Audio Signal Generator with Specific Requirements

[MarkF]

When you get to building a Function Generator with a DDS chip, I can offer the circuit I did with a AD9834 DDS and PIC18F2550.  It has Sine, Triangle, Square, Sine Sweep Up & Down, and PWM.  Frequency range is 1Hz to 15MHz.

That function generator looks good. Is this a kit that you sell?

No.  It's a one-of.  You're welcome to build it.
You might want to make at least one change.  (i.e. A SMD oscillator instead of the through-hole)

Edit:. I will provide the Gerber's and Hex file.
I don't have a BOM, but could wack one together if you really need it.

[gbaddeley]

You’ve hit a couple of common snags. Old designs using 741 (went out of common use 20+ years ago) and old test equipment (uses unobtainable custom values and parts) and wien bridges that really only work with specific bulbs. Try finding more recent designs.

[gbaddeley]

sound-au.com, I built project 86. There are a couple of other oscillators under Test Equipment.

[TimFox]

Back to the G-R 1313A.  If you really want to clone it, you would probably do better using a good operational amplifier, running on +30V or +/-15V power, instead of the discrete circuit used in the original.  With luck and some creativity about the thermal stabilization, you could make it work.

[Frankentronics]

You’ve hit a couple of common snags. Old designs using 741 (went out of common use 20+ years ago) and old test equipment (uses unobtainable custom values and parts) and wien bridges that really only work with specific bulbs.

To be honest, I actually am very attracted to old stuff, perhaps because I'm also (getting) old, LOL. At home I have tube radios, reel to reel players, and even three Tefifon players (forgotten technology).

The 741 designs were just what I was finding on the internet and even if they are old (which I know now) I'm guessing they still should work.

Regarding the bulbs, I have a few different ones and some circuits specified some parameters and I did find a match. Some circuits just said to use any bulb within whatever voltage range (or I forgot whatnot). I would still like to put together a bulb oscillator, at least for fun and learning.

So, basically, even if I'm testing circuits with a 741 or with a bulb I'd still like to put together circuits that work out, even if they never end up in an enclosure. I guess at the end of the day, to have a reliable piece of test equipment, I might have to look into more upgraded stuff.

Back to the G-R 1313A.  If you really want to clone it, you would probably do better using a good operational amplifier, running on +30V or +/-15V power, instead of the discrete circuit used in the original.  With luck and some creativity about the thermal stabilization, you could make it work.

I'd like to clone that circuit and last night I already started compiling a cart on DigiKey. For now I put all the semiconductors in that cart and I was going to compile the caps today.

I would also love to put together an upgraded clone that uses and OP Amp, but my knowledge of electronics is so limited that I would need a lot of guidance on the way. In fact, I wouldn't even know where to start. But let's say the goal would be to build a circuit that produces a natural sine wave (not a converted square wave) and to have the control pot be able to dial in the entire spectrum, without having a separate decade switch. And if it's an upgrade I might want to have a multi turn pot do that job.

I'm thinking, I should probably start a separate thread just on building the General Radio 1313A.

Purchase standard-value C0G/NP0 capacitors lower than the required value and measure them.

I'm not finding anything listed as COG or NPO on DigiKey. I'm sure they are stocking them, but they list their caps as follows:

Aluminum - Polymer Capacitors (10941 items)
Aluminum Electrolytic Capacitors (126208 items)
Capacitor Networks, Arrays (3022 items)
Ceramic Capacitors (741074 items)
Electric Double Layer Capacitors (EDLC), Supercapacitors (1733 items)
Film Capacitors (69732 items)
Mica and PTFE Capacitors (9102 items)
Niobium Oxide Capacitors (476 items)
Silicon Capacitors (267 items)
Tantalum - Polymer Capacitors (13141 items)
Tantalum Capacitors (104071 items)
Thin Film Capacitors (4197 items)
Trimmers, Variable Capacitors (2659 items)

Thanks...

[MarkF]

Look at Digikey Cermanic Capacitors for example.
C0G, NP0 is the capacitor dielectric or temperature coefficient.
You will see most common C0G, N0G, X7R, Y5V, ZSU  for example.

