I also check your new gen from 38 post.

http://forum.allaboutcircuits.com/attachment.php?attachmentid=61172&d=1383876504

And here you have the results:
C1 = 100nF; F = 4.95Khz

C1 = 10nF; F = 24.27Khz

 

Wow! Thanks Jony! You have just made me very happy
I guess theres something wrong with my built. I will rebuild it again.
Just out of curiosity, did you match your transistors beta? If so how well were they matched? And did you use bypass cap?
Thanks so much again for helping me testing these out!

 

Just out of curiosity, did you match your transistors beta? If so how well were they matched? And did you use bypass cap?

No, I don't match transistor and also don't use any bypass cap.
But I had a very tightly packed layout on the breadboard.
And tomorrow I try my more traditional design. Triangle wave generator based on Schmitt trigger.
Or maybe use this trick with diodes.
http://www.all-electric.com/schematic/eticircuits/555-triangle-with-independent-slopes.htm

 

As I promised today I build and test this circuit



First I use C1 = 100nF





Read waveform is voltage at Q3 collector. And blue one voltage across C1 capacitor.

Next, I change C1 capacitor to 10nF



Now as you can see we have a phase shift and also notice that Schmitt trigger threshold voltage also change.

Now I put C1 = 1nF



And circuit still "work" as an oscillator even without C1 capacitor.




And since circuit self oscillation frequency is relatively low, this confirms that the circuit is too slow to be able to generate high frequency triangle wave.

 

And circuit still "work" as a oscillator even without C1 capacitor. ... And since circuit self oscillation frequency is relatively low, this confirms that the circuit is too slow to be able to generate high frequency triangle wave.

Jony13: nice work. Here are some observations...

There are 2 limitations on this circuit:

First, the capacitances of the transistors and the prototype tend to add up very quickly. A quick rule of thumb is to assume that there is 15 pF between every pair of contacts on a solderless breadboard. If you use long wires, then there will also be a lot of unexpected coupling between different parts of the circuit.

The easy solution to fix this distributed capacitance effect is to run the circuit at higher currents using lower value resistors.

The second thing that slows down the circuit is the time delay in the transistors when they turn off after saturation. For instance, the 2N2907 takes as much as 80 ns to turn off after saturation and the 2N2222 can take 225 ns (!). There are also other turn on and turn off delays -- and they all add up. :-(

One solution to this is to use transistors that have better turn-on, turn-off and saturation delays such as the 2N2369 and 2N5771 that were mentioned in earlier posts.

Another (partial) solution is to keep the transistors from saturating by putting a germanium (such a s a 1N34A) or Schottky diode (such as a BAT54) from the collector to base of the offending transistor. This diode has the anode to the base and cathode to the collector of an NPN transistor.

 

Or maybe use this trick with diodes.
http://www.all-electric.com/schemati...ent-slopes.htm

Jony130: How about this version?

 

I know of an even simpler ne555 based triangular wave oscillator: http://i.imgur.com/mknvmsK.png

 

I know of an even simpler ne555 based triangular wave oscillator: http://i.imgur.com/mknvmsK.png

They got a nice little class D amp too

 

Jony13: nice work. Here are some observations...

There are 2 limitations on this circuit:

First, the capacitances of the transistors and the prototype tend to add up very quickly. A quick rule of thumb is to assume that there is 15 pF between every pair of contacts on a solderless breadboard. If you use long wires, then there will also be a lot of unexpected coupling between different parts of the circuit.

The easy solution to fix this distributed capacitance effect is to run the circuit at higher currents using lower value resistors.

The second thing that slows down the circuit is the time delay in the transistors when they turn off after saturation. For instance, the 2N2907 takes as much as 80 ns to turn off after saturation and the 2N2222 can take 225 ns (!). There are also other turn on and turn off delays -- and they all add up. :-(

One solution to this is to use transistors that have better turn-on, turn-off and saturation delays such as the 2N2369 and 2N5771 that were mentioned in earlier posts.

