Trouble with high frequency astable multivibrator

[Exidor]

I am having trouble getting an astable multivibrator to oscillate higher than 480 KHz. I added a "baker clamp" which gave some improvement, but didn't get me to the target frequency of 816.5 KHz. I am on an iPad and can't post a schematic. I use 2sc6082-1e transistors and bat41 diodes for the clamp. Any help would be appreciated.

 

[phyzguy]

We'd have to see the schematic to be able to make an assessment, but probably the parasitic resistances and capacitances are larger than you think. Have you tried measuring any of these? By "parasitic" I mean unintended resistances and capacitances.

 

[Exidor]

All I can do is email the schematic, unless there is a schematic editor available at this site?

 

[Exidor]

https://www.physicsforums.com/cid:EA100AAB-8C63-496F-85F5-3B2AFE713B16

 

[phyzguy]

I can't see that image. Take a screen shot, then you can upload the .png file using the upload button.

 

[berkeman]

I am having trouble getting an astable multivibrator to oscillate higher than 480 KHz. I added a "baker clamp" which gave some improvement, but didn't get me to the target frequency of 816.5 KHz.

Welcome to the PF.

Why do you want to use a multivibrator circuit? That's one of the last choices I would make for designing an oscillator. What other topologies have you looked into?

 

[Exidor]

I wanted to avoid inductors. I need a two phase square wave at 816.5 KHz. There would be high voltage around so I wanted to go with transistors instead of IC's. The output signals drive a transistor amplifier, which would feed the primary of a Tesla coil. I could use a CD4046 and an inverter, but the chips might get burnt out.

 

[berkeman]

How precise is your frequency requirement? That's a frequency that I don't recognize. I'd use a shielded Arduino circuit, based on what you've said so far, but that's just me...

 

[Exidor]

I like to keep it as simple and cheap as possible. I don't know how sensitive the frequency is. I imagine it is pretty finicky.

 

[Tom.G]

The transistor is not fast enough. According to pg.2 of the datasheet, the switching times are:
Ton = 52nS
Tstg = 560nS
Tfall = 37nS
-----------------
. . . . . =649nS for each transistor; double that for two transistors = 1298nS or ≅1.3μS cycle time. That gives 770kHz maximum. Your transistors a little better than the datasheet so are operating a little faster.

Pick faster transistors.

 

[Exidor]

I was going by the f sub T 195 MHz. Thanks!

 

[Exidor]

Does anybody have any suggestions for power transistors that would switch well at a little under 1 MHz? I need them in a TO 220 package.

 

[Exidor]

 

[Exidor]

 

[Tom.G]

Take a look https://www.mouser.com/Semiconductors/Discrete-Semiconductors/Transistors/Bipolar-Transistors-BJT/_/N-ax1sh?P=1z0z73rZ1z0y00jZ1z0y1znZ1z0zl9uZ1z0zl5lZ1z0zlrjZ1z0zpefZ1z0zls6Z1z0zls5Z1y95l6hZ1z0y4diZ1z0w6ylZ1z0wa61Z1z0w6i4Z1z0w0jdZ1z0w6iwZ1yyzhtoZ1yyzhgqZ1yyyk2w for a several possibilities to choose from.

You haven't said what the load is on the multivibrator so some/all of those above may not be appropriate. You may want to choose a transistor that will work for both the multi and the output stages. (reduces the number of spares you will need )

Cheers,
Tom

p.s. The Fairchild https://www.mouser.com/ProductDetail/ON-Semiconductor-Fairchild/KSC5603DTU?qs=sGAEpiMZZMshyDBzk1%2fWi4G1GLBZKHK1sLGWiiwqkM0%3d looks like a possibility.

 

[Exidor]

The second schematic is the load of the first, with the astable multivibrator replacing the 816 KHz ac source. It's the astable multivibrator driving an H bridge amplifier with a 5 uH inductor as load. The amplifier may need revision.

 

[Tom.G]

The amplifier may need revision.

Yup. Connected that way you will lose all four transistors as soon as power is applied.

 

[Exidor]

I have had it running for a few minutes, watching the power consumption. It heats up a bit. The transistors have a high current and power dissipation rating. How would you change it?

 

[Tom.G]

The present output stage may function adequately as long as there it has a sufficient input signal. With no signal, there is a DC path from +30v, thru E-B junction, 400Ω, B-E junction, -30V. This forward biases all four transistors so that they short the 30V supply.

