EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: SArepairman on August 31, 2014, 02:50:27 am

Title: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: SArepairman on August 31, 2014, 02:50:27 am
So I built this nice charge pump to divide my voltage reference... the only problem is... its soooooo bloody slow. it seriously takes like 5 min to get to the proper value... this was a hard problem to predict.

I am not sure what to do now, I thoought it was a golden bullet for divider voltage references through out a system. I guess I need to replace it with a chopper amplifier + thermally matched resistors 


perhaps I can include a comparator to switch this on while the user waits..
https://www.youtube.com/watch?v=Q2I7FN7fiKU (https://www.youtube.com/watch?v=Q2I7FN7fiKU)
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: c4757p on August 31, 2014, 03:05:08 am
it seriously takes like 5 min to get to the proper value...

Five minutes? Uh - how are you using it, and how close do you expect it to get? ???
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: SArepairman on August 31, 2014, 03:09:23 am
I am using it in a divide by four (2x divide by two)

made with 4x high quality polypropylene 1uF capacitors + 10nF switching PP Cap (vishay type 63V signal grade)
I feed it 4, it got to 0.8 in like.. 9 minutes

slow as molasses going up hill

going into a LTC1151 chopper amplifier (15pA bias current )

10V supply on the lt6943

not to mention its over counting, 1.2 v in a div x4 circuit
i hope that something is wrong?
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: rs20 on August 31, 2014, 04:01:04 am
Maybe other people are smarter than me, but I need to see a schematic to have any confident opinion on this. Is the 10nF switch cap the one that's alternately being connected to one side of the lt6943 and then the other?  Because 10nF is 100,000 times smaller than 1uF, even in an ideal world it's going to take hundreds of thousands of cycles before it approaches equilibrium. Any reason not to use a bigger switched cap?
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: SArepairman on August 31, 2014, 04:40:01 am
a ha. it was a floating pin.

I forgot a trace going from the output of the first divider into the bottom side switch.

(7 to 11 was DC)

It had me fooled because it kinda almost worked...

now its speedy and it works great!

the only thing that was missing from the circuit was a piece of kynar wire!
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: c4757p on August 31, 2014, 04:41:44 am
a ha. it was a floating pin.

Dammit!

I suspected as such, but was unsure and didn't want to talk out of my ass. :P
Title: Re: beware of lt6943 charge pump divider circuits (SLOOOOOOOOOOOOOOOW)
Post by: SArepairman on August 31, 2014, 04:43:00 am
it was kind of a pita to bodge that wire in, because its a tiny smd chip surrounded by massive (slightly molten ) capacitors.

if i ever use one of these again I am going to surround it with the capacitors like a castle wall and put some decals and king arthur on the rampart