Hello all, I just wanted to let you know that I've finally built and tested the circuit, and here's my verdict.

Ronv's circuit is more efficient (93.6%) than Richard's (91.4%) but there are other differences:

Ronv's circuit:

1. The output voltage varies significantly if the load is less than 100 ma
   
2. The output voltage varies significantly if the voltage at the source is a little less or a little more than 12V
3. Good efficiency at the 555, power dissipated is 48 mW

Richar's circuit

1. The output voltage is very stable even if the load goes down to 1 mA
2. The output voltage is very stable even if the voltage at the source varies between 12V and 14V
3. Efficiency is not so great at the 555, with a dissipated power of 62 mW, according to LTspice

I tried to tweak Ronv's circuit a little, see if I could stabilize its output voltage when the load varied, but I couldn't accomplish much. I guess I'm still too much of a rookie to fully understand what I'm doing.

So I decided to build Richard's circuit, adding a 100K resistor in parallel with the load at the output so that the voltage would remain stable even when no load were present, and VOILA!... it worked perfectly... with probably the minor setback of the 555 getting really hot to the touch... I mean, I haven't actually placed a thermometer on the thing, but I can hardly keep my finger on it for a second or so before feeling the urge to remove it.
I'll see if I can add a transistor to the circuit to help it switch the MOSFET's gate, instead of driving it directly from the 555's output pin... maybe that will help cool things a bit.

Finally, I want to thank all who helped me with their comments and observations on this little project of mine, it's been a very educational experience.

Regards, and again, thank you.

 

Hello all, I just wanted to let you know that I've finally built and tested the circuit, and here's my verdict.

1. Efficiency is not so great at the 555, with a dissipated power of 62 mW, according to LTspice

.
.
.
So I decided to build Richard's circuit, adding a 100K resistor in parallel with the load at the output so that the voltage would remain stable even when no load were present, and VOILA!... it worked perfectly... with probably the minor setback of the 555 getting really hot to the touch... I mean, I haven't actually placed a thermometer on the thing, but I can hardly keep my finger on it for a second or so before feeling the urge to remove it.
I'll see if I can add a transistor to the circuit to help it switch the MOSFET's gate, instead of driving it directly from the 555's output pin... maybe that will help cool things a bit.

The 555 should not get hot dissipating 62 mW!

Bob Pease "5 second rule": If you can hold your finger on a hot device for 5 seconds, the heat sink is about right, and the case temperature is about 85 deg C.

This circuit is running at a low frequency so I would be a bit surprised if the capacitance of the MOS-FET is causing enough current draw from the 555 output to make the 555 run hot. I suspect there is some other problem lurking:

Maybe a subtle miss-wire?
Is the 555 really running on 12 volts? It is only rated for 16 volts maximum.
A ground loop putting an over-voltage on part of the 555 circuit?
There are inductive transients getting into the 555 pins?

 

The 555 should not get hot dissipating 62 mW!

Bob Pease "5 second rule": If you can hold your finger on a hot device for 5 seconds, the heat sink is about right, and the case temperature is about 85 deg C.

This circuit is running at a low frequency so I would be a bit surprised if the capacitance of the MOS-FET is causing enough current draw from the 555 output to make the 555 run hot. I suspect there is some other problem lurking:

Maybe a subtle miss-wire?
Is the 555 really running on 12 volts? It is only rated for 16 volts maximum.
A ground loop putting an over-voltage on part of the 555 circuit?
There are inductive transients getting into the 555 pins?

Hello again, Richard.

To answer your questions:

• The 62 mW is reported by LTspice, but I haven't actually measured it on my circuit
• Yours truly is using a crappy'ol ISR740 mosfet, which has a Qg of 63 nC, and an Rds of 0.55 Ω
• Rest assured, the 555 is running on regulated 12V
• I'm using four electrolytic 1 µF @ 350V caps at the output instead of the ceramic recommended one (way too expensive)... and high voltage spikes are shown at the output, just as Ronv told me would happen, and LTspice shows in the sim, and my scope read when I tested the circuit... I don't know if that's important...
• I've placed no heat sink on the 555
  
• Yes, I guess I could hold my finger on the thing for about 5 seconds while gently reciting the Lord's prayer... asking him to spare my index's fingerprint.

I can assure you there's no ground loop in the circuit nor miss-wiring. And as for inductive transients, if you tell me how to look for them I'll be more than happy to oblige... I have a scope, if you think it will be necessary...

Thanks for the follow up

 

There a a couple of possibilities.
You could add a little resistance between the output of the 555 and the gate of the FET. Often the inductance of the layout can cause oscillation here. You can scope the gate and maybe see it. You want to keep the leads short from the output to the gate and the ground for the FET and the 555.
The second is the snubber - R6 and C3. If you scope the drain you will see some high frequency oscillations. You can play with the R and C to minimize these. (larger C smaller R) But not to far as the power in the resistor will go up.

