Twitter   Patreon WIGGLING 'LITE' IN GUEST MODE

Author LM13700 as divider
555x555
 I finally found an application for 2Q division, for which the LM13700 linearizing diodes plus a v to i circuit seem pretty ideal. Anybody actually rig up the LM13700 as a divider? Any tips / gotchas?
guest
 there are 2 different ways to do this. the first is as you describe, where you control the current to the diodes. the second is to put the LM13700 in the feedback loop of an opamp. the diode control method is a less mathematically accurate divider, but has a higher bandwidth and is more stable. basically, it works well with a decent diode current, but as the diode current drops, the gain doesnt go to infinity, and instead tops off at the gain which you would get if you just didnt use the diodes at all. this can be an advantage, depending upon the application. the feedback case works more ideally, but can have stability issues, and definitely loses control at the zero control voltage case, as this makes gain infinite.
555x555
 Yeah, the behavior around zero is actually why I want to use the diodes. The application I'm thinking of is only 2 quadrant, and I want zero and "negatives" (clamped to zero) to go into gentle saturation type behavior. I've been playing with a simulator and the only tricky thing seems to be that the diodes require compliance of the current source to be above zero V, since they're draining to ground. I used a V to I converter with -x to a current mirror on the top rail, but it strikes me there may be a better way to do this. More out of completeness/curiosity than practicality, the data sheet claims that the key condition to maintain the integrity of the division would be to keep Id>|Is|/2. The idea just being, basically, that the current being pulled/pushed by the transistors is provided by the diodes, yeah? But I'm thinking that maybe that absolute value isn't quite exactly correct when, as is normally the case, one side of the drive is held at or near ground and only the other allowed to vary. Not sure...
555x555
 555x555 wrote: the only tricky thing seems to be that the diodes require compliance of the current source to be above zero V

...aaaaaaaand I just realized that I don't have to do this since the input is differential and I can just throw something less than ground on that node...I think...
guest
 yeah, if you want to use the opamp/transistor current source, you can either move its reference voltage above ground, or move the OTA reference voltage below ground. you can also use the second half of the 13700 as your current source, or a simple transistor current mirror.
guest
 and by the way, both the CA3280 and BA6110 have diode current sources internally that you can control.
 Page 1 of 1