Maker Pro
Maker Pro

Compression by the resistor

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Can't get this to oscillate

Hi everybody,

I'm trying to breadboard one of my first circuits which is this one: http://www.vk2zay.net/article/196 (the first on the page). Attached is a picture of the breadboarded circuit and the measurement I get on my oscilloscope.

I have checked all the connections. I have checked my transistors using the dedicated slot and the hfe position on my multimeter. The readings I get are 368 (BC557) and 354 (BC547). I equivalents to the original transistors depicted in the circuit and found those.

I am not very savvy yet but all I want is to get better and understand my mistakes. Am I missing something obvious? If not what other debugging options do I have?

Thank you for your time!

Edit: Please don't judge me for labelling those resistors, I'm colorblind :)
 

Attachments

  • osc.jpg
    osc.jpg
    45.9 KB · Views: 132
  • bread1.jpg
    bread1.jpg
    146.1 KB · Views: 140
  • bread2.jpg
    bread2.jpg
    95.8 KB · Views: 167

CDRIVE

Hauling 10' pipe on a Trek Shift3
May 8, 2012
4,960
Joined
May 8, 2012
Messages
4,960
Insure that the NPN & PNP aren't swapped.

What model scope are you using?

Chris

Edit: The scope's vert input should be set to 1V/Div with direct measurement and 100mV/Div when using a 10x probe.
 
Last edited:

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Hi Chris, thanks for you quick answer!

I just solved the issue. The BC557 was faulty. I found out by testing the voltages across the capacitors one by one. The whole right section with R2, R3 and Q2 seemed in a "reasonable" range (in an intuitive manner more than a mathematical one actually), so did the voltage across R1 but the voltage across Q1 (BC557) was really low (0.74). I swapped it for another one in my box (so glad I got 4 of each from the shop!) and I get the right signal at the output.

Thanks for the advice about the scope's range though. It helped a lot to know what the right setting was in order to rule out a misuse of my metering tools. I'd also forgotten to set it to AC... It is a handheld Velleman HPS140. All I could afford but it seems like a fine device!

Thanks very much again for you help.

Edit: By the way, how can my transistor pass the "hFE" test if it is faulty?
 
Last edited:

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Good morning, afternoon, evening depending on where on the globe you are. It is late in the night here and I am learning electronics the hard way!

I finally got my relaxation oscillator to oscillate. I found it on this website: http://www.vk2zay.net/article/196. It seems like a very good way for me to learn how to make an oscillator for my long term synth DIY project. I really like the fact that the author starts with a really simple implementation and builds from there explaining why components get added each time.

Now before I continue with perfecting the circuit I would like to attenuate the output signal to a line level to try and hear it through my sound card. I took my biggest resistor (1M) and hooked it to the output, knowing that it wouldn't be enough to tame all 12V to the 2V that I need for a line level. I took all the 10M resistors that I had (6 of them) and hooked them in series to that output at C1. That's a total of 6M which is still not enough but:

Two unexpected things are happening:

- When I look through the oscilloscope at the output signal, not only do the highest peaks come down but the lowest come up by the same factor. A bit the way a compressor acts in an audio chain.
- When I hook up the output to the tip of my jack lead (to the sound card) I hear a buzz and a faint clicking (I know this clicking is the oscillation because I have a pot in series with R1 that allows me to change the frequency and it changes the frequency of the click). If I hook the sleeve of the jack lead to the ground of the circuit the buzz stops, but the click goes away too.

I have been Googling for the last hour and a half a way to tame a 12v output to a line level output without success. Am I right to assume that all I need is to use resistors and connect tip to out and sleeve to ground? What am I doing wrong :(

P.S: Here is a link to a Java simulation of the circuit: http://www.falstad.com/circuit/#$+1....0+0.8+2+-1 o+0+64+0+35+40.0+3.90625E-4+3+-1
 

Attachments

  • Screen Shot 2013-04-18 at 2.17.41 AM.png
    Screen Shot 2013-04-18 at 2.17.41 AM.png
    11.2 KB · Views: 197

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
You need two resistors wired as a voltage divider. Use a 10K to the output, a 1K to ground, and take your output from the junction of the two.

