Converting UC3842 to constant off time controller?

Hello,
Sorry to bother, I wondered if any reader remembers which page of Christophe Basso's book shows how to convert a uc3842 into a constant off time controller?
Ive been looking for the last 30 mins and cant find it...its not in the index.
I only ask because I feel sure that others will benefit from this too.

See page 9: https://dc195.4shared.com/doc/0soglPxp/preview.html

Thanks FvM,
That looks the best, "lowest components count" way of doing it....

isn't it strange that this more complicated version is offered on the web?....(also attached)


**broken link removed**

I din't try the circuit, so I can't say if there are possible pitfalls involved with it (or with the other one).

I think the venables on ethat you kindly attached looks very good, however, of course, the off time is governed by how long it takes for the cap to charge up to the upper reference point, and then discharge from there down to the lower reference point...................these 'points' have a tolerance, thus the dead time will have a tolerance....also, when the timin cap is being discharged by the uc3842 chip, that discharge current has a tolerance.

So basically,your article gives a dead time comprising the charge up time of the cap, plus the discharge time down to the lower reference voltage............I also think its unsafe to discharge the cap directly with the output transistor of the LM139....a limiting resistor would surely be needed here?

It amazes me that off-the-shelf constant off time pwm controllers simply don't exist other than HV9910B and NCL30105......Why are there so few?

I also think its unsafe to discharge the cap directly with the output transistor of the LM139....a limiting resistor would surely be needed here?

Click to expand...

No, it has safe current limit of about 20 mA.

The LM139 datasheet states that 20mA is the maximum allowable current in the output transistor........but the current is not limited to that......if it was limited to 20mA , then there would have been no need to put this into the "absolute maximum" ratings table?

I still think an external limiting resistor will be needed.

LM139 DATASHEET;
https://www.ti.com/lit/ds/symlink/lm139.pdf

The output transistor saturation voltage is said to be up to 700mV.