[SOLVED] UC3843 boost converter (boost from 3.7 to 80v)

[TuralH]

I want to assemble a dc-dc converter that will increase the voltage from 3.7 volts to 80 volts/
I need to increase it using uc3843, but due to the fact that this PWM controller operates from 12-28 volts, I assembled another boost converter on mc34063 increased it to 27-30 volts and using LM7812 supplied 12 volts, in Ltspice the circuit works and the expected voltage is obtained at the output, but in reality and in practice will it be possible to increase the voltage using uc3843 and external NMOS from 3.7 to 80?
The duty cycle is: D≈83.76% simulated in LTspice and with 91khz of frequency. Of course there are ready-made DC-DC converters that can do this, but I would like this circuit to work too.
I will be glad for your help. Photos of the simulation are attached below.

 

[std_match]

I think the uc3843 requires minimum 9V DC in to start, so it seems to be a bad choice for this application.

 

[TuralH]

I think the uc3843 requires minimum 9V DC in to start, so it seems to be a bad choice for this application.

yes, You are right, for uc3843 starting i use mc34063 for boosting from 3.7 to 27-30v and with lm7812 make 12v for VCC, but 3.7v i supplied inductor of uc3843 for boost from 3.7 to 80v

 

[Easy peasy]

3.7 VDC to 80VDC needs 2 or 3 stages to get you there, 1st you need a small converter that will run off 3.7VDC to get you up to 12V say to run a proper control IC,

a small self oscillating flyback can do this,

then you need to look at a cascaded boost ( 2 fets 2 diodes - same gate drive ) to go from 3.7 to 17V, and thence to 80V, 17 being the geometric mean of 17 & 80,
thus the boost in each case is 4.7 x - a much more manageable figure, 4.7 = D/( 1-D ) therefore D still needs to be > 82.5% on the gate of each fet to achieve your ends . . . .

goooooooooooooood luck !

 

[TuralH]

3.7 VDC to 80VDC needs 2 or 3 stages to get you there, 1st you need a small converter that will run off 3.7VDC to get you up to 12V say to run a proper control IC,

a small self oscillating flyback can do this,

then you need to look at a cascaded boost ( 2 fets 2 diodes - same gate drive ) to go from 3.7 to 17V, and thence to 80V, 17 being the geometric mean of 17 & 80,
thus the boost in each case is 4.7 x - a much more manageable figure, 4.7 = D/( 1-D ) therefore D still needs to be > 82.5% on the gate of each fet to achieve your ends . . . .

goooooooooooooood luck !

thank You for Your help and advice. I tried to do a gradual increase, but as I connected the second converter, the voltage of the first dropped to 3V (but it should have been 12V) did you have any experience? For example, I did this, connected two mc34063 in series, but nothing worked, can you send a source or an approximate circuit

 

[betwixt]

What current are you needing at 80V?
Do consider that boosting the voltage by 80/3.7 = 21.6 times you also need to draw a current somewhat more than 21.6 times greater from the 3.7V supply.

Chopping and stepping up with a transformer might be a suitable solution.

Brian.

 

[TuralH]

current consumption is maximum 1-2A, I wanted to use cascade boost, but the fact is that there is no information on the network or at least a similar circuit, only scientific tests about cascade boost converter

 

[BradtheRad]

Besides the helpful replies above, consider interleaving two or more twin boost converters. That is another method to allow 21X step-up V. Be advised you still need to supply 22 to 44 Amperes from little 3.7V batteries.

--- Updated Monday at 4:07 PM ---

Did I say 22-44A? It has to be really more like 70 A from little 3.7V batteries.

 

[TuralH]

Besides the helpful replies above, consider interleaving two or more twin boost converters. That is another method to allow 21X step-up V. Be advised you still need to supply 22 to 44 Amperes from little 3.7V batteries.

--- Updated Monday at 4:07 PM ---

Did I say 22-44A? It has to be really more like 70 A from little 3.7V batteries.

I tried to connect 2-3 converters, increasing them step by step, but when I connected the second converter, the voltage of the first one dropped to 3V. Is this possible due to the common ground?

 

[KlausST]

Hi,

Another approach for
cascaded design with equal factors:
80V/3.7V = 21.9
two stages: F x F = 21.9 -->
F = sqrt(21.9) = 4.64

first stage: 3.7V x 4.64 = 17.2V
second stage: 17.2V x 4.64 = 80V
**
for three stages:
F x F x F = 21.9
F = 21.9 ^(1/3) = 2.80

Klaus

 

[TuralH]

Hi,

Another approach for
cascaded design with equal factors:
80V/3.7V = 21.9
two stages: F x F = 21.9 -->
F = sqrt(21.9) = 4.64

first stage: 3.7V x 4.64 = 17.2V
second stage: 17.2V x 4.64 = 80V
**
for three stages:
F x F x F = 21.9
F = 21.9 ^(1/3) = 2.80

Klaus

thanks for the advice, theoretically this works, but in practice the first converter should increase from 3.7V to 11-17V, its output becomes equal to 3.0-.3.2V, do you have a practical experiment with cascaded boost? if so, can you send me a diagram or with which IC can I assemble this topology?