Votage divider

Bob
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Hey Bob, thanks!

I'd come across Voltage dividers but never saw them... for what they are I guess!
This makes hell of a lot of sense now.

I still get this massive buzz though, the ratio signal / noise is maybe a tad better but still ridiculous. Should I use higher values for the resistors?
Also, again when I connect the ground everything disappears. Am I plugging this right? Junction of the two resistors = tip, ground = sleeve?

Edit: I'm also interested in knowing what cause the raise of the low peaks with my makeshift solution of the single resistor from my first message :)
 
Last edited:

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Where in your circuit are you taking the output from? What frequency is it oscillating at?

On and AC setting I would expect the low peaks to go up and the high peaks to go down, i.e. toward 0 volts. If the scope is on DC I would expect both to go towards zero as well, which would be down.

Bob
 

CDRIVE

Hauling 10' pipe on a Trek Shift3
May 8, 2012
4,960
Joined
May 8, 2012
Messages
4,960
the voltage across Q1 (BC557) was really low (0.74). I swapped it for another one in my box (so glad I got 4 of each from the shop!) and I get the right signal at the output.


Edit: By the way, how can my transistor pass the "hFE" test if it is faulty?

It was more likely a bad connection. A bad connection in either node indicated by a red X would produce a voltage reading much like what you describe. I wouldn't toss the transistor until I was sure it was bad.

Chris
 

Attachments

  • UnijuncEquiv.JPG
    UnijuncEquiv.JPG
    33.1 KB · Views: 149

CDRIVE

Hauling 10' pipe on a Trek Shift3
May 8, 2012
4,960
Joined
May 8, 2012
Messages
4,960
Don't expect clean tonal audio from your osc. The Sawtooth waveform contains a very fast falling edge that's rich in harmonics. Raspy audio is inherent with a sawtooth. Also, your osc freq in your other topic is only ~36Hz.

Any test, observations, etc that you observe without the ground to the Line Input jack connected are spurious and meaningless.

Chris
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Where in your circuit are you taking the output from?
I am taking the output at the junction of C1 and Q1's emitter.
What frequency is it oscillating at?
Chris actually answered this question: 36Hz. More about this in my answer to him a bunch of lines down...
On and AC setting I would expect the low peaks to go up and the high peaks to go down, i.e. toward 0 volts. If the scope is on DC I would expect both to go towards zero as well, which would be down.
OK this makes sense

Don't expect clean tonal audio from your osc. The Sawtooth waveform contains a very fast falling edge that's rich in harmonics. Raspy audio is inherent with a sawtooth. Also, your osc freq in your other topic is only ~36Hz.
I am working on a sawtooth oscillator because I want to get to build a VCO for a subtractive synthesiser and the more harmonics the better!
I realise now that 36Hz is very low and I shouldn't be able to hear it. Still, the meter on my sound card's software indicates no level at all when I connect the tip and the sleeve so there is no signal at all coming through. The voltage is way to high for a line level though so maybe some sort of protection cuts it off in the sound card.
Any test, observations, etc that you observe without the ground to the Line Input jack connected are spurious and meaningless.
Chris
Will remember this :)

My long term project is to build modules for a synthesiser that are all made of discrete components, i.e. no IC's. I am starting to understand what a tough endeavour this is for a beginner like me. Maybe I should play with the mighty 555 a bit first, get my head around simpler concepts before diving back into the dark and twisted world of pure analogue circuitry :rolleyes:
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
It was more likely a bad connection. A bad connection in either node indicated by a red X would produce a voltage reading much like what you describe. I wouldn't toss the transistor until I was sure it was bad.