 

There a a couple of possibilities.
You could add a little resistance between the output of the 555 and the gate of the FET. Often the inductance of the layout can cause oscillation here. You can scope the gate and maybe see it. You want to keep the leads short from the output to the gate and the ground for the FET and the 555.
The second is the snubber - R6 and C3. If you scope the drain you will see some high frequency oscillations. You can play with the R and C to minimize these. (larger C smaller R) But not to far as the power in the resistor will go up.

I'll check for those things and then I'll get back to you. Thanks for the recommendations.

 

A couple more thoughts.

I did a quick mental calculation and I don't think that the capacitance of the FET is enough to cause the 555 to run hot.

What frequency is the output of the 555 running at?

Is it a CMOS 555?

Maybe the 555 was damaged in a previous use.

At the temperature you are describing, I would not expect the 555 to last for long.

 

I found this circuit after doing some goggling, built it, tested it... and it worked wonderfully.

So I tweaked it a little bit by first powering it up with 12V instead of 9V, and then I changed the 220K resistor above the 1K pot with a 60K resistor... and voila! I got the 75V output that I had been looking for (but that's another story).

Unfortunately, I can only get about 15-20 mA out of it...

Question: what changes would be necessary for it to deliver up to 100ma without having to use an entirely different circuit? I'm trying to avoid using a transformer by all means possible.
Would adding another IR740 in parallel help? Or using a higher value inductor? or increasing the capacitor at its output from 4.7µ to 10µf?

Or all the above???

A forward converter might give you more current than that flyback converter.

You'd be far better starting with a chip specifically designed for the job - the MC34063 is particularly convenient, as there are several calculator utilities online for working out the best component values.

 

A couple more thoughts.

I did a quick mental calculation and I don't think that the capacitance of the FET is enough to cause the 555 to run hot.

What frequency is the output of the 555 running at?

Is it a CMOS 555?

Maybe the 555 was damaged in a previous use.

At the temperature you are describing, I would not expect the 555 to last for long.

Richard,

I've done some measurements with my scope and here are the results:

• mosfet gate 01.png (image 3/6) shows bursts of 12V, lasting 3.2ms, every 8ms (125 hz), apparent duty cycle of 40%
• mosfet gate 02.png (image 4/6) shows a burst's detail (with a horizontal grid division of 10 µs) clearly showing a switching frequency of approximately 600 Khz
• the trans col images (5/6 and 6/6) show readings at the transistor's collector, very similar to the mosfet readings when it comes to frequency and duration.
  
• the mosfet drain 01 and 02 images (1/6 and 2/6) show readings using the same time parameters as the previous images
• I've physically measured the temperature at the 555 and it's running at 58°C (137°F)

The datasheet says that the 555 can be operated at up to 85°C, so we're kind of safe on that side... I guess I was being too much of a wimp when I told you I almost fried my finger...

My uneducated guess is that the 555 is being overclocked at each burst, and that's why it's running too hot, since the datasheet says that 500 Khz is the top frequency for this device... but I could be wrong, of course.

 

A forward converter might give you more current than that flyback converter.

You'd be far better starting with a chip specifically designed for the job - the MC34063 is particularly convenient, as there are several calculator utilities online for working out the best component values.

Thanks for the tip, Ian.
Could you share an example of the circuit you're suggesting?

 

Something very strange going on.
It should be running at 40 - 50 KHz.
Did you try any of the fixes?
If you did, with no results, you can disconnect the gate from the 555 and ground the gate. Then scope the output of the 555 and see what the frequency is.

 

Something very strange going on.
It should be running at 40 - 50 KHz.
Did you try any of the fixes?
If you did, with no results, you can disconnect the gate from the 555 and ground the gate. Then scope the output of the 555 and see what the frequency is.

Here's what the scope shows at the 555's trigger pin, with the gate disconnected and grounded.
It's a signal of approx 29 Khz, with an amplitude of around 11V

 

Yes, I did place a 47Ω resistor at the gate and it didn't make a difference...
I'm guessing something has to be done to the BC547's base to damper the triggering a little bit... maybe a 10 nf cap in parallel with the 150K resistor, or something...

 

errrrr... nope... a 220 nf cap placed there was a BAD idea... the cap overheated and kapooothhh!
fortunately nothing happened to the rest of the circuit...

 

Thanks for the tip, Ian.
Could you share an example of the circuit you're suggesting?

There's a few online that only need finding.

 

Do you have the resistor and capacitor across the FET?

 

Ok... here are a couple of images of my project... which by the way, is working beautifully thanks to all of you...

The first one is a picture of the complete assembly. The left half of the circuit is the 12V-to-75V flyback converter, and the other half contains an mcu that controls a 556 that I'm using to generate the output pulses.
The other image is a scope reading that shows the 75 pulse (yellow trace) that is triggered by a 556 that is in turn being triggered by the mcu (green trace).
The circuit's output is running through a 732 Ω resistor. I only switched it on for a few seconds, since the load resistor is only 1/8 W... and I'm not in the mood for getting a high through electronic smoke inhalation today...