 

[Easy peasy]

Search images on the net for cascaded boost - it works very well - is your 3.7 V source capable of supplying the current needed ?

--- Updated Monday at 8:06 PM ---

1 - 2 amps from 3.7 V = 7.4 watts max = 80V and 90mA max, if you require 2A at 80V this is 44A from your 3.7V source.

 

[TuralH]

Search images on the net for cascaded boost - it works very well - is your 3.7 V source capable of supplying the current needed ?

--- Updated Monday at 8:06 PM ---

1 - 2 amps from 3.7 V = 7.4 watts max = 80V and 90mA max, if you require 2A at 80V this is 44A from your 3.7V source.

I searched on the Internet, mainly on the ResearchGate website there are theoretical circuits, but I need practical ones, yes, I use a 20A power supply

--- Updated Monday at 8:10 PM ---

Search images on the net for cascaded boost - it works very well - is your 3.7 V source capable of supplying the current needed ?

--- Updated Monday at 8:06 PM ---

1 - 2 amps from 3.7 V = 7.4 watts max = 80V and 90mA max, if you require 2A at 80V this is 44A from your 3.7V source.

at the output I have a capacitor of 2200 80 Volts, I need to charge this capacitor but with 3.7 volts, previously I used the Chinese analogue of lm2577, but due to the fact that its maximum is 40 Volts, I had to use uc3843 with external NMOS, but for the experiment i need to boost from 3.7

 

[dick_freebird]

I think you want a fly back for the ratio and a 5V technology (though this is close to the bone, snubber will be key to survival?). But the lowest "distribution bus" I've seen is 5V and in such cases I've picked 4V VINmin and 3.8V UVLO.

We made a stab at 3.3V input versions but got no traction from the customers. And you'd be on the other wrong side, for VIN, on that (3.0 to 3.6 for a 10% supply).

If you don't need fixed frequency then perhaps a basic bodge of Vref, comparator, low side switch set up to chatter when below Vref (hysteretic control) is your pony. Use a logic level FET or switching type NPN or eGan with appropriate gate network.

 

[TuralH]

I think you want a fly back for the ratio and a 5V technology (though this is close to the bone, snubber will be key to survival?). But the lowest "distribution bus" I've seen is 5V and in such cases I've picked 4V VINmin and 3.8V UVLO.

We made a stab at 3.3V input versions but got no traction from the customers. And you'd be on the other wrong side, for VIN, on that (3.0 to 3.6 for a 10% supply).

If you don't need fixed frequency then perhaps a basic bodge of Vref, comparator, low side switch set up to chatter when below Vref (hysteretic control) is your pony. Use a logic level FET or switching type NPN or eGan with appropriate gate network.

thanks for the advice and method, yes, I don’t need a fixed frequency, my main goal is to achieve 74-80 volts of output voltage, and charge a capacitor with a nominal value of 2200uF and 80V,

 

[KlausST]

but I need practical ones, yes, I use a 20A power supply

how .. do you think a 20A supply will work ... while you need 44A?

Klaus

 

[TuralH]

how .. do you think a 20A supply will work ... while you need 44A?

Klaus

I have 3 more power supplies, I can connect them in parallel and I also have no load at the output, the main task is to charge the capacitor

 

[Easy peasy]

" cascaded boost pdf " 2 sec search on google, click on images :

--- Updated Tuesday at 4:55 AM ---

tie the gates together with 22E to each mosfet - single pwm.

--- Updated Tuesday at 4:59 AM ---

Don't forget your 1st inductor has to handle 20 or 40 amp DC during charging - plus the design ripple current

I seriously doubt you have the requisite chokes for this ?

The 1st mosfet must handle 50A easily and be at least 40V . . . .

 

[TuralH]

" cascaded boost pdf " 2 sec search on google, click on images :

View attachment 195428

--- Updated Tuesday at 4:55 AM ---

tie the gates together with 22E to each mosfet - single pwm.

--- Updated Tuesday at 4:59 AM ---

Don't forget your 1st inductor has to handle 20 or 40 amp DC during charging - plus the design ripple current

I seriously doubt you have the requisite chokes for this ?

The 1st mosfet must handle 50A easily and be at least 40V . . . .

I found this circuit in ResearchGate and assembled it; the only thing I used was a gate drive from different PWM controllers, so it didn’t work, which means I have to use one PWM controller for two mosfets also my goal is charging 2200uF output capacitor and I(out) for charging will be i think about 178mA , then i need about 4.5A input for charging capacitor only.

 

[BradtheRad]

I have 3 more power supplies, I can connect them in parallel and I also have no load at the output, the main task is to charge the capacitor

Frequent advice is to beware of paralleling power supplies. The risk is that one gets 'weaker', sags in voltage (for whatever reason), thus obliging the other 'strong' supplies to send current through it in reverse. This is surely counterproductive. Steering diodes are one solution although that subtracts voltage from already low 3.7V supplies.

In a reply above I jumped prematurely to the conclusion you had battery power (Li-ion type) because 3.7V is common for Li-ion cells.

Did you rule out taking 80VDC from house voltage? With a few added components? Of course that brings its own hurdles.