Comments on my sawtooth generator design?

[keantoken]

Hi all.

I've integrated much of what I know into this scheme, and I find myself pleased with the trickyness of the circuit. It was made with inspiration from the PUT oscillator here: http://www.4qdtec.com/putpr.html .

D5 is supposed to compensate for Q1's Vbe, but I could probably come up with a better way of doing this. Anyways, I was more interested in the rest of the circuit.

The greatest care is taken to provide constant current to the cap, though I cheated a little using the SPICE current source. Oh well. Anyways, the significance of using the 680p (I should probably make it larger, but if I do the circuit has problem discharging it in time, causing problems for high frequency operation) cap is that 680 is roughly the Vbe drop of a transistor and so I can easily make a 680u CCS and get an approximate 200KHz saw wave (although the implementation of this is easily challenged).

The essential aim of this circuit is simple: generate a harmonically accurate sawtooth wave, for use in sweep circuits such as curve tracers and also to examine the output of an audio amplifier. The circuit should be simple, sweet, accurate and not a hassle to use. It should operate at a good, high frequency, and should handle different supply voltages without trouble. I still have some details to work out, namely the CCS to be used, and I might replace the zener with an LM431, and also I want to change Q1 to an RF type so it can compete with Q6 and Q8.

So my obvious question is: What did I do wrong? Is this practical? The circuit simulates quite well, but I haven't built it.

I know that there are probably hundreds of better, cheaper ways I could do this, but I am just interested in making a sawtooth generator with discrete components, using the PUT pair described in the link at the beginning of this post.

I have always been fascinated by all the simple resistors and tuning networks in precision circuits like in Tek scopes, combined with the amazing simplicity and low component count. I hope to emulate that here. Ambitious, I know. But I can't stop trying - it's so fun!

Thank you for looking,
- keantoken

[lineup]

hello keantoken.
I love you ..

As you know I started this topic, with some interest shown!

>Solid State >Sine wave - Square & Triangle wave generator using Transistors / OP-Amps
Sine wave - Square & Triangle wave generator using Transistors / OP-Amps

I also posted some hints on the ELEKTUUR, ELEKTOR Rising Sawtooth circuit
which is very nice to see how amplifier performs ---
--- if it is slow or have overshoot etc.
http://www.diyaudio.com/forums/showt...63#post1597163


I suggest you have a look at this circuit with adjustable frequency.
It is one very useful Design!
For Audio amplifier testing
And go from there ...

[keantoken]

Okay, my thoughts when I looked at that circuit were that there are devices causing the current through the capacitors to be nonlinear...

I would have to simulate the circuit to know how it behaves, which I might do soon.

So since you suggest starting from a different circuit, what is wrong with the one I posted?

It doesn't set discharge time, it is simply meant to make a saw wave and do it well.

- keantoken

[lineup]

nothing wrong
just posted the other for a compare, if you wish

I am sure your circuit has good quality
I begin to trust your knowledge more and more, keantoken

I can hardly cope with some of your stuff .. seems you are ahead
thanks to all hours you spent in study and design

wehn comes to sqare or sawtooth generators, not only the circuit topology
but the transistor choice is an important factor
We want fast transistors, and with fast we this time mean very short rise and fall time
this is a parameter shown for switching type transistors which normally is not same type as fast low noise devices used for audio

of course how high frequency you aim for with your generator
will set the requirements

[sawreyrw]

keantoken,

I thought you might be interested in this circuit, since you like discrete circuit. It is the horizontal sweep circuit for a type 321 Tektronix CRO. I have inserted the .asc file below.