In the middle of the picture, a bit to the right, is a patch that I was forced to add (an excellent high and low side mosfet driver designed by JDT) since I couldn't make the stupid mosfet driver chip that I selected for switching the output pulses work, no matter how many times I tweaked it, and how thoroughly I studied its application notes... it's a bummer, because this extra driver circuit adds a few dollars to my design.

I'm now only left with the mystery as to why the 555 is running so hot, and how to fix it...

 

Hello,

A couple of good points already made in this thread so i might repeat a couple because it looks like they were ignored anyway

First, for ANY boost circuit there is a maximum resistance for the inductor and switch and if this resistance is exceeded it means you will not ba able to get the required output no matter what else you do. That means the resistance of the switch and inductor must be below a certain value which depends on the input voltage and desired output voltage. The input voltage source impedance is also part of this, so that has to be low enough also.
The reason for this is because the inductor charges up though these resistances and if it can not charge up enough the output will never be able to get high enough. We can calculate this resistance if you like, or you can use a simulator.

Second, the switching time of the MOSFET must be fast. The rise an fall must be fast because that eats up energy available at the input. This means that a small MOSFET driver might be needed. The 555 may not be able to drive the MOSFET fast enough.

These points require careful thought to get from 9vdc to 170vdc, it doesnt happen by magic. To get from 9v to 12v doesnt take much thought, but the higher the output/.input ratio the more careful the design has to be done.

The inductance for inductors in series adds. So two inductors each 100uH in series will give you 200uH. No change in current ability, but the series resistance doubles so if each has 0.1 ohms then the total now is 0.2 ohms.
Inductors in parallel increase the current capability and decrease the series resistance, but the inductance decreases. Two 100uH inductors in parallel will give you 50uH but the current rating will be double and the series resistance halved.

 

The 555 may not be able to drive the MOSFET fast enough.

Thanks for your observations MrAl.

You might very well have a point regarding the MOSFET switching speed and its probably needing a special driver, I'll look into that.
As for the inductor, I'm using three 100 µH inductors in parallel to obtain a 33 µH inductance. Here's the datasheet of said inductors. I'm planning on replacing them with a single 33 µH inductor.

Thanks again

 

Hello again,

Well 33uH doesnt sound like much. The frequency has to be fairly high, maybe 100kHz or better. That's because if the frequency is too low the inductor current has to rise too high to get the output to work right. Paralleling inductors might work, but the inductance still has to be high enough.

Also, there is an inherent problem with regular boost circuits that relates to the duty cycle and main resistances. In a buck circuit, if the output voltage falls lower then the control circuit increases the duty cycle and that picks the output back up again, and this continues as needed. In a boost circuit the inductor 'on' duty cycle also increases in this way but now there is a limit. If we go over that duty cycle limit, the output goes DOWN. So the feedback changes from negative feedback to positive feedback, and the output goes too low.
This duty cycle limit can be calculated approximately knowing all the main resistances from:
PeakDMax=1-sqrt((Ron+Rin+Lesr)/Ro)

and since the duty cycle is 1-Vin/Vout we have:
Ron+Rin+Lesr<Ro*Vin^2/Vout^2

where
Ron is the switch resistance,
Rin is the input resistance (internal to the battery in most cases),
Lesr is the inductor total series resistance,
Ro is the output resistance.

The first equation above states the maximum duty cycle that can be used before the converter output drops, and the second equation states that the sum of those three resistances must be less than the output resistance times the square of Vin/Vout. If these conditions are met then the converter can work, but if not met then it wont be able to get up to the required output voltage.

Because of the output drop with large duty cycle sometimes a means to limit this has to be incorporated. If not, the converter may not be able to start up properly and even if it does start it may not be able to get up to the required voltage. Of course this means there has to be some minimum 'off' time, such as 5 percent of the total period. A so called "slow start" circuit may help here too.

 

Thanks for your observations MrAl.

You might very well have a point regarding the MOSFET switching speed and its probably needing a special driver, I'll look into that.
As for the inductor, I'm using three 100 µH inductors in parallel to obtain a 33 µH inductance. Here's the datasheet of said inductors. I'm planning on replacing them with a single 33 µH inductor.

Thanks again

A few years ago I cobbled together a simple boost converter with a bipolar 555 (CMOS has faster edges but less drive capability). The objective was to get higher voltages from a 4.8V Ni-Cd or Ni-Mh pack, the initial Vcc came via the flyback inductor and its diode but was boosted once the 555 got going.

The measured Vcc turned out to be just over 30V - the datasheet said 18V max!. It crossed my mind to sense any voltage under 18V to switch on a shunt transistor to ground the reset pin, but the chip blew before I got round to it, I had to go and do something else, and somehow never got back to that experiment.

IIRC; the MOSFET I used was an IRF640 - no special reason other than it was laying about on the bench - an IRF540 probably would've been better.