Simplest, though it may not work:

• do not provide a DC path between Bases of the transistors
• feed each transistor Base thru its own capacitor
• connect a resistor between each Base and Emitter (value calculated so Base leakage current does not turn on the transistor at highest operating tempurature)

Best:
Push-Pull Output stage. from: http://www.ecircuitcenter.com/Circuits/pushpull/pushpull.htm

Also try a Google search: https://www.google.com/search?&q=push+pull+output+transistor+configuration

Cheers,
Tom

EDIT: For your H-Bridge configuration, you will need to supply an out-of-phase driving signal to the two sides of the H-bridge. Either take the signals from opposite side of the multi or use a phase inverter stage.


EDIT: After sleeping on the push-pull configuration, I realized it too is problematic for this application. It introduces problems with drive levels to the output stage and changes the resonant frequency and circuit Q of the load. In other words it's a significant redesign of the overall system. Please ignore the push-pull suggestion.
Sorry,
Tom

 

[Svein]

Basic oscillator IC: https://en.wikipedia.org/wiki/555_timer_IC

Circuit example:
https://www.electronics-tutorials.ws/waveforms/tim58a.gif

 

[Exidor]

I can't get the 555 to high enough frequency. I need to be able to tune it.

 

[vortextor]

Try with Schmitt trigger IC, like CD4093 (four NAND gates) from Texas.
Needs only an RC group.This IC works up to 20V.
I made once a 100MHz multivibrator with SN74AHC1G14 Schmitt trigger gate,
unfortunately this admits only max. 5V

 

[Tom.G]

Please see my edit to post #19.
Tom

 

[phyzguy]

Even with the variable resistor set to 0, you still have the 1KOhm resistor charging the 2nF capacitor. Won't this give a period on the order of (1E3 * 2E-9) * 2 = 4 uSec. This is a 250 kHz frequency. This is just a "back of the envelope" calculation, but it seems that it shouldn't be surprising it doesn't run faster. To make it run faster, you need to reduce the 1 kOhm resistors, or reduce the 2nF capacitors, or both. Have you tried this?

 

[Exidor]

I already ordered transistors for the AMV. Wouldn't the 555 be more vulnerable to high voltage than BJT's?
I would think CMOS chips might have trouble with high voltage too.

 

[phyzguy]

I already ordered transistors for the AMV. Wouldn't the 555 be more vulnerable to high voltage than BJT's?
I would think CMOS chips might have trouble with high voltage too.

Was this addressed to me? Did you see my post #24? Do you have a response to this?

 

[Exidor]

Sorry, I thought you were talking about the 555 circuit. It doesn’t work with smaller capacitors. I don’t have those smaller resistors on hand.

 

[rbelli1]

I would think CMOS chips might have trouble with high voltage too.

Put over-voltage suppression (TVS, MOV, Zener, Gas discharge, etc.) on your power supply and outputs an you can use any technology you like. If the output voltage of the coil or the back EMF of the primary hits your transistors they will be damaged no matter which ones you buy.

BoB

 

[phyzguy]

Sorry, I thought you were talking about the 555 circuit. It doesn’t work with smaller capacitors. I don’t have those smaller resistors on hand.

No, I'm talking about the multi-vibrator. Aren't those discrete components that you can change?

 

[Exidor]

Below 2 nF the AMV won't oscillate with the current transistors. It is breadboarded. I don't have the necessary power resistors on hand to try.
I am operating at 30V.
I was thinking of using Schottky diodes on the Tesla primary coil (H bridge output) to quench any high voltage feedback.
Sorry for the confusion, I am addressing everybody in the thread.

 

[phyzguy]

Below 2 nF the AMV won't oscillate with the current transistors. It is breadboarded. I don't have the necessary power resistors on hand to try.

Can you just short out the 1 KOhm resistors and rely on the variable resistors only? This would test out whether reducing the RC would speed up the oscillator.

 

[Exidor]

Wouldn’t that damage the pot?

 

[phyzguy]

Wouldn’t that damage the pot?

Why should it? Before you have 1 KOhm plus whatever the pot is set for, which ranges from 0-1KOhm. You want to try reducing the RC time constant. You said you can't go smaller than 2nF on the capacitor. If you set the pots at the highest setting, you will have 1 KOhm, which is the value you had before with the pots turned to minimum. Then you can try slowly reducing the pot setting and see if the frequency increases. According to your schematic, there is only about 20 mA flowing down that leg. The pot should be able to handle this, even if the current increases. Do you know the power rating on the pot?

 

[Exidor]

I could give that a try. It’s a 10k pot @ 1 watt. The schematic editor wouldn’t allow me to change the value of the pot.

 

[phyzguy]

I could give that a try. It’s a 10k pot @ 1 watt. The schematic editor wouldn’t allow me to change the value of the pot.

It's a 10K pot? Your schematic shows it at 1K. What's the resistance when turned to the lowest setting?