Rick

Version 4
SHEET 1 1020 680
WIRE 688 -64 32 -64
WIRE 32 -48 32 -64
WIRE 144 0 128 0
WIRE 496 0 144 0
WIRE 688 0 688 -64
WIRE 688 0 560 0
WIRE -160 16 -240 16
WIRE -16 16 -80 16
WIRE 256 32 160 32
WIRE 272 32 256 32
WIRE 368 32 272 32
WIRE -240 48 -240 16
WIRE 368 48 368 32
WIRE 896 64 768 64
WIRE 960 64 896 64
WIRE 624 80 544 80
WIRE 768 80 768 64
WIRE 960 80 960 64
WIRE 32 96 32 48
WIRE 128 96 128 0
WIRE 128 96 32 96
WIRE 160 96 160 32
WIRE 624 96 624 80
WIRE 544 112 544 80
WIRE 368 128 320 128
WIRE 320 144 320 128
WIRE -240 160 -240 128
WIRE 400 160 336 160
WIRE 480 160 400 160
WIRE 256 176 256 32
WIRE 688 192 688 0
WIRE 720 192 688 192
WIRE 768 192 768 160
WIRE 768 192 720 192
WIRE 768 208 768 192
WIRE 32 224 0 224
WIRE 128 224 128 96
WIRE 128 224 112 224
WIRE 160 224 160 176
WIRE 160 224 128 224
WIRE 192 224 160 224
WIRE 544 256 544 208
WIRE 704 256 544 256
WIRE 960 256 960 160
WIRE 960 256 912 256
WIRE 912 272 912 256
WIRE 448 304 368 304
WIRE 256 320 256 272
WIRE 336 320 336 160
WIRE 336 320 256 320
WIRE 448 320 448 304
WIRE 256 336 256 320
WIRE 368 336 368 304
WIRE 544 336 544 256
WIRE 960 336 960 256
WIRE 256 448 256 416
WIRE 368 448 368 416
WIRE 368 448 256 448
WIRE 544 448 544 416
WIRE 768 448 768 304
WIRE 768 448 544 448
WIRE 960 448 960 416
WIRE 960 448 768 448
FLAG 624 96 0
FLAG 912 272 0
FLAG 448 320 0
FLAG 320 144 0
FLAG 896 64 Vcc
FLAG 0 224 Vcc
FLAG 272 32 +10
FLAG -240 160 0
FLAG 720 192 out
FLAG 144 0 in
FLAG 400 160 emitter
SYMBOL pnp 480 208 M180
SYMATTR InstName Q2
SYMATTR Value 2N5401
SYMBOL pnp 704 304 M180
SYMATTR InstName Q3
SYMATTR Value 2N5401
SYMBOL npn 192 176 R0
SYMATTR InstName Q4
SYMATTR Value 2N2222
SYMBOL res 528 320 R0
SYMATTR InstName R2
SYMATTR Value 27k
SYMBOL res 128 208 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName Rt
SYMATTR Value 20k
SYMBOL cap 560 -16 R90
WINDOW 0 0 32 VBottom 0
WINDOW 3 32 32 VTop 0
SYMATTR InstName Ct
SYMATTR Value 500pF
SYMBOL res 752 64 R0
SYMATTR InstName R4
SYMATTR Value 100k
SYMBOL voltage 960 64 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 45
SYMBOL voltage 960 320 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V2
SYMATTR Value 45
SYMBOL res 240 320 R0
SYMATTR InstName R1
SYMATTR Value 100k
SYMBOL voltage 368 320 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V3
SYMATTR Value 10
SYMBOL voltage 368 32 R0
WINDOW 3 34 74 Left 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR Value 10
SYMATTR InstName V4
SYMBOL res 144 80 R0
SYMATTR InstName R5
SYMATTR Value 18meg
SYMBOL voltage -240 32 R0
WINDOW 3 24 104 Invisible 0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR Value PULSE(-0 10 0 .1u .1u 8.9u 12u)
SYMATTR InstName V5
SYMBOL pjf -16 -48 R0
SYMATTR InstName J1
SYMATTR Value 2N5462
SYMBOL res -64 0 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R6
SYMATTR Value 1K
TEXT -216 -160 Left 0 !.tran 20u
TEXT -216 -128 Left 0 !;op

[Conrad Hoffman]

For some very nice discrete saw and triangle generators, download the manual and schematics for one of the Wavetek signal generators like the 185 from the usual suspects. With different range capacitors they go from fractional Hz to 5 Mhz. Very linear and very clever designs.