See Alan's video on capacitors (at time 7:45 for ceramic caps)

   

[TimFox]

The capacitors are C-Zero-G or N-P-Zero, not Oh.  The "0" refers to nominal 0 ppm/K temperature co-efficient.  NP0 is the old commercial term, C0G is the more modern term.  These will be found under capacitors-ceramic in most catalogs.  C0G is available in leaded or surface-mount, as "MLCC" monolithic construction.  Small values of NP0 can be found as disc ceramic leaded construction.

[ArthurDent]

One note about trying to clone the GR-1313 oscillator is that it uses a dual section variable air capacitor that may be hard to find unless you disassemble an old AM radio from 40 years ago. I don't see where the capacitor range is specified in the manual and the shape of the blade may have a very unusual shape to give a more linear frequency change per degree rotation.  See photo below.

https://api.utmel.com/Upload/Images/Article/b0cadc01-6e26-4734-8c09-ddfd0da89b86.jpg

[TimFox]

The dual-capacitor plates in the 1313 are, in fact, an interesting shape since the scale is logarithmic (decades per degree of rotation) to cover the full audio range in 180 deg of rotation.
I own one of these and had some trouble getting it to work, since I could only find a manual for the "normal" Wien-bridge (multi-range) unit at the time.  It now works, but it takes an inordinate amount of time to achieve a stable oscillation, after which it seems to work well.  I was intrigued by the circuit used to get the full range without resistor switching.

[Frankentronics]

One note about trying to clone the GR-1313 oscillator is that it uses a dual section variable air capacitor that may be hard to find unless you disassemble an old AM radio from 40 years ago. I don't see where the capacitor range is specified in the manual and the shape of the blade may have a very unusual shape to give a more linear frequency change per degree rotation.

Well, that's probably the most challenging part to obtain. I can see that the schematic says C401A and C401B, which means it's a 2 gang variable air cap. And as you said, the range is not specified anywhere. I guess one way to find out what the range might be is to build the entire circuit, minus the C401 cap, then swap some caps in and out of circuit to see what range is needed to get 10Hz to 50KHz out of the circuit. But then what would be the next step? If those fins are really odd shape then the only way to obtain one might be to build one from scratch. I would actually be able to do that as I have a mini machine shop, but I wouldn't know what the physical specifications of the needed cap would be.

[TimFox]

If you have the tooling to build such a capacitor, and you know the capacitance for a discrete set of frequencies that covers your range, you can fit a curve and do the geometry for the required plate overlap area as a function of rotation angle.  It shouldn’t be more difficult than Kepler deriving his laws of planetary motion...

[Frankentronics]

I think I made some progress on the GR-1313 clone project.

A friend of mine just let em have a tube radio chassis that has a 3 gang tuning capacitor with asymmetrical fins. I didn't remove the capacitor from the chassis yet, but here are some pictures.





Would it be safe to say that even if it's not the proper value capacitor, I should still be able to make that circuit work and get some decent frequency range out of that capacitor? Then if the circuit works out with this cap I can look into figuring out the best value for the tuning cap and work on obtaining (or making) one?

Also, I am thinking of building this circuit, as well.


http://www.seekic.com/circuit_diagram/Signal_Processing/Oscillator_Circuit/20_Hz_TO_200_kHz.html

But I do not have any 2N4340 JFETS. I do have some other JFETS in my bin MPF102, J201 and J112.

What would be the process of figuring out if any of those FETS would be suitable replacements for that circuit? And of so, which one would be the best choice?

Thanks...

[bob91343]

A DIY variable capacitor project has some pitfalls.  The first to come to mind is the means of connecting the rotor (usually to ground) with low impedance and minimal noise.  In commercial units there is a clip that does that job.  Look closely at any unit and find that clip, usually at the end opposite the ball bearing.  High performance units for high grade lab gear use several clips per section.  Sometimes they are made of beryllium copper.

End play can create unstable capacitance.  The material of the plates and frame is important, as its temperature coefficient of size affects the oscillator.  The nonconducting supports also affect capacitance and losses.  One of the rotor plates is usually slotted to allow for fine adjustment of tracking.

The balls in the bearing need to be free of lumps, which could cause bumpy tuning.

The shape of the plates will affect the dial calibration of course.  So that needs to be planned carefully to obtain a linear dial.  This will be compromised by trimmer capacitors but generally not too seriously.

High grade units have an adjustable rear bearing to obtain smoothest operation.