I have been a bad boy and I have unplugged components without unplugging my power supply. I thought I'd popped my transistor by doing so but I tried again following your advice and got this same transistor working again... ...then not. When the oscillator stops working and I measure the transistor's hFE I get a really low value. What seems to solve the problem is time. If I stop using this transistor for 15mn then test it again the hFE comes back up and when I put the transistor back in the circuit the oscillator works again. Is this possible? Could overheating be an explanation?
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Just a little update:

I did a simulation of the circuit with the Falstad java applet (you can see it here: Simulation) with the voltage divider and I realised that it makes the oscillation stop.
Does this mean that I need to isolate my output somehow?
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Hi Robyn,

That oscillator circuit generates a "curved sawtooth" oscillation at the output, but the point in the circuit where the output is taken from is a "high impedance" point in the circuit. You can think of it as being "weak" or "flimsy". "Loading" it, by connecting resistors with values like 10k, will affect the circuit and cause it to stop oscillating.

This is why you kill the oscillator when you connect a voltage divider to it, and when you connect the ground of your sound card. A sound card has a typical input resistance of around 50k and this is too much load for the oscillator to drive.

So you do need to use a circuit called a buffer, which places almost no load on the oscillator circuit and provides an output that you can connect load resistors (such as a voltage divider with 10k and 1k resistors) to. That oscillator, as designed, is only intended to demonstrate the action of the oscillator itself; its output impedance is too high to be useful without a buffer of some kind.

Alternatively, you can use a voltage divider made from very high value resistors, which put very little load on the oscillator, but the final output will have an even higher impedance. It's not a good general approach.

The simplest type of buffer is called an "emitter follower". It's made from a single transistor; in this case an NPN such as a 2N3904 or BC547B, or any small-signal NPN transistor. A transistor with a high current gain, such as a BC547C/548C/549C, provides more buffering. Connect the transistor as follows.

Collector: connect to the positive supply rail.
Base: connect to the oscillator capacitor, i.e. the point currently labelled as the output.
Emitter: this is the output of the buffer. You need to connect it to the 0V rail through a resistor; you can use the voltage divider, as this provides the "emitter load" and also attenuates the voltage.

That's all you need to do. Just one transistor, and the two resistors that form the voltage divider.

I hope thiis is clear. If in doubt, look up "emitter follower" on Wikipedia.

Edit: BTW I'm not sure if you've realised this yet, but the "massive buzz" you hear is the actual oscillation of the circuit. A 36 Hz sawtooth sounds like a buzz. If you want the oscillator to run at a higher frequency, so it sounds like a tone, you need to reduce the value of the capacitor in the oscillator. The frequency is proportional to the reciprocal of the capacitor value, so for example, halving the capacitor value (from 100 nF to 47 nF) will double the frequency. A capacitor around 6.8 nF (also called 0.0068 uF) will give a frequency of around 500 Hz, which is a medium-pitched tone.

Edit 2: If you want an actual sawtooth waveform, rather than the curved waveform generated by this oscillator, you can replace the 100k resistor that charges the capacitor with a "constant current source" (see also Wikipedia). But there are better oscillator designs around; I wouldn't use this one except as an initial experiment.

Edit 3: That oscillator has a design error: it applies a large voltage across the PNP's base-emitter junction in the reverse direction, causing the junction to break down. (Any reverse base-emitter voltage above about 7V does this, because the transistor's base-emitter junction behaves like a zener diode.) This can damage the transistor. Using a power supply voltage of 6V will avoid this problem but you may need to change some resistor values to make it oscillate reliably.
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
See my post on your other thread on this subject. That oscillator has a design error and will damage the PNP. Use a power supply voltage of 6V. You may need to change resistor values to make it oscillate reliably with 6V power supply.

davenn, would you merge the two threads please? They are on exactly the same subject.

Now merged :)
 
Last edited by a moderator:

BobK

Jan 5, 2010
7,682
Joined
Jan 5, 2010
Messages
7,682
Kris's post has a lot of good info. I think if you want to build a synthesizer you should look into op amp circuits. They are much easier to design with and much more well behaved than the circuit you are using. For example, you would not have the output impedance problem with op amps, they could directly output a line level signal. And a op amp sawtooth oscillator is very simple.