[keantoken]

Thank you. Will do.

Sawreyrw, this circuit is very strange looking to me, even though it seems to be fairly common. Very few parts, but apparently good enough performance for horizontal sweep. It was confusing at first, but it turned out to be just a simple feedback circuit... I think...

I will have to compare the circuits I find with my own so see what I'm up against.

Most of the saw generators I see use a pulse input to discharge the cap. While this makes frequency control trivial, the threshold voltage is harder to control. Granted, one can always amplify the output to whatever voltage range you need.

My designs sense a threshold voltage and discharge when this is reached. So we get constant voltage output, but frequency control is more difficult. The frequency stability is dependent upon the cap at this point. Maybe the first method is easier because tolerance is trivially corrected with amplification.

I have been thinking of trying to implement some kind of RC filter between the CCS and the output, that would automatically tune the generator to the desired frequency without trouble. We could also do some kind of PLL, but I like to start simple and determine whether we need anything more.

Are there any TO92 constant current source ICs? Surely they must have them for dimes. It would save me a lot of trouble. Inventing your own is fun, but gets annoying when you want to focus on other things.

Thank you for all your input. I will see if I can find the wavetek schematics. I will have to order a bunch of FETs soon so I can experiment with my more interesting circuits.

- keantoken

[sawreyrw]

keantoken,

I don't think the circuit I posted is strange; I think it is quite innovative. This circuit was designed when there were no ICs and it was desirable to minimize the number of active devices to minimize cost. As far as I'm concerned it does a lot with just 4 transistors. I didn't describe it in my earlier post, because I wanted to see what your reaction would be. Q2, Q3, and Q4 form a high gain linear amplifier. Q3 and Q4 are emitter followers. They have a high input impedance and a voltage gain of just less than 1. Q2 is a common emitter amplifier. It's gain is gm*Rload, where gm=.04*Ic. One 'trick' this circuit uses is that the base of Q4 is held at approximately 0 volts because the base of Q2 (emitter of V4) sits at Vbe2.

The high gain amplifier along with Rt and Ct form an integrator. The current through Rt is essentially constant (during rampdown) and the current through Ct is the same. A constant current through a cap causes the cap voltage to change at a constant rate.

Of course, the FET resets the circuit.

Regarding TO92 current sources, there are constant current diodes available. What they really are is FETs operating at Idss, because the gate and source are shorted internally. I don't think their tolerance and temperature performance are very good.

Rick

[Conrad Hoffman]

The new Linear Tech current source seems ideal for this, but I don't think they've reached distribution yet.

[keantoken]

Originally posted by sawreyrw
keantoken,

I don't think the circuit I posted is strange; I think it is quite innovative. This circuit was designed when there were no ICs and it was desirable to minimize the number of active devices to minimize cost. As far as I'm concerned it does a lot with just 4 transistors. I didn't describe it in my earlier post, because I wanted to see what your reaction would be. Q2, Q3, and Q4 form a high gain linear amplifier. Q3 and Q4 are emitter followers. They have a high input impedance and a voltage gain of just less than 1. Q2 is a common emitter amplifier. It's gain is gm*Rload, where gm=.04*Ic. One 'trick' this circuit uses is that the base of Q4 is held at approximately 0 volts because the base of Q2 (emitter of V4) sits at Vbe2.

The high gain amplifier along with Rt and Ct form an integrator. The current through Rt is essentially constant (during rampdown) and the current through Ct is the same. A constant current through a cap causes the cap voltage to change at a constant rate.

Of course, the FET resets the circuit.

Regarding TO92 current sources, there are constant current diodes available. What they really are is FETs operating at Idss, because the gate and source are shorted internally. I don't think their tolerance and temperature performance are very good.

Rick

No circuit is strange, but to me it was. I see similar "strange" circuits a lot in the Tek manuals and other test equipment schematics.

So an Integrator is what we call it. I could tell as much as that it was a high-gain amplifier designed to keep the current through the cap constant.

Maybe my reaction wasn't too great . I will have to play with this circuit and see what I can do with it.

- keantoken