Another (partial) solution is to keep the transistors from saturating by putting a germanium (such a s a 1N34A) or Schottky diode (such as a BAT54) from the collector to base of the offending transistor. This diode has the anode to the base and cathode to the collector of an NPN transistor.

Although my first circuit linearity is not true, it does how ever oscillates at 20kHz+ as proven and shown by Jony130 (not using 2N2369 and 2N5771). So there must be some other factors that slow down both other circuits, my new one and Jony130's

PS: I still havent rebuild my first circuit yet, as I have not the time to do so. Once I'm able to reproduces Jony130's results, I'm planning to see how fast I can push its limit to. As of now, I'm still trying to find a way to produce true linearity for my fist circuit. I like discrete stuffs ; I learned so much more about transistor from this.

 

Version 4

1. No more current mirrors!
2. Sharp square wave to trigger current source.
3. Added current sense transistor Q8 to monitor and correct Q5's base current
4. Added schottky diode across Q6's base and collector to prevent it from saturating. This solved the problem of non-linear discharging!

I think I've nailed it! Although at higher frequency, changing cap to 1nF, distortion begins to show, might have to shove a little more current into it. Still trying to overcome this.

Anyway, thats it for now. Thanks to all who have been following this thread, and those who have been responding, helping and thus providing me encouragement to further explore.

As always, please provide feedback!

 

Here is the output of the current running into and out of cap C1

 

4. Added schottky diode across Q6's base and collector to prevent it from saturating. This solved the problem of non-linear discharging!

Great! I only wish I could take credit for predicting this result. )

 

Another (partial) solution is to keep the transistors from saturating by putting a germanium (such a s a 1N34A) or Schottky diode (such as a BAT54) from the collector to base of the offending transistor. This diode has the anode to the base and cathode to the collector of an NPN transistor.

You the man! RichardO

 

I finally assembled my version 4 together on a breadboard, using same transistors from version 1, other components value are exactly the same as on the schematic, only exception is R4, it is now changed to 220k and since I didnt have a BAT54 handy, it was omitted.

Worked on the first try!

Here is the result:

Edit: C1 = 1nF

 

... since I didnt have a BAT54 handy, it was omitted.

The Schottky diode is only going to have an effect at higher frequencies. Considering your experience with the earlier circuit, "higher" may not be very high.

The BAT54 is a surface mount part. For testing in solderless breadboards, you may want to use the BAT46 in a leaded package. Check that it works in simulation before getting a lot of them...

 

The Schottky diode is only going to have an effect at higher frequencies. Considering your experience with the earlier circuit, "higher" may not be very high.

The BAT54 is a surface mount part. For testing in solderless breadboards, you may want to use the BAT46 in a leaded package. Check that it works in simulation before getting a lot of them...

I noticed that BAT54 is a surface mount part, so I went ahead and ordered some 1n34A.

 

Non-linearity still exists when discharging C1 via Q6 base. If I substitute R3/R4 with a constant current source, this will limits Q6 Ic at a constant rate, thus its Ib will remain at a constant rate, this would work nicely at lower freq, but I cant seem to get rid of the current spike (red). While I understand that, this is due to voltage not develop fast enough at Q5's base when Q6 turns off, the reason for this is still unclear to me , increasing current source helps, but will slow down the discharging rate.

Anyone have any suggestion?
Is there a trick to keep Q6's Ic constant without the use of a current source?

Blue line - Q6's base
Red line - current in and out of C1

 

Here is my latest work, version 6.

50% reduction in power consumption!

This is an overall improvement over all previous version, I think i'm going to stick with this new method of discharging, and charging the cap. In this version, CCS (constant current source) are all being used as reference currents to be amplified at the charge and discharge states, therefore, power consumption being used by these are only a few µA. There are still distortions to be corrected but I'm quite happy with it now.

Please help me finish this, I'm getting a little bored doing it by myself