Bob
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Hey neighbour Kris!

Thank you so much for your replies, and the way you formulate them :) I am currently trying out your solution on the other thread.

Sorry about the double thread. I wanted to keep things clean by separating two issues from the one circuit. I guess I over categorised things!
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Seeing as most of the issue dealt with in this thread is due to the "flimsiness" ;) of the design of the circuit I don't mind seeing this thread deleted. It will avoid making the other (way more interesting) thread messy...
 

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
OK guys,

If there was a [SOLVED] button to click, I'd click it:

I have put together the emitter follower buffer and I understand how it works.
I have also learnt what a voltage divisor does and how.
And I have realised that I am actually setting myself back by trying to work with this implementation of a sawtooth oscillator.

I feel like I am torturing this poor PNP. So, Operational Amplifiers here I come.

Kris, Bob and Chris, thank you so much for your support. Electronics is a immense unknown universe for me in which you guys are interstellar lighthouses.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Hi Robyn... or is it Rob?

I read further on the page you linked to, and the author goes into a lot of detail on improvements that can be made to the design. He mentions the zener breakdown of the PNP transistor, but doesn't mention that it does temporary and/or permanent damage to the actual transistor. I'm not sure if he suggests adding an emitter follower as a buffer. But it's a pretty interesting page.

Re the threads, this is the one that should stay. It deals mainly with the impedance and loading issues but I've also raised the PNP zener breakdown issue here. The other thread only deals with the PNP deterioration.

I was interested to read your description of the damage. The transistor deteriorates and stops oscillating; it then shows a low current gain, but recovers after 15 minutes to the point where the gain has increased and it will oscillate again! Is that what you found?

There was another thread on this subject recently: https://www.electronicspoint.com/emitter-base-breakdown-t257654.html

My sister is moving to Melbourne later this month. (She's currently in Geelong.) I've been there a few times too, and I would live there if I could.

Edit: Yes, I agree with Bob and you, you should look into op-amps. I used to have an old Korg monophonic analogue synth, and the board was absolutely packed with 4558 dual op-amp ICs! Op-amps are very versatile.
 
Last edited:

Robyn

Apr 17, 2013
34
Joined
Apr 17, 2013
Messages
34
Hi Robyn... or is it Rob?
I read further on the page you linked to [...] it's a pretty interesting page.
It's Robyn, well actually Robin, but people have been spelling it with a Y here so I went on with :) I'm actually from south of France but I've been in Melbourne for 3 and a half years now.
I really like the fact that this page goes through the whole process of troubleshooting the weaknesses of a simple discrete design. I've always been looking for this kind of approach as I believe it is the best way to learn. This particular case is a bit to extreme though as I am not sure that any of my PNP's are functional :rolleyes:. Which brings me to your next point...

I was interested to read your description of the damage. The transistor [...] recovers after 15 minutes
I am not 100% sure that this is what happens. It seems like it but I am still fragile with my handling of metering tools so I would double check this finding (if you have a couple of PNP's to spare!). I enjoy this way of learning though, trial / error / observation / conclusion. It helps my artist brain make sense of the equations by being able to "feel" what happens in the circuit. I was watching a documentary about Robert Moog recently and he mentioned this intuitive approach. I guess I got a glimpse of what he meant :)

My sister is moving to Melbourne...
Well if she is ever around Gertrude St. in Fitzroy she should come to Southpaw. I am barista there, mainly on weekends and I would be happy to shout her a coffee!

Edit: Yes, I agree with Bob and you, you should look into op-amps...
I just watched this video:
. Also going to check the two books I bought (Electronics for dummies & Electronics demystified) for infos about our 5 legged friends (seems like they come more in 10 legged twins these days!). Thank you again very much, would be great to meet you in person if you pass by Melbourne again.

Edit: Another great op amp video: http://www.youtube.com/watch?feature=endscreen&NR=1&v=OMJ9WGrRf6A

Edit 2: Oh my god this all makes so much sense!

Edit 3: :D
 
Last edited:
Top