dBm into 50 ohms it looks entirely suitable to me.
Look up Smith charts they are helpful when figuring out matching issues.
A lot of the ingrained assumptions that apply to audio stuff go out the window above a few Mhz, you pretty much start thinking about power gain not voltage gain, and pretty much every node becomes a complex impedance.
have fun, and you might want to investigate the ARRL handbook (what you are driving looks a lot like an electrically short aerial) as far as the match is concerned.
Regards, Dan.
The laser itself will have some phase noise which I fear (unless it is a very good single longitudinal mode job) will swamp your modulation.
Of course you could use a lock in or phase sync. detector to lift the result out of the noise but you do need to account for the optical noise in your experimental noise budget.
Something a little funky with that amp spec, 50W into 50R is 50V RMS (so more like 70 odd peak), so in no case can the thing swing a mere 26V (but it may have a power rail somewhere in the 24 - 28V region, this does not limit the output swing as the impedance at the collector of the output device will be MUCH lower then the impedance at the final output.
Assuming 44dBm into 50 ohms it looks entirely suitable to me.
Look up Smith charts they are helpful when figuring out matching issues.
A lot of the ingrained assumptions that apply to audio stuff go out the window above a few Mhz, you pretty much start thinking about power gain not voltage gain, and pretty much every node becomes a complex impedance.
have fun, and you might want to investigate the ARRL handbook (what you are driving looks a lot like an electrically short aerial) as far as the match is concerned.
Regards, Dan.
The input requirements (1v) come from the lock in which sends the same signal to the laser and the EOM, which is on a side path via a beam splitter off the main beam into the cryostat - I'm afraid I don't know much more about the optical side than that. I think they also have some sort of fancy detector behind the EOM that feeds a signal back to the thing driving the laser, assuming the EOM works correctly (my job).
I've had the '99 and '09 ARRL handbooks in front of me since the beginning. I wish they had a chapter on driving physics experiments 😛
Do you have any objections to the circuit spec'ed above... I.e. ~800nH inductor in parallel with a 250ohm, 100w resistor all in series with the EOM and a small (0-30pF) cap for final adjustment? On PSPICE it does indeed produce 200v pk-pk with 50w dissipated across the resistor, and mathematically the complex impedance is, within tune-able error, zero.
I've had the '99 and '09 ARRL handbooks in front of me since the beginning. I wish they had a chapter on driving physics experiments 😛
Do you have any objections to the circuit spec'ed above... I.e. ~800nH inductor in parallel with a 250ohm, 100w resistor all in series with the EOM and a small (0-30pF) cap for final adjustment? On PSPICE it does indeed produce 200v pk-pk with 50w dissipated across the resistor, and mathematically the complex impedance is, within tune-able error, zero.
- Status
- Not open for further replies.