[Frankentronics]

A DIY variable capacitor project has some pitfalls.  The first to come to mind is the means of connecting the rotor (usually to ground) with low impedance and minimal noise.  In commercial units there is a clip that does that job.  Look closely at any unit and find that clip, usually at the end opposite the ball bearing.  High performance units for high grade lab gear use several clips per section.  Sometimes they are made of beryllium copper.

End play can create unstable capacitance.  The material of the plates and frame is important, as its temperature coefficient of size affects the oscillator.  The nonconducting supports also affect capacitance and losses.  One of the rotor plates is usually slotted to allow for fine adjustment of tracking.

The balls in the bearing need to be free of lumps, which could cause bumpy tuning.

The shape of the plates will affect the dial calibration of course.  So that needs to be planned carefully to obtain a linear dial.  This will be compromised by trimmer capacitors but generally not too seriously.

High grade units have an adjustable rear bearing to obtain smoothest operation.

And I thought this would be a relatively simple project for a machinist. Of course, a lot of work, but still doable.

I'm surprised to hear about the beryllium. That's the same stuff that in some magnetrons in microwave ovens. It's important to make sure not to chip the beryllium parts as fractured parts can release beryllium particles into the air and it's nasty stuff if you inhale it. They really used some dangerous stuff in old equipment, even in children's toys.

I think it's common knowledge that in the era of tube radios those guys were real experts. Well, this is not to say that people aren't experts now (or during other decades) but it's interesting to see that in the early days of electronics those people designed really good circuits and they were in fact real experts. That's one reason why I like old equipment.

Thanks for alerting me to that possibility of beryllium being present in those capacitors. I really had no idea.

Thanks...

[Frankentronics]

So, I removed the tuning capacitor from that radio chassis and measured a range of about 130p to 560p on each of the gangs. I built that circuit that I mentioned above and it is a complete failure. I don't get an oscillation whatsoever. Just a flat line on the scope.



It seems that no matter what oscillator circuit I try to build I just don't have any luck. Although, this time I did substitute the JFET's because I do not have any 2N4340 in stock. So I tried with MPF102, and J112 JFETs to see what I get. But all I get is a completely flat line.

Is there something wrong with the schematic?

Now that I went through the trouble of obtaining a tuning capacitor I really want to build a circuit that makes use of it. If the circuit I was using is no good, could anyone suggest another simple circuit that I could build using components that I am likely to have. I basically have what I believe should be the "usual suspects" in terms of transistors and OP Amps, and the MPF102, J201, J112 JFETs.

Thanks...

[TimFox]

Two comments about the variable capacitor:
1.  These tuning capacitors for the MW broadcast band were usually nominal 365 pF.  Your range of 130 to 560 pF is suspicious, since the minimum value is far too large.  The mica compression trimmers typically found on the side of each section should not be so large as 100 pF (typically 30 pF).  As a typical example, the single-gang unit in this link https://www.tubesandmore.com/products/capacitor-365pf-variable-single-section  does not have built-in trimmers, and has a range of 14.9 to 384.2 pF, for a delta of 369.3 pF.
2.  How did you insulate the case of the three-gang capacitor from ground? The case is common to the two gangs, and will connect to the gate of Q1 in your circuit.

[Frankentronics]

Two comments about the variable capacitor:
1.  These tuning capacitors for the MW broadcast band were usually nominal 365 pF.  Your range of 130 to 560 pF is suspicious, since the minimum value is far too large.  The mica compression trimmers typically found on the side of each section should not be so large as 100 pF (typically 30 pF).  As a typical example, the single-gang unit in this link https://www.tubesandmore.com/products/capacitor-365pf-variable-single-section  does not have built-in trimmers, and has a range of 14.9 to 384.2 pF, for a delta of 369.3 pF.

I don't see any damage on the capacitor and I also carefully used compressed air to clean it. Perhaps the measurements are not correct due to the method I used to measure. I recently watched a YouTube video that explained that you will red different values, depending on what frequency is used by the measuring device. Is is possible that my BK Precision DMM is simply using the wrong frequency for measuring this type of capacitor?

However, the measurements were quite consistent between the gangs, which leads me to believe that the unit is OK. I think if there was damage to the unit, each gang would read something completely different.

2.  How did you insulate the case of the three-gang capacitor from ground? The case is common to the two gangs, and will connect to the gate of Q1 in your circuit.

You bring up a good point. In fact, I was a bit unclear about how this should be connected, so I actually tried various combinations. Perhaps that's the reason why I get no waveform.

What I did was, connect the lead wire from the rotor to the gate of Q1. Then I connected one of the gangs to ground and the other gang to what is marked as point A of the selector switch.

Did I do that correctly?

Thanks...

[gbaddeley]

The frame of the big ganged variable cap is a fairly high impedance point, connected to the gate of a fet? Good luck with that! A Small plastic ganged cap from a transistor radio might work better.

[TimFox]

The frame of the capacitor is, in fact, a problem for this circuit, since it goes to a high-impedance node.  It must be carefully insulated, including the shaft through an insulated coupling to the knob, to avoid difficulties.
I would suspect the DMM method of measuring small capacitances.  Does your unit have the mica compression trimmers mounted?  Note that in the photograph I showed, there is provision on the side, but they were not mounted.  If the screws on the trimmers are tightened all the way, they may be damaged or at an excessive capacitance.  For this application, I would remove them entirely.
Did you check the DC voltages inside the circuit?  The voltage from the negative rail (ground) to each FET drain should be roughly between 1/3 and 2/3 of the positive supply voltage.  If the Idss of your FETs is totally different from the original ones, the quiescent point may be way off.

[Frankentronics]

I read some more on the internet about building these kinds of audio oscillators and came across an old article that said this was difficult to do and the many people get discouraged. Clearly I am not making great progress so I guess that's what they meant.

I measured the DC voltages between ground and drains of the FETs. When the circuit is powered up by 18V I get 2.47V at Q1 and 2.72V at Q2. So, it is not between 1/3 and 2/3 of supply voltage. That's when I use MPF102 JFETs.

When I switch to J112 JFETs I get 2.16V at Q1 and 2.58V at Q2.

So, perhaps one thing that is definitely wrong is the fact that I am not using the correct 2N4340 JFETs.


I am having trouble understanding what the issue is with the frame of the capacitor. I searched the internet to make sense of it and I looked at many images but I can't find any similar capacitors that do not have a metal frame. Is this circuit perhaps designed for that other kind of variable cap that comes in a plastic frame?

My capacitor does have trim pots mounted on top. One of the gangs does not and the other two gangs do. I connected the gang without the trim pot to the B side, as seen on the schematic, and one of the gangs that has the trim pot to the A side.

But I really wish I could make something work, now that I went through the trouble of removing the capacitor from that radio chassis. I can put it back into the chassis, but I would much rather use it to build an audio oscillator that uses this cap.

I am attaching an image that shows the capacitor and how I connected it to the breadboard.

There must be some other relatively simple circuit that can use this capacitor.

Thanks...

[Cubdriver]

I can't quite tell from the image - are you connecting to the frame?  Do a resistance check between your connections and the cap plates, and make sure that one side connects to the stator plates, and the other to the rotor (obviously, do this carefully so as not to bend or otherwise damage the plates).

-Pat

[Frankentronics]

I can't quite tell from the image - are you connecting to the frame?  Do a resistance check between your connections and the cap plates, and make sure that one side connects to the stator plates, and the other to the rotor (obviously, do this carefully so as not to bend or otherwise damage the plates).

-Pat

I did a continuity check when I was cleaning the capacitor and trying to measure the capacitance (as well as possible on my DMM). The rotor plates are connected to the rotor shaft, which is also electrically connected to the frame. There are three leaf springs, mounted on the frame, which provide secure electrical connections of the shaft. Each leaf spring extends to the bottom and ends with a soldering tab. There is a black lead wire soldered to all three leaf spring soldering tabs. Of course, these leaf springs are also touching the metal frame, but so is the shaft itself. I connected that black lead wire to the gate of Q1, as per the schematic.

Each stator is isolated (I checked continuity and capacitance by turning the shaft) and each stator has one white lead wire. Only two of the stators have trim caps. I connected the lead wire of the stator without the trim cap to ground and the lead wire of one of the other stators to point A on the schematic.

Regarding the trim cap. Those have a solder tab. So, I tried connecting that solder tab to point A and I also tried to test the circuit without that connection(as I don't know if the trimmer is already internally connected).

I believe I did all those connections correctly. Here is the schematic again (for convenience).



The plates are not bent nor damaged. Although I do know that in some old radios the outer rotor plates are slotted and in fact bent at the factory for accurate tracking. I did not attempt to straighten these outer rotor plates. Only the outer (slotted) rotor plate of the gang without the trim cap is a bit bent. (I hope I'm explaining clearly).

I still don't understand what is the technical issue with the metal frame as I am still under the impression that all of those types of capacitors, from that period, are built in the same way. At least that how it looks visually, but I never did any continuity tests on any of them.

Thanks...

[TimFox]

Virtually all multi-gang capacitors of that type have all rotors connected to a common frame that would be grounded in the original application but must be insulated for the Wien bridge circuit.
The drain voltages you measure indicate that the FETs are fully ON and incapable of amplification.  You can verify that by measuring both drain and source voltages.  To measure all three voltages on Q1, select the 20k resistors.  If the FETs can’t amplify, then oscillation is impossible.

[Frankentronics]

Virtually all multi-gang capacitors of that type have all rotors connected to a common frame that would be grounded in the original application but must be insulated for the Wien bridge circuit.

Alright, I completely dismantled the capacitor. There's some light surface rust on the shell, so I'll have to clean it off with hydrochloric acid, tomorrow.

The shaft can easily be isolated from the frame, by replacing the steel balls with non conductive ones. I happen to have acrylic balls of the exact same diameter, which was 1/8 inch. There are 9 balls on the face side and a single ball at the rear. That one will eventually have to be a glass ball because there is a lot of pressure on it.

I also removed all the leaf springs.

Tomorrow I'll finish the work and post some sequential pictures.

If this circuit doesn't work I at least hope to be able to use this modified capacitor for another circuit, perhaps even when I get to making that GR-1313  clone. But I really want a simpler circuit to work out first, because I want to make sure that I'm capable of putting together an oscillator of this type.

The drain voltages you measure indicate that the FETs are fully ON and incapable of amplification.  You can verify that by measuring both drain and source voltages.  To measure all three voltages on Q1, select the 20k resistors.  If the FETs can’t amplify, then oscillation is impossible.

Once I reassemble the capacitor I'll be able to measure all the voltages.

I guess I still need to resolve the FETs. Mouser has 2N4340 FETS in stock for $12.16. Is there any chance that MPF102 or J112 FETs can work by changing some of the resistors in the circuit?

Thanks...

[AVGresponding]

A DIY variable capacitor project has some pitfalls.  The first to come to mind is the means of connecting the rotor (usually to ground) with low impedance and minimal noise.  In commercial units there is a clip that does that job.  Look closely at any unit and find that clip, usually at the end opposite the ball bearing.  High performance units for high grade lab gear use several clips per section.  Sometimes they are made of beryllium copper.

End play can create unstable capacitance.  The material of the plates and frame is important, as its temperature coefficient of size affects the oscillator.  The nonconducting supports also affect capacitance and losses.  One of the rotor plates is usually slotted to allow for fine adjustment of tracking.

The balls in the bearing need to be free of lumps, which could cause bumpy tuning.

The shape of the plates will affect the dial calibration of course.  So that needs to be planned carefully to obtain a linear dial.  This will be compromised by trimmer capacitors but generally not too seriously.

High grade units have an adjustable rear bearing to obtain smoothest operation.

And I thought this would be a relatively simple project for a machinist. Of course, a lot of work, but still doable.

I'm surprised to hear about the beryllium. That's the same stuff that in some magnetrons in microwave ovens. It's important to make sure not to chip the beryllium parts as fractured parts can release beryllium particles into the air and it's nasty stuff if you inhale it. They really used some dangerous stuff in old equipment, even in children's toys.

I think it's common knowledge that in the era of tube radios those guys were real experts. Well, this is not to say that people aren't experts now (or during other decades) but it's interesting to see that in the early days of electronics those people designed really good circuits and they were in fact real experts. That's one reason why I like old equipment.

Thanks for alerting me to that possibility of beryllium being present in those capacitors. I really had no idea.

Thanks...

Beryllium is alloyed to softer metals like copper and gold in order to stiffen them without work hardening them.
In this case that would be to make the copper clips retain their springiness over time.

I should think the chance of releasing beryllium particles to be vanishingly small, unless you decide to use an angle grinder